issue_owner_repo listlengths 2 2 | issue_body stringlengths 0 261k ⌀ | issue_title stringlengths 1 925 | issue_comments_url stringlengths 56 81 | issue_comments_count int64 0 2.5k | issue_created_at stringlengths 20 20 | issue_updated_at stringlengths 20 20 | issue_html_url stringlengths 37 62 | issue_github_id int64 387k 2.46B | issue_number int64 1 127k |
|---|---|---|---|---|---|---|---|---|---|
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-23 04:38:48 +0200
From: Hans <<hans_0_>>
To: @sjoerdmullender
Version: -- development
Last updated: 2009-04-07 01:11:06 +0200
## Comment 7941
Date: 2007-07-23 16:38:48 +0200
From: Hans <<hans_0_>>
The following set of queries result in a runtime error:
xquery>pf:del-doc("notes_consumer02.xml")
more>
xquery>pf:add-doc("notes_template.xml", "notes_consumer02.xml", "notes_consumer02.xml", 10)
xquery>doc("notes_consumer02.xml")
<?xml version="1.0" encoding="utf-8"?>
<notes>
<objectnote xid="4" unixtime="1180000000" user="hans">This is my first note</objectnote>
</notes>
xquery>let $doc := doc("notes_consumer02.xml") return $doc
<?xml version="1.0" encoding="utf-8"?>
<notes>
<objectnote xid="4" unixtime="1180000000" user="hans">This is my first note</objectnote>
</notes>
xquery>let $doc := doc("notes_consumer02.xml")
more>let $orig := $doc/notes/objectnote[@xid="14"]
more>let $on := element objectnote { attribute xid {"14"}, attribute user {"unknown"}, attribute unixtime {"1185198696"}, text {"nootje"}, element original { $orig } }
more>return if (exists($orig)) then (do replace $orig with $on) else (do insert $on as last into $doc/notes)
xquery>let $doc := doc("notes_consumer02.xml") return $doc
<?xml version="1.0" encoding="utf-8"?>
<notes>
<objectnote xid="4" unixtime="1180000000" user="hans">This is my first note</objectnote>
<objectnote unixtime="1185198696" user="unknown" xid="14">nootje<original/></objectnote></notes>
xquery>let $doc := doc("notes_consumer02.xml")
more>let $orig := $doc/notes/objectnote[@xid="14"]
more>let $on := element objectnote { attribute xid {"14"}, attribute user {"unknown"}, attribute unixtime {"1185198696"}, text {"nootje, pinda"}, element original { $orig } }
more>return if (exists($orig)) then (do replace $orig with $on) else (do insert $on as last into $doc/notes)
MAPI = monetdb@botswana.xiraf:50000
QUERY = let $doc := doc("notes_consumer02.xml")
ERROR = !ERROR: interpret: no matching MIL operator to 'seqbase(BAT[oid,int], oid)'.
!MAYBE YOU MEAN:
! seqbase(BAT[void,any::1], oid) : BAT[void,any::1]
! seqbase(BAT[void,any::1]) : oid
! seqbase(BAT[oid,any::1]) : oid
!ERROR: interpret_params: key(param 1): evaluation error.
xquery>let $doc := doc("notes_consumer02.xml") return $doc
<?xml version="1.0" encoding="utf-8"?>
<notes>
<objectnote xid="4" unixtime="1180000000" user="hans">This is my first note</objectnote>
<objectnote unixtime="1185198696" user="unknown" xid="14">nootje<original/></objectnote></notes>
xquery>let $doc := doc("notes_consumer02.xml") "14"}, attribute user {"unk
more>let $orig := $doc/notes/objectnote[@xid="14"] "14"}, attribute user {"unk
more>let $on := element objectnote { attribute xid {"14"}, attribute user {"unknown"}, attribute unixtime {"1185198696"}, text {"nootje, pinda"}}
more>return if (exists($orig)) then (do insert $on as last into $doc/notes, do delete $orig) else (do insert $on as las
t into $doc/notes)
MAPI = monetdb@botswana.xiraf:50000
QUERY =
ERROR = !ERROR: CMDdelete_bat_bun: operation failed.
xquery>let $doc := doc("notes_consumer02.xml") return element { "result" } { $doc }
<result><notes>
<objectnote user="hans" unixtime="1180000000" xid="4">This is my first note</objectnote>
<objectnote xid="14" user="unknown" unixtime="1185198696">!ERROR: XML Generation: tmp_2231 BAT does not have a 9 head.
MAPI = monetdb@botswana.xiraf:50000
QUERY = let $doc := doc("notes_consumer02.xml") return element { "result" } { $doc }
ERROR = !ERROR:
!ERROR: XML Generation: NULL BAT has a 9 head, but tail is NULL.
!ERROR:
!ERROR: xquery_print_result_loop: operation failed.
xquery>let $doc := doc("notes_consumer02.xml") return $doc
<?xml version="1.0" encoding="utf-8"?>
<notes>
<objectnote xid="4" unixtime="1180000000" user="hans">This is my first note</objectnote>
<objectnote unixtime="1185198696" user="unknown" xid="14"></objectnote></notes>
## Comment 7942
Date: 2007-07-25 17:42:11 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Hans,
without the respective (sample) documents, there is no (reasonable) way to reproduce, analyze, locate, fix, test this (potential) bug.
Hence, could you please provide us with the documents?
Thanks!
Stefan
## Comment 7943
Date: 2007-07-25 22:10:58 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Actually, the document is implicit in the report. Look at the output of the doc() function at the start.
I can indeed reproduce the problem. It looks like the call to seqbase on line 2454 of pf_support.mx is the one that is failing. Needless to say, it shouldn't. More investigation is needed.
## Comment 7944
Date: 2007-07-26 13:08:39 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in CVS (stable).
Added test pathfinder/tests/BugTracker/Tests/replace-corrupts.SF-1758902*.
## Comment 7945
Date: 2009-02-16 01:02:21 +0100
From: @drstmane
re-opened as test output differs from stable one with both ALG & MPS:
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d.1-Fedora8/tests_BugTracker/replace-corrupts.SF-1758902.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTestsg103/GNU.64.64.d.1-Fedora8/tests_BugTracker/replace-corrupts.SF-1758902.out.00.html
## Comment 7946
Date: 2009-04-07 12:26:44 +0200
From: @peterboncz
is this something serious, Sjoerd?
## Comment 7947
Date: 2009-04-07 13:11:06 +0200
From: @sjoerdmullender
Looking carefully at what's going on in the test, I believe the actual output is correct, so I approved it.
Closing.
## Comment 7948
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1758902 at http://sourceforge.net/support/tracker.php?aid=1758902
| XQUF replace function seems to corrupt the database | https://api.github.com/repos/MonetDB/MonetDB/issues/1410/comments | 0 | 2020-11-30T09:36:57Z | 2024-06-27T11:25:43Z | https://github.com/MonetDB/MonetDB/issues/1410 | 753,316,435 | 1,410 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-21 08:36:54 +0200
From: Bharani <<bharani_monetdb>>
To: @njnes
Version: -- development
Last updated: 2007-08-12 11:36:45 +0200
## Comment 7933
Date: 2007-07-21 08:36:54 +0200
From: Bharani <<bharani_monetdb>>
I am trying to execute this sql as prepared statement
update reason set failurereason=? where ID=?
reason is a table with two columns
ID is integer
failurereason is a string
During the batch execution i see
exec 3('whatever',2);
in the batch and then i get
MALException:(unknown):wrong argument for prepared statement (expected int instead of char)
Looks like the types are getting swapped or dosnt maintain the proper order
Thanks
Bharani
## Comment 7934
Date: 2007-07-25 17:45:18 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
fixed/extended administrative info.
## Comment 7935
Date: 2007-07-26 10:49:16 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
Hi bharani,
Is it possible to provide a JDBC logfile for this case? Our testing does a very similar test every night (e.g. Test_PStypes) which doesn't fail. I suspect you case is slightly different, for instance in the order of columns. Is ID earlier defined than failurereason in the create statement for instance? A log of the prepare (anonymised data is fine, I'm looking for the return of the actual prepare, and the exec commands) would help to get some grip on this issue.
## Comment 7936
Date: 2007-07-26 10:59:37 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
never mind, I changed the test to match your statement, and indeed the server returns the variables in the wrong order for both M4 and M5.
## Comment 7937
Date: 2007-07-26 12:29:28 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I tried to tarzan around in Niels' jungle, but I can't find it. So I guess Niels should solve this. It looks like the arguments of the where_clause are added first to the params list, resulting in this wrong order.
## Comment 7938
Date: 2007-07-26 12:45:12 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
TEST ADDED
clients/src/java/tests/Test_PStypes
sql/src/jdbc/tests/Tests/Test_PStypes
modified existing test program to include the update statement as in this bug
## Comment 7939
Date: 2007-08-12 23:36:45 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed checked into the release branch. The problem was that the parameters were handled only in the semantic phase. Now we allready handle them in the parser which solves the order problem.
## Comment 7940
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1757923 at http://sourceforge.net/support/tracker.php?aid=1757923
| SQL: Prepared statement type mismatch | https://api.github.com/repos/MonetDB/MonetDB/issues/1409/comments | 0 | 2020-11-30T09:36:55Z | 2024-06-27T11:25:42Z | https://github.com/MonetDB/MonetDB/issues/1409 | 753,316,408 | 1,409 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-20 09:28:41 +0200
From: vdwel <<vdwel>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-09-29 10:23:18 +0200
## Comment 7928
Date: 2007-07-20 09:28:41 +0200
From: vdwel <<vdwel>>
When adding two documents to the same collection, throught the web management interface, the attributes of the second document are not displayed.
They are not displayed when the document is being viewed through the web interface. They also seem to be gonne when querying through the client prompt.
The versio I am using the the latest Windows version.
## Comment 7929
Date: 2007-08-01 15:43:50 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
I cannot reproduce this using the MonetDB4-XQuery-i686-0.18.2.msi installer on a clean system.
## Comment 7930
Date: 2007-09-09 11:08:28 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
It seems this can be closed, because there has not been
an update and it is not reproducible. I suggest to
close with 'Works for me'
## Comment 7931
Date: 2007-09-29 22:23:18 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
I cannot reproduce this one with the latest MonetDB/XQuery 0.20.0 release candidate version on Linux, either.
Hence, closing.
Feel free to re-open or submit a bug report in case the problem still persists for you, preferably attaching the respective documents.
## Comment 7932
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1757327 at http://sourceforge.net/support/tracker.php?aid=1757327
| XQ: attibutes in 2nd doc don't show | https://api.github.com/repos/MonetDB/MonetDB/issues/1408/comments | 0 | 2020-11-30T09:36:52Z | 2024-06-28T13:31:21Z | https://github.com/MonetDB/MonetDB/issues/1408 | 753,316,377 | 1,408 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-17 03:19:38 +0200
From: thijs <<thijswes>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-07-31 10:22:53 +0200
## Comment 7920
Date: 2007-07-17 15:19:38 +0200
From: thijs <<thijswes>>
Hi,
I found MonetDB/XQuery has trouble with the printing of a document/node sequence when special characters occur in comment nodes.
For example, the attached test.xml looks like this
<?xml version="1.0" encoding="utf-8"?>
<test>
<!--
<comment/>
abc
& def
-->
</test>
Looking at the tables in the DB I see the comment string is correctly parsed (including the &), but when the node is printed output stops just before &, the rest of the document/node/node sequence is not printed.
E.g.,
doc("..../test.xml")
gives the following output:
<?xml version="1.0" encoding="utf-8"?>
<test>
<!--
<comment/>
abc
The problem only occurs when a line *starts* with '&'
## Comment 7921
Date: 2007-07-18 10:26:30 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Undoubtedly this has to do with the Mapi protocol where lines starting with & have a meaning.
## Comment 7922
Date: 2007-07-18 10:36:55 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
it goes wrong in multiple ways it seems:
RD 1184747690872: inserting prompt
TD 1184747698266: write final block: 17 bytes
TX 1184747698266: sdoc("test.xml")
RD 1184747698310: read final block: 189 bytes
RX 1184747698310: &1 0 1 1 1
% xquery.result table_name
% xml name
% varchar type
% 60 length
[ "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n<test>\n\t<!--
<comment/>
abc
& def
-->\n</test>\n" ]
For some reason the newlines in the comment are not escaped. I can imagine that in the = mode MapiClient uses the same sort of thing happens therefore having the & as first char on the line or something.
## Comment 7923
Date: 2007-07-18 10:39:09 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I think its not Mapi but the serialiser
## Comment 7924
Date: 2007-07-25 17:47:12 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
tagged summary as "XQ".
## Comment 7925
Date: 2007-07-30 15:14:51 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed with a test in pathfinder/tests/BugTracker/Tests/comment-print-bug.SF-1755466.*
This will be checked in as soon as Sourceforge lets me.
## Comment 7926
Date: 2007-07-31 10:22:53 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Checked in in stable branch.
## Comment 7927
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1755466 at http://sourceforge.net/support/tracker.php?aid=1755466
| XQ: printing special characters in comment nodes | https://api.github.com/repos/MonetDB/MonetDB/issues/1407/comments | 0 | 2020-11-30T09:36:49Z | 2024-06-27T11:25:40Z | https://github.com/MonetDB/MonetDB/issues/1407 | 753,316,342 | 1,407 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-15 11:10:54 +0200
From: Aaron Evans <<aaronevans>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2007-07-17 12:57:07 +0200
## Comment 7912
Date: 2007-07-15 23:10:54 +0200
From: Aaron Evans <<aaronevans>>
While trying to perform many consecutive XQuery updates using the JDBC driver, I eventually receive a NPE:
Exception in thread "SendThread" java.lang.NullPointerException
at nl.cwi.monetdb.jdbc.MonetConnection$SendThread.run(MonetConnection.java:2115)
This constructor looks wrong to me:
public SendThread(BufferedMCLWriter conn) {
super("SendThread");
setDaemon(true);
this.out = out;
start();
}
It looks like SendThread should probably use the argument in the constructor.
## Comment 7913
Date: 2007-07-16 11:39:33 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
Indeed that looks like it is the case. I applied this patch:
diff -u -r1.16 MonetConnection.java
--- src/nl/cwi/monetdb/jdbc/MonetConnection.java 5 Jun 2007 10:31:41 -0000 1.16
+++ src/nl/cwi/monetdb/jdbc/MonetConnection.java 16 Jul 2007 09:25:09 -0000
@@ -2088,7 +2088,7 @@
*
* @param monet the socket to write to
*/
- public SendThread(BufferedMCLWriter conn) {
+ public SendThread(BufferedMCLWriter out) {
super("SendThread");
setDaemon(true);
this.out = out;
Can you check?
## Comment 7914
Date: 2007-07-16 11:43:51 +0200
From: Aaron Evans <<aaronevans>>
Logged In: YES
user_id=1845614
Originator: YES
Yes, that patch eliminates the NPE.
## Comment 7915
Date: 2007-07-16 12:22:16 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
Fixed in CVS. Thanks!
## Comment 7916
Date: 2007-07-17 10:42:23 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Shouldn't we also add test for this one?
Aaron,
could you provides us with your experiment (data & update queries)?
Thanks in advance!
Stefan
## Comment 7917
Date: 2007-07-17 10:53:40 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
yes, sorry. The batching test doesn't do this any more since it optimises on the buffer size.
## Comment 7918
Date: 2007-07-17 12:57:07 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
TEST ADDED - meltdown
sql/src/jdbc/tests/Tests/Test_Clargequery
## Comment 7919
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1754452 at http://sourceforge.net/support/tracker.php?aid=1754452
| NullPointerException in MonetConnection | https://api.github.com/repos/MonetDB/MonetDB/issues/1406/comments | 0 | 2020-11-30T09:36:47Z | 2024-06-27T11:25:39Z | https://github.com/MonetDB/MonetDB/issues/1406 | 753,316,302 | 1,406 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-11 09:45:37 +0200
From: thijs <<thijswes>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2008-06-03 03:36:26 +0200
## Comment 7897
Date: 2007-07-11 09:45:37 +0200
From: thijs <<thijswes>>
fn:sum() causes a merged_union error on some data. I could not pinpoint the exact pre-conditions for this error to occur, but the following simple query (on the attached small document) fails:
let $res := doc("/....../tmp.xml")
for $f in distinct-values(data($res//@id))
return ($f,fn:sum($res//link[@id=$f]/@score))
The same query with a fn:count instead of a fn:sum does not fail and produces the expected output.
I tried the query on the X.18 release.
My environment:
FC6
gcc 4.1.1,
flex version 2.5.4
bison (GNU Bison) 2.3
autoconf (GNU Autoconf) 2.59
automake (GNU automake) 1.9.6
ltmain.sh (GNU libtool) 1.5.22 (1.1220.2.365 2005/12/18 22:14:06)
## Comment 7898
Date: 2007-07-11 09:53:10 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Thijs,
What is the exact error (-message)?
Could you please try with property checking enabled (`Mserver --debug=10 ...`) and report on the results?
Stefan
## Comment 7899
Date: 2007-07-11 09:59:30 +0200
From: thijs <<thijswes>>
Logged In: YES
user_id=1305406
Originator: YES
Sorry, intended to include this, but forgot. The error message is the following:
MAPI = monetdb@localhost:50000
QUERY = let $res := doc("/ufs/thijs/tmp/tmp.xml")
ERROR = !ERROR: merged_union: tail of first BAT must be sorted.
!ERROR: CMDmerged_union: operation failed.
With property checking enabled, the query runs fine and produces the expected results.
## Comment 7900
Date: 2007-07-11 10:19:14 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
Looks like a chk_order() in mps is missing.
## Comment 7901
Date: 2007-07-11 10:37:21 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
Hi Thijs,
Could you please apply the following diff, recompile, and check if your query then works?
-- Thanks, Jan
Index: mil/milprint_summer.c
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/compiler/mil/milprint_summer.c,v
retrieving revision 1.391
diff -c -r1.391 milprint_summer.c
*** mil/milprint_summer.c 27 Jun 2007 13:46:45 -0000 1.391
--- mil/milprint_summer.c 11 Jul 2007 08:35:55 -0000
***************
*** 4661,4667 ****
"var iter_grp := iter.materialize(ipik);\n"
"iter := iter_grp.tunique();\n"
"var iter_aggr := {%s}(item%s, iter_grp, iter).tmark(0@0);\n"
! "iter := iter.hmark(0@0);\n"
"ipik := iter;\n"
"pos := 1@0;\n"
"kind := %s;\n",
--- 4661,4667 ----
"var iter_grp := iter.materialize(ipik);\n"
"iter := iter_grp.tunique();\n"
"var iter_aggr := {%s}(item%s, iter_grp, iter).tmark(0@0);\n"
! "iter := iter.hmark(0@0).chk_order();\n"
"ipik := iter;\n"
"pos := 1@0;\n"
"kind := %s;\n",
## Comment 7902
Date: 2007-07-11 11:04:26 +0200
From: thijs <<thijswes>>
Logged In: YES
user_id=1305406
Originator: YES
Yes, this fixes the bug. Thanks.
## Comment 7903
Date: 2007-07-11 13:08:33 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
Stefan,
assigning the bug to you as you (in comparison to me) certainly have a stable version running :)
Could you please check in the below change.
Thanks,
Jan
## Comment 7904
Date: 2007-07-11 15:18:16 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Fixed in CVS (XQuery_0-18 release branch; propagation to development trunk pending).
Added respective test in
pathfinder/tests/BugTracker/Tests/merged-union_error_with_fn-sum.SF-1751684.*
## Comment 7905
Date: 2008-06-02 00:02:26 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Re-opening:
While working fine with the MPS back-end, the tests fails with the ALG
back-end, mainly as fn:sum() seems to calculate different results;
cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTestsG103/GNU.64.64.d-Fedora8/tests_BugTracker/merged-union_error_with_fn-sum.SF-1751684.out.00.html
See also
[ 1636588 ] Predicate selects too few nodes
https://sourceforge.net/tracker/index.php?func=detail&aid=1636588&group_id=56967&atid=482468
## Comment 7906
Date: 2008-06-02 12:06:17 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
It turned out that this was a problem introduced on December 6.th, 2007. I changed back the implementation of unordered positions and approved the algebra specific output.
## Comment 7907
Date: 2008-06-03 11:26:53 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
The modified standard stable output makes the MPS test fail, now.
Is the output correct and MPS wrong, or vice versa?
In the former case, do we still care about MPS's correctness?
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTestsG103/GNU.64.64.d-Fedora8/tests_BugTracker/id-function_fails_to_locate_some_kinds_of_nodes.SF-1698498.out.00.html
## Comment 7908
Date: 2008-06-03 11:30:22 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
sorry, wrong URL; correct one is
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.32.d-Fedora8/tests_BugTracker/merged-union_error_with_fn-sum.SF-1751684.out.00.html
## Comment 7909
Date: 2008-06-03 12:01:10 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
MPS is incorrect.
I won't fix MPS.
## Comment 7910
Date: 2008-06-03 15:36:26 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Algebra is right and output is fixed.
For the MPS bug see
[ 1983577 ] PF/mps: fn:sum() default return value has wrong type
https://sourceforge.net/tracker/index.php?func=detail&aid=1983577&group_id=56967&atid=482468
## Comment 7911
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1751684 at http://sourceforge.net/support/tracker.php?aid=1751684
| XQ: merged_union error with fn:sum() | https://api.github.com/repos/MonetDB/MonetDB/issues/1405/comments | 0 | 2020-11-30T09:36:44Z | 2024-06-27T11:25:39Z | https://github.com/MonetDB/MonetDB/issues/1405 | 753,316,272 | 1,405 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-11 09:40:44 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-07-25 06:29:08 +0200
## Comment 7893
Date: 2007-07-11 09:40:44 +0200
From: @drstmane
Since Thursday June 28 2007, i.e., after checkins on Wednesday June 27 2007, most pathfinde algebra tests fail, e.g.:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qa04alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qj01alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qj02alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qj03alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qj04alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_XMark_alg/q02.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_XMark_alg/upd.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_R/Q13alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_R/Q13xalg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_R/Q18alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_XMP/Q04alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/item_sequence_creation.SF-1543871a.alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/item_sequence_creation.SF-1543871a.alg0.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/item_sequence_creation.SF-1543871b.alg0.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/constructor_with_computed_QName.SF-1548043.alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/path_steps_with_name_tests_on_transsient_nodes.SF-1610247.alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/add_qnames.SF-1611353.alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_value_comparison.SF-1576463.alg.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_value_comparison.SF-1576463.alg.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/nested_element_construction.SF-1712578_alg.out.00.html
## Comment 7894
Date: 2007-07-11 10:18:38 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
There exists already a bug report: [ 1741459 ]
## Comment 7895
Date: 2007-07-25 18:29:08 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Bug is covered by more specific/recent bug reports
[ 1760459 ] PF/alg: tests fails since 2007/07/11
http://sourceforge.net/tracker/index.php?func=detail&aid=1760459&group_id=56967&atid=482468
[ 1741459 ] PF: ws_create inconsistently adds a collection node
http://sourceforge.net/tracker/index.php?func=detail&aid=1741459&group_id=56967&atid=482468
## Comment 7896
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1751681 at http://sourceforge.net/support/tracker.php?aid=1751681
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| PF/alg: most algebra tests fail after checkins on 2007/06/27 | https://api.github.com/repos/MonetDB/MonetDB/issues/1404/comments | 0 | 2020-11-30T09:36:42Z | 2024-06-28T13:31:21Z | https://github.com/MonetDB/MonetDB/issues/1404 | 753,316,245 | 1,404 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-03 02:31:36 +0200
From: @peterboncz
To: @sjoerdmullender
Version: -- development
Last updated: 2009-11-12 10:55:53 +0100
## Comment 7891
Date: 2007-07-03 14:31:36 +0200
From: @peterboncz
on windows, the MonetDB dbfarm is located in:
c:\Documents and Settings\..user..\Application Data\MonetDB4
when MonetDB is deinstalled, this directory is not emptied or moved away. When a user installs a new MonetDB version, the net effect is that it starts using the repository left there by the old MonetDB!
This can cause all kinds of nasty (hard to understand/reproduce bugs) as the logical database format may change version after version.
it also causes confusion with 64-bits windows. A 64-bits windows server exits immediately with an error upon seeing the 32-bits repo. The CMD windows closes so quickly that this error is illegible. The user (me too) just thinks the 64-bits product does not work at all; ie that it crashed.
my suggested solution is to remove the old repositories at *deinstall* time. For me, removing would only entail recursively removing the contents of the dbfarm/*/bat/* directories. The deinstaller should tell the user that certain dbfarm subdirs (repositories) are going to be thrown away, and offer a Cancel button with the suggestion to start MonetDB/XQuery, use the admin console to backup his/her databases.
Please do not remove the dbfarm/* directories, because these backups are found in the dbfarm/*/xml/ subdir.
## Comment 7892
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1747068 at http://sourceforge.net/support/tracker.php?aid=1747068
| XQ: windows de-install should warn and empty bat/ dirs | https://api.github.com/repos/MonetDB/MonetDB/issues/1403/comments | 0 | 2020-11-30T09:36:39Z | 2024-06-28T07:36:58Z | https://github.com/MonetDB/MonetDB/issues/1403 | 753,316,217 | 1,403 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-03 12:13:15 +0200
From: @peterboncz
To: @sjoerdmullender
Version: -- development
Last updated: 2007-07-13 10:51:00 +0200
## Comment 7883
Date: 2007-07-03 12:13:15 +0200
From: @peterboncz
The 0.18 windows installer (32-bits) does not install the images of the admin console. This impacts the readbility of the admin pages.
suggested fix:
the windows installer should copy the .gif/.jpg/.png images in the
runtime/xrpc
directory should be copied to the
share/MonetDB/xrpc/export
directory.
I did not check an RPM Linux install; but there this may also have been forgotten.
(ps I noted that in the 18.0 binary, the new MapiClient help text is not yet displayed -- presumably something for 18.1?)
## Comment 7884
Date: 2007-07-03 12:19:52 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
I also noted that the XQuery installer also installs SQL. Fine with me, but maybe not intended.
## Comment 7885
Date: 2007-07-03 18:15:03 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Peter,
are you referring to / using the 0.18.0 installer or the 0.18.2 installer?
## Comment 7886
Date: 2007-07-03 19:10:45 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
Stefan,
Today (13:30) I just clicked download on the website, and expected getting the latest version.
That was version 18.0.
Now (19:00), I see that 18.2 is there on the page. It does have the images! It also has the welcome message!
Somebody changed the website downlad page, lately?
It does appear, though, that the version number was not upgraded. That is, the MonetDB4 MIL server says at startup still says:
Monet Database Server V4.18.0
which just recently added a bit of numbering confusion for me. But it appears to be 18.2 really.
Peter
## Comment 7887
Date: 2007-07-03 20:28:11 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Sjoerd produced the bugfix release on June 19 and uploaded it to SF.
I noticed only today (triggered by this bug report) that the nobody apparently had updated the website accordingly.
(Hence, I did it today.)
more details are probably in the CVS logs.
## Comment 7888
Date: 2007-07-03 23:45:41 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
... had to leave without finishing my reply. Sorry.
Here's the remainder:
The version numbers follow our new (agreed on) version numbering and release scheme:
In order to reduce release work, we agreed on releasing each part separately instead of (always) all parts together;
hence, version numbers may run "out of sync", as it appears here,
causing the confusion as anticipated during discussions before the above agreement.
Due to (significant?) bug-fixes in pathfinder right after (< 6 days!) the *.18.0 ("Mars") release,
a bug-fix release of Pathfinder (aka. MonetDB/XQuery) was apparently necessary, increasing its version number to 0.18.2,
while all other parts remained unmodified as MonetDB Common 1.18.0, MonetDB Server 4.18.0 / MonetDB Server 5.0.0, SQL 2.18.0, Clients 1.18.0.
A possible solution to the version number confusion could be that all parts (not only MonetDB4/5 Server, but also MonetDB Common (GDK), SQL, Pathfinder, Clients) clearly state their respective version number in the welcome message during server/client startup or when the respective module is loaded.
## Comment 7889
Date: 2007-07-13 10:51:00 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
The only "bug" seems to have been that our website still pointed to the old version.
If you want to have all relevant version numbers displayed, file a RFE.
## Comment 7890
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1747017 at http://sourceforge.net/support/tracker.php?aid=1747017
| XQ: admin console images not installed | https://api.github.com/repos/MonetDB/MonetDB/issues/1402/comments | 0 | 2020-11-30T09:36:37Z | 2024-06-27T11:25:36Z | https://github.com/MonetDB/MonetDB/issues/1402 | 753,316,178 | 1,402 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-01 05:08:05 +0200
From: @drstmane
To: @mlkersten
Version: -- development
Last updated: 2007-07-02 07:43:41 +0200
## Comment 7879
Date: 2007-07-01 17:08:05 +0200
From: @drstmane
With MonetDB5, test
MonetDB5/src/mal/Tests/tst670.mal
fails since Friday June 29 2007, i.e., after checkins on Thursday June 24 2007.
For details, see
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/G
NU.64.64.d-Fedora6/src_mal/tst654.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/G
NU.64.64.d-Fedora6/src_mal/tst655.out.00.htm
## Comment 7880
Date: 2007-07-01 17:09:36 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Correct URL is:
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/GNU.64.64.d-Fedora6/src_mal/tst670.out.00.html
## Comment 7881
Date: 2007-07-02 07:42:45 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This test aimed at handling a suicide call, i.e. a clients.quit()
in a nested MAL program, console or prelude file. It is solved
by catching an internal exception at the proper place to turn
it into a real action
## Comment 7882
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1746076 at http://sourceforge.net/support/tracker.php?aid=1746076
| M5: test tst670 fails since 2007.06.29 | https://api.github.com/repos/MonetDB/MonetDB/issues/1401/comments | 0 | 2020-11-30T09:36:34Z | 2024-06-27T11:25:35Z | https://github.com/MonetDB/MonetDB/issues/1401 | 753,316,150 | 1,401 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-07-01 05:05:44 +0200
From: @drstmane
To: @mlkersten
Version: -- development
Last updated: 2007-07-02 07:46:17 +0200
## Comment 7876
Date: 2007-07-01 17:05:44 +0200
From: @drstmane
With MonetDB5, tests
MonetDB5/src/mal/Tests/tst654.mal
MonetDB5/src/mal/Tests/tst655.mal
fail since Monday June 25 2007, i.e., after checkins on Sunday June 24 2007.
For details, see
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/GNU.64.64.d-Fedora6/src_mal/tst654.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/GNU.64.64.d-Fedora6/src_mal/tst655.out.00.html
## Comment 7877
Date: 2007-07-02 07:46:17 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Reverted the patch of the Column definition in Tablet.
This structure requires at least on Column definition
at all times. The structure is allocated with the
precise number of (inlined) Columns array.
## Comment 7878
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1746075 at http://sourceforge.net/support/tracker.php?aid=1746075
| M5: tests tst654 & tst655 fail since 2007.06.25 | https://api.github.com/repos/MonetDB/MonetDB/issues/1400/comments | 0 | 2020-11-30T09:36:32Z | 2024-06-27T11:25:34Z | https://github.com/MonetDB/MonetDB/issues/1400 | 753,316,116 | 1,400 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-28 06:34:25 +0200
From: Jeremy Volkman <<jvolkman>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2008-12-12 12:13:52 +0100
## Comment 7868
Date: 2007-06-28 06:34:25 +0200
From: Jeremy Volkman <<jvolkman>>
Attached is a sample XML document and a query that causes MonetDB to segfault at the following location:
PFll_descendant (result=0xafdf6b10, iter_bat=0x819c100, ctx_bat=0x8c5c5c0, pre_size=0x8877840, cand_bat=0x0, _one_iter=0xafdf6b60 "",
_one_ctx=0xafdf6b70 "\001", _min_iter=0xafdf6b80, _max_iter=0xafdf6b90, _no_iter_order=0xafdf6ba0 "\001", _kind_test=0xafdf6bb0 "\200")
at ll_staircasejoin.mx:299
299 *(oid*)dst = *(oid*)BUNtail(iter_bat,iter);
### Steps to reproduce:
1. Extract archive.
2. shred_doc(<path-to-test.xml>,"test.xml");
3. MapiClient -l x test.xq
## Comment 7869
Date: 2007-06-28 06:37:20 +0200
From: Jeremy Volkman <<jvolkman>>
Logged In: YES
user_id=1702210
Originator: YES
The host system is Ubuntu Linux 7.04 on an Intel Pentium M w/ 2GiB RAM
## Comment 7870
Date: 2007-07-03 12:03:01 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
I will have a look at this one
## Comment 7871
Date: 2007-07-03 12:06:19 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
renamed with XQ: convention; recategorized into Pathfinder
## Comment 7872
Date: 2007-07-12 11:15:33 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Jeremy,
the segfault is due to an intermediate result of 10000*100000 = 1G elements, i.e., 8GB/16GB (32/64 bit system), that apparently cannot be allocated/memory-mapped.
Well, indeed, we should not crash/segfault, but give a proper error message in that case --- investigating...
However, a closer look at your query reveals the reason for such huge intermediate result, and rises the question, whether the query is indeed hwat you intended:
<result>
{
count(collection("test.xml")//b[count(//c) > 2])
^^^^^^^^^^^^^^
}
</result>
Due to the asolute path (//c), "count(//c) > 2" is independant/invariant of the context!??
Unfortunately, Pathfinder does (can?) not (yet?) recognize such invariant and factor it our of the loop of context nodes.
Hence, normal bulk-processing in MonetDB/XQuery evaluates the //c (yielding 100000 elements) for all 10000 //b elements in bulk ("loop-lifted"), yielding an intermediate result of 10000 * 100000 = 1G elements as mentioned above.
In case you actually wanted to count all c elements below each b element, replacing count(//c) by count(.//c) works just fine:
<result>
{
count(collection("test.xml")//b[count(.//c) > 2])
^
}
</result>
Stefan
## Comment 7873
Date: 2007-07-12 13:55:03 +0200
From: Jeremy Volkman <<jvolkman>>
Logged In: YES
user_id=1702210
Originator: YES
Stefan,
Thanks for the follow up. The query I submitted was a result of me playing around with MonetDB/XQuery (which is blazingly fast given a valid query). It was indeed a mistake (I was going for ".//c" as you suggested), but as you said an invalid/inexecutable query should not crash the server. I suppose I should have pointed out the insanity of the query with the report -- sorry about that. :)
-Jeremy
## Comment 7874
Date: 2008-01-22 15:36:06 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
The segfault occurs when filling a (exceptionally huge) intermediate-result-BAT, that has been memory-mapped with holes larged than disk capacity, reaces/exceed disk capacity.
The only(?) way to prevent this, would be to restrict memory-mapped files to available disk capacity (which of course might change between creating the memory-mapped file and filling it ...)
## Comment 7875
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1744590 at http://sourceforge.net/support/tracker.php?aid=1744590
| XQ: segfault in descendant step | https://api.github.com/repos/MonetDB/MonetDB/issues/1399/comments | 0 | 2020-11-30T09:36:29Z | 2024-06-28T07:36:58Z | https://github.com/MonetDB/MonetDB/issues/1399 | 753,316,088 | 1,399 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-26 12:24:16 +0200
From: Joris van Rantwijk <<jvrantwijk>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2007-07-03 10:02:21 +0200
## Comment 7863
Date: 2007-06-26 12:24:16 +0200
From: Joris van Rantwijk <<jvrantwijk>>
If the result of an XQuery is an attribute node, the default serializer in Pathfinder returns something which is almost, but not quite, an XQuery expression. I think that quoting and escaping the attribute value would make the result more useful.
For example, the following query:
<aap q="noot " mies" />/@q
returns the following result:
attribute q { noot " mies }
while the following might be better:
attribute q { "noot " mies" }
## Comment 7864
Date: 2007-07-03 12:01:15 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Jan, maybe have a look at this one?
## Comment 7865
Date: 2007-07-03 12:07:00 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
renamed bug with XQ: in title
## Comment 7866
Date: 2007-07-03 22:02:21 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
I aligned the serialization of attributes in mode DM with Joris' suggestion. (I am however a little bit puzzled that mode DM is the default mode.)
Added test BugTracker/attribute_serialization.SF-1743442
## Comment 7867
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1743442 at http://sourceforge.net/support/tracker.php?aid=1743442
| XQ: illegal XML syntax in serialized attributes | https://api.github.com/repos/MonetDB/MonetDB/issues/1398/comments | 0 | 2020-11-30T09:36:27Z | 2024-06-27T11:25:32Z | https://github.com/MonetDB/MonetDB/issues/1398 | 753,316,063 | 1,398 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-26 11:55:56 +0200
From: Joris van Rantwijk <<jvrantwijk>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-08-03 10:43:47 +0200
## Comment 7850
Date: 2007-06-26 11:55:56 +0200
From: Joris van Rantwijk <<jvrantwijk>>
XML attribute values containing an ampersand give unexpected query results. In the query response, the ampersand is apparently escaped twice into &
Example document aap.xml:
<aap>
This is ok: &
But this is bad: <noot q="&" />
</aap>
The XQuery doc("aap.xml") now returns:
<aap>
This is ok: &
But this is bad: <noot q="&"/>
</aap>
Note that the problem only occurs with shredded documents, not with nodes embedded in the XQuery. This suggests that the bug is somewhere in the shredder.
I enabled the debug statements in shred_start_element in shredder.c. This showed that val at that point already contains "&" (I was expecting "&").
## Comment 7851
Date: 2007-07-03 11:59:36 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
shredder, entities.. so assigning to Jan
## Comment 7852
Date: 2007-07-03 12:07:34 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
renamed bug with XQ: in title
## Comment 7853
Date: 2007-07-03 14:05:36 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
This seems to be a known bug in libxml software. I googled on this and found for instance this discussion in a debian forum:
http://bugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=397395
I cannot do anything to fix it in the shredder. We just have to wait for the fix in libxml2.
## Comment 7854
Date: 2007-07-03 17:12:27 +0200
From: Joris van Rantwijk <<jvrantwijk>>
Logged In: YES
user_id=1480851
Originator: YES
I agree that libxml is throroughly messed up in this respect. But I'm not convinced that the discussion you refer to is about exactly the same issue. I found an option in libxml2 which seems to make it behave better.
Patch attached, it fixes the ampersand issue and apparently does not cause regression on test cases.
File Added: pathfinder_ampersand_fix.diff
## Comment 7855
Date: 2007-07-04 10:23:50 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
I have been trying the XML_PARSE_NOENT option before when using user defined entities. But it caused very strange behaviour. It doubled the output for user defined entities. Try this old entity bug (there was regression on entities_dtd.SF-1642663 and entities_dtd.SF-1642665:-) example with the XML_PARSE_NOENT option:
<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE foo [
<!ELEMENT foo (bar)>
<!ELEMENT bar (PCDATA)>
<!ENTITY eacute "é" ><!-- small e, acute accent -->
]>
<foo>
<bar>want to see a nice e with a acute accent:é</bar>
</foo>
You will see that is emits TWO eacutes here.
I am still convinced there is no easy way to solve this problem with the shredder. Indeed was the example not exactly your problem but it was very closely related. It also only happens with the & entity. All other entities are handled correct. Maybe we should post a libxml2 bug report for this (how?)
## Comment 7856
Date: 2007-07-06 18:24:07 +0200
From: Joris van Rantwijk <<jvrantwijk>>
Logged In: YES
user_id=1480851
Originator: YES
Hello Jan,
I found a relevant libxml2 bug report: http://bugzilla.gnome.org/show_bug.cgi?id=172638
The libxml2 people seem to believe that & is a perfectly reasonable thing to do unless you explicitly set XML_PARSE_NOENT. I disagree, but I think it unlikely that they would change this.
I tried your é case, and in fact it seems that this libxml2 bug is now fixed.
When I run it against libxml2 2.6.20 it gives me two eacutes, but when I run it against libxml2 2.6.27 it works as expected.
Joris.
## Comment 7857
Date: 2007-07-26 10:39:08 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
The problem appears to be the same with dev-libs/libxml2-2.6.29
## Comment 7858
Date: 2007-07-26 11:00:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Anybody interested in / volunteering to add(ing) a proper test for this one to CVS?
Our nightly testing pool offers several version of libxml2 (2.6.10, 2.6.16, 2.6.27, 2.6.28, 2.6.29), and coudl hence provide a simple check, whether this bug indeed depeds on the libxml2 version, respectively in with version it might be fixed ...
## Comment 7859
Date: 2007-07-26 12:18:52 +0200
From: Joris van Rantwijk <<jvrantwijk>>
Logged In: YES
user_id=1480851
Originator: YES
Please note that the reported bug does not (as far as I can tell) depend on the libxml2 version.
It is rather the proposed resolution (XML_PARSE_NOENT) which is known to break an existing testcase (entities_dtd.SF-1642663 according to Jan Flokstra) on some versions of libxml2.
## Comment 7860
Date: 2007-08-02 14:14:27 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
It seems that attribute values are returned as character entities, so I added code to translate those to UTF-8. I also added a test, so now we will see whether this fix fixes the problem on all platforms.
The test is /pathfinder/tests/BugTracker/Tests/attr-entity-bug.SF-1743433.*
I'll leave this open until after tonight's testing.
## Comment 7861
Date: 2007-08-03 10:43:46 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Last night's testing showed the bug is fixed.
## Comment 7862
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1743433 at http://sourceforge.net/support/tracker.php?aid=1743433
| XQ: problem with ampersand in attribute value | https://api.github.com/repos/MonetDB/MonetDB/issues/1397/comments | 0 | 2020-11-30T09:36:24Z | 2024-06-27T11:25:31Z | https://github.com/MonetDB/MonetDB/issues/1397 | 753,316,037 | 1,397 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-25 09:54:21 +0200
From: @drstmane
To: @njnes
Version: -- development
Last updated: 2007-07-04 10:46:04 +0200
## Comment 7844
Date: 2007-06-25 09:54:21 +0200
From: @drstmane
Since Monday June 4 2007, i.e., after checkins on Saturday June 2 2007 and/or Sunday une 3 2007 (due to system maintenance at CWI/SARA there was no testing on Sunday une 3 2007),
JDBC tests Test_JdbcClient & Test_Sbatching fail with both SQL/4 & SQL/5 on
CentOS4.4 G.32.32.d
RedHat4WS G.64.64.d
SUSE10.2 G.32.32.d
Test_JdbcClient produces extra (unexpected?) output
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.32.32.d-CentOS4.4/src_jdbc_tests/Test_JdbcClient.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d-RedHat4WS/src_jdbc_tests/Test_JdbcClient.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.32.32.d-SUSE10.2/src_jdbc_tests/Test_JdbcClient.out.00.html
Test_Sbatching runs into a timeout after 60 seconds (used to finish without problems within well below 10 seconds before)
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.32.32.d-CentOS4.4/src_jdbc_tests/Test_Sbatching.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d-RedHat4WS/src_jdbc_tests/Test_Sbatching.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.32.32.d-SUSE10.2/src_jdbc_tests/Test_Sbatching.err.00.html
## Comment 7845
Date: 2007-06-25 15:14:16 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Some usefull comment from Fabian:
"
I cannot login, so I'll do it this way:
the JdbcClient test fails, because the Sbatching test doesn't clean up
it's mess (because it times out).
"
## Comment 7846
Date: 2007-07-03 11:30:15 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I cannot find any checkins related to JDBC on Saturday June 2 2007, only some M5 checkins that seemingly look harmless.
On Sunday june 3 2007 release was prepared/done, so lots of propagation and version number commits that don't feel like they can be the cause of action.
Unfortunately I don't have access to the CentOS or SuSE machines that currently still seem to timeout. It might be related to the Mtest change to use hostname instead of localhost again, which routes the traffic over the interface, and hence aggressive firewalls and real packet transportation may slow down the actual speed.
## Comment 7847
Date: 2007-07-03 11:59:39 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Indeed, the actual problem seems to be a significant (up to factor 10!) slow-down of test Test_Sbatching:
========
C4.4 D8.8.0 D8.9.0 F6 F6 F6 F6 F6 F6 G1.12.9 R4 S10.2 S5.10
G32/32d G64/64d G32/32d G32/32d G64/32d G64/64d G64/64s n64/32d n64/64d G32/32d G64/64d G32/32d G32/32d
Mtest10..2007.06.01_04-31-24..time
Test_Sbatching 10.440 8.936 1.992 1.450 1.558 1.385 1.476 1.472 6.225 17.578 0.465 2.557 Test_Sbatching
Mtest10..2007.06.02_04-21-32..time
Test_Sbatching 10.603 4.523 1.973 1.457 1.977 1.650 1.672 1.529 6.238 17.847 0.476 2.283 Test_Sbatching
Mtest10..2007.06.04_00-58-07..time
Test_Sbatching 18.343 9.232 5.988 6.111 19.052 60,387 60,255 22.413 Test_Sbatching
Mtest10..2007.06.05_00-09-07..time
Test_Sbatching 60,485 23.277 8.759 6.042 5.987 5.920 6.058 5.767 20.805 60,395 60,199 22.641 Test_Sbatching
Mtest10..2007.06.06_00-09-08..time
Test_Sbatching 60,465 19.886 8.811 6.502 6.040 5.965 6.047 5.898 13.847 60,381 60,405 22.503 Test_Sbatching
G32/32d G32/32d G32/32d G64/32d G64/64d G64/64s n64/32d n64/64d G32/32d G64/64d G32/32d G32/32d
C4.4 D8.9.0 F6 F6 F6 F6 F6 F6 G1.12.9 R4 S10.2 S5.10
========
I extended the timeout to 2 mins --- let's see tomorrow, whether this helps.
## Comment 7848
Date: 2007-07-04 10:46:04 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed by Niels' yesterday's changes.
Thanks!
## Comment 7849
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1742695 at http://sourceforge.net/support/tracker.php?aid=1742695
| SQL: some JDBC tests fail since 2007/06/04 on some platforms | https://api.github.com/repos/MonetDB/MonetDB/issues/1396/comments | 0 | 2020-11-30T09:36:21Z | 2024-06-27T11:25:30Z | https://github.com/MonetDB/MonetDB/issues/1396 | 753,316,000 | 1,396 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-23 07:35:35 +0200
From: @drstmane
To: @njnes
Version: -- development
Last updated: 2007-07-03 07:49:09 +0200
## Comment 7840
Date: 2007-06-23 19:35:35 +0200
From: @drstmane
Various SQL tests (see below) produce different (wrong?) results with SQL/5 since Friday June 22 2007, i.e., after changes on Thursday June 21 2007; with SQL/4, these tests still work fine.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_benchmarks_tpch/02.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_benchmarks_tpch/04.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_benchmarks_tpch/21.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_benchmarks_tpch/22.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_benchmarks_tpch/01-22.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_test/zones.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_test_BugDay_2005-10-06_2.9.3/having.SF-922614.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_test_BugDay_2005-10-06_2.9.3/custom_function_in_where.SF-937859.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_test_BugTracker/crash_on_delete.SF-1639108.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.32.32.d-Gentoo1.12.9/src_test_bugs/groupby_having-bug-sf-947600.out.00.html
## Comment 7841
Date: 2007-06-23 19:37:23 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The "problematic" platforms are:
Gentoo1.12.9 G.32.32.d
SUSE10.2 G.32.32.d
SunOS5.10 G.32.32.d
## Comment 7842
Date: 2007-07-03 07:49:09 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed by Niels' yesterday's checkin.
Thanks!
## Comment 7843
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1742121 at http://sourceforge.net/support/tracker.php?aid=1742121
| SQL/5: tests fail since 2007/06/22 on some platforms | https://api.github.com/repos/MonetDB/MonetDB/issues/1395/comments | 0 | 2020-11-30T09:36:18Z | 2024-06-27T11:25:29Z | https://github.com/MonetDB/MonetDB/issues/1395 | 753,315,967 | 1,395 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-22 02:36:24 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-08-12 07:11:54 +0200
## Comment 7833
Date: 2007-06-22 14:36:24 +0200
From: Jan Rittinger <<jan.rittinger>>
In pathfinder.mx PROC ws_create a bogus COLLECTION node is added.
This is however only done for a subset of PRE_* columns (PRE_CONT and NID_RID are ignored). If the first change to the working set container adds its nodes with append() the pre numbers are not aligned anymore.
I propose to also add a dummy value for the ignored BATs. Then it shouldn't matter whether insert() or append() is used. As I have no idea how collections are represented I leave the solution to somebody else.
## Comment 7834
Date: 2007-06-29 18:11:54 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: YES
I think all failing algebra queries relate to this bug!
## Comment 7835
Date: 2007-07-03 12:02:45 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
the bogus collection is a super-root node for all temporarily constructed XML fragments
thanks for noting -- will fix
## Comment 7836
Date: 2007-07-06 18:17:43 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: YES
Peter, does the following patch do the correct thing? - Or is something else needed?
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/runtime/pathfinder.mx,v
retrieving revision 1.370
diff -r1.370 pathfinder.mx
1366a1367,1368
> ws.fetch(PRE_CONT).fetch(WS).append(oid_nil);
> ws.fetch(NID_RID).fetch(WS).append(0@0);
## Comment 7837
Date: 2007-07-17 10:36:44 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
JanR's suggested patch seems to fix most algebra tests
without causing any (obvious) harm.
However, some algebra tests still fail with
"fatal error: attribute `item' referenced in projection not found"
"halted in .../pathfinder/compiler/algebra/logical.c (PFla_project_), line 867"
in the XQuery compiler.
## Comment 7838
Date: 2007-08-12 19:11:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
I consider the originally reported bug fixed by JanR's patch.
No extra test required as the problem is covered by the existing algebra tests.
The remaining algebra failures are covered by
[ 1760459 ] PF/alg: tests fails since 2007/07/11
http://sourceforge.net/tracker/index.php?func=detail&aid=1760459&group_id=56967&atid=482468
Peter, Jan,
please correct me, in case I'm wrong.
## Comment 7839
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1741459 at http://sourceforge.net/support/tracker.php?aid=1741459
| PF: ws_create inconsistently adds a collection node | https://api.github.com/repos/MonetDB/MonetDB/issues/1394/comments | 0 | 2020-11-30T09:36:16Z | 2024-06-27T11:25:28Z | https://github.com/MonetDB/MonetDB/issues/1394 | 753,315,930 | 1,394 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-22 11:40:33 +0200
From: @drstmane
To: @njnes
Version: -- development
Last updated: 2007-07-01 04:58:47 +0200
## Comment 7829
Date: 2007-06-22 11:40:33 +0200
From: @drstmane
Since Tuesday Juen 12 2007, i.e., after checkins on Monday June 11 2007, test MonetDB5/src/optimizer/Tests/remote00.py fails with a timeout after the default 60 seconds --- before, it used to successfully finish within less than 5 seconds.
cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/GNU.64.64.d-Fedora6/src_optimizer/remote00.err.00.html
## Comment 7830
Date: 2007-06-22 12:37:15 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This is collateral damage after trimming the client exit code.
An area that requires attention and a clean(er) solution.
## Comment 7831
Date: 2007-07-01 16:58:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Apparently, this tests seems to work, again, since Friday June 29 2007, i.e., after checkins on Thursday June 28 2007.
Closing.
## Comment 7832
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1741387 at http://sourceforge.net/support/tracker.php?aid=1741387
| M5: test remote00.py times out since 2007/06/12 | https://api.github.com/repos/MonetDB/MonetDB/issues/1393/comments | 0 | 2020-11-30T09:36:13Z | 2024-06-27T11:25:27Z | https://github.com/MonetDB/MonetDB/issues/1393 | 753,315,883 | 1,393 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-20 07:59:20 +0200
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2007-07-20 03:11:03 +0200
## Comment 7825
Date: 2007-06-20 19:59:20 +0200
From: @drstmane
Since Saturday June 16 2007, i.e., after checkins on Friday June 15 2007,
pathfinder tests
tests/BugTracker/Tests/wrong_result_duplication.SF-1549292.xq
and
tests/XQuery/Tests/transient_upward_steps.xq
fail with
ERROR = !ERROR: no matching MIL operator to 'leftfetchjoin(BAT[oid,void], BAT[oid,oid])'.
and
ERROR = !ERROR: no matching MIL operator to 'insert(BAT[oid,str], void, str)'.
respectively; cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/wrong_result_duplication.SF-1549292.err.00.html
and
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_XQuery/transient_upward_steps.err.00.html
## Comment 7826
Date: 2007-07-18 21:19:44 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The problem seems to be triggered by this checkin:
========
$ cvs log -N -r1.23.2.3 compiler/include/mil_opt.h
RCS file: /cvsroot/monetdb/pathfinder/compiler/include/mil_opt.h,v
Working file: compiler/include/mil_opt.h
head: 1.25
branch:
locks: strict
access list:
keyword substitution: kv
total revisions: 42; selected revisions: 1
description:
----------------------------
revision 1.23.2.3
date: 2007/06/15 12:45:48; author: boncz; state: Exp; lines: +3 -3
- now that I think of it, reduce by a further factor 16 (256KB)
=============================================================================
$ cvs diff -r1.23.2.{2,3} compiler/include/mil_opt.h
Index: compiler/include/mil_opt.h
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/compiler/include/mil_opt.h,v
retrieving revision 1.23.2.2
retrieving revision 1.23.2.3
diff -u -r1.23.2.2 -r1.23.2.3
--- compiler/include/mil_opt.h 15 Jun 2007 12:35:24 -0000 1.23.2.2
+++ compiler/include/mil_opt.h 15 Jun 2007 12:45:48 -0000 1.23.2.3
@@ -24,7 +24,7 @@
* 2000-2005 University of Konstanz and (C) 2005-2007 Technische
* Universitaet Muenchen, respectively. All Rights Reserved.
*
- * $Id: mil_opt.h,v 1.23.2.2 2007/06/15 12:35:24 boncz Exp $
+ * $Id: mil_opt.h,v 1.23.2.3 2007/06/15 12:45:48 boncz Exp $
*/
include <stdio.h>
@@ -45,9 +45,9 @@
*/
define OPT_STMTS 32767 /* <65535 if first-use further out than this amount of stamements, dead codes survises anyhow */
-define OPT_VARS 1024 /* don't try dead code elimintation above this amount of live variables */
+define OPT_VARS 256 /* don't try dead code elimintation above this amount of live variables */
define OPT_REFS 63 /* keep track of usage dependencies; may omit some, which results in surviving dead code */
-define OPT_CONDS 1024 /* maximum if-nesting */
+define OPT_CONDS 256 /* maximum if-nesting */
define OPT_SEC_PROLOGUE 0
define OPT_SEC_QUERY 1
========
re-increasing OPT_VARS to (e.g.) 512 make the tests work correctly, again;
hence, I guess there is a bug in the live-variable-limit-handling in the
MIL code optimizer.
Unfortunately I have not (yet?) managed to locate the actual problem...
## Comment 7827
Date: 2007-07-20 15:11:03 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in CVS. There was indeed a problem when there were too many variables.
No new test needed.
## Comment 7828
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1740502 at http://sourceforge.net/support/tracker.php?aid=1740502
| PF: live-variable-limit-handling in mil_opt.c buggy | https://api.github.com/repos/MonetDB/MonetDB/issues/1392/comments | 0 | 2020-11-30T09:36:10Z | 2024-06-27T11:25:26Z | https://github.com/MonetDB/MonetDB/issues/1392 | 753,315,854 | 1,392 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-18 10:13:29 +0200
From: bendeguz <<bendeguz>>
To: @njnes
Version: -- development
Last updated: 2007-10-03 09:38:25 +0200
## Comment 7816
Date: 2007-06-18 22:13:29 +0200
From: bendeguz <<bendeguz>>
Dear MonetDB Developers,
After my EXPLAIN bug report I have checked the TRACE, PROFILE,DEBUG availability for the JDBC client. The result is the same.
### Steps to reproduce:
1) Connect the database with mjclient
2) Issue the following queries:
debug select count(*) from tables;
trace select count(*) from tables;
profile select count(*) from tables;
Example scanario:
monetdb-> debug select count(*) from tables;
Error: protocol violation, unexpected line: mdb> mdb.start()
monetdb-> trace select count(*) from tables;
Error: protocol violation, unexpected line: breakpoint on select count(*) from tables; not set
monetdb-> profile select count(*) from tables;
Connection warning: mdb Symbol not found
monetdb->
Regards,
B.
## Comment 7817
Date: 2007-06-18 22:31:25 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Thanks for remarking this.
Please use the C-version instead, i.e. mclient
More details will follow.
## Comment 7818
Date: 2007-07-16 20:05:13 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
details:
mdb is an interactive debugger that should never be started for an application like connection such as a JDBC connection. There is no way how JDBC can deal with an interactive debugger prompt.
The interactive commands you described hence need to be disabled for JDBC connections.
## Comment 7819
Date: 2007-09-06 08:26:56 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This limitation has been added to the documentation.
The role of mjclient has been reduced.
The mjclient code could also trap SQL statements
starting with these keywords and issue a warning.
## Comment 7820
Date: 2007-09-06 09:35:22 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
nono. This is about JDBC and ODBC, which are application interfaces. The SQL server should make it technically IMPOSSIBLE to allow interactive stuff for such connections. IOW, only mclient should be allowed to access it.
## Comment 7821
Date: 2007-09-06 09:42:52 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
It depends on SQL to be able to detect the client protocol.
The documentation has been updated to reflect the limitation.
## Comment 7822
Date: 2007-09-06 09:46:59 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
A (business) application should never get stuck because the server suddenly expects it to be interactive, while it uses S mode, which explicitly states that the application is NOT interactive. (Hence also no MORE prompt.)
## Comment 7823
Date: 2007-10-03 09:38:25 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Added a test in src/test/BugTracker/jdbc_no_debug.SF-1739356
The parser currently returns a correct error message as it cannot support these modes in line mode.
## Comment 7824
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1739356 at http://sourceforge.net/support/tracker.php?aid=1739356
| TRACE,PROFILE,DEBUG should be impossible for JDBC and ODBC | https://api.github.com/repos/MonetDB/MonetDB/issues/1391/comments | 0 | 2020-11-30T09:36:07Z | 2024-06-27T11:25:25Z | https://github.com/MonetDB/MonetDB/issues/1391 | 753,315,832 | 1,391 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-18 10:06:27 +0200
From: bendeguz <<bendeguz>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2009-11-16 11:07:51 +0100
## Comment 7804
Date: 2007-06-18 22:06:27 +0200
From: bendeguz <<bendeguz>>
Dear MonetDB Developers,
The EXPLAIN query statement does not work from the JDBC client. The same query works from the command line client (mclient)
### Steps to reproduce:
1) Connect the database from a jdbc client (squirrelsql or mjclient)
2) issue the following query:
explain select count(*) from tables;
Check the result and the error message:
a) Squirrelmail:
Error: java.sql.SQLException: protocol violation, unexpected line: function user.s11_1():void;, SQL State:
null, Error Code: 0
b) mjclient:
Error: protocol violation, unexpected line: function user.s4_3():void;
Regards,
B.
## Comment 7805
Date: 2007-07-16 17:48:30 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
Fixed in CVS, thanks!
You need to update MonetDB5, clients and sql packages for this to work.
## Comment 7806
Date: 2007-07-17 10:44:00 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Shouldn't we add a test for this one?
A simple
"explain select 1;"
via JDBC might be enough ...
## Comment 7807
Date: 2007-07-17 10:54:55 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
yes, but it depends a bit on if EXPLAIN will be continued to be supported.
## Comment 7808
Date: 2007-07-17 10:58:00 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Well, once EXPLAIN-support via JDBC is discontinued, the correct expected output for such test will be a proper error message ;-)
## Comment 7809
Date: 2007-07-17 13:13:32 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
TEST ADDED - meltdown
sql/src/test/BugTracker/Tests/explain.SF-1739353
## Comment 7810
Date: 2009-08-13 14:28:25 +0200
From: @drstmane
re-opened as the test now fails with
Error on line 1: protocol violation, unexpected line: function user.s1_1{autoCommit=true}():void;
In case this is indeed the expected behavior, i.e., JDBC does not / no longer support EXPLAIN, then we should consider approving this error message in stable.err (and the empty stable.out).
## Comment 7811
Date: 2009-08-13 14:30:12 +0200
From: @drstmane
see
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_BugTracker/explain.SF-1739353.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_BugTracker/explain.SF-1739353.err.00.html
## Comment 7812
Date: 2009-08-14 09:35:50 +0200
From: @grobian
JDBC by itself does not understand any EXPLAIN stuff. What happens, is that the server sends the MAL code without any leading marker (!, , [, ^, %) which upsets the MCL driver, which is very strict. IMO, EXPLAIN output should be a tuple result, e.g. all rows should be written as tuples ([ "blabla;" ]), or if that's easier, as XQuery rows (= blabla;), but a header should always be there.
## Comment 7813
Date: 2009-11-07 15:34:50 +0100
From: @drstmane
Now that JDBC does (seem to) understand EXPLAIN (and PLAN, etc.),
I added the stable output.
Once testing confirms that it works fine on all platform,
we can close this one.
## Comment 7814
Date: 2009-11-16 23:07:51 +0100
From: @drstmane
Tests work fine with multi-threaded server;
cf., e.g.,
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_BugTracker/explain.SF-1739353.out.00.html
minor(?) output differences with single-threaded server are to be checked/documented/fixed elsewhere;
cf., e.g.,
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/Int.64.64.d.1-Windows6.0/src_test_BugTracker/explain.SF-1739353.out.00.html
## Comment 7815
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1739353 at http://sourceforge.net/support/tracker.php?aid=1739353
| EXPLAIN support from JDBC | https://api.github.com/repos/MonetDB/MonetDB/issues/1390/comments | 0 | 2020-11-30T09:36:05Z | 2024-06-27T11:25:24Z | https://github.com/MonetDB/MonetDB/issues/1390 | 753,315,806 | 1,390 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-15 01:38:20 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-06-21 10:51:17 +0200
## Comment 7800
Date: 2007-06-15 13:38:20 +0200
From: @romulogoncalves
Dear developers,
I run this query:
select objID1 from "Match" where objID1 not in (select objID1 from PhotoObjAll);
against the skyserver data:
table "Match" has:
[ 25836 ] rows
and
table PhotoObjAll has:
[ 200276 ]
The mserver crashed (SQL-5 stable):
[New Thread 46912513495968 (LWP 29046)]
MonetDB Server v5.0.0
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs dynamically linked
dbname:skyserver
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /ufs/goncalve/scratch/MonetDB/Stable/MonetDB5/etc/monetdb5.conf)
[New Thread 1082132800 (LWP 29049)]
[New Thread 1090525504 (LWP 29055)]
[Thread 1090525504 (LWP 29055) exited]
>[New Thread 1090525504 (LWP 29059)]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1090525504 (LWP 29059)]
0x00002aaabbdc950e in sql_select (sql=0x6e42a78, scp=0x6d38a08, sn=0x6e452c8, ek={type = 0 '\0', card = 3 '\003', reduce = 0 '\0'})
at /ufs/goncalve/MonetDB/Stable/sql/src/server/sql_select.mx:4032
4032 list_append(rl, stmt_join(stmt_dup(grp->ext), foundsubset, cmp_equal));
(gdb)
Regards,
Romulo
## Comment 7801
Date: 2007-06-15 13:54:22 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
Add a proper test:
just run the example.sql in attach.
Regards,
Romulo
File Added: example.sql
## Comment 7802
Date: 2007-06-21 10:51:17 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug can be closed. It was fixed by Niels checkin at 10:07.
The bug is covered by:
sql/src/test/BugTracker/select_id_not_in_another_select.SF-1737777.sql
Regards,
Romulo
## Comment 7803
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1737777 at http://sourceforge.net/support/tracker.php?aid=1737777
| select id not in another select... | https://api.github.com/repos/MonetDB/MonetDB/issues/1389/comments | 0 | 2020-11-30T09:36:02Z | 2024-06-27T11:25:23Z | https://github.com/MonetDB/MonetDB/issues/1389 | 753,315,776 | 1,389 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-15 11:55:21 +0200
From: @peterboncz
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-06-15 04:03:18 +0200
## Comment 7787
Date: 2007-06-15 11:55:21 +0200
From: @peterboncz
I noticed that on a windows laptop, PF compilation times are quite high now (300 msec).
The following images show:
- before-query: MXQ just started, taking 30MB RAM
- after-query: after running the query '1'
- after-42: restarted query, now after running the query '42'
- after-mil42: after running a MIL query print(42);
The query '42' is a latency test in MXQ, it does not invoke the pathfinder compiler, nor executes MIL, it just sends back '42'. The difference in 'PEAK MEMORY CONSUMPTION' between the queries '1' and '42' is striking, in case of '42' it does not changes (30MB), in case of '1' it increases to 168MB.
The difference between the two is PF comilation and MIL execution. To show the cuplrit is not MIL execution, the after_mil42 experiment shows again only 31MB of peak usage with a mil query.
Is the PF embedded memory allocation allocating in 128MB chunks? In case of windows, that apparently hurts quite a bit. I tried changing gdk_mem_bigsize, but that does not influence the problem.
Peter
## Comment 7788
Date: 2007-06-15 11:55:52 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
File Added: after_query.JPG
## Comment 7789
Date: 2007-06-15 11:56:28 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
File Added: after_42.JPG
## Comment 7790
Date: 2007-06-15 11:57:14 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
File Added: after_mil42.JPG
## Comment 7791
Date: 2007-06-15 12:17:56 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
Peter, I don't understand what you mean with PF compilation times are quite high *now*.
As far as I know nothing has changed in the memory allocation for embedded evaluation. It still allocates 4MB chunks (define SA_BLOCK (4*1024*1024)). If you want to debug the query try 'echo 1 | pf -T' and check where most of the time is used (assuming that libgc was not available on windows during built and thus Niels old memory allocation scheme -- the same as for embedded -- is used).
## Comment 7792
Date: 2007-06-15 12:56:59 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Since pf is not included in the installer (I'll add it), I did the experiment on my VMware instance. Here are the timing results:
C:\Program Files\CWI\MonetDB4>echo 1 | pf.bat -T > pfout.txt
parsing: 000ms
module import: 000ms
normalization: 000ms
semantical analysis: 016ms
XML Schema import: 000ms
core mapping: 000ms
core simplification: 000ms
type checking: 000ms
core tree optimization: 000ms
MIL code output: 05s 594ms
## Comment 7793
Date: 2007-06-15 13:32:53 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
I checked the behaviour on the Linux version and met a single malloc (in addition to the first one over 4MB). This malloc is issued in mil_opt.c and requests 139 MB!
As you know mil_opt.c best, Peter, please have a look yourself.
Here is the gdb output:
(gdb) bt
0 0xffffe410 in __kernel_vsyscall ()
1 0xb75697d0 in raise () from /lib/libc.so.6
2 0xb756aea3 in abort () from /lib/libc.so.6
3 0xb756301b in __assert_fail () from /lib/libc.so.6
4 0xb227a2e1 in mem_alloc (pa=0x849b9c0, sz=139846592) at /local_tmp/work2/pathfinder/compiler/mem/mem.c:108
5 0xb227a339 in PFmalloc_ (n=139846584, file=0xb234c678 "/local_tmp/work2/pathfinder/compiler/mil/mil_opt.c", func=0xb234c6c2 "opt_open", line=875)
at /local_tmp/work2/pathfinder/compiler/mem/mem.c:151
6 0xb22452ff in opt_open (optimize=1) at /local_tmp/work2/pathfinder/compiler/mil/mil_opt.c:875
7 0xb22604e0 in PFprintMILtemp (c=0xaf8c8b48, optimize=1, module_base=0, num_fun=0, timing=681457785, prologue=0xb004f3e4, query=0xb004f3e0, epilogue=0xb004f3dc,
url=0x8a25831 "1\n", standoff=false) at /local_tmp/work2/pathfinder/compiler/mil/milprint_summer.c:11582
8 0xb2185c6c in PFcompile_MonetDB (xquery=0x8a25831 "1\n", url=0x8a25831 "1\n", prologue=0xb005032c, query=0xb004f464, epilogue=0xb0050324, options=0)
at /local_tmp/work2/pathfinder/compiler/compile.c:792
9 0xb217fe36 in xquery_compile_exec (ctx=0x804e120, url=0x8a25831 "1\n", is_url=0, prologue=0xb005032c, query=0xb0050328, epilogue=0xb0050324, nsurl=0x0)
at /local_tmp/work2/pathfinder/runtime/pathfinder.mx:4270
10 0xb2183130 in xquery_prepare (ctx=0x804e120, usec=6838848, query=0x8a25831 "1\n") at /local_tmp/work2/pathfinder/runtime/pathfinder.mx:5611
11 0xb2183f4d in xquery_client_engine (mc=0xb28a4c18) at /local_tmp/work2/pathfinder/runtime/pathfinder.mx:5753
12 0xb28a150d in mapi_client_engine (FC=0xb28a4c18) at /local_tmp/work2/MonetDB4/src/mapi/mapi.mx:501
13 0xb77fe34b in start_thread () from /lib/libpthread.so.0
14 0xb75fe65e in clone () from /lib/libc.so.6
(gdb)
## Comment 7794
Date: 2007-06-15 13:50:51 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
In mil_opt.h we have an interesting comment:
typedef struct {...} opt_t; /* should take ~1.5MB of RAM resources */
But when something of the size of opt_t is allocated, it takes 139 MB!
That's the alloc in mil_opt.c that Jan talks about.
## Comment 7795
Date: 2007-06-15 14:25:30 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Peter's fix for the bug "[ 1730617 ] XQ: 32 docs" involved, among other changes, the increase of the value of OPT_CONDS in mil_opt.h from 32 to 32768. It is this change that caused the huge increase in memory consumption.
The struct opt_var_t has 2 arrays of char of size 2*OPT_CONDS. There are 1024 (OPT_VARS) of such structs in opt_t. A quick sum learns that these two arrays alone account for 134217728 bytes, whereas before the size used was 131072 bytes.
Was this huge increase really necessary? Can we do with a more modest increase?
## Comment 7796
Date: 2007-06-15 14:31:42 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Sorry, it is 32767, not 32768. That doesn't change the argument, though.
## Comment 7797
Date: 2007-06-15 14:43:48 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
I changed OPT_CONDS from 32767 to 1024, and the time went down significantly. The MIL code output phase now consistently scores below 300 ms (after a few runs to get VMware to swap in the virtual machine).
Is there really a reason to have such a huge value for OPT_CONDS? I understand it is the number of nested conditionals you can have in the code? A program with even close to that many, even generated code, deserves no better than not getting optimized.
## Comment 7798
Date: 2007-06-15 16:03:18 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
With Peter's checkin to lower the value of OPT_CONDS and OPT_VARS both to 256, the timing is now a bit better:
C:\Program Files\CWI\MonetDB4>echo 1 | pf.bat -T > pfout.txt
parsing: 000ms
module import: 000ms
normalization: 000ms
semantical analysis: 000ms
XML Schema import: 000ms
core mapping: 000ms
core simplification: 000ms
type checking: 000ms
core tree optimization: 000ms
MIL code output: 016ms
Closing as fixed.
## Comment 7799
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1737735 at http://sourceforge.net/support/tracker.php?aid=1737735
| 128MB RAM needed to compile '1'? | https://api.github.com/repos/MonetDB/MonetDB/issues/1388/comments | 0 | 2020-11-30T09:35:59Z | 2024-06-27T11:25:22Z | https://github.com/MonetDB/MonetDB/issues/1388 | 753,315,748 | 1,388 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-15 11:31:34 +0200
From: @peterboncz
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-15 12:41:02 +0200
## Comment 7784
Date: 2007-06-15 11:31:34 +0200
From: @peterboncz
I tested the windows installer for XQuery (32-bits).
It works, but in the admin console, all images are missing.
these are the guf/jpeg/png images inside the pathfinder/runtime/xrpc directory, that shoule be copied in the share/MonetDB/xrpc installation directory.
apparently, this is not yet specifified in the installer.
## Comment 7785
Date: 2007-06-15 12:40:39 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in CVS.
## Comment 7786
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1737714 at http://sourceforge.net/support/tracker.php?aid=1737714
| admin console images missing | https://api.github.com/repos/MonetDB/MonetDB/issues/1387/comments | 0 | 2020-11-30T09:35:56Z | 2024-06-27T11:25:21Z | https://github.com/MonetDB/MonetDB/issues/1387 | 753,315,714 | 1,387 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-15 10:25:23 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-06-15 02:00:53 +0200
## Comment 7779
Date: 2007-06-15 10:25:23 +0200
From: @romulogoncalves
Dear developers,
Just run this example and you will see the mserver crashing and also fkey violations that I think are wrong.
Note: The example was ran in a server from the cvs branch(debug compilation) and in the stable branch (optimized and strcit compilation) both have the same problem.
create table t2 (id int , name varchar(1024), CONSTRAINT pk_t2_id PRIMARY KEY (id));
create table t3 (age int , address int, CONSTRAINT pk_t3_age PRIMARY KEY (age));
create table t1 (id int , age int);
insert into t1 values(1, 10);
insert into t1 values(2, 11);
insert into t1 values(3, 12);
insert into t1 values(4, 13);
insert into t1 values(5, 14);
insert into t2 values(10, 'monetdb');
insert into t2 values(11, 'moetdb');
insert into t2 values(12, 'montdb');
insert into t2 values(13, 'monetb');
insert into t2 values(14, 'metdb');
insert into t3 values(1, 101);
insert into t3 values(2, 118);
insert into t3 values(3, 108);
insert into t3 values(4, 18);
insert into t3 values(5, 1);
alter table t1 add constraint fk_t1_id_t2_id FOREIGN key(id) references t2(id);
alter table t1 add constraint fk_t1_age_t3_age FOREIGN key(age) references t3(age);
drop table t1;
drop table t3;
drop table t2;
The errors are:
MAPI = monetdb@localhost:50000
QUERY = alter table t1 add constraint fk_t1_id_t2_id FOREIGN key(id) references t2(id);
ERROR = !SQLException:assert:INSERT INTO: FOREIGN KEY constraint 't1.fk_t1_id_t2_id' violated
MAPI = monetdb@localhost:50000
QUERY = alter table t1 add constraint fk_t1_age_t3_age FOREIGN key(age) references t3(age);
ERROR = !SQLException:assert:INSERT INTO: FOREIGN KEY constraint 't1.fk_t1_age_t3_age' violated
and the mserver crash:
>[New Thread 1090525504 (LWP 964)]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1090525504 (LWP 964)]
0x00002aaabc09dc1e in drop_table (sql=0x11290d8, qname=0x139e3c8,
drop_action=0) at /ufs/goncalve/MonetDB/sql/src/server/sql_schema.mx:1696
1696 return sql_error(sql, 02,
(gdb) bt
0 0x00002aaabc09dc1e in drop_table (sql=0x11290d8, qname=0x139e3c8,
drop_action=0) at /ufs/goncalve/MonetDB/sql/src/server/sql_schema.mx:1696
1 0x00002aaabc09e3d8 in schemas (sql=0x11290d8, s=0x139e498)
at /ufs/goncalve/MonetDB/sql/src/server/sql_schema.mx:1846
2 0x00002aaabc0b6cce in semantic (sql=0x11290d8, scp=0x0, s=0x139e498)
at /ufs/goncalve/MonetDB/sql/src/server/sql_semantic.mx:1090
3 0x00002aaabc0b6f4f in output_semantic (sql=0x11290d8, scp=0x0, s=0x139e498)
at /ufs/goncalve/MonetDB/sql/src/server/sql_semantic.mx:1185
4 0x00002aaabc06ccc7 in sql_symbol2stmt (c=0x11290d8, sym=0x139e498)
at /ufs/goncalve/MonetDB/sql/src/backends/monet5/sql.mx:817
5 0x00002aaabc07ab02 in SQLparser (c=0x604938)
at /ufs/goncalve/MonetDB/sql/src/backends/monet5/sql_scenario.mx:1002
6 0x00002aaaaaaf7ffe in runScenarioBody (c=0x604938)
at /ufs/goncalve/MonetDB/MonetDB5/src/mal/mal_scenario.mx:553
7 0x00002aaaaaaf81cc in runScenario (c=0x604938)
at /ufs/goncalve/MonetDB/MonetDB5/src/mal/mal_scenario.mx:595
8 0x00002aaaaaac7a7f in MSserveClient (dummy=0x604938)
at /ufs/goncalve/MonetDB/MonetDB5/src/mal/mal_session.mx:495
9 0x0000003974606305 in start_thread () from /lib64/libpthread.so.0
10 0x0000003973acd50d in clone () from /lib64/libc.so.6
11 0x0000000000000000 in ?? ()
(gdb)
Regards,
Romulo
## Comment 7780
Date: 2007-06-15 11:00:53 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The reason why the mserver crash is because the drop of table t1 removes the fkey over the pkey(age) in t3, but when we do the drop table t3 the drop is looking for the fkey in pkey->keys which was not clean properly.
Regards,
Romulo
## Comment 7781
Date: 2007-06-15 11:10:21 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
Test added in the stable branch:
sql/src/test/BugTracker/Tests/alter_table_add_fkey.SF-1737668.sql
Regards,
Romulo
## Comment 7782
Date: 2007-06-15 14:00:53 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
It can be closed. It was fixed and checkin by Niels at 13:45
Regards,
Romulo
## Comment 7783
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1737668 at http://sourceforge.net/support/tracker.php?aid=1737668
| Alter table to add Fkey.... | https://api.github.com/repos/MonetDB/MonetDB/issues/1386/comments | 0 | 2020-11-30T09:35:54Z | 2024-06-27T11:25:20Z | https://github.com/MonetDB/MonetDB/issues/1386 | 753,315,685 | 1,386 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-11 11:43:29 +0200
From: @sjoerdmullender
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-15 03:51:56 +0200
## Comment 7776
Date: 2007-06-11 11:43:29 +0200
From: @sjoerdmullender
When installing the clients RPMs, I get the following messages. If we require a user monetdb, we should create one, and if not, we should assign the files to root. (Actually, we should most likely assign all files to root always and only run any services under a to-be-created account monetdb.)
rpm -i clients/Linux/rpm/RPMS/i686/*
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
warning: user monetdb does not exist - using root
warning: group monetdb does not exist - using root
## Comment 7777
Date: 2007-06-15 15:51:56 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: YES
Fixed in CVS. Done for all generated RPMs.
## Comment 7778
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1734867 at http://sourceforge.net/support/tracker.php?aid=1734867
| clients RPM requires user monetdb | https://api.github.com/repos/MonetDB/MonetDB/issues/1385/comments | 0 | 2020-11-30T09:35:51Z | 2024-06-27T11:25:19Z | https://github.com/MonetDB/MonetDB/issues/1385 | 753,315,655 | 1,385 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-09 07:01:16 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-06-09 08:25:56 +0200
## Comment 7772
Date: 2007-06-09 19:01:16 +0200
From: @romulogoncalves
Dear developers,
check this example:
> create table t1 (id int, name varchar(1024), PRIMARY KEY(id));
>
> create table t2 (id int DEFAULT 13, age int, PRIMARY KEY (ID), FOREIGN
> KEY(id) REFERENCES t1(id) ON UPDATE SET DEFAULT ON DELETE SET DEFAULT);
>
> insert into t1 values(1, 'monetdb');
>
> insert into t2 values(1, 23);
>
> update t2 set id = 2 where id =1;
>
> select * from t1;
> select * from t2;
>
> drop table t2;
> drop table t1;
>
The execution result is:
> [goncalve@amelia Tests]$ mclient <example.sql
> [ 1 ]
> [ 1 ]
> [ 1 ]
> % sys.t1, sys.t1 table_name
> % id, name name
> % int, varchar type
> % 1, 7 length
> [ 1, "monetdb" ]
> % sys.t2, sys.t2 table_name
> % id, age name
> % int, int type
> % 1, 2 length
> [ 2, 23 ]
>
The update should have been rejected because the new value in the fkey (t2.id) does not exist in the primary key (t1.id).
Regards,
Romulo
## Comment 7773
Date: 2007-06-09 19:16:48 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
Test added in:
sql/src/test/BugTracker/Tests/update_fkey_with_value_not_in_the_pkey.SF-1734143.sql
Regards,
Romulo
## Comment 7774
Date: 2007-06-09 20:25:56 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
allways check updates on fkeys (not only on restricted keys)
## Comment 7775
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1734143 at http://sourceforge.net/support/tracker.php?aid=1734143
| Update fkey with a values not in the pkey... | https://api.github.com/repos/MonetDB/MonetDB/issues/1384/comments | 0 | 2020-11-30T09:35:49Z | 2024-06-27T11:25:18Z | https://github.com/MonetDB/MonetDB/issues/1384 | 753,315,622 | 1,384 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-08 08:28:33 +0200
From: gat <<gat74>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-08 10:43:48 +0200
## Comment 7769
Date: 2007-06-08 20:28:33 +0200
From: gat <<gat74>>
Connecting to demo database (provided by regular setup) starts to do something (consuming more and more memory) until after 2 and half minutes PowerBuilder crashes.
WinXP SP2, PB 10.0, ODBC 3.525, MonetDB ODBC driver 2.16.02.01.
PB specialities: during connect to any database it tries to get list of tables, and create 5 own tables for storing meta data.
## Comment 7770
Date: 2007-06-08 22:43:48 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Thanks for reporting this.
This is a duplicate of bug [ 1724989 ] Cannot connect to MonetDB via ODBC from PowerBuilder.
The bug has been fixed in CVS and the fix will be in the upcoming release.
## Comment 7771
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1733744 at http://sourceforge.net/support/tracker.php?aid=1733744
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| Cannot connect via ODBC (from Powerbuilder 10) | https://api.github.com/repos/MonetDB/MonetDB/issues/1383/comments | 0 | 2020-11-30T09:35:46Z | 2024-06-28T07:36:58Z | https://github.com/MonetDB/MonetDB/issues/1383 | 753,315,591 | 1,383 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-08 07:34:51 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-06-09 01:13:18 +0200
## Comment 7766
Date: 2007-06-08 19:34:51 +0200
From: @romulogoncalves
Dear developers,
To see the bug run this example:
create table t1 (id int, name varchar(1024), PRIMARY KEY(id));
insert into t1 values(1, 'monetdb');
insert into t1 values(2, 'mon');
update t1 set id = 10 where id =1;
update t1 set id = 10 where id =2;
update t1 set id = 12 where id =3;
select * from t1;
drop table t1;
you will get:
% sys.t1, sys.t1 table_name
% id, name name
% int, varchar type
% 2, 7 length
[ 10, "monetdb" ]
[ 10, "mon" ]
t1.id is pkey so the second update should not be allowed.
It seems the constraint check is not done for pkeys.
Regards,
Romulo
## Comment 7767
Date: 2007-06-09 13:13:18 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed, we checked the update violations on the updates bats, it not on the full columns.
Added a test to the BugTracker directory update_pkey_bug.SF-1733712.sql
## Comment 7768
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1733712 at http://sourceforge.net/support/tracker.php?aid=1733712
| Update pKey with the same value for different rows... | https://api.github.com/repos/MonetDB/MonetDB/issues/1382/comments | 0 | 2020-11-30T09:35:43Z | 2024-06-27T11:25:15Z | https://github.com/MonetDB/MonetDB/issues/1382 | 753,315,562 | 1,382 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-08 08:23:58 +0200
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-08 10:43:11 +0200
## Comment 7760
Date: 2007-06-08 08:23:58 +0200
From: @drstmane
After checkins on 2007/06/05, test src/test/BugDay_2005-10-06_2.8/Tests/MapiClient-dump.SF-905851
fails --- possibly just a question of approving the new output...
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d-Fedora6/src_test_BugDay_2005-10-06_2.8/MapiClient-dump.SF-905851.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql5/.mTests103/GNU.64.64.d-Fedora6/src_test_BugDay_2005-10-06_2.8/MapiClient-dump.SF-905851.out.00.html
## Comment 7761
Date: 2007-06-08 10:11:53 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
The table dump contains ON DELETE RESTRICT ON UPDATE RESTRICT which I didn't expect since I don't see them in the table create statement. Niels, is this expected?
## Comment 7762
Date: 2007-06-08 10:18:03 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
RESTRICT is indeed the default, ie the output can be approved
## Comment 7763
Date: 2007-06-08 10:26:38 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Alternatively, since it is the default, should it be suppressed?
## Comment 7764
Date: 2007-06-08 10:43:11 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed by not dumping the default action.
No test needed.
## Comment 7765
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1733236 at http://sourceforge.net/support/tracker.php?aid=1733236
| SQL: test MapiClient-dump.SF-905851 fails | https://api.github.com/repos/MonetDB/MonetDB/issues/1381/comments | 0 | 2020-11-30T09:35:40Z | 2024-06-27T11:25:14Z | https://github.com/MonetDB/MonetDB/issues/1381 | 753,315,525 | 1,381 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-07 09:34:49 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-06-22 07:48:53 +0200
## Comment 7756
Date: 2007-06-07 09:34:49 +0200
From: @drstmane
Since
===================================================================
2007/06/02 - yingying: runtime/xrpc_server.mx,1.48
- added namespace check in count_node() and get_elem_pre_by_name() as
indicated in the TODO note.
- removed TODO notes in get_type(), since they do not apply any more.
===================================================================
xrpc_server fails to compile with icc
========
/export/scratch1/monet/monet.ntv.64.64.d.5248/pathfinder/runtime/xrpc_server.mx(688): error 593: variable "qn_prefix" was set but never used
BAT *qn_loc = NULL, *qn_prefix = NULL, *qn_uri = NULL;
^
compilation aborted for xrpc_server.c (code 2)
========
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.ntv.64.64.d-RedHat4WS/make-k.out
Obviously, the simple solution is to remove variable "qn_prefix" completely, however, I'm not sure, whether this is correct.
Currently, this only happens in the development trunk, since we (have to) use gcc in the release branch (to build proper RPMs).
However, since the code base is identical between release branch and development trunk, I consider this a release bug.
## Comment 7757
Date: 2007-06-22 05:46:04 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Any news on this one?
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.ntv.64.32.d-Fedora6/make-k.out
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.ntv.64.64.d-Fedora6/make-k.out
## Comment 7758
Date: 2007-06-22 19:48:53 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Ok, I felt free to remove the unused variable "qn_prefix" in CVS (XQuery_0-18 branch).
Hope, this does not break anything ...
## Comment 7759
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1732600 at http://sourceforge.net/support/tracker.php?aid=1732600
| PF: xrpc_server fails to compile with icc | https://api.github.com/repos/MonetDB/MonetDB/issues/1380/comments | 0 | 2020-11-30T09:35:38Z | 2024-06-27T11:25:13Z | https://github.com/MonetDB/MonetDB/issues/1380 | 753,315,484 | 1,380 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-07 09:23:00 +0200
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2007-06-13 10:20:45 +0200
## Comment 7737
Date: 2007-06-07 09:23:00 +0200
From: @drstmane
Since 2007/06/06, i.e., after checkins on 2007/06/05, PFtijah test fail on Itanium (64-bit running RedHat4WS Linux), as shredding (indexing?) documents fails with
========
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples remains marked radix-clustered on 32 bits; not checked!
!ERROR: TJoffsetindex: index-bat must be ordered on tail.
========
cf.,
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-RedHat4WS/modules_pftijah/loadthesis.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-RedHat4WS/modules_pftijah/colltest1.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-RedHat4WS/modules_pftijah/createcoll0.out.00.html
etc.
Since test fail only on our only big-endian 64-bit machine, the problem *might* be due to some incorrect treatment of OIDs and/or INTs (and/or incorrect mixture of them; recall: OID != INT).
Furthermore, the problems *might* be triggered by the shredder changes on 2007/06/05, or (less likely?) due to the fact that we run testing with --disable-assert in the release branch since 2007/06/05.
## Comment 7738
Date: 2007-06-07 09:43:22 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
I have seen the first two warnings before during shredding. When I do 3 seperate document-shreds (without pftijah!!!)into one pf collection I got exactly the same 2 warnings on a 32bit machine. I changed this to one shred with 3 docs into one pf collection and this worked fine (testcoll0.milS). I plan to make a test with bug-report for this. The last warning is new. I did not change anything in the shredder document generation part so that cannot be the cause. I will have a look at it.
## Comment 7739
Date: 2007-06-07 11:24:29 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
I found the checking point where the ERROR occurs and I'm a bit puzzled by the bitwise & 1 with the BATtordered() check. I think this is Henning's code (not available the rest of the week) which possibly stole it from some exiting code. I think I will just delete this & 1 and look what happens tonight.
int CMDoffsetindex ( BAT** result, BAT* tid, int* res_size )
[..........]
BATcheck(tid, name);
if (!(BATtordered(tid) & 1))
{
/* BUG1732596 this test currently fails on Itanium */
GDKerror("%s: index-bat must be ordered on tail.\n", name);
return GDK_FAIL;
}
## Comment 7740
Date: 2007-06-07 12:14:59 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
IMHO, your latest change in pathfinder/modules/pftijah/pftijah.mx is not correct.
Though being of type BIT (which is physically a signed char, hence 8 bit
wide), the ordered flags use all 8 bits, only the lowest one (1) indicates
sortedness, the highest one (1<<7=128) indicates reverse sortedness, and the
remaining ones are "mis-"used for radix clustering.
Hence, checking for sortedness shoudl alwasy be done with
BAT[ht]ordered(b)&1 --- regardless of the endianess, this should alwasy
check for the lowest bit, now matter whether this is located left or right
--- anyway, endianess is about byte orde, not about bit order, right??
Hence, if the your change does fix the Itanium problems, than the bug is
IMHO caused somewhere else (wrong setting of the ordered flag?), and only
appears here...
Which value does (int)BATtordered(tid) have in your case on your machine?
Stefan
## Comment 7741
Date: 2007-06-07 13:22:55 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
OK my first guess was very wrong. To me it looked like an optimized debug if:-)
I had a look at at the code where the MIL offsetindex(), which is where the
error occurs, is used. I do not think pftijah has a chance to screw up
things here. My main suspects are the ssort() and the hmark() functions.
The error occurs to the first arg to the offsetindex() function. The are only 2
offsetindex() calls in pftijah. They are:
file:pftijah:mx
PROC _buildIRindex(str ftiName, BAT[str,bat] collBat) : void :=
{
...............
var tmp := tids.kdiff(bat("tj_" + ftiName + "_pfpre"));
BUG:1732596: possible cause(1) here:
tmp is a BAT[oid,oid]
tmp := tmp.reverse().ssort();
var termindex := tmp.hmark(0@0).offsetindex(bat(_tj_TermBat(ftiN
ame)).count() + 1);
................
tmp := tids.semijoin(bat("tj_" + ftiName + "_pfpre"));
BUG:1732596: possible cause(2) here:
tmp is a BAT[oid,oid]
tmp := tmp.reverse().ssort();
var tagindex := tmp.hmark(0@0).offsetindex(bat(_tj_TagBat(ftiNam
e)).count() + 1);
...............
In both cases the tails are ascend sorted (BAT.ssort().hmark()). What MIL
statements should I add to check if I'm right. I do not have a clue here.
## Comment 7742
Date: 2007-06-08 09:10:53 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
First I propchecked the pftijah generated bats. They were OK. I added an assert_order() to both offsetindex() uses, eg:
var tmp := tids.kdiff(bat("tj_" + ftiName + "_pfpre"));
BUG:1732596: possible cause(1) here:
tmp is a BAT[oid,oid]
tmp := tmp.reverse().ssort();
var termindex := tmp.hmark(0@0).assert_order().offsetindex(bat(_tj_TermBat(ftiName)).count() + 1);
And as expected the error is now:
tj_add2collection("thesis",path+"/xml-docs/thesis.xml","thesis.xml",true);
Elapsed time = 001ms 210us [016us/node]
Shredded 1 XML document (thesis.xml), total time after commit=0.023s
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples remains marked radix-clustered on 32 bits; not checked!
!ERROR: assert_order((null)): tail was not ordered.
=!ERROR: CMDassert_order: operation failed.
!ERROR: interpret_params: offsetindex(param 1): evaluation error.
~ quit();
First thing that puzzles me is the assert_order(null). Does hmark() have a (null) result?
I still do not see the problem. When I have any [oid,oid] bat X and I ssort() it I get a bat with a sorted head. When I hmark()this bat this sorted head is moved to the tail. So I have a sorted tail. What could be wrong with initial bat X (a pftijah generated bat) that I do not en up with a sorted tail?
I will remove the assert_order() and will print the entire BAT tonight in offsetindex().
## Comment 7743
Date: 2007-06-08 09:32:33 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
I suppose the problem could be in ssort. Could you try with sort instead?
## Comment 7744
Date: 2007-06-08 09:36:45 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
> First thing that puzzles me is the assert_order(null). Does hmark() have a (null) result?
No, it only means that the result of hmark() does not have a logical name:
MonetDB4/src/modules/plain/bat.mx: GDKerror("assert_order(%s): tail was not ordered.\n", BBP_logical(input->batCacheid));
Which command triggers the inital BATpropcheck warnings?
The ssort() ?
If so, why/how does this relate to the changes on 2007/06/05 (see initial report)?
## Comment 7745
Date: 2007-06-09 10:08:58 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
As expected the tail is indeed not sorted. Going to run the test with sort() instead of ssort() now.
tj_add2collection("thesis",path+"/xml-docs/thesis.xml","thesis.xml",true);
Elapsed time = 001ms 215us [016us/node]
Shredded 1 XML document (thesis.xml), total time after commit=0.023s
!ERROR:TJoffsetindex: batordered()=64
-----------------
t h name
void oid type
-----------------
[ 0@0, 1@0 ]
[ 1@0, 2@0 ]
[ 2@0, 3@0 ]
[ 3@0, 4@0 ]
[ 4@0, 5@0 ]
[ 5@0, 6@0 ]
[ 6@0, 6@0 ]
[ 7@0, 10@0 ]
[ 8@0, 7@0 ]
[ 9@0, 8@0 ]
[ 10@0, 8@0 ]
[ 11@0, 9@0 ]
[ 12@0, 9@0 ]
[ 13@0, 8@0 ]
[ 14@0, 9@0 ]
[ 15@0, 9@0 ]
[ 16@0, 14@0 ]
[ 17@0, 12@0 ]
[ 18@0, 11@0 ]
[ 19@0, 13@0 ]
[ 20@0, 8@0 ]
[ 21@0, 4@0 ]
[ 22@0, 8@0 ]
[ 23@0, 6@0 ]
[ 24@0, 4@0 ]
[ 25@0, 4@0 ]
[ 26@0, 6@0 ]
[ 27@0, 8@0 ]
[ 28@0, 6@0 ]
[ 29@0, 4@0 ]
[ 30@0, 6@0 ]
[ 31@0, 4@0 ]
[ 32@0, 11@0 ]
[ 33@0, 9@0 ]
[ 34@0, 11@0 ]
[ 35@0, 8@0 ]
[ 36@0, 8@0 ]
[ 37@0, 11@0 ]
[ 38@0, 8@0 ]
[ 39@0, 9@0 ]
[ 40@0, 11@0 ]
[ 41@0, 2@0 ]
[ 42@0, 8@0 ]
[ 43@0, 4@0 ]
[ 44@0, 15@0 ]
[ 45@0, 14@0 ]
[ 46@0, 12@0 ]
[ 47@0, 13@0 ]
[ 48@0, 9@0 ]
[ 49@0, 8@0 ]
[ 50@0, 17@0 ]
[ 51@0, 7@0 ]
[ 52@0, 17@0 ]
[ 53@0, 17@0 ]
[ 54@0, 11@0 ]
[ 55@0, 6@0 ]
[ 56@0, 17@0 ]
[ 57@0, 16@0 ]
[ 58@0, 2@0 ]
[ 59@0, 2@0 ]
[ 60@0, 11@0 ]
[ 61@0, 17@0 ]
[ 62@0, 11@0 ]
[ 63@0, 17@0 ]
[ 64@0, 2@0 ]
[ 65@0, 2@0 ]
[ 66@0, 8@0 ]
[ 67@0, 11@0 ]
[ 68@0, 6@0 ]
[ 69@0, 8@0 ]
[ 70@0, 12@0 ]
[ 71@0, 18@0 ]
[ 72@0, 18@0 ]
[ 73@0, 10@0 ]
[ 74@0, 11@0 ]
[ 75@0, 2@0 ]
[ 76@0, 4@0 ]
[ 77@0, 6@0 ]
[ 78@0, 9@0 ]
[ 79@0, 8@0 ]
[ 80@0, 11@0 ]
[ 81@0, 15@0 ]
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_504(324)[oid,oid] with 82 tuples remains marked radix-clustered on 32 bits; not checked!
!ERROR: TJoffsetindex: index-bat must be ordered on tail.
=!ERROR: CMDoffsetindex: operation failed.
## Comment 7746
Date: 2007-06-10 19:17:25 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
The version with sort() instead of ssort() worked correct!!! Only two tests failed on another location and another ssort() in CMDmergeIndex.
So a buggy ssort causes the problems. I restored the ssorts(). In the sundaynight test I will print the input BAT of the ssort and its property bat. After this I think a gdk hacker should have a look at the problem.
## Comment 7747
Date: 2007-06-11 10:05:25 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
I had a look at the ssort impl. but would not know where to start. I hope the author (Sjoerd) can have a look at the code. I now print in the testsets ( loadthesis.milS ) the output bat and its properties. I can also print the input bat if you like.
## Comment 7748
Date: 2007-06-11 10:36:58 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Yes please, that would be helpful.
## Comment 7749
Date: 2007-06-11 14:22:36 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
Ok,now I also print the ssort() input bat and it's .info() bat
## Comment 7750
Date: 2007-06-11 18:09:42 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
I can reproduce the error on our ia64 machine (the one that is used in the testweb) when I compile MonetDB/XQuery with optimization on, but I cannot reproduce the error when I compile with optimization off (and debug on). Could it be a compiler error?
## Comment 7751
Date: 2007-06-12 10:03:27 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Niels, is it possible to upgrade the compiler on titan, or alternatively, cross compile with a newer compiler?
Otherwise I think we must close this bug with a Won't Fix because Somebody Else's Problem.
## Comment 7752
Date: 2007-06-12 14:26:40 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
I now have proof positive that it's the compiler: I compiled everything with --enable-optimize, but gdk_ssort I compiled separately with an extra -O0 at the end of the command line. In this configuration the tests (both the pftijah test and the test no.171 from MonetDB4) pass with flying colors.
So, Jan, can you undo your debugging changes? And thanks for the effort.
## Comment 7753
Date: 2007-06-13 10:20:45 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
In the end this was not a bug in our code but in the compiler. After upgrading the compiler on the Itanium (RHEL4) from gcc 3.4.6 to gcc 4.2.0 the bug disappeared.
## Comment 7754
Date: 2007-06-13 11:15:41 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
Gefeliciteerd!!!
## Comment 7755
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1732596 at http://sourceforge.net/support/tracker.php?aid=1732596
| PFtijah tests fail on Itanium | https://api.github.com/repos/MonetDB/MonetDB/issues/1379/comments | 0 | 2020-11-30T09:35:35Z | 2024-06-27T11:25:13Z | https://github.com/MonetDB/MonetDB/issues/1379 | 753,315,449 | 1,379 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-07 08:48:41 +0200
From: @drstmane
To: Romulo Goncalves <<goncalve>>
Version: -- development
Last updated: 2007-06-11 11:47:00 +0200
## Comment 7733
Date: 2007-06-07 08:48:41 +0200
From: @drstmane
Just to document this properly:
With both SQL4 and SQL5, most tests in src/test/Update_Delete_action have been failing since they have been added.
Is there any plan to fix this before the release?
## Comment 7734
Date: 2007-06-07 09:38:51 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
Well, I hope to finish them until Monday, but until now it was impossible to get some free time for that.
I will talk with Niels after TTT and see what we can do.
Sorry fro keep theses tests red for so long.
Regards,
Romulo
## Comment 7735
Date: 2007-06-11 11:47:00 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
The checkin at 11:20 fix the last problems related with these tests. All the tests are now green for FC6.
The night tests will tell us if they are also green for another platforms.
So, I consider this bug close.
Regards,
Romulo
## Comment 7736
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1732582 at http://sourceforge.net/support/tracker.php?aid=1732582
| SQL: most test in src/test/Update_Delete_action fail | https://api.github.com/repos/MonetDB/MonetDB/issues/1378/comments | 0 | 2020-11-30T09:35:32Z | 2024-06-27T11:25:12Z | https://github.com/MonetDB/MonetDB/issues/1378 | 753,315,424 | 1,378 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-06 09:16:44 +0200
From: Maurice van Keulen <<mvankeulen>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-06-06 02:14:42 +0200
## Comment 7729
Date: 2007-06-06 09:16:44 +0200
From: Maurice van Keulen <<mvankeulen>>
let $node := <a><b><c/></b><d><e/></d></a>
for $desc in $node/descendant::*
return count($desc/../preceding::*)
produces error
ERROR = !ERROR: PFll_preceding: ctx_bat must be ordered on tail.
!ERROR: PFll_preceding: operation failed.
(Suse Linux, checkout of 24 may)
## Comment 7730
Date: 2007-06-06 09:29:36 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Maurice,
the "usual" question to make (hopefully) debugging simpler and hence faster:
did you run Mserver with property checking disbaled (default) or enabled (--debug=10)?
If the former, could you please also try the latter?
Thanks!
Stefan
## Comment 7731
Date: 2007-06-06 14:14:42 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
fixed in CVS (XQuery_0-18 release branch):
in case the input to a preceding or following step is ordered on item|iter,
make sure we first re-order it on iter|item
before handing it to ll_preceeding()/ll_following().
test added in pathfinder/tests/BugTracker/Tests/error_in_preceeding_axis.SF-1731850.
## Comment 7732
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1731850 at http://sourceforge.net/support/tracker.php?aid=1731850
| Error in preceding axis | https://api.github.com/repos/MonetDB/MonetDB/issues/1377/comments | 0 | 2020-11-30T09:35:30Z | 2024-06-27T11:25:11Z | https://github.com/MonetDB/MonetDB/issues/1377 | 753,315,398 | 1,377 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-05 06:42:33 +0200
From: @peterboncz
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2007-09-12 12:58:54 +0200
## Comment 7725
Date: 2007-06-05 18:42:33 +0200
From: @peterboncz
<a t='1.1'>/t = 1
should return false, because in a general comparision of an untypedAtomic (atomized text node) and a numeric, the untypedAtomic should be cast to xs:double
the error is the PF currently casts to the other type (here xs:integer). This is only correct for the non-numerical types (we support of these only xs:boolean and xs:string).
## Comment 7726
Date: 2007-08-02 21:25:53 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: YES
I think I meant
<a t='1.1'>/@t = 1
Anyway, this one yields 'true''while it shouldn't
## Comment 7727
Date: 2007-09-12 12:58:54 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
This is now fixed in CVS.
Added test wrong_numeric_cast_in_general_comparison.SF-1731555.
## Comment 7728
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1731555 at http://sourceforge.net/support/tracker.php?aid=1731555
| XQ: wrong numeric cast in general comparison | https://api.github.com/repos/MonetDB/MonetDB/issues/1376/comments | 0 | 2020-11-30T09:35:27Z | 2024-06-27T11:25:10Z | https://github.com/MonetDB/MonetDB/issues/1376 | 753,315,373 | 1,376 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-04 10:21:21 +0200
From: alxtoth <<alxtoth>>
To: @njnes
Version: -- development
Last updated: 2007-06-05 03:16:39 +0200
## Comment 7721
Date: 2007-06-04 22:21:21 +0200
From: alxtoth <<alxtoth>>
Hi,
Unfortunately, MonetDB Server v5.1.0 still hangs on a table with 2304000 rows with SQL (same as previous version):
select sum(figure * cast( (13 <= dim_1 and dim_1 < 33) as integer) ) as c_1_7 from fact_1_3;
--------
CPU goes 100% for mserver5 process
macu:~ alex$ mserver5 --dbinit="include sql;sql.start();"
MonetDB Server v5.1.0
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled for powerpc-apple-darwin8.9.0/32bit with 32bit OIDs dynamically linked
dbname:demo
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /Users/alex/MonetFix/etc/monetdb5.conf)
>\q
MALexit: clients still active
MALexit: clients still active
MALexit: server forced exit 0 finishing 0 claimed 1 waiting
-------
Attached is the script to generate the random database including for example fact_1_3 used in offending SQL statements.
BR, Alex
## Comment 7722
Date: 2007-06-05 14:47:41 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
There was a problem in the M5 handling of bit to int conversion.
still needs a proper test.
## Comment 7723
Date: 2007-06-05 15:16:39 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
added test to M5/tests/BugTracker/Tests/batcast.SF-1730987
## Comment 7724
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1730987 at http://sourceforge.net/support/tracker.php?aid=1730987
| cast of expression hangs | https://api.github.com/repos/MonetDB/MonetDB/issues/1375/comments | 0 | 2020-11-30T09:35:24Z | 2024-06-27T11:25:09Z | https://github.com/MonetDB/MonetDB/issues/1375 | 753,315,342 | 1,375 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-04 01:24:29 +0200
From: Wouter Alink <<alink>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-06-06 06:15:24 +0200
## Comment 7707
Date: 2007-06-04 13:24:29 +0200
From: Wouter Alink <<alink>>
Given a document:
<?xml version="1.0"?>
<result/>
And the query:
element { "result" } { for $i at $k in (()
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
, doc("/tmp/result.xml")/result/*
) return element { name($i) } { $i/@*, $i/*, attribute { "id" } { $k } } }
Which has 32 doc-statements. Returns:
ERROR = !ERROR: merged_union: BAT 4 (1 BUNs) must have the same size as BAT 2 (2 BUNs).
!ERROR: CMDmerged_union: operation failed.
Note that this does only occur with 32 or more doc-statements, 31 or less works fine. Bug can be reproduced in 0.12 and 0.17 (was built last week).
Maybe this is helpful: it might be a bug in the standoff-code, as in 0.12 the the error message appears after [ " extracting regions...\n" ] has been printed, and " extracted 0 regions" is never printed.
## Comment 7708
Date: 2007-06-04 18:12:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Added test in pathfinder/tests/BugTracker/Tests/Attic/32_docs.SF-1730617* (XQuery_0-18 release branch).
Seems to work fine for me using 31, 32, & 99 doc() statements
with the latest XQuery_0-18 release branch.
Returns
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><result/></XQueryResult>
## Comment 7709
Date: 2007-06-04 22:36:23 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
stefan, i don't know what is different in my setup, but i get the same error-message as above with a very fresh checkout of the 0.18 release.
I compiled using monedbt_install.sh with: --enable-xquery --enable-optimise --nightly=stable
(i tweaked the script to also include "--enable-oid32")
## Comment 7710
Date: 2007-06-04 22:46:43 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Could you try with property checking enabled, i.e., `Mserver --debug=10`?
I tried both with and without property checking and the added tests work fine in both cases.
Further, I honestly cannot say, whether the monetdb_install.sh (and alike) scripts do "automatically" use the correct version after the version number increment, or whether they have to be updated.
Fabian and/or Martin should know ...
Just check, what your Mserver's welcome message says...
## Comment 7711
Date: 2007-06-04 22:55:06 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
just compiled monetdb with --enable-debug and --enable-oid64, same result, the bug persists
[wouter@wouter-PC bin]$ ./Mserver --dbinit="module(pathfinder);"
Monet Database Server V4.18.0
Copyright (c) 1993-2007, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
XRPC administrative console at http://localhost:50183/admin
MonetDB-4.18-D>
## Comment 7712
Date: 2007-06-04 22:57:59 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
your message crossed mine --debug=10 does not make a difference either, same error.
## Comment 7713
Date: 2007-06-04 23:15:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Did you run my test?
If not, could you please check whether it does what you intended?
I'm using a from-scratch compilation of the latest *_?-18 branches configured with both "--enable-debug --disable-optimize" and "--disable-debug --enable-optimize" on 64-bit FC6 (default 64-bit OIDs) compiled with gcc 4.1.1.
## Comment 7714
Date: 2007-06-04 23:45:45 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
it's getting stranger. indeed stefan, the query you checked in works, i was quite sure i checked it before i submitted it, but i don't have access to that machine now.
but:
- if i leave out the result-steps in the query i get the same error-message again.
- if the query with the result steps is used, but then on a document <result><aap/></result> (the difference is the child-element 'aap'), then it will with the same error-message as well.
- and finally, if the same query is used, but the last two 'doc("/tmp/result.xml")/result/*' are replaced with 'doc("/tmp/result.xml")/*' (which i ran into by mistake) i get the error-message: ERROR = !ERROR: err:XPTY0004: illegal qname ''
## Comment 7715
Date: 2007-06-05 08:36:44 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
modified test (omitting the /result step)
work fine with 31 doc() statements,
but fails as reported for 32 doc() statements
...!??
## Comment 7716
Date: 2007-06-06 15:30:45 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
This appears to be related to the MIL code optimizer that is supposed to optimize (shrink) the MIL code generated by milprint_summer:
========
.../pathfinder/tests/BugTracker/Tests $ sed 's|\$TSTSRCDIR|'"$PWD"'|g' 32_docs.SF-1730617_s00.xq.in | pf | Mserver --dbinit='module(pathfinder);' --set monet_prompt=''
^ ^^^
| |default: -O1
| 31 doc() statements
--------
Monet Database Server V4.18.0
Copyright (c) 1993-2007, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
XRPC administrative console at http://localhost:50001/admin
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><result><result id="1"/><result id="2"/><result id="3"/><result id="4"/><result id="5"/><result id="6"/><result id="7"/><result id="8"/><result id="9"/><result id="10"/><result id="11"/><result id="12"/><result id="13"/><result id="14"/><result id="15"/><result id="16"/><result id="17"/><result id="18"/><result id="19"/><result id="20"/><result id="21"/><result id="22"/><result id="23"/><result id="24"/><result id="25"/><result id="26"/><result id="27"/><result id="28"/><result id="29"/><result id="30"/><result id="31"/></result></XQueryResult>
========
.../pathfinder/tests/BugTracker/Tests $ sed 's|\$TSTSRCDIR|'"$PWD"'|g' 32_docs.SF-1730617_s01.xq.in | pf | Mserver --dbinit='module(pathfinder);' --set monet_prompt=''
^ ^^^
| |default: -O1
| 32 doc() statements
--------
Monet Database Server V4.18.0
Copyright (c) 1993-2007, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
XRPC administrative console at http://localhost:50001/admin
!ERROR: merged_union: BAT 4 (1 BUNs) must have the same size as BAT 2 (2 BUNs).
!ERROR: CMDmerged_union: operation failed.
========
.../pathfinder/tests/BugTracker/Tests $ sed 's|\$TSTSRCDIR|'"$PWD"'|g' 32_docs.SF-1730617_s01.xq.in | pf -O0 | Mserver --dbinit='module(pathfinder);' --set monet_prompt=''
^ ^^^
|
| 32 doc() statements
--------
Monet Database Server V4.18.0
Copyright (c) 1993-2007, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
XRPC administrative console at http://localhost:50001/admin
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><result><result id="1"/><result id="2"/><result id="3"/><result id="4"/><result id="5"/><result id="6"/><result id="7"/><result id="8"/><result id="9"/><result id="10"/><result id="11"/><result id="12"/><result id="13"/><result id="14"/><result id="15"/><result id="16"/><result id="17"/><result id="18"/><result id="19"/><result id="20"/><result id="21"/><result id="22"/><result id="23"/><result id="24"/><result id="25"/><result id="26"/><result id="27"/><result id="28"/><result id="29"/><result id="30"/><result id="31"/><result id="32"/></result></XQueryResult>
========
However, I could not yet find/locate the potential limitation in pathfinder/mil/mil_opt.c ...
*Might* be related to
define OPT_CONDS 32 /* maximum if-nesting */
in compiler/include/mil_opt.h
and/or an insufficient check of this limitation
in compiler/mil/mil_opt.c
and/or compiler/mil/milprint_summer.c .
However, simply increasing OPT_CONDS to 33 does not seem to help / change anything ...
## Comment 7717
Date: 2007-06-06 16:02:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Simply disabling code-motion (OPT_CODEMOTION)
does not seem to help, either ...
## Comment 7718
Date: 2007-06-06 16:15:14 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Correction:
Setting OPT_CONDS to 33 in compiler/include/mil_opt.h does "help" --- provided one forces a recompile of both compiler/mil/mil_opt.c and compiler/mil/milprint_summer.c
(
apparently, the dependency of compiler/mil/mil_opt.c and compiler/mil/milprint_summer.c on compiler/include/mil_opt.h is not noticed properly;
however, I don't know how to easily/properly do this in compiler/mil/Makefile.ag ... Niels?
)
Hence, it indeed appears that the OPT_CONDS=32 limit is not handled properly in
compiler/mil/mil_opt.c and/or compiler/mil/milprint_summer.c ...
## Comment 7719
Date: 2007-06-06 18:15:24 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
FIXED
## Comment 7720
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1730617 at http://sourceforge.net/support/tracker.php?aid=1730617
| XQ: 32 docs | https://api.github.com/repos/MonetDB/MonetDB/issues/1374/comments | 0 | 2020-11-30T09:35:21Z | 2024-06-27T11:25:08Z | https://github.com/MonetDB/MonetDB/issues/1374 | 753,315,313 | 1,374 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-04 10:56:40 +0200
From: Wouter Alink <<alink>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2007-06-05 12:33:21 +0200
## Comment 7700
Date: 2007-06-04 10:56:40 +0200
From: Wouter Alink <<alink>>
when walking through a query result which looks like:
"aap",
"aap",
....
"aap"
and using JDBC's rs.next(), JDBC says:
java.sql.SQLException: protocol violation, unexpected line: syntax error, unexpected $end on line 1
after about 250 rows.
This bug can be reproduced by typing:
for $i in 1 to 500 return "aap";
on a jdbcclient prompt. notice that this bug only occurs when using jdbc.
## Comment 7701
Date: 2007-06-04 15:01:56 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
On my (slightly) outdated M4/XQuery I get:
% JdbcClient -lxquery -Xdebug
Welcome to the MonetDB interactive JDBC terminal!
Type \q to quit, \h for a list of available commands
auto commit mode: on
monetdb-> for $i in 1 to 500 return "aap";
monetdb=>
Error: parse error: syntax error, unexpected ;, expecting $end on line 1, column 32 (next token is `;')
parse error: XQuery parsing failed
halted in /ufs/fabian/scratch/monetdb/current/xquery/compiler/parser/parser.y (PFparse), line 2902
monetdb->
I'll update and retry.
## Comment 7702
Date: 2007-06-04 18:45:10 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
fabian, the semi-colon should not have been there (i think it was used as a query separator in 0.12)
i didn't report that i only have witnessed this bug in 0.12, my apologies.
## Comment 7703
Date: 2007-06-04 21:52:36 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Added test in pathfinder/tests/BugTracker/Tests/Attic/JDBC_250_results.SF-1730556.*
for $i in 1 to 250 return "aap"
works fine, but
for $i in 1 to 251 return "aap"
fails with
Exception in thread "main" java.lang.AssertionError: block 0 should have been fetched by now :(
at nl.cwi.monetdb.jdbc.MonetConnection$ResultSetResponse.getLine(MonetConnection.java:1455)
at nl.cwi.monetdb.jdbc.MonetResultSet.absolute(MonetResultSet.java:191)
at nl.cwi.monetdb.jdbc.MonetResultSet.relative(MonetResultSet.java:2023)
at nl.cwi.monetdb.jdbc.MonetResultSet.next(MonetResultSet.java:1990)
at nl.cwi.monetdb.util.XMLExporter.dumpResultSet(XMLExporter.java:338)
at nl.cwi.monetdb.client.JdbcClient.executeQuery(JdbcClient.java:802)
at nl.cwi.monetdb.client.JdbcClient.processInteractive(JdbcClient.java:586)
at nl.cwi.monetdb.client.JdbcClient.main(JdbcClient.java:446)
## Comment 7704
Date: 2007-06-05 09:19:43 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
the latter crash feels kind of unintuitive, but it is correct, as XQuery over JDBC is just a hack, because XQuery doesn't support result "sets" which are fetchable in chunks. I'll look if I can get the error message a bit less confusing.
## Comment 7705
Date: 2007-06-05 12:33:21 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
The exact details are as follows:
The XQuery server "violates" the contract of the protocol by sending more tuples than the blocksize requested by JDBC. Since the XQuery server doesn't support this, it is sort of normal. I now ignore the preferred block size of JDBC for non-SQL languages.
## Comment 7706
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1730556 at http://sourceforge.net/support/tracker.php?aid=1730556
| XQ/JDBC: 250 results | https://api.github.com/repos/MonetDB/MonetDB/issues/1373/comments | 0 | 2020-11-30T09:35:19Z | 2024-06-27T11:25:07Z | https://github.com/MonetDB/MonetDB/issues/1373 | 753,315,287 | 1,373 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-04 10:42:53 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2009-11-24 10:31:31 +0100
## Comment 7687
Date: 2007-06-04 10:42:53 +0200
From: Wouter Alink <<alink>>
The attached query looks like:
"aap",
"aap",
"aap",
"aap",
... (1000 times)
"aap",
"aap"
When given to Mserver, it says: Segmentation Fault (0.12 as well as 0.17 (build last week)).
This bug should not have high priority.
## Comment 7688
Date: 2007-06-04 18:20:39 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Running this through plain pf (0.18), i.e.,
{ for i in `seq 1 999` ; do echo '"aap",' ; done ; echo '"aap"' ; } | pf
makes pf grow to 2.7 GB virt (1.8 GB res) on my 64-bit FC6 machine;
hence, this might well be a memory and/or burg problem,
especially on 32-bit machines.
## Comment 7689
Date: 2007-06-04 18:21:56 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
for got to mention:
pf "eventually" produces some MIL plan; no error.
## Comment 7690
Date: 2007-06-04 21:19:01 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Debugging releaved that the segfault is indeed caused by burg running out of stack space due to too deep recursion.
See also
[ 1607210 ] XQ: server-side compilation crash (member benchmark)
http://sourceforge.net/tracker/index.php?func=detail&aid=1607210&group_id=56967&atid=482468
## Comment 7691
Date: 2007-06-05 18:31:40 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7692
Date: 2007-06-05 18:31:48 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7693
Date: 2007-06-05 20:44:58 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Test added in pathfinder/tests/BugTracker/Tests/Attic/crash_on_concatenated_query.SF-1730547.*
Thanks to Peter's latest fix, it now correctly triggers
"ERROR = !fatal error: aborted too deep recursion"
instead of crashing the server.
## Comment 7694
Date: 2008-11-19 14:42:56 +0100
From: Jan Rittinger <<jan.rittinger>>
Hmm I'm not sure if the 'correct' output is indeed the correct one:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/Int.64.32.d.1-Fedora8/tests_BugTracker/crash_on_concatenated_query.SF-1730547.out.00.html
@ Stefan could you please verify.
## Comment 7695
Date: 2008-11-19 14:52:18 +0100
From: @drstmane
It was the "expected" output, since this test used to trigger a "ERROR = !fatal error: aborted too deep recursion" (and hence produce no output) on all platforms.
Apparently, this is no longer the case but now some platforms are able to run the test without error, i.e., producing some (correct? -> to be checked! ...) output.
Maybe, we need to split this test into two, a small one that works on all platforms and a huge one that fails (i.e., triggers a reasonable error message) on all platforms ...
## Comment 7696
Date: 2009-11-17 10:15:22 +0100
From: @drstmane
Instead of a "ERROR = !fatal error: aborted too deep recursion"
the test now (again) triggers a timeout or segfault on all platform:
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d.1-Fedora10/tests_BugTracker/crash_on_concatenated_query.SF-1730547.err.00.html
## Comment 7697
Date: 2009-11-21 15:34:54 +0100
From: @drstmane
===================================================================
2009/11/21 - stmane: pathfinder/runtime/pathfinder.mx,1.462.4.12
use more conservative stack limit (PFmaxstack)
for too deep recusion check (PFrecursion_fence())
to prevent segfault (due to too deep recursion and
consequent stack overflow) in tests
tests/BugTracker/Tests/collection_management.SF-1726599.*
tests/BugTracker/Tests/crash_on_concatenated_query.SF-1730547.*
cf.,
https://sourceforge.net/tracker/?func=detail&aid=1607210&group_id=56967&atid=482468
https://sourceforge.net/tracker/?func=detail&aid=1730547&group_id=56967&atid=482468
===================================================================
2009/11/21 - stmane:
pathfinder/tests/BugTracker/Tests/crash_on_concatenated_query.SF-1730547.timeout,1.1.2.1
extended timeout for slow(er) systems
===================================================================
Report can be closed once nightly testing confirms the fix for all platforms.
## Comment 7698
Date: 2009-11-24 10:31:31 +0100
From: @drstmane
Test now works fine on all platforms.
Closing.
## Comment 7699
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1730547 at http://sourceforge.net/support/tracker.php?aid=1730547
| XQ: Mserver crashes on concatenated query | https://api.github.com/repos/MonetDB/MonetDB/issues/1372/comments | 0 | 2020-11-30T09:35:17Z | 2024-06-27T11:25:06Z | https://github.com/MonetDB/MonetDB/issues/1372 | 753,315,266 | 1,372 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-03 04:02:01 +0200
From: @peterboncz
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-15 03:53:12 +0200
## Comment 7684
Date: 2007-06-03 16:02:01 +0200
From: @peterboncz
Something cygwin-ish seems to have gone quite wrong in the only available xquery download for windows (4.16.2), as starting the server yields a popup windowd with an error like:
the application could not be started because cygwin1.dll could not be found
It seems to be related to xrpc. I would guess that the xrpc code uses a library feature that is only provided in cygwin (not in win32) and gets linked in.
All in all: it is high time to switch to native builds!
Monet Database Server V4.16.2
Copyright (c) 1993-2007, CWI. All rights reserved.
Compiled for i386-pc-mingw32/32bit with 32bit OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
!ERROR: DL_open: dlopen(C:\Program Files\CWI\MonetDB-XQuery\lib\MonetDB4\lib_xrp
c_server-0.dll,6) -> Error: Kan opgegeven module niet vinden.
!ERROR: .
!ERROR: TBL_openmod: DL_open: Kan opgegeven module niet vinden.
!ERROR: .
!ERROR: TBL_loadmod: module load _xrpc_server cancelled.
!ERROR: moduleClient: module(_xrpc_server) load error.
!ERROR: interpret: no matching MIL operator to 'rpcd_start(int, bit, str)'.
!MAYBE YOU MEAN:
! rpcd_start(bit, str) : void
! rpcd_start(int, str) : void
! rpcd_start(int, bit) : void
! rpcd_start(int) : void
! rpcd_start() : void
MonetDB>
On my machine (32-bits win xp) I do have cygwin installed. Maybe a different version? In any case, the distributed windows bnary should not depend on windows.
## Comment 7685
Date: 2007-06-15 15:53:12 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in 0.18.0.
## Comment 7686
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1730234 at http://sourceforge.net/support/tracker.php?aid=1730234
| main windows download broken (please move back to 4.16.0) | https://api.github.com/repos/MonetDB/MonetDB/issues/1371/comments | 0 | 2020-11-30T09:35:14Z | 2024-06-27T11:25:05Z | https://github.com/MonetDB/MonetDB/issues/1371 | 753,315,235 | 1,371 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-01 03:29:47 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2007-11-15 10:59:11 +0100
## Comment 7671
Date: 2007-06-01 15:29:47 +0200
From: @grobian
When loading the tpch load files, the server crashes on Solaris 10/Sparc as follows:
>(build-sol10:~) fabian% gdb mserver5
GNU gdb 6.6
This GDB was configured as "sparc-sun-solaris2.10"...
(gdb) r
Starting program: /scratch/monetdb-fabian/current/program-sun4u/bin/mserver5
warning: Temporarily disabling breakpoints for unloaded shared library "/usr/lib/ld.so.1"
warning: Lowest section in /lib/libpthread.so.1 is .dynamic at 00000074
warning: Lowest section in /lib/libdl.so.1 is .dynamic at 00000094
MonetDB Server v5.0.0_beta2_1
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled for sparc-sun-solaris2.10/32bit with 32bit OIDs dynamically linked
dbname:demo
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /scratch/monetdb-fabian/current/program-sun4u/etc/monetdb5.conf)
include sql;
[New LWP 1]
[New LWP 2]
>include sql;
>[New LWP 3]
Program received signal SIGSEGV, Segmentation fault.
[Switching to LWP 3]
0xfa50ff70 in libiconv () from /scratch/gentoo/lib/libiconv.so.2
(gdb) bt
0 0xfa50ff70 in libiconv () from /scratch/gentoo/lib/libiconv.so.2
1 0xfa63439c in strIconv (res=0x9d29b0,
org=0x9d9d58 "/scratch/monetdb-fabian/current/sql/src/benchmarks/tpch/SF/region.tbl", f=0x9d8130 "UTF-8", t=0x9d8220 "646")
at /scratch/monetdb-fabian/current/five/src/modules/atoms/str.mx:1561
2 0xfa63443c in STRIconv (res=0x9d29b0, o=0x9d2980, fp=0x9d29c0, tp=0x9d29a0)
at /scratch/monetdb-fabian/current/five/src/modules/atoms/str.mx:2111
3 0xff29be28 in runMALsequence (cntxt=0x246c8, mb=0x9d9af0, startpc=9,
stoppc=10, stk=0x9d28e0, env=0x0, pcicaller=0x0)
at /scratch/monetdb-fabian/current/five/src/mal/mal_interpreter.mx:1048
4 0xff29d1dc in reenterMAL (cntxt=0x246c8, mb=0x9d9af0, startpc=3072,
stoppc=36, stk=0x9d28e0, env=0x9da020, pcicaller=0x0)
at /scratch/monetdb-fabian/current/five/src/mal/mal_interpreter.mx:318
5 0xf9f95f08 in OPTevaluate (mb=0x9d9af0, stk=<value optimized out>,
p=<value optimized out>)
The crash appears to be in the libiconv library.
(build-sol10:~) fabian% equery depends libiconv
[ Searching for packages depending on libiconv... ]
dev-libs/glib-2.12.11 (virtual/libiconv)
net-dns/bind-tools-9.4.1 (idn? dev-libs/libiconv)
sys-devel/gcc-4.1.2 (virtual/libiconv)
sys-devel/gettext-0.16.1 (virtual/libiconv)
virtual/libiconv-0 (!elibc_glibc? dev-libs/libiconv)
From the glib package:
(build-sol10:~) fabian% ldd /scratch/gentoo/usr/bin/glib-genmarshal
libglib-2.0.so.0 => /scratch/gentoo/usr/lib/libglib-2.0.so.0
libintl.so.8 => /scratch/gentoo/usr/lib/libintl.so.8
libiconv.so.2 => /scratch/gentoo/lib/libiconv.so.2
libc.so.1 => /lib/libc.so.1
libgcc_s.so.1 => /scratch/gentoo/usr/lib/gcc/sparc-sun-solaris2.10/4.1.2/libgcc_s.so.1
libgcc_s.so.1 => /scratch/gentoo/lib/libgcc_s.so.1
libm.so.2 => /lib/libm.so.2
/platform/SUNW,Ultra-250/lib/libc_psr.so.1
(build-sol10:~) fabian% /scratch/gentoo/usr/bin/glib-genmarshal
bla
/dev/stdin:2: error: unexpected character `\12', expected character `:'
(build-sol10:~) fabian%
Conclusion, glib can use libiconv correctly and without a problem. So the library appears to be correct.
A similar problem occurs when libpcre is being used.
(build-sol10:~) fabian% ldd /scratch/gentoo/usr/bin/pcregrep
libpcre.so.0 => /scratch/gentoo/usr/lib/libpcre.so.0
libc.so.1 => /lib/libc.so.1
libgcc_s.so.1 => /scratch/gentoo/usr/lib/gcc/sparc-sun-solaris2.10/4.1.2/libgcc_s.so.1
libm.so.2 => /lib/libm.so.2
/platform/SUNW,Ultra-250/lib/libc_psr.so.1
(build-sol10:~) fabian% pcregrep '^\[.*gcc' system
[ebuild R ] virtual/gcc-0
(build-sol10:~) fabian%
However, when mserver uses the library, a crash occurs in libpcre's pcre_compile method, which pcregrep also uses in the same way mserver does (I tracked/traced this).
It looks as if mserver is doing something fundamentally wrong on Solaris/Sparc which causes these crashes in external libraries. Maybe the same happens on Solaris/x86, but there it is not tested as the libiconv and libpcre libraries aren't installed. I can try on my own Solaris/amd64 box later perhaps to see what happens there.
## Comment 7672
Date: 2007-07-02 14:10:44 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
what is the status of this bug?
## Comment 7673
Date: 2007-07-02 14:43:32 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
can try tomorrow when barn.project is alive and kicking again
## Comment 7674
Date: 2007-07-03 15:14:11 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
seems to be the same:
http://homepages.cwi.nl/~fabian/mtest-current-build-sol10/five/mTests/
## Comment 7675
Date: 2007-09-09 11:11:01 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Is this bug still valid in the latest release? current?
## Comment 7676
Date: 2007-09-10 09:37:14 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
I didn't test recently because I didn't disable the hwcounters on Sparc.
## Comment 7677
Date: 2007-10-05 21:55:29 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
didn't we fix this recently (alignment problem)
## Comment 7678
Date: 2007-10-05 21:57:37 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
that fix (during VLDB) wasn't sufficient.
## Comment 7679
Date: 2007-10-05 22:00:02 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
m5 ? What is the current problem (and how can I repeat it)?
## Comment 7680
Date: 2007-11-10 23:34:10 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This seems to be solved with the latest actions of Fabian
## Comment 7681
Date: 2007-11-11 01:25:11 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
as far as I know it still does weird stuff.
## Comment 7682
Date: 2007-11-15 10:59:11 +0100
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
now also pcre is fixed. On our itanium platform this could easily be solved using prctl, which enables signals on unaligned memory access. The problem was solved by aligning the patterns with a size_t instead of a sht.
## Comment 7683
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1729455 at http://sourceforge.net/support/tracker.php?aid=1729455
| Solaris 10/Sparc: weird crashes on external library usage | https://api.github.com/repos/MonetDB/MonetDB/issues/1370/comments | 0 | 2020-11-30T09:35:11Z | 2024-06-27T11:25:04Z | https://github.com/MonetDB/MonetDB/issues/1370 | 753,315,199 | 1,370 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-06-01 02:55:10 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2007-06-22 08:33:13 +0200
## Comment 7668
Date: 2007-06-01 14:55:10 +0200
From: @grobian
Various SQL (loading) tests on our Sun5.10/x86 machine in the testweb have a failing assertion:
Assertion failed: ss->access == ST_READ, file /export/scratch1/monet/monet.GNU.32.32.d.8612/MonetDB/src/common/stream.mx, line 2495, function bs_read
i.e.:
src/benchmarks/wisconsin/load
src/benchmarks/tpch/check0
For the record: it seems that the build-sol10 (Solaris 10/Sparc) has no problems doing the wisconsin benchmark with mtest.
## Comment 7669
Date: 2007-06-22 08:33:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
All M4, M5, SQL4 & SQL5 tests seem to work fine also on Solaris 10/x86, now.
Hence, I feel free to close this one.
## Comment 7670
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1729443 at http://sourceforge.net/support/tracker.php?aid=1729443
| stream.mx: assertion fails on Solaris 10/x86 | https://api.github.com/repos/MonetDB/MonetDB/issues/1369/comments | 0 | 2020-11-30T09:35:08Z | 2024-06-27T11:25:03Z | https://github.com/MonetDB/MonetDB/issues/1369 | 753,315,171 | 1,369 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-29 11:27:46 +0200
From: @grobian
To: @mlkersten
Version: -- development
Last updated: 2007-05-29 12:38:06 +0200
## Comment 7664
Date: 2007-05-29 11:27:46 +0200
From: @grobian
I stumbled upon a little bug while messing around with functions in M5:
% alias run_five
$INSTALL_DIR/bin/mserver5
% run_five
MonetDB Server v5.0.0_beta2_1
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled for x86_64-pc-linux-gnu/64bit with 64bit OIDs dynamically linked
dbname:demo
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /ufs/fabian/scratch/monetdb/current/program-x86_64/etc/monetdb5.conf)
>return();
mserver5: /ufs/fabian/scratch/monetdb/current/five/src/mal/mal_instruction.mx:2015: pushInstruction: Assertion `p->argv[0]>=0' failed.
Abort
## Comment 7665
Date: 2007-05-29 11:32:18 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
Added test:
tests/BugTracker/Tests/assertion_on_return.SF-1727339.mal
## Comment 7666
Date: 2007-05-29 12:38:06 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This statement now correctly generates a warning
that the return variable has not been set.
## Comment 7667
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1727339 at http://sourceforge.net/support/tracker.php?aid=1727339
| M5: Assertion crash on simple return | https://api.github.com/repos/MonetDB/MonetDB/issues/1368/comments | 0 | 2020-11-30T09:35:05Z | 2024-06-27T11:25:02Z | https://github.com/MonetDB/MonetDB/issues/1368 | 753,315,140 | 1,368 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-28 04:57:09 +0200
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-07-13 10:22:26 +0200
## Comment 7653
Date: 2007-05-28 16:57:09 +0200
From: Wouter Alink <<alink>>
(yesterday's MonetDB head)
it looks like inserting parts of large documents (more than 4K elements?) into another document does not work...
[wouter@vzzzbx framework]$ /opt/MonetDB-4.17/bin/MapiClient -lxq -p50170
xquery>for $i in doc("monetdb-sources.xml")/XIRAF/*
more>return do insert $i
more> into doc("test.xml")/XIRAF/device[@name="pc01"]
more>MAPI = monetdb@localhost:50170
QUERY = for $i in doc("monetdb-sources.xml")/XIRAF/*
ERROR = !ERROR: CMDfind: value not found.
!ERROR: CMDfind: operation failed.
xquery>count(doc("monetdb-sources.xml")//*)
more>6977
xquery>
I attached an XML-file monetdb-sources.xml, but i guess any (large) document will do for testing. (I could reproduce the bug with documents with 7K to 200K elements).
test.xml could look like:
<XIRAF>
<device name="pc01"/>
</XIRAF>
## Comment 7654
Date: 2007-06-04 10:35:17 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Hi Sjoerd, could you take a look at this one?
## Comment 7655
Date: 2007-06-04 17:35:23 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
It works for me.
I added a test to the stable branch. We'll see whether it fails anywhere tomorrow.
Wouter, I shredded the to-be-inserted document without holes (last arg to pf:add-doc == 0). Is that ok?
## Comment 7656
Date: 2007-06-06 14:22:34 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Test seems to work fine everywhere:
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/insert_large_doc.SF-1726954.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/insert_large_doc.SF-1726954.err.00.html
## Comment 7657
Date: 2007-06-06 19:49:27 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
i cannot reproduce the bug any more either with the 0.18 branch. answer to sjoerds question: i added the to-be-inserted-doc as read-only too.
## Comment 7658
Date: 2007-07-12 22:53:38 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
I can finally reproduce this bug with 0.18.2? (installed using the monetdb-super-tarball, the version still says 4.18.0 ?!?) using the already attached document. strange enough, the error only occurs when 'test.xml' does _not_ contain whitespace between the xiraf-element and the device-element:
<XIRAF><device name="pc01"/></XIRAF>
and i had to insert the data twice. The first the update succeeds. The second time the error-message is returned:
xquery>pf:add-doc("/home/wouter/Desktop/monetdb.xml","bigdoc.xml","bigdoc.xml",0)
more>xquery>pf:add-doc("/home/wouter/Desktop/test.xml","target.xml","target.xml",10)
more>xquery>for $i in doc("bigdoc.xml")/XIRAF/* return do insert $i into doc("target.xml")/XIRAF/device[@name="pc01"]
more>xquery>for $i in doc("bigdoc.xml")/XIRAF/* return do insert $i into doc("target.xml")/XIRAF/device[@name="pc01"]
more>MAPI = monetdb@localhost:50182
QUERY = for $i in doc("bigdoc.xml")/XIRAF/* return do insert $i into doc("target.xml")/XIRAF/device[@name="pc01"]
ERROR = !ERROR: CMDfind: value not found.
!ERROR: CMDfind: operation failed.
xquery>
## Comment 7659
Date: 2007-07-13 09:09:33 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I'm afraid you have XQuery 0.18.0 as I could never find the 0.18.2 sources here at CWI, so I assumed it was a Windows only thing. Correct me if I'm wrong.
PS.: that the server announces itself as 4.18(.0) has nothing to do with the pathfinder module. It would perhaps be a good idea if the pathfinder module would announce itself as well upon load, thereby reducing version number confusion. Remember that MonetDB is not a product, it is a collection of items that seem to be tightly integrated, but sold as completely independent pieces.
## Comment 7660
Date: 2007-07-13 09:19:12 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Fabian,
since Sjoerd produced the official packages for both the initial *.18.0 ("Mars") release and the MonetDB/XQuery 0.18.0 ("Mars 2") bugfix release from/in his private space instead of simply copying nightly builds, the only reference packages and source are those on SF at
http://sourceforge.net/project/showfiles.php?group_id=56967&package_id=219621
and the respective tags in CVS.
Wouter,
MonetDB/XQuery 0.18.0 (still) builds on / uses the MonetDB 4.18.0 server, hence, you correctly see this message in the *server* welcome message.
We plan to extend the welcome message to give the version numbers of all packages involved (i.e., MonetDB Common (GDK), MonetDB server (M4 or M5), SQL module and/or pathfinder module).
Stefan
## Comment 7661
Date: 2007-07-13 09:22:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Making this re-opend bug report more visible.
Test tests/BugTracker/Tests/insert_large_doc.SF-1726954.* should be extended accordingly or a new test should be added.
## Comment 7662
Date: 2007-07-13 10:22:26 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in CVS.
I also updated the test with Wouter's new information. This time I was able to reproduce the problem.
## Comment 7663
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1726954 at http://sourceforge.net/support/tracker.php?aid=1726954
| XQuery: inserting (large) doc into other doc | https://api.github.com/repos/MonetDB/MonetDB/issues/1367/comments | 0 | 2020-11-30T09:35:02Z | 2024-06-27T11:25:01Z | https://github.com/MonetDB/MonetDB/issues/1367 | 753,315,089 | 1,367 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-28 04:15:32 +0200
From: rafb <<rafb>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-06-04 10:55:56 +0200
## Comment 7647
Date: 2007-05-28 16:15:32 +0200
From: rafb <<rafb>>
The following query
let $out :=
for $t in doc("thumbnail.xml")//thumbnail
return element copy { $t/@* }
for $j in $out/@* return name($j)
on this document
<thumbnail xstart="82189222612" xend="82189226360" name="_thumbnail.jpg" xid="313298"/>
results in this error messages
MAPI = monetdb@localhost:50000
QUERY = let $out :=
ERROR = !ERROR: mposjoin: missing matches. The result requires the same number of tuples as the first two input arguments: 0 < 4.
!ERROR: CMDmposjoin: operation failed.
Raoul
## Comment 7648
Date: 2007-05-29 00:10:11 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
added test in pathfinder/tests/BugTracker/Tests/attribute_access_mposjoin_error.SF-1726928.*
seems to work fine with the latest CVS development trunk (HEAD) (4.17.1/0.17.1)
## Comment 7649
Date: 2007-05-29 10:27:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_access_mposjoin_error.SF-1726928.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_access_mposjoin_error.SF-1726928.err.00.html
## Comment 7650
Date: 2007-05-29 12:34:33 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
for convenience:
marked as XQuery / pathfinder bug
## Comment 7651
Date: 2007-06-04 10:55:56 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
closing, as the problems seems to be fixed in the code base for the upcoming 0.18 release;
cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_access_mposjoin_error.SF-1726928.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/attribute_access_mposjoin_error.SF-1726928.err.00.html
(with the 0.18 release due in about a week from now, we won't do any fixes on the old 0.16 release, anymore.)
## Comment 7652
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1726928 at http://sourceforge.net/support/tracker.php?aid=1726928
| XQuery: Attribute access causes mposjoin error | https://api.github.com/repos/MonetDB/MonetDB/issues/1366/comments | 0 | 2020-11-30T09:35:00Z | 2024-06-27T11:25:00Z | https://github.com/MonetDB/MonetDB/issues/1366 | 753,315,052 | 1,366 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-27 10:32:38 +0200
From: Wouter Alink <<alink>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2009-11-21 11:42:21 +0100
## Comment 7639
Date: 2007-05-27 22:32:38 +0200
From: Wouter Alink <<alink>>
Today's MonetDB 4.17
Given a small document: <aap id="1"/>
shredding it once, deleting it, and shredding it again corrupts the database.
[wouter@localhost test]$
/opt/MonetDB-4.17/bin/MapiClient -lxq -p50170
xquery>pf:add-doc("/home/wouter/test.xml","doc1.xml")
more>xquery>pf:del-doc("doc1.xml")
more>xquery>pf:add-doc("/home/wouter/test.xml","doc1.xml")
more>xquery>pf:collection("doc1.xml")
more><aap id="1"/><aap id="1"/>
xquery>
At first sight the document still looks intact:
xquery>doc("doc1.xml")
more><?xml version="1.0" encoding="utf-8"?>
<aap id="1"/>
But then it looks like the document's document node has gone, because:
xquery>doc("doc1.xml")/aap
more>
xquery>
gives an empty result.
## Comment 7640
Date: 2007-05-28 01:16:47 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
yes indeed, it seems that:
- pf:del-doc() does not cleanup empty-collections, while MIL delete_doc() does this
as a result, the second add-doc adds the document to the previous collection
- there seems to be a bug in index filtering for deleted documents in a multi-doc collection
as the functinality is new, so it is not that surprising
will fix soon
## Comment 7641
Date: 2007-05-28 12:55:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Test added in pathfinder/tests/BugTracker/Tests/collection_management.SF-1726599*
## Comment 7642
Date: 2007-05-29 10:26:18 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/collection_management.SF-1726599.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/collection_management.SF-1726599.out.00.html
## Comment 7643
Date: 2007-06-08 02:23:43 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7644
Date: 2009-11-16 22:20:49 +0100
From: @drstmane
re-opened as the test fails on all platforms at least since 2009.04.09
cf.,
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d.1-Fedora10/tests_BugTracker/collection_management.SF-1726599.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.32.32.d.1-Debian4.0/tests_BugTracker/collection_management.SF-1726599.err.00.html
## Comment 7645
Date: 2009-11-21 11:42:21 +0100
From: @drstmane
===================================================================
2009/11/20 - boncz:
pathfinder/tests/BugTracker/Tests/collection_management.SF-1726599.stable.err,1.3.24.1
pathfinder/tests/BugTracker/Tests/collection_management.SF-1726599.stable.out,1.3.24.1
approving test output
===================================================================
## Comment 7646
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1726599 at http://sourceforge.net/support/tracker.php?aid=1726599
| XQuery: collection management broken | https://api.github.com/repos/MonetDB/MonetDB/issues/1365/comments | 0 | 2020-11-30T09:34:57Z | 2024-06-27T11:25:00Z | https://github.com/MonetDB/MonetDB/issues/1365 | 753,315,027 | 1,365 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-27 07:54:25 +0200
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-06-08 02:26:02 +0200
## Comment 7632
Date: 2007-05-27 19:54:25 +0200
From: Wouter Alink <<alink>>
A few remarks (certainly not high-priority issues) on the webbased management interface (based on today's MonetDB 4.17):
- pressing "View DB Environment" as the first button (after starting mserver) results in an error (a soap-message with an error inside which reads something like: !ERROR: BUNdelHead, access denied). This problem does not occur when pressing another button first.
- pressing "Restart DB Server" does stop the server, but does not start it again.
- When shredding a document it is not clear that a new name and a name for the collection are required. When pressing 'Add' an error 'xs:string is expected to have a simple type' is returned.
- after shredding a document ("test.xml" in collection "test1"), it can be view using "List All Documents" -> "View", but not when using "List Collections" -> "View" -> "View". In that case an error "404 Not Found" is given.
## Comment 7633
Date: 2007-05-27 20:21:34 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
some more remarks:
- when deleting a collection ("List Collections" -> "Delete") the documents in the collection are deleted, but the collection will still be shown.
- when again adding a new document (in a collection with a different name), both collections are shown, but the "docs" column is mixed up (it says the first collection holds 1 document and the second "UNKNOWN" documents, while in reality the first collection was deleted and the second holds 1 document)
## Comment 7634
Date: 2007-05-28 01:37:12 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Restart DB server=> I would like to have a start shell script that is just a while(true) loop being documented as the proper way to start Mserver. Killing the server can be done by killing the script.
Sjoerd, any thoughts on how to get such a script into the distribution?
all other issues are fixed. Note that the UNKNOWN in the docs is because the admin guis gives back also info on non-committed (=busy) document management operations. That was correct behavior.
## Comment 7635
Date: 2007-05-29 12:28:02 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
We could perhaps use the mguardian script from M5 and adapt it to M4+PF.
Another idea is to have a built-in restart command that execv()s the program (i.e. Mserver) with the arguments originally given.
## Comment 7636
Date: 2007-06-01 00:48:42 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
Hi Wouter,
About your two additional remarks: by deletion, values of the corresponding collection/document tuples are nil-ed but not removed. So they are included in the soap response message. the admin console replace the 'nil' values with 'unknown'. I think it might be better to filter out those nil-ed tuples and do not display them.
Regards,
Jennie
## Comment 7637
Date: 2007-06-08 02:26:02 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed, except for the restart
## Comment 7638
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1726548 at http://sourceforge.net/support/tracker.php?aid=1726548
| XQuery: Administrative Console | https://api.github.com/repos/MonetDB/MonetDB/issues/1364/comments | 0 | 2020-11-30T09:34:55Z | 2024-06-27T11:24:59Z | https://github.com/MonetDB/MonetDB/issues/1364 | 753,314,988 | 1,364 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-25 12:11:41 +0200
From: Manuel Mayr <<mayrm>>
To: @njnes
Version: -- development
Last updated: 2007-12-26 02:15:08 +0100
## Comment 7622
Date: 2007-05-25 12:11:41 +0200
From: Manuel Mayr <<mayrm>>
Dear developers,
I tried to execute some of the (slightly adapted) SQL scripts from pathfinders SQL Back-End on monet5 and it leads to a
segmentation fault (with set optimization='off').
The test instance was a 3MB xmlgen-document.
I load the whole scenario to reproduce the bug,
on my website to avoid overcharching the mail traffic ;) I attached the gdb output at the end of the report.
http://www-db.in.tum.de/~mayrm/bug_report.tar.gz
Regards,
Manuel
(gdb) run --dbinit="include sql;"
Starting program: /home/mayrm/develop/bin/mserver5 --dbinit="include sql;"
Failed to read a valid object file image from memory.
[Thread debugging using libthread_db enabled]
[New Thread -1217702208 (LWP 4930)]
MonetDB Server v5.0.0_beta2_1
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled for i686-pc-linux-gnu/32bit with 32bit OIDs dynamically linked
dbname:demo
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /home/mayrm/develop//etc/monetdb5.conf)
[New Thread -1283089488 (LWP 4937)]
>[New Thread -1294316624 (LWP 4957)]
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
= TODO: common/optimize.c: push down head binop
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1294316624 (LWP 4957)]
0xb7cad70a in BAThashjoin (l=0x0, r=0x88147a8, estimate=2147483648)
at gdk_relop.mx:405
405 bunfastins(bn, BUNhead(l, p), BUNtail(r, w));
## Comment 7623
Date: 2007-05-25 12:36:41 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
It seems I cannot repeat your crash. Loading the data and executing the query just gives me an empty result. Which version of M5/sql did you use?
## Comment 7624
Date: 2007-05-25 12:39:33 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
`mserver5 --version` may also give relevant information
## Comment 7625
Date: 2007-05-25 13:11:12 +0200
From: Manuel Mayr <<mayrm>>
Logged In: YES
user_id=1629542
Originator: YES
Well even an empty result should not be correct.
With DB2 i got exactly one tuple.
Maybe there is some difference in our main memory setup,
that leads to our different observations.
I've 2GB of main memory, and during the execution
of the script monet makes an intense use of it.
I observed that exactly before the seg fault occurs
monet begins to allocate lots of memory.
Best Regards,
Manuel
I updated the sources yesterday.
mserver5 --version
MonetDB Server v5.0.0_beta2_1
Copyright (c) 1993-2007 CWI, all rights reserved
Compiled by: mayrm@notekemper21
Compilation: gcc -g -O2 -Wall -Wextra -std=c99 -Werror-implicit-function-declaration -Werror -Wpointer-arith -Wdeclaration-after-statement
Linking: /usr/bin/ld
## Comment 7626
Date: 2007-05-27 18:32:08 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
There are no records with name 'person' or 'people', ie the first part a0000 allready has 0 rows.
So the query seems to work.
## Comment 7627
Date: 2007-05-27 20:46:38 +0200
From: Manuel Mayr <<mayrm>>
Logged In: YES
user_id=1629542
Originator: YES
ouutch, I think the D_sql_exp.csv_names I sent you
was not correct.
http://home.in.tum.de/~mayrm/bug_report.tar.gz
^^
this should contain the correct names file.
I added also the corresponding D.xml document.
## Comment 7628
Date: 2007-05-29 12:35:12 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
for convenience:
tagged as SQL bug.
## Comment 7629
Date: 2007-11-10 23:22:06 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Is this bug still open, or cast in a test?
## Comment 7630
Date: 2007-12-26 14:15:08 +0100
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
First part of handing range joins is added now. The query now at least ends (still way to slow).
A new test is added in the src/test/xquery directory
## Comment 7631
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1725456 at http://sourceforge.net/support/tracker.php?aid=1725456
| SQL: Segmentation fault with query. | https://api.github.com/repos/MonetDB/MonetDB/issues/1363/comments | 0 | 2020-11-30T09:34:52Z | 2024-06-27T11:24:56Z | https://github.com/MonetDB/MonetDB/issues/1363 | 753,314,952 | 1,363 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-25 09:27:45 +0200
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2007-06-04 04:35:18 +0200
## Comment 7616
Date: 2007-05-25 09:27:45 +0200
From: @drstmane
On the 1GB Solaris 10/x86 machine in our testing pool,
PFtijah tests test_nllr_manyqueries & test_lms_manyqueries fail with
========
Assertion failed: p_command, file /export/scratch1/monet/monet.GNU.32.32.d.1901/pathfinder/modules/pftijah/nexi_generate_plan.c, line 184, function CAS_plan_gen
========
(this was a segfault before I added the assert()'s yesterday)
after Mserver has grown to >1.2 GB, eating up all physical memory and swap of the machine.
While Solaris 10/x86 is admittedly not our most prominent target platform, this reveals two (IMHO major) problems in the PFtijah code.
(1)
In many places, the code does not check at all (except from the assert()'s I added yesterday) whether alloc() (in any of its flavours, e.g., malloc(), calloc(), GDKmalloc()) was successful (i.e., returned a valid pointer to allocated memory) or failed (returning a NULL pointer).
(2)
PFtijah (at least for these two tests) does requires larger amounts of memory but (in addition to (1)) does not switch to explicit use of virtual memory and/or memory-mapped files in case the physical-/swap- memory is (about to be) exceeded.
Consistently using GDK*alloc() instead of mixing GDK*alloc() with plain *alloc() *might* improve this situation.
## Comment 7617
Date: 2007-05-25 09:29:15 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Cf.,
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/modules_pftijah/test_lms_manyqueries.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/modules_pftijah/test_nllr_manyqueries.err.00.html
## Comment 7618
Date: 2007-05-25 09:31:03 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Possibly related:
[ 1664237 ] PF/tijah: test_nllr fails on Solaris
http://sourceforge.net/tracker/index.php?func=detail&aid=1664237&group_id=56967&atid=482468
## Comment 7619
Date: 2007-05-25 14:47:05 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
Changed all old calloc()'s from Vojkans code and their corresponding free()'s into GDK calls.
Added a couple of error handlers where appropriate and left some asserts() in place.
There are still a couple of OS mallocs left in the flex generated code and the snowball generated code but I see other flexers/parsers also still use the OS memory managers.
Let's see what happens in the testWeb.
## Comment 7620
Date: 2007-06-04 16:35:18 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
There are now clean errors generated when all memory is gone. After this I made the testset a little bit smaller make it also run on the small Solaris machine. I now consider the bug fixed and closed.
## Comment 7621
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1725352 at http://sourceforge.net/support/tracker.php?aid=1725352
| PFtijah: failing alloc() causes assertion failure (segfault) | https://api.github.com/repos/MonetDB/MonetDB/issues/1362/comments | 0 | 2020-11-30T09:34:49Z | 2024-06-27T11:24:56Z | https://github.com/MonetDB/MonetDB/issues/1362 | 753,314,913 | 1,362 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-25 09:22:54 +0200
From: Colin Foss <<cgfoss>>
To: @njnes
Version: -- development
Last updated: 2007-05-27 05:15:31 +0200
## Comment 7612
Date: 2007-05-25 09:22:54 +0200
From: Colin Foss <<cgfoss>>
During COPY INTO of several large input files (280K lines each) with a table defined with six column compound primary key an error is produced:
"ERROR = !MALException:batmtime.EQ:Unexpected input type"
If I drop the primary key and use only a unique index (with the same columns) the data loads properly.
If I load only the file that produced the error, the error does not occur until the next COPY INTO of another file.
MonetDB5 was built using the monetdb-install.sh script and http://monetdb.cwi.nl/testing/projects/monetdb/Current/.DailyBuilds./20070524/SourceTarBalls/
i.e. --enable-sql --nightly=current
I have attached a tar file containing an awk script to generate the data and some shell scripts to run the test.
First, run "generate_data.csh". This will produce four files called file00.tab - file03.tab in the current directory.
Next, edit the pkey_copy_1.sql and pkey_copy_2.sql files and update the file path to match your system.
Finally, run "invalid_input_type_bug.csh" to see the error I have reported.
My work-around is to define an identical unique index to the desired primary key.
## Comment 7613
Date: 2007-05-25 16:58:33 +0200
From: Colin Foss <<cgfoss>>
Logged In: YES
user_id=420513
Originator: YES
I'm running on CentOS 4.4 64bit.
Linux 2.6.9-22.ELsmp 1 SMP Sat Oct 8 21:32:36 BST 2005 x86_64 x86_64 x86_64 GNU/Linux
## Comment 7614
Date: 2007-05-27 17:15:31 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed problem in batmtime module.
Test is added on the M5 level (src/modules/kernel/Tests/time01.mal)
## Comment 7615
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1725349 at http://sourceforge.net/support/tracker.php?aid=1725349
| large COPY INTO with primary key | https://api.github.com/repos/MonetDB/MonetDB/issues/1361/comments | 0 | 2020-11-30T09:34:46Z | 2024-06-27T11:24:55Z | https://github.com/MonetDB/MonetDB/issues/1361 | 753,314,873 | 1,361 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-24 05:42:16 +0200
From: Franco Guidoli <<feg624>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-05-24 05:57:37 +0200
## Comment 7609
Date: 2007-05-24 17:42:16 +0200
From: Franco Guidoli <<feg624>>
Hi all: I'm trying to connect to MonetDB via ODBC from PowerBuilder, but during connection it gets into a "loop" (that's what I think after viewing the ODBC driver log which I'm sending now).
I've tested that in a WinXP box and in a Win2k box.
Thanks in advance.
Franco
## Comment 7610
Date: 2007-05-24 17:57:37 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Thanks for reporting this. (The reason to report this here is so that we can track the bug and have something to refer to in the What's Fixed section in a new release.)
I have fixed the bug in the CVS head.
I also changed odbcsample1.c to use SQLFetchScroll instead of SQLFetch in one of the loops so that the fix gets tested.
## Comment 7611
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1724989 at http://sourceforge.net/support/tracker.php?aid=1724989
| Cannot connect to MonetDB via ODBC from PowerBuilder†| https://api.github.com/repos/MonetDB/MonetDB/issues/1360/comments | 0 | 2020-11-30T09:34:43Z | 2024-06-27T11:24:54Z | https://github.com/MonetDB/MonetDB/issues/1360 | 753,314,851 | 1,360 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-24 01:50:09 +0200
From: Erik_Aleman <<erik_aleman>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2007-05-25 02:11:43 +0200
## Comment 7604
Date: 2007-05-24 13:50:09 +0200
From: Erik_Aleman <<erik_aleman>>
The following code demonstrates the problem I ran into:
let $data := <aap><b/><c/></aap>
for $i in $data/*
order by "" descending
return $i
-bash-3.1$ /opt/MonetDB/bin/MapiClient -lxq -huganda < test.xq
ERROR: interpret: no matching MIL operator to 'CTrefine_rev(BAT[oid,oid], str)'.
MAYBE YOU MEAN:
CTrefine_rev(BAT[oid,any], BAT[oid,any]) : BAT[oid,oid]
-bash-3.1$
## Comment 7605
Date: 2007-05-24 22:05:43 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Added test in pathfinder/tests/BugTracker/Tests/order-by_error.SF-1724774.*
Also the latest development version fails as reported.
Though the current error message is indeed a bug,
I actually don't know for sure, whether this is valid XQuery,
and if so, what the correct / expected result should be.
## Comment 7606
Date: 2007-05-24 23:18:42 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
some background info, the query reported was derived from the following query:
let $doc := <aap><a name="aap"/><a name="beer"/><a/></aap>
for $i in $doc/*
order by if ($i/@name) then exactly-one($i/@name) cast as xs:string else "" descending
return $i
i guess this looks like a valid query. and this query seems to produce valid output. only when for every item in the list the else-clause is returned, then the query produces the described bug, so for example:
let $doc := <aap><a/><a/><a/></aap>
for $i in $doc/*
order by if ($i/@name) then exactly-one($i/@name) cast as xs:string else "" descending
return $i
would produce the bug.
note also that in 'ascending' order the query works fine.
## Comment 7607
Date: 2007-05-25 02:11:43 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7608
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1724774 at http://sourceforge.net/support/tracker.php?aid=1724774
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: Error while using 'order by' | https://api.github.com/repos/MonetDB/MonetDB/issues/1359/comments | 0 | 2020-11-30T09:34:41Z | 2024-06-27T11:24:53Z | https://github.com/MonetDB/MonetDB/issues/1359 | 753,314,815 | 1,359 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-23 06:24:08 +0200
From: alxtoth <<alxtoth>>
To: @mlkersten
Version: -- development
Last updated: 2007-05-29 11:40:37 +0200
## Comment 7597
Date: 2007-05-23 18:24:08 +0200
From: alxtoth <<alxtoth>>
monetdb helper uses "ps" arguments unsupported on Mac OS X.
monetdb --stop
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
No server stopped
--
In case of "monetdb --start" , it enters endless loop
----
Version: MonetDB Server v5.0.0_beta2_1
Mac OS X 10.4.9
autoconf (GNU Autoconf) 2.59
automake (GNU automake) 1.9.6
libtool= Apple Computer, Inc. version cctools-622.5
gcc=powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build
5367)
## Comment 7598
Date: 2007-05-23 18:55:09 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
unfortunately, the MonetDB helper scripts are highly experimental, and not written with any cross-platform intentions at all for the moment.
## Comment 7599
Date: 2007-05-23 23:56:10 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
This request belongs to the long-standing issue
to create a real and portable Mguardian.
## Comment 7600
Date: 2007-05-25 01:04:33 +0200
From: alxtoth <<alxtoth>>
Logged In: YES
user_id=1737757
Originator: YES
Few suggestions:
-replace " ps --no-heading --format pid -p $pid " with " ps -o pid -p $pid | tail -n 1 "
-replace " ps -p $pid -o comm= " with " ps c -o command -p $pid | tail -n "
This ps syntax runs on Mac (possibly more BSDs), and on Linux (Red Hat 2.6.9-42.0.10.ELsmp )
Still, there is somewhere else a ps, because there is error message, and infinite loop.
BR Alex
------------------------------------
$monetdb --start --dbinit="include sql;sql.start();"
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
Starting database server 'demo'... ok
mguardian working on /Users/alex/MonetFix//var/MonetDB5/dbfarm
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
macu:~/Desktop/MonetDB-Venus-SuperBall-SR1/MonetDB5-server/MonetDB5/src/tools alex$ usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
usage: sleep seconds
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
!FATAL: GDKlockHome: Database lock '.gdk_lock' denied
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
usage: sleep seconds
ps: illegal option -- -
usage: ps [-aACcehjlmMrSTuvwx] [-O|o fmt] [-p pid] [-t tty] [-U user]
ps [-L]
usage: sleep seconds
## Comment 7601
Date: 2007-05-25 07:20:44 +0200
From: alxtoth <<alxtoth>>
Logged In: YES
user_id=1737757
Originator: YES
Forgot to atach the modified files, from MonetDB5-server/MonetDB5/src/tools/
BR, Alex
File Added: mguardian_monetdb.zip
## Comment 7602
Date: 2007-05-29 11:40:36 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Alex,
Thanks for the patches! They have been included in the source tree.
At some point in the future we (will) have to redesign the mguardian
into a C-program that has better functionality. And the monetdb script
should be re-cast into a simple web-application to control the
state of a collection of Mservers.
## Comment 7603
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1724358 at http://sourceforge.net/support/tracker.php?aid=1724358
| monetdb helper uses "ps" arguments unsupported on Mac | https://api.github.com/repos/MonetDB/MonetDB/issues/1358/comments | 0 | 2020-11-30T09:34:38Z | 2024-06-27T11:24:52Z | https://github.com/MonetDB/MonetDB/issues/1358 | 753,314,787 | 1,358 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-23 02:49:09 +0200
From: alxtoth <<alxtoth>>
To: @njnes
Version: -- development
Last updated: 2007-05-29 02:00:07 +0200
## Comment 7592
Date: 2007-05-23 14:49:09 +0200
From: alxtoth <<alxtoth>>
Test script bellow has transactions, and the one statement outside any transaction is not executed (INSERT negative numbers, easy to spot).
MonetDB does not execute all the SQL in batch mode, and does not report error message. Confusingly, code runs fine in interactive mode.
Save the script bellow as 1.sql
-----------------
start transaction;
create table tbl_xy (x int , y int);
insert into tbl_xy values (0,0);
insert into tbl_xy values (1,1);
commit;
start transaction;
insert into tbl_xy values (2,2);
insert into tbl_xy values (3,3);
insert into tbl_xy values (4,4);
commit;
start transaction;
insert into tbl_xy values (5,5);
insert into tbl_xy values (6,6);
insert into tbl_xy values (7,7);
commit;
insert into tbl_xy values (-1, -100);
start transaction;
insert into tbl_xy values (8,8);
insert into tbl_xy values (9,9);
insert into tbl_xy values (10,10);
commit;
----------------
Execution of script:
macu:~/Desktop alex$ mclient -lsql -umonetdb ~/Desktop/1.sql
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
[ 1 ]
--------------------
Looking for negative numbers, none found:
macu:~/Desktop alex$ mclient -lsql -umonetdb
sql>select * from tbl_xy;
% sys.tbl_xy, sys.tbl_xy table_name
% x, y name
% int, int type
% 2, 2 length
[ 0, 0 ]
[ 1, 1 ]
[ 2, 2 ]
[ 3, 3 ]
[ 4, 4 ]
[ 5, 5 ]
[ 6, 6 ]
[ 7, 7 ]
[ 8, 8 ]
[ 9, 9 ]
[ 10, 10 ]
Version: MonetDB Server v5.0.0_beta2_1
## Comment 7593
Date: 2007-05-24 00:21:51 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
The problem here is related to the use of file mode, ie mclient file instead of mclient < file.
In this mode the full file is send in 'one statement'. So auto commits are only done over this one (multi statement) statement. I know this sounds vage but its a compromise for the jdbc batch mode. We will
have more discussions to solve the issue correctly.
## Comment 7594
Date: 2007-05-29 12:35:51 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
for convenience:
marked as SQL bug.
## Comment 7595
Date: 2007-05-29 14:00:07 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed. The file mode is now correctly handled by the sql reader.
Added test to src/tests/BugTracker/Tests/auto_commit.SF-1724151
## Comment 7596
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1724151 at http://sourceforge.net/support/tracker.php?aid=1724151
| SQL: mixing auto-commit with explicit transactions | https://api.github.com/repos/MonetDB/MonetDB/issues/1357/comments | 0 | 2020-11-30T09:34:35Z | 2024-06-27T11:24:51Z | https://github.com/MonetDB/MonetDB/issues/1357 | 753,314,754 | 1,357 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-23 01:55:35 +0200
From: @sjoerdmullender
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2007-05-24 09:30:20 +0200
## Comment 7588
Date: 2007-05-23 13:55:35 +0200
From: @sjoerdmullender
Trying to compile the attached file (using "pf tpox1idx.xq") with the 1.46 revision of coreopt.brg, all available memory and swap space is consumed. Using the previous revision (1.45) does not have this problem.
## Comment 7589
Date: 2007-05-24 01:00:46 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
A simple test
========
n=0 ; t=7 ; for h in `grep -n '^};$' tpox1idx.xq | cut -d: -f1` ; do n=$[n+1] ; echo $n ; head -7 tpox1idx.xq > /tmp/$n.xq ; head -$h tpox1idx.xq | tail -$[$h-$t-1] | tee -a /tmp/$n.xq | head -1 ; pf /tmp/$n.xq > /tmp/$n.mil ; echo $n ; read ; t=$h ; done
========
reveals that the following two function declarations seem to trigger the problem:
========
declare function tpox:get_security_price($sym as xs:string) as xs:anyNode*
{
for $s in pf:text(pf:collection("security.xml"), $sym)/parent::s:Symbol/parent::s:Security[./parent::security]
return
<print>The open price of the security "{$s/s:Name/text()}" is {$s/s:Price/s:PriceToday/s:Open/text()} dollars</print>
};
========
declare function tpox:cust_sold_security($zip as xs:string) as xs:anyNode*
{
for $cust in pf:text(pf:collection("custacc.xml"), $zip)/parent::c:PostalCode/parent::c:Address/parent::c:Addresses/parent::c:Customer[./parent::custacc]
for $ord in (for $tmp in $cust/c:Accounts/c:Account/@id/fn:string(.)
return pf:attribute(pf:collection("order.xml"), $tmp, "", "Acct"))/self::o:Order[@Side = "2" and ./parent::o:FIXML/parent::order]
order by exactly-one($cust/c:Name/c:LastName/text())
return
<Customer>
{$cust/c:Name/c:LastName/text()} -
{$cust/c:Addresses/c:Address[@primary = "Yes"]/c:Phones/c:Phone[@primary = "Yes"]}
</Customer>
};
========
## Comment 7590
Date: 2007-05-24 09:30:20 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
Only mark a pattern rewritten if a rewrite was applied.
This bug should be fixed now. Sorry for the inconvenience!
## Comment 7591
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1724118 at http://sourceforge.net/support/tracker.php?aid=1724118
| PF: coreopt optimization eats all available memory and swap | https://api.github.com/repos/MonetDB/MonetDB/issues/1356/comments | 0 | 2020-11-30T09:34:33Z | 2024-06-27T11:24:50Z | https://github.com/MonetDB/MonetDB/issues/1356 | 753,314,727 | 1,356 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-23 07:47:05 +0200
From: rafb <<rafb>>
To: Jens Teubner <<teubner>>
Version: -- development
Last updated: 2007-05-28 12:56:45 +0200
## Comment 7577
Date: 2007-05-23 07:47:05 +0200
From: rafb <<rafb>>
The following query is valid, I think, but generates a warning. The warning is subsequently treated as an error and the query is never executed.
Here's the query:
let $input := (
let $t := "icro"
let $input :=
(let $t := "soft"
for $f in doc("files.xml")//file[@name[contains(., $t)]]
return $f)
for $f in $input[@name[contains(., $t)]]
return $f)
return $input
Here is the result:
MAPI = monetdb@localhost:50000
QUERY = let $input := (
ERROR = !warning: variable reuse: at (4,10-4,11): $t
I'll attach the input document (files.xml).
Renaming one $t solves the problem. I can do such renaming at query-generation time, but it requires a little bit of extra code to generate unique temporaries. Also, treating warnings as errors is probably not what you want.
Raoul
## Comment 7578
Date: 2007-05-23 07:48:33 +0200
From: rafb <<rafb>>
Logged In: YES
user_id=1797522
Originator: YES
Fixed typo in summary.
## Comment 7579
Date: 2007-05-23 11:21:43 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Originator: NO
Variable re-usage should indeed be a warning only (although
it had been discussed in the W3C Working Group to make this
a real error).
The Pathfinder compiler correctly handles this situation as
a warning only. So this appears to be a server-side XQuery
problem only.
## Comment 7580
Date: 2007-05-23 11:25:46 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
ERROR = !warning: variable reuse: at (4,10-4,11): $t
That string clearly indicates an error, not a warning (note the starting !).
In the latest Mapi protocol, a warning can be issued using a (which is also known as a comment, or info).
My guess is that what happens here is that the client just aborts because it sees an error, while the server did not intend this to happen.
## Comment 7581
Date: 2007-05-23 11:49:32 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
As a side note for info (possibly just off-topic):
The Mapi protocol has then apparently been changed without keeping GDK (and the complete server side) aligned / in sync:
MonetDB/src/gdk/gdk.mx says
========
define GDKWARNING "!WARNING: "
define GDKERROR "!ERROR: "
define GDKMESSAGE "!OS: "
define GDKFATAL "!FATAL: "
========
and GDKerror & GDKwarning use GDKERROR & GDKWARNING, respectively.
## Comment 7582
Date: 2007-05-23 11:52:20 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
as far as I can tell, ! has always indicated an error, so in that respect nothing has changed
## Comment 7583
Date: 2007-05-25 12:04:51 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Jens: would you mind signaling warnings with
warning: ...
?
## Comment 7584
Date: 2007-05-27 10:18:13 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Originator: NO
Peter,
the `!' is not generated within the Pathfinder compiler code.
The error string generated by the Pathfinder compiler is defined
in compiler/oops.c:80. The runtime module code around
runtime/pathfinder.mx:4281 then adds the `!'. This happens
unconditionally, i.e., regardless whether the compiler produced
an error or just a warning.
I see two ways of fixing this with reasonable effort:
(1) In the Pathfinder compiler, indicate warnings explicitly
(e.g., with `warning'). The runtime code would then
interpret the `' and not add a `!'. This feels a bit
dirty to me.
(2) Completely suppress warnings in the Pathfinder compiler
if it is invoked in server-side mode. This might make
less trouble, but users will no longer get (helpful)
warnings.
Any opinions?
## Comment 7585
Date: 2007-05-28 00:56:45 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed by instead of ! prefixing with for lines atring with 'warning'
this is also a hack, but only affects the runtime
## Comment 7586
Date: 2007-05-28 12:04:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Bug if covered by tests
pathfinder/tests/XQuery/Tests/let_for[23].xq
Cf.:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora6/tests_XQuery/let_for2.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora6/tests_XQuery/let_for2.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora6/tests_XQuery/let_for3.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora6/tests_XQuery/let_for3.err.00.html
## Comment 7587
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1723931 at http://sourceforge.net/support/tracker.php?aid=1723931
| XQuery compiler warning treated as error | https://api.github.com/repos/MonetDB/MonetDB/issues/1355/comments | 0 | 2020-11-30T09:34:30Z | 2024-06-27T11:24:49Z | https://github.com/MonetDB/MonetDB/issues/1355 | 753,314,699 | 1,355 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-23 02:09:31 +0200
From: Venkatesh Mantha <<dreambubble>>
To: @njnes
Version: -- development
Last updated: 2007-05-24 12:02:52 +0200
## Comment 7574
Date: 2007-05-23 02:09:31 +0200
From: Venkatesh Mantha <<dreambubble>>
The following query is crashing the server.
select count(*)
from (
select col2, count(*) as row_count
from data_1to10
group by col2
order by row_count desc, col1
) sq
I have attached a PERL script which can be used to generate the required data for testing. I have also attached a table definition script and the load script.
Please use the attached PERL script to generate the data as follows:
./massrecs.pl > <filename>
You will need to edit the "test_load_tables.sql" script to update the filename according to the name given.
Please let me know if you are unable to execute the test for any reason.
Thanks,
- Venkatesh
## Comment 7575
Date: 2007-05-24 00:02:52 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed. Its not allowed to use non group by and non selection columns, in order by.
This is now checked correctly.
added a test groupby_orderby_nonselected.SF-1723863 to src/test/BugTracker
## Comment 7576
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1723863 at http://sourceforge.net/support/tracker.php?aid=1723863
| SQL Query Crashes Server - CRITICAL | https://api.github.com/repos/MonetDB/MonetDB/issues/1354/comments | 0 | 2020-11-30T09:34:27Z | 2024-06-27T11:24:48Z | https://github.com/MonetDB/MonetDB/issues/1354 | 753,314,675 | 1,354 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-22 11:28:07 +0200
From: alxtoth <<alxtoth>>
To: @njnes
Version: -- development
Last updated: 2007-05-24 12:04:07 +0200
## Comment 7571
Date: 2007-05-22 23:28:07 +0200
From: alxtoth <<alxtoth>>
MonetDB does not evaluate rand() for every row. For example when generating a join with random column, all the values from column are constant. Script to reproduce:
create function rand() returns int external name mmath.rand;
create table nr (nr int);
insert into nr values (1);
insert into nr values (2);
insert into nr values (3);
insert into nr values (4);
insert into nr values (5);
insert into nr values (6);
insert into nr values (7);
insert into nr values (8);
insert into nr values (9);
insert into nr values (10);
sql>select rand(), nr from nr;
% ., sys.nr table_name
% rand, nr name
% int, int type
% 9, 2 length
[ 530511967, 1 ]
[ 530511967, 2 ]
[ 530511967, 3 ]
[ 530511967, 4 ]
[ 530511967, 5 ]
[ 530511967, 6 ]
[ 530511967, 7 ]
[ 530511967, 8 ]
[ 530511967, 9 ]
[ 530511967, 10 ]
--------
Version: MonetDB Server v5.0.0_beta2_1
Mac OS X 10.4.9
autoconf (GNU Autoconf) 2.59
automake (GNU automake) 1.9.6
libtool= Apple Computer, Inc. version cctools-622.5
gcc=powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build 5367)
## Comment 7572
Date: 2007-05-24 00:04:07 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Fixed by introducing functions with side effects.
Added a test in src/tests/BugTracker/groupby_orderby_nonselected.SF-1723863.sql
## Comment 7573
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1723791 at http://sourceforge.net/support/tracker.php?aid=1723791
| rand() is not executed for every row | https://api.github.com/repos/MonetDB/MonetDB/issues/1353/comments | 0 | 2020-11-30T09:34:25Z | 2024-06-27T11:24:47Z | https://github.com/MonetDB/MonetDB/issues/1353 | 753,314,638 | 1,353 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-22 08:16:30 +0200
From: @drstmane
To: Manuel Mayr <<mayrm>>
Version: -- development
Last updated: 2007-05-24 07:16:57 +0200
## Comment 7566
Date: 2007-05-22 08:16:30 +0200
From: @drstmane
Last night's testing revealed (see testing result email):
When assertions are disabled (`configure --disable-assert ...`) --- default when producing RPMs (`make rpm`) ---,
the new compiler/sql/lalg2sql.brg fails to compile with:
========
gcc:
/export/scratch1/monet/monet.GNU.32.32.d.25029/pathfinder/compiler/sql/lalg2sql.brg: In function 'llscj_attr_axis':
/export/scratch1/monet/monet.GNU.32.32.d.25029/pathfinder/compiler/sql/lalg2sql.brg:3977: warning: unused parameter 'frag'
make[6]: *** [libsql_la-lalg2sql.lo] Error 1
========
icc:
/export/scratch1/monet/monet.ntv.64.64.d.25029/pathfinder/compiler/sql/lalg2sql.brg(3977): error 869: parameter "frag" was never referenced
const PFla_op_t * frag, struct PFsql_alg_ann_t *ann)
^
compilation aborted for lalg2sql.c (code 2)
make[6]: *** [libsql_la-lalg2sql.lo] Error 1
========
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.Makes3/index_short.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.GNU.64.64.d-Fedora6/rpm.out
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.ntv.64.64.d-RedHat4WS/rpm.out
1) Is 'frag' indeed required as parameter in the signature of 'llscj_attr_axis()' (e.g., for "canonical" reasons?
2) If so, is it correct, that 'llscj_attr_axis()' does not use 'frag' at all (except from the assert)?
3) If so, you should indecate the compiler that you know what you're doing by adding a "fake usage" of 'frag': `(void)frag; /* pacify compiler */`, either next to the assert (incase that is indeed required), or instead of the assert (incase that is not required, here).
a) In case 1) yields a "No", 'frag' (and the assert) should simply be removed.
b) In case 2) yields a "No", the code obviously needs to be fixed.
## Comment 7567
Date: 2007-05-22 09:11:41 +0200
From: Manuel Mayr <<mayrm>>
Logged In: YES
user_id=1629542
Originator: NO
You're of course right,
I don't see a reason why llscj_attr should use frag as parameter,
i fixed the buggy part of the code, and it works fine for me.
Thx for your advice,
regards
Manuel
## Comment 7568
Date: 2007-05-22 13:34:28 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Thanks for the quick fix --- we should close this one, once tomorrows testing confirms the fix.
No need to thank me; thank the compilers and the nightly testing!
I just repeated what they told us ...
## Comment 7569
Date: 2007-05-24 07:16:57 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Testing confirmed fix.
Thanks, again!
## Comment 7570
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1723217 at http://sourceforge.net/support/tracker.php?aid=1723217
| PF: compiler/sql/lalg2sql.brg fails to compile w/o assertion | https://api.github.com/repos/MonetDB/MonetDB/issues/1352/comments | 0 | 2020-11-30T09:34:22Z | 2024-06-27T11:24:46Z | https://github.com/MonetDB/MonetDB/issues/1352 | 753,314,598 | 1,352 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-20 08:41:10 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-25 12:28:57 +0200
## Comment 7564
Date: 2007-05-20 20:41:10 +0200
From: @drstmane
After some of the checkins on Friday May 18 2007,
test tests/W3C_use_cases/XQUF/R/Tests/Q4.milS
produces error message
"!ERROR: merged_union: BAT 4 (0 BUNs) must have the same size as BAT 2 (1 BUNs)."
instead of
"!ERROR: err:FORG0005: function fn:exactly-one expects exactly one value."
While the new error message might be "correct", the old one is IMHO more informative; hence it might be worth checking why the error message has changed.
I haven't had a closer look at it, yet, but might be able to do so tomorrow...
## Comment 7565
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1722257 at http://sourceforge.net/support/tracker.php?aid=1722257
| PF: tests/W3C_use_cases/XQUF/R/Q4 err-msg changed on May 18 | https://api.github.com/repos/MonetDB/MonetDB/issues/1351/comments | 0 | 2020-11-30T09:34:19Z | 2024-06-27T11:24:45Z | https://github.com/MonetDB/MonetDB/issues/1351 | 753,314,568 | 1,351 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-20 10:51:22 +0200
From: @drstmane
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2007-05-25 02:17:05 +0200
## Comment 7560
Date: 2007-05-20 10:51:22 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_WebSite/books1.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_WebSite/books2.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_WebSite/books3.out.00.html
incomplete result
See
[ 1712461 ] PF: several tests fail since 2007.05.03
http://sourceforge.net/tracker/index.php?func=detail&aid=1712461&group_id=56967&atid=482468
for details
(I was too quick to close that one...)
## Comment 7561
Date: 2007-05-25 00:01:34 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
perhaps this bug is caused by the fact that the query tries to load a document ("http://monetdb.cwi.nl/XQuery/files/bib.xml") from the internet. do all test-machines have access to the internet? if not, then the document needs to be shredded before the test begins.
what seems strange is that the query doesn't report an error saying it cannot find the document, instead it assumes a different (empty?) document.
## Comment 7562
Date: 2007-05-25 02:17:05 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
wouter was right, temporary read-only documents now get indexed
in the past they did not get an index
with the value-indxing, this policy was changed. but the index creation would make all read-only indexes persistent. this ends up pollution your repository with a 1_nsloc_nid andrary 1_vx_hsh_nid bats, where 1 is a temporary docid. this in turn causes a subsequent query to use those bogus bats instead.
## Comment 7563
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1722078 at http://sourceforge.net/support/tracker.php?aid=1722078
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| PF: WebSite tests books[123] fail since 2007.05.03 | https://api.github.com/repos/MonetDB/MonetDB/issues/1350/comments | 0 | 2020-11-30T09:34:17Z | 2024-06-27T11:24:44Z | https://github.com/MonetDB/MonetDB/issues/1350 | 753,314,545 | 1,350 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-19 10:54:12 +0200
From: @drstmane
To: Maurice van Keulen <<mvankeulen>>
Version: -- development
Last updated: 2007-05-25 10:42:24 +0200
## Comment 7554
Date: 2007-05-19 10:54:12 +0200
From: @drstmane
Before this info gets lost:
The latest checkin (see below) causes the following tests to fail with "!ERROR: Division by zero":
tests/BugTracker/Tests/predicate_selects_too_few_nodes.SF-1636588.xq
tests/BugTracker/Tests/id-function_fails_to_locate_some_kinds_of_nodes.SF-1698498.xq
tests/BugsViaSourgeforce/ID.1652527.xq
tests/BugsViaSourgeforce/ID.1648854.xq
===================================================================
2007/05/19 - boncz: pathfinder/compiler/mil/milprint_summer.c,1.382
pathfinder/runtime/pathfinder.mx,1.362
pathfinder/runtime/pf_support.mx,1.241
- plug a known remaining hole in the index maintenance
=> if a document from a multi-doc collection was deleted (del-doc)
its nodes would still be found by the indices (value+nsloc).
until we implement some vacuum cleaner, deleted document nodes leak
for now, we post-filter the index lookup using a new ws_docfilter()
function. This one now always makes sure that the documents from
which the returned nodes stem are added to the OPEN_DOCID bats in the ws.
- accelerated attribute lookup and fn:root() for multi-doc collections
because in the past we needed for those to swizzle/sort the entire
FRAG_ROOT bat (in case of TPOX millions of docs=buns).
now get_root only examines the *open* documents (OPEN_DOCID in the ws).
it still needs to join with FRAG_ROOT, but this one can be kept sorted
*both* on head (docid) and on tail (nid) - both docid and nid are
assigned increasingly and never change.
- in the temporary container, FRAG_ROOT head contains 0@0 now iso nil
(to keep it also sorted).
- some more accurate decision making in htordered_thetajoin
===================================================================
## Comment 7555
Date: 2007-05-19 16:39:43 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
yes, indeed a problem was left with the recognition of fragment roots in the temp document container.
Errors are gone, but one diference remains in the output of tests/BugsViaSourgeforce/ID.1652527.xq
This is one of Maurice Probabiistic XML tests.
I think it has to do with the following. With these changes, and previous ones, there is a change in the implementation of fn:id(id,node), regarding temporary XML fragments:
In the past, for finding temporary nodes by id it did not matter $node was in the *same* temporary XML fragment, as long as it was temporary.
Now, the restriction is again that $node is from the same fragment. I think this is more in line with the specs (although they are vague on this issue -- presumably because XPath 1.0 did not contsruct nodes).
Maurice, three questions:
(1) can you live with the new semantics?
(2) is the output correct given the new semantics?
(3) could you approve the current output?
## Comment 7556
Date: 2007-05-20 02:49:48 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
yes, indeed a problem was left with the recognition of fragment roots in the temp document container.
Errors are gone, but one diference remains in the output of tests/BugsViaSourgeforce/ID.1652527.xq
This is one of Maurice Probabiistic XML tests.
I think it has to do with the following. With these changes, and previous ones, there is a change in the implementation of fn:id(id,node), regarding temporary XML fragments:
In the past, for finding temporary nodes by id it did not matter $node was in the *same* temporary XML fragment, as long as it was temporary.
Now, the restriction is again that $node is from the same fragment. I think this is more in line with the specs (although they are vague on this issue -- presumably because XPath 1.0 did not contsruct nodes).
Maurice, three questions:
(1) can you live with the new semantics?
(2) is the output correct given the new semantics?
(3) could you approve the current output?
## Comment 7557
Date: 2007-05-20 11:09:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
see also
[ 1712451 ] PF: test fails on assert_order() since 2007.03.31
http://sourceforge.net/tracker/index.php?func=detail&aid=1712451&group_id=56967&atid=482468
and
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1652527.out.00.html
## Comment 7558
Date: 2007-05-25 10:42:24 +0200
From: Maurice van Keulen <<mvankeulen>>
Logged In: YES
user_id=654674
Originator: NO
Yes, I can live with the semantics that in "id($i,$node)", $i has to be an identifier of a node in the same fragment of $node. I also agree that this is closer to the XQuery standard. About the mentioned tests:
* ID.1652527: most probably indeed the result of the new semantics
* ID.1648854: gives correct output
* id-function_fails_to_locate_some_kinds_of_nodes.SF-1698498: more elborate version of previous test that checks pf:nid/id functionality for both transient and stored document nodes
* predicate_selects_too_few_nodes.SF-1636588: Not related to id semantics.
Closing the bug.
## Comment 7559
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1721701 at http://sourceforge.net/support/tracker.php?aid=1721701
| PF: "!ERROR: Division by zero" after latest checkins | https://api.github.com/repos/MonetDB/MonetDB/issues/1349/comments | 0 | 2020-11-30T09:34:15Z | 2024-06-27T11:24:43Z | https://github.com/MonetDB/MonetDB/issues/1349 | 753,314,514 | 1,349 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-17 09:28:05 +0200
From: Venkatesh Mantha <<dreambubble>>
To: @njnes
Version: -- development
Last updated: 2007-05-19 02:16:37 +0200
## Comment 7548
Date: 2007-05-17 21:28:05 +0200
From: Venkatesh Mantha <<dreambubble>>
Hi,
Please find attached a simple JAVA program that crashes the MonetDB server. This happens only when a "bind" variable is used in conjunction with CAST function.
You will need JDK and MonetDB JDBC jar file to compile and execute this program. You will also need to edit the program to update the connect information and the query based on some test table in your database.
Regards,
- Venkatesh
## Comment 7549
Date: 2007-05-18 09:37:35 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
execution of
PREPARE select id from data_1to10 where id > CAST(? AS INTEGER) limit 10;
seems to fail according to the bug reporter, not tested.
## Comment 7550
Date: 2007-05-18 09:48:51 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
And the argument setting is:
stmt.setString(1, "100");
## Comment 7551
Date: 2007-05-18 14:19:18 +0200
From: Venkatesh Mantha <<dreambubble>>
Logged In: YES
user_id=1794901
Originator: YES
You will need to edit the query so that you the table and column information exists. You can run this against any simple test table. The program fails "as such" because the table I used wouldn't not exist in your database.
Regards,
- Venkatesh
## Comment 7552
Date: 2007-05-19 14:16:37 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
fixed the crash. We tried to get the type of a '?' parameter, which at compile time isn't known yet.
As this is only needed for the error message we postpone it.
Added test script to BugTracker bind_with_cast.SF-1720934.sql
## Comment 7553
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1720934 at http://sourceforge.net/support/tracker.php?aid=1720934
| Bind Variable with CAST function Crashes Server - CRITICAL | https://api.github.com/repos/MonetDB/MonetDB/issues/1348/comments | 0 | 2020-11-30T09:34:12Z | 2024-06-27T11:24:42Z | https://github.com/MonetDB/MonetDB/issues/1348 | 753,314,488 | 1,348 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-16 10:29:04 +0200
From: Venkatesh Mantha <<dreambubble>>
To: @njnes
Version: -- development
Last updated: 2007-05-16 10:44:50 +0200
## Comment 7545
Date: 2007-05-16 22:29:04 +0200
From: Venkatesh Mantha <<dreambubble>>
WITH t AS (
SELECT * FROM tables )
SELECT * FROM t
If I execute the query again, it gives the following error:
Variable 't' allready declared
If I change the alias from "t" to "t2" it works for the first time and then gives the error on second execution.
- Venks
## Comment 7546
Date: 2007-05-16 22:44:50 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Added missing stack_push_frame and stack_pop_frame around the with queries.
Added the test to BugTracker with_only_once.SF-1720293.sql
## Comment 7547
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1720293 at http://sourceforge.net/support/tracker.php?aid=1720293
| WITH queries works only once! | https://api.github.com/repos/MonetDB/MonetDB/issues/1347/comments | 0 | 2020-11-30T09:34:09Z | 2024-06-27T11:24:41Z | https://github.com/MonetDB/MonetDB/issues/1347 | 753,314,453 | 1,347 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-15 11:50:44 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-05-16 09:40:04 +0200
## Comment 7541
Date: 2007-05-15 11:50:44 +0200
From: @drstmane
The new loop-lifted following step causes XPath q09 & XPathMark Q10 to fail with SEGFAULT or property error (inconsistent descriptor) on selected platforms:
XPath q09:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XPath/q09.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XPath/q09.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/ntv.64.64.d-RedHat4WS/benchmarks_XPath/q09.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/ntv.64.64.d-RedHat4WS/benchmarks_XPath/q09.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/benchmarks_XPath/q09.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/benchmarks_XPath/q09.err.00.html
XPathMark Q10 (only all 32-bit systems):
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XPathMark/Q10.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XPathMark/Q10.err.00.html
...
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/benchmarks_XPathMark/Q10.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SunOS5.10/benchmarks_XPathMark/Q10.err.00.html
## Comment 7542
Date: 2007-05-15 13:41:16 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
(Hopefully) fixed in CVS:
When extending the result bug, make sure that we also update the cursors accordingly.
To be closed, once nightly testing confirms the fix.
## Comment 7543
Date: 2007-05-16 09:40:04 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Testing confirmed fix.
## Comment 7544
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1719164 at http://sourceforge.net/support/tracker.php?aid=1719164
| PF: new loop-lifted following step still buggy? | https://api.github.com/repos/MonetDB/MonetDB/issues/1346/comments | 0 | 2020-11-30T09:34:07Z | 2024-06-27T11:24:41Z | https://github.com/MonetDB/MonetDB/issues/1346 | 753,314,426 | 1,346 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-14 04:17:26 +0200
From: @yzchang
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-17 12:36:42 +0200
## Comment 7536
Date: 2007-05-14 16:17:26 +0200
From: @yzchang
(In Current branch)
I don't know if the following is an expected behaviour:
$ mapiclnt -lx
"a.xml" is an updatable document in MXQ
xquery>doc("a.xml")
more><?xml version="1.0" encoding="utf-8"?>
<hello lang="en">
<world/>
</hello>
delete a node:
xquery>do delete doc("a.xml")/hello/world
deletion succeeded:
xquery>doc("a.xml")/hello
more><hello lang="en">
</hello>
Is this error an expected behaviour?
xquery>doc("a.xml")/hello/world
more>MAPI = monetdb@localhost:50000
QUERY = doc("a.xml")/hello/world
ERROR = !ERROR: interpret: if bit(nil): don't know what to do!
Or should the above query just return empty result?
Because select a not existing node returns empty
result:
xquery>doc("a.xml")/hello/doesnotexist
more>
xquery>
Regards,
Jennie
## Comment 7537
Date: 2007-05-17 01:27:41 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
filtering of nslc index deltas on only deltas that concern unq (ie the worthwhile ones), took head iso tail of unq in kintersect
this was a bit too quick hack applied by Sjoerd and me
## Comment 7538
Date: 2007-05-17 12:34:09 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
For completeness:
added test in pathfinder/tests/BugTracker/Tests/accessing_renamed_inserted_deleted_node.SF-1718622-1718635-1718709.*
fix confirmed.
## Comment 7539
Date: 2007-05-17 12:36:42 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
restored "Closed" & "Fixed" after accidentially undoing them.
## Comment 7540
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1718709 at http://sourceforge.net/support/tracker.php?aid=1718709
| PF: select deleted node results in error | https://api.github.com/repos/MonetDB/MonetDB/issues/1345/comments | 0 | 2020-11-30T09:34:04Z | 2024-06-27T11:24:40Z | https://github.com/MonetDB/MonetDB/issues/1345 | 753,314,393 | 1,345 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-14 03:07:13 +0200
From: @yzchang
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-17 12:36:34 +0200
## Comment 7530
Date: 2007-05-14 15:07:13 +0200
From: @yzchang
(This happens in the Current branch, I did not check the Stable branch)
Shred the following doc:
<hello lang="en">
<world/>
</hello>
into Mserver:
MonetDB>shred_doc("/ufs/zhang/xml/a.xml", "a.xml", "test", 10LL);
shredding with holes: pagesize=16384, fillsize=14746
Elapsed time = 001ms 495us [249us/node]
Shredded 1 XML document (a.xml), total time after commit=0.029s
Then start a MapiClient:
$ mapiclnt -lx
insert a new node:
xquery>do insert <cwi/> as first into doc("a.xml")/hello
more>
insertion succeeded:
xquery>doc("a.xml")
more><?xml version="1.0" encoding="utf-8"?>
<hello lang="en"><cwi/>
<world/>
</hello>
delete one of the children nodes of "hello":
xquery>do delete doc("a.xml")/hello/world
more>
deletion succeeded, the node is gone:
xquery>doc("a.xml")
more><?xml version="1.0" encoding="utf-8"?>
<hello lang="en"><cwi/>
</hello>
Select the node 'cwi' returns empty result:
xquery>doc("a.xml")/hello/cwi
more>
but the node 'cwi' is really there:
xquery>doc("a.xml")/hello
more><hello lang="en"><cwi/>
</hello>
Again, this problem disappears after start Mserver:
$ mapiclnt -lx
xquery>doc("a.xml")/hello/cwi
more><cwi/>
xquery>
Regards,
Jennie
## Comment 7531
Date: 2007-05-14 15:57:46 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Possibly (most probably?) related to
1718622 'PF: strange behaviour of "rename"'
http://sourceforge.net/tracker/index.php?func=detail&aid=1718622&group_id=56967&atid=482468
and
1712461 'PF: several tests fail since 2007.05.03'
http://sourceforge.net/tracker/index.php?func=detail&aid=1712461&group_id=56967&atid=482468
## Comment 7532
Date: 2007-05-17 01:38:22 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
- nsloc index maintenance code forgot when pruning about new qnames
## Comment 7533
Date: 2007-05-17 12:34:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
For completeness:
added test in pathfinder/tests/BugTracker/Tests/accessing_renamed_inserted_deleted_node.SF-1718622-1718635-1718709.*
fix confirmed.
## Comment 7534
Date: 2007-05-17 12:36:34 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
restored "Closed" & "Fixed" after accidentially undoing them.
## Comment 7535
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1718635 at http://sourceforge.net/support/tracker.php?aid=1718635
| PF: node's sibling not accessable after deletion of the node | https://api.github.com/repos/MonetDB/MonetDB/issues/1344/comments | 0 | 2020-11-30T09:34:01Z | 2024-06-27T11:24:39Z | https://github.com/MonetDB/MonetDB/issues/1344 | 753,314,348 | 1,344 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-14 02:54:55 +0200
From: @yzchang
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-17 12:36:28 +0200
## Comment 7522
Date: 2007-05-14 14:54:55 +0200
From: @yzchang
After I have renamed a node successfully, I can not select this node by its new name, instead, I can select this node by its old name. This strange behaviour disappears after Mserver has be restarted.
This happens only in the Current branch, not in the Stable branch.
More precisely, shred the following document:
<hello lang="en">
<world/>
</hello>
in Mserver:
MonetDB>shred_doc("/ufs/zhang/xml/a.xml", "a.xml", "test", 10LL);
shredding with holes: pagesize=16384, fillsize=14746
Elapsed time = 002ms 083us [347us/node]
Shredded 1 XML document (a.xml), total time after commit=0.034s
Then start a MapiClient:
$ mapiclnt -lx
the original doc:
xquery>doc("a.xml")
more><?xml version="1.0" encoding="utf-8"?>
<hello lang="en">
<world/>
</hello>
rename a node
xquery>do rename doc("a.xml")/hello/world into "cwi"
more>
rename succeeded:
xquery>doc("a.xml")
more><?xml version="1.0" encoding="utf-8"?>
<hello lang="en">
<cwi/>
</hello>
select the node by its new name, the result is empty:
xquery>doc("a.xml")/hello/cwi
more>
select the node by its old name, we get the node:
xquery>doc("a.xml")/hello/world
more><cwi/>
After restarting Mserver, this strange behaviour disappears:
$ mapiclnt -lx
xquery>doc("a.xml")/hello/world
more>
xquery>doc("a.xml")/hello/cwi
more><cwi/>
Regards,
Jennie
## Comment 7523
Date: 2007-05-14 14:59:07 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
looks very much the same as this one:
1712461 "PF: several tests fail since 2007.05.03"
http://sourceforge.net/tracker/index.php?func=detail&aid=1712461&group_id=56967&atid=482468
... an index update/maintenance problem?
## Comment 7524
Date: 2007-05-14 15:53:52 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: YES
Peter also said it is probably a index problem.
## Comment 7525
Date: 2007-05-14 15:57:50 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
See also
1718635 "PF: node's sibling not accessable after deletion of the node"
http://sourceforge.net/tracker/index.php?func=detail&aid=1718635&group_id=56967&atid=482468
## Comment 7526
Date: 2007-05-17 01:43:00 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
indeed fixed by FIX of 1718635 "PF: node's sibling not accessable after deletion of the node"
## Comment 7527
Date: 2007-05-17 12:34:22 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
For completeness:
added test in pathfinder/tests/BugTracker/Tests/accessing_renamed_inserted_deleted_node.SF-1718622-1718635-1718709.*
fix confirmed.
## Comment 7528
Date: 2007-05-17 12:36:28 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
restored "Closed" & "Fixed" after accidentially undoing them.
## Comment 7529
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1718622 at http://sourceforge.net/support/tracker.php?aid=1718622
| PF: strange behaviour of "rename" | https://api.github.com/repos/MonetDB/MonetDB/issues/1343/comments | 0 | 2020-11-30T09:33:59Z | 2024-06-27T11:24:38Z | https://github.com/MonetDB/MonetDB/issues/1343 | 753,314,316 | 1,343 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-13 07:31:40 +0200
From: @mlkersten
To: @njnes
Version: -- development
Last updated: 2007-05-13 08:14:19 +0200
## Comment 7519
Date: 2007-05-13 19:31:40 +0200
From: @mlkersten
-- test derived from experience Cubulus OLAP tools
create table t( i decimal(10,4) default null);
insert into t values('1.12345');
insert into t values('1.1234');
insert into t values('1.123');
insert into t values('1.12');
insert into t values('1.1');
insert into t values('1.');
select * from t;
drop table t;
This script generates a floating point exception on
the first insert statement. And produces wrong
results when you start at the next.
insert into t values('1.1234');
insert into t values('1.123');
insert into t values('1.12');
insert into t values('1.1');
insert into t values('1.');
select * from t;
&2 1
>&2 1
>&2 1
>&2 1
>&2 1
>&1 0 5 1 5
% sys.t table_name
% i name
% decimal type
% 12 length
[ 1.1234 ]
[ 1.1230 ]
[ 0.2240 ]
[ 0.0330 ]
[ 0.0040 ]
## Comment 7520
Date: 2007-05-13 20:14:19 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
fixed dec2dec convert function.
Added test decimal.SF-1718253 to BugTracker directory
## Comment 7521
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1718253 at http://sourceforge.net/support/tracker.php?aid=1718253
| SQL: decimal handling errors | https://api.github.com/repos/MonetDB/MonetDB/issues/1342/comments | 0 | 2020-11-30T09:33:56Z | 2024-06-27T11:24:37Z | https://github.com/MonetDB/MonetDB/issues/1342 | 753,314,289 | 1,342 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-13 12:01:49 +0200
From: @yzchang
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2008-06-02 11:49:45 +0200
## Comment 7512
Date: 2007-05-13 12:01:49 +0200
From: @yzchang
This happens both in the Current and the Stable branch.
The following query:
put(<hello/>, "/tmp/hello.xml")
causes Mserver to segmentation fault, while this query:
put(<hello/>, "tmp/hello.xml")
works fine.
I can understand that only relative paths are allowed to be used in put, but giving an error message would be better than the segmentation fault.
Temporarily assign to Peter, since he might be the one who has implemented put(). Please reassign if I'm wrong.
Regards,
Jennie
## Comment 7513
Date: 2007-05-16 18:22:50 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
For completeness:
Test added as pathfinder/tests/BugTracker/Tests/put_to_tmp.SF-1718066*
put(<hello/>, "tmp/hello.xml")
works fine
put(<hello/>, "/tmp/hello.xml")
produces expected error
fn_put: file URI '/tmp/hello.xml' must be a relative path (1 such errors).
## Comment 7514
Date: 2008-06-01 22:37:56 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Re-opened, as the test fails (again):
The first (correct) fn:put() call now produces and empty XML fragment
========
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult></XQueryResult>
========
as result, while it used to produce no output/result at all; see
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU.64.64.d-Fedora8/tests_BugTracker/put_to_tmp.SF-1718066.out.00.html
In case this is an intended change, the new output must be approved.
Otherwise the (new) bug must be fixed.
## Comment 7515
Date: 2008-06-02 00:06:24 +0200
From: @lsidir
Logged In: YES
user_id=1856546
Originator: NO
The xquery update specs say that the second argument must be a valid xs:anyURI:
"If $uri is not a valid lexical representation of the xs:anyURI type, a dynamic error is raised [err:FOUP0002]. If $uri is a relative URI Reference, it is resolved relative to the value of the base URI property in the static context."
Why do we raise an error if it is not a relative URI? Am I missing something?
Also for the second issue of what to print as a result, my understanding is that it should print the result of the query:
"The fn:put function has no effect on the current query."
## Comment 7516
Date: 2008-06-02 00:14:18 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Lefteris: relative paths are only allowed because of security reasons: we do not want to give an open license to overwrite any file in the server computer. Therefore, it also checks against .. in the paths. So i think the error should still be given for absolute paths.
as for the change in output: probably correct indeed, as fn:put now is a normal function, not anymore an updating function (update functions produce an empty output).
## Comment 7517
Date: 2008-06-02 23:49:45 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Approved the new output as correct.
Closing.
## Comment 7518
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1718066 at http://sourceforge.net/support/tracker.php?aid=1718066
| PF: put() to /tmp crashes Mserver | https://api.github.com/repos/MonetDB/MonetDB/issues/1341/comments | 0 | 2020-11-30T09:33:53Z | 2024-06-27T11:24:36Z | https://github.com/MonetDB/MonetDB/issues/1341 | 753,314,253 | 1,341 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 11:56:24 +0200
From: @drstmane
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2007-05-10 05:23:59 +0200
## Comment 7510
Date: 2007-05-09 23:56:24 +0200
From: @drstmane
icc -no-gcc -DHAVE_CONFIG_H -I. -I/export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt -I../../.. -I/export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt -I../../include -I/export/scratch1/monet/
monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/../../include -I/export/scratch1/monet/monet.ntv.64.64.d.23868/MonetDB/.ntv.64.64.d-RedHat4WS/include/MonetDB -I/export/scratch1/monet/monet.ntv.64.64.d.23868/MonetDB/.ntv.64.64.d-RedHat4WS/i
nclude/MonetDB/common -I/export/scratch1/monet/monet.ntv.64.64.d.23868/MonetDB/.ntv.64.64.d-RedHat4WS/include/MonetDB/gdk -DLIBOPT -we140 -wd1418 -Wall -w2 -c99 -mp1 -O2 -restrict -unroll -tpp2 -mcpu=itanium2 -DHWCOUNTERS -DHW_Linux -DHW_ia64 -we2
66 -Werror -wd1418,1419,279,981,810,193,111,1357,1572,1599 -c /export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/opt_thetajoin.c -KPIC -DPIC -o .libs/libopt_la-opt_thetajoin.o
/export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/opt_thetajoin.c(1044): error 188: enumerated type mixed with another type
modified = modify_binary_op (p, PFla_and, false, 0);
^
/export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/opt_thetajoin.c(1047): error 188: enumerated type mixed with another type
modified = modify_binary_op (p, PFla_or, false, 0);
^
/export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/opt_thetajoin.c(1093): error 188: enumerated type mixed with another type
PFalg_comp_t comp = 0;
^
compilation aborted for /export/scratch1/monet/monet.ntv.64.64.d.23868/pathfinder/compiler/algebra/opt/opt_thetajoin.c (code 2)
make[8]: *** [libopt_la-opt_thetajoin.lo] Error 1
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.Makes3/index_short.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.ntv.64.64.d-RedHat4WS/make-k.out
## Comment 7511
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1716158 at http://sourceforge.net/support/tracker.php?aid=1716158
| PF: compiler/algebra/opt/opt_thetajoin.c icc-compil. fails | https://api.github.com/repos/MonetDB/MonetDB/issues/1340/comments | 0 | 2020-11-30T09:33:51Z | 2024-06-27T11:24:35Z | https://github.com/MonetDB/MonetDB/issues/1340 | 753,314,208 | 1,340 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 02:35:35 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 10:21:32 +0200
## Comment 7506
Date: 2007-05-09 14:35:35 +0200
From: @romulogoncalves
Just run:
sql>create table t1 as select id from tables order by id asc with data;
sql>select * from t1;
% sys.t1 table_name
% id name
% oid type
% 1 length
sql>
The table should contain the tables ids and is not the case.
Regards,
Romulo
## Comment 7507
Date: 2007-05-10 10:21:32 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
the order by in the with data part wasn't handled correctly. This is fixed now.
## Comment 7508
Date: 2007-05-10 11:28:22 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug is fixed and the test to cover it is:
sql/src/test/BugTracker/Test/create_table_with_data.SF-1715679.sql
Regards,
Romulo
## Comment 7509
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715679 at http://sourceforge.net/support/tracker.php?aid=1715679
| CREATE TABLE WITH DATA... | https://api.github.com/repos/MonetDB/MonetDB/issues/1339/comments | 0 | 2020-11-30T09:33:48Z | 2024-06-27T11:24:34Z | https://github.com/MonetDB/MonetDB/issues/1339 | 753,314,170 | 1,339 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 02:28:47 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 09:46:15 +0200
## Comment 7502
Date: 2007-05-09 14:28:47 +0200
From: @romulogoncalves
Just run:
[goncalve@amelia ~]$ mclient
sql>create LOCAL TEMPORARY table t1 (id int) ON COMMIT DROP;
sql>create GLOBAL TEMPORARY table t1 (id int) ON COMMIT DROP;
you will get:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1090525504 (LWP 10596)]
0x0000003973a6ed1a in _int_malloc () from /lib64/libc.so.6
(gdb) up
1 0x0000003973a7086d in malloc () from /lib64/libc.so.6
(gdb) up
2 0x00002aaaab405593 in GDKmallocmax (size=232, maxsize=0x41000a38,
emergency=1) at /ufs/goncalve/MonetDB/MonetDB/src/gdk/gdk_utils.mx:1046
1046 GDKmalloc_prefixsize(s, size);
(gdb) up
3 0x00002aaaab405723 in GDKmalloc (size=232)
at /ufs/goncalve/MonetDB/MonetDB/src/gdk/gdk_utils.mx:1067
1067 return GDKmallocmax(size, &size, 1);
(gdb) up
4 0x00002aaab8de96e2 in create_sql_table (name=0x1140c98 "t1", type=0,
system=0 '\0', persistence=2, commit_action=3)
at /ufs/goncalve/MonetDB/sql/src/storage/bat/bat_store.mx:2259
2259 sql_table *t = NEW(sql_table);
(gdb)
Regards,
Romulo
## Comment 7503
Date: 2007-05-10 09:46:15 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
write after free problem. fixed in cvs.
## Comment 7504
Date: 2007-05-10 11:27:35 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug is fixed and the test to cover it is:
sql/src/test/BugTracker/Test/local_global_temporary_table.SF-1715672.sql
Regards,
Romulo
## Comment 7505
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715672 at http://sourceforge.net/support/tracker.php?aid=1715672
| LOCAL AND GLOBAL temporary table | https://api.github.com/repos/MonetDB/MonetDB/issues/1338/comments | 0 | 2020-11-30T09:33:45Z | 2024-06-27T11:24:33Z | https://github.com/MonetDB/MonetDB/issues/1338 | 753,314,134 | 1,338 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 02:25:14 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 09:44:38 +0200
## Comment 7498
Date: 2007-05-09 14:25:14 +0200
From: @romulogoncalves
Just run:
[goncalve@amelia ~]$ mclient
sql>create TEMPORARY table t1 (id int) ON COMMIT DROP;
sql>select * from tables;
You will get:
[New Thread 1082132800 (LWP 10423)]
warning: please don't forget to set your vault key!
(see /ufs/goncalve/scratch/MonetDB/MonetDB5/etc/monetdb5.conf)
>[New Thread 1090525504 (LWP 10438)]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1090525504 (LWP 10438)]
0x0000003973a6ed1a in _int_malloc () from /lib64/libc.so.6
(gdb)
Regards,
Romulo
## Comment 7499
Date: 2007-05-10 09:44:38 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
it was a write after free problem. Fixed now.
## Comment 7500
Date: 2007-05-10 11:26:00 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug is fixed and the test to cover it is:
sql/src/test/BugTracker/Test/temporary_tables.SF-1715669.sql
Regards,
Romulo
## Comment 7501
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715669 at http://sourceforge.net/support/tracker.php?aid=1715669
| Temporary tables.... | https://api.github.com/repos/MonetDB/MonetDB/issues/1337/comments | 0 | 2020-11-30T09:33:43Z | 2024-06-27T11:24:32Z | https://github.com/MonetDB/MonetDB/issues/1337 | 753,314,111 | 1,337 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 02:16:33 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 10:41:29 +0200
## Comment 7494
Date: 2007-05-09 14:16:33 +0200
From: @romulogoncalves
Just run the example:
sql>select now() LIKE '12:14';
The Mserver5 will crash:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1090525504 (LWP 9767)]
0x00002aaaab40b272 in ATOMcmp (t=255, l=0xffffffff00000000, r=0xb2ff202a07470)
at /ufs/goncalve/MonetDB/MonetDB/src/gdk/gdk_atoms.mx:1016
1016 return (l == r) ? 0 : atom_CMP(l, r, t);
(gdb)
Regards,
Romulo
## Comment 7495
Date: 2007-05-10 10:41:29 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
crash was allready gone (probaly because of the write after free problems). But also the type checking wasn't complete. This is now fixed.
## Comment 7496
Date: 2007-05-10 11:29:18 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug is fixed and the test to cover it is:
sql/src/test/BugTracker/Test/like_for_time_values.SF-1715659.sql
Regards,
Romulo
## Comment 7497
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715659 at http://sourceforge.net/support/tracker.php?aid=1715659
| LIKE for time values... | https://api.github.com/repos/MonetDB/MonetDB/issues/1336/comments | 0 | 2020-11-30T09:33:40Z | 2024-06-27T11:24:31Z | https://github.com/MonetDB/MonetDB/issues/1336 | 753,314,069 | 1,336 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 01:56:58 +0200
From: @grobian
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2008-08-05 02:10:12 +0200
## Comment 7490
Date: 2007-05-09 13:56:58 +0200
From: @grobian
algebra.h:5:27: error: algebra.proto.h: No such file or directory
In file included from /ufs/fabian/scratch/monetdb/current/four/src/modules/plain/algebra.mx:1399:
algebra.h:7: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'ptr'
algebra.h:8: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'ptr'
/ufs/fabian/scratch/monetdb/current/four/src/modules/plain/algebra.mx:1458: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'BATmin'
/ufs/fabian/scratch/monetdb/current/four/src/modules/plain/algebra.mx:1458: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'BATmax'
/ufs/fabian/scratch/monetdb/current/four/src/modules/plain/algebra.mx:1341: error
just restarting the compilation makes it continue. I'm using make -j2.
## Comment 7491
Date: 2007-10-21 20:26:23 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Most likely still the same?
## Comment 7492
Date: 2008-08-05 02:10:12 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
since this path is obsolete and it is unclear
if the bug persists, it is closed until further notice.
## Comment 7493
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715646 at http://sourceforge.net/support/tracker.php?aid=1715646
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-monetdb4@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| M4: parallel building broken | https://api.github.com/repos/MonetDB/MonetDB/issues/1335/comments | 0 | 2020-11-30T09:33:38Z | 2024-06-28T07:36:58Z | https://github.com/MonetDB/MonetDB/issues/1335 | 753,314,043 | 1,335 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-09 11:11:42 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 12:21:47 +0200
## Comment 7487
Date: 2007-05-09 11:11:42 +0200
From: @romulogoncalves
With this example:
create table t1(id int, name varchar(1024), age int );
alter table t1 alter id set DEFAULT RANK() OVER (PARTITION BY id, age ORDER BY id DESC);
I get:
[goncalve@amelia ~]$ mclient < example.sql
MAPI = monetdb@localhost:50000
QUERY = alter table t1 alter id set DEFAULT RANK() OVER (PARTITION BY id, age ORDER BY id DESC);
ERROR = !unknown table element (18034648)->token = DEFAULT
[ 1 ]
From yesterday I learn that this kind of messages are not accepted.
Regards,
Romulo
## Comment 7488
Date: 2007-05-10 00:21:47 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug can be close it was fixed by the checkin of Romulo at 00:20.
The bug is covered by:
sql/src/test/BugTracker/Tests/set_default_value.SF-1715553.sql
Regards,
Romulo
## Comment 7489
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1715553 at http://sourceforge.net/support/tracker.php?aid=1715553
| Set default value..... | https://api.github.com/repos/MonetDB/MonetDB/issues/1334/comments | 0 | 2020-11-30T09:33:35Z | 2024-06-27T11:24:29Z | https://github.com/MonetDB/MonetDB/issues/1334 | 753,314,013 | 1,334 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-08 12:07:22 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-08 07:33:09 +0200
## Comment 7476
Date: 2007-05-08 12:07:22 +0200
From: @romulogoncalves
I created the table:
create table t1 (id int CHECK (f1() = 1));
MAPI = monetdb@localhost:50000
QUERY = create table t3 (id int CHECK (f1() = 1));
ERROR = !unknown constraint (20874664)->token = CHECK
I think the example says everything.
Regards,
Romulo
## Comment 7477
Date: 2007-05-08 12:19:52 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
it tells you that this feature isn't supported!
So this is not a bug report but a feature request which Niels kindly
ignore.
Regards,
Romulo
## Comment 7478
Date: 2007-05-08 12:22:11 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I disagree. The message could have been in Chinese, I don't think it's clear at all.
## Comment 7479
Date: 2007-05-08 12:43:52 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
Well, If the message error is weak it is not a bug.
We already agree that the error system of M5-SQL has to be improved.
The problem is that we need time.
I do not close the bug, it should be closed by Niles. He is the big boss of SQL module.
Regards,
Romulo
## Comment 7480
Date: 2007-05-08 17:29:41 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Test file added to repository. 'BugConstraints/check.sql'
## Comment 7481
Date: 2007-05-08 17:34:21 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
Why did you add a test file if Niels said that it is not a bug?
Regards,
Romulo
## Comment 7482
Date: 2007-05-08 17:36:21 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
It *is* a bug, because the message produced is completely garbage. If the feature is not supported, then the parser shouldn't accept it, or an appropriate message conveyed.
## Comment 7483
Date: 2007-05-08 17:48:11 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
I aggree with you,
I will change the parser. It is an implemented feature not handle correctly by Niels.
Regards,
Romulo
## Comment 7484
Date: 2007-05-08 19:33:09 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
This bug can be closed. The checkin at 19:31 by Romulo contains the fix.
The test to cover it is:
sql/src/test/BugConstraints/Tests/check_constraint.SF-1714829
Regards,
Romulo
PS: The test check.sql added by Martin was removed.
## Comment 7485
Date: 2007-05-08 22:23:20 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
From Romulo's last comment, I read that this bug has been closed, because the problem has been "Fixed".
Please correct me in case I'm wrong.
## Comment 7486
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714829 at http://sourceforge.net/support/tracker.php?aid=1714829
| SQL: CHECK CONSTRAINT... | https://api.github.com/repos/MonetDB/MonetDB/issues/1333/comments | 0 | 2020-11-30T09:33:33Z | 2024-06-27T11:24:28Z | https://github.com/MonetDB/MonetDB/issues/1333 | 753,313,989 | 1,333 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-08 11:29:29 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-08 06:32:40 +0200
## Comment 7471
Date: 2007-05-08 11:29:29 +0200
From: @romulogoncalves
create table t1(id int, name varchar(1024), age int);
create table t2(id int, age int);
create function f1()
returns int
BEGIN
return 1;
END;
create view v1 as select * from t1 where id = f(1);
This example crash the mserver. I know that it should be f1(), however it should not crash the mserver it should give an error message.
Regards,
Romulo
## Comment 7472
Date: 2007-05-08 14:22:21 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
fixed, ie added check of result of subquery.
## Comment 7473
Date: 2007-05-08 17:35:07 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Added test file to the repository
## Comment 7474
Date: 2007-05-08 18:32:40 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug was fixed by Niels check at 14:22
the test to cover is
sql/src/test/BugTracker/view_creation_using_func.SF-1714814.sql
Regards,
Romulo
PS: The test added by Martin will be removed....
## Comment 7475
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714814 at http://sourceforge.net/support/tracker.php?aid=1714814
| SQL: View creation using a function.... | https://api.github.com/repos/MonetDB/MonetDB/issues/1332/comments | 0 | 2020-11-30T09:33:30Z | 2024-06-27T11:24:27Z | https://github.com/MonetDB/MonetDB/issues/1332 | 753,313,965 | 1,332 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-08 11:17:20 +0200
From: @romulogoncalves
To: @njnes
Version: -- development
Last updated: 2007-05-10 05:28:52 +0200
## Comment 7462
Date: 2007-05-08 11:17:20 +0200
From: @romulogoncalves
Just run the example:
create table t1(id int, name varchar(1024), age int);
create table t2(id int, age int);
create PROCEDURE p1(id int, age int)
BEGIN
insert into t2 values(id, age);
END;
create PROCEDURE p1()
BEGIN
declare id int, age int;
set id = 1;
set age = 23;
call p1(id, age);
END;
create trigger test_0 after insert on t2
BEGIN ATOMIC
insert into t1 values(1, 'monetdb', 24);
call p1();
END;
insert into t2 values(0, 24);
There is a infinite loop due the call of p1(). The infinite loops created by triggers (an update operation from one trigger which fires another trigger) were killed. However, when the update comes from a procedure the infinite loop is not avoided.
Regards,
Romulo
## Comment 7463
Date: 2007-05-08 17:04:25 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
A defense line is to call language.assertSpace(N)
where the stack space is limited to N*variables in current block.
This could could be included in the start of a (recursive)
function body, or those that call at least one other SQL function/proc.
## Comment 7464
Date: 2007-05-08 20:31:22 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
When I set off the optimizers I get:
>set optimizer='off';
>insert into t2 values(0, 24);
!TypeException:user.p1[21]:'user.p1' undefined in: _39:any := user.p1()
&2 1
!MALException:user.p1[0]:Syntax error in script
>
What does it mean?
Regards,
Romulo
## Comment 7465
Date: 2007-05-08 22:34:08 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
The lastest error is not confirmed when all optimizers are on.
It looks more like a typo.
[note, I also added cleanup code]
set optimizer='off';
create table t1(id int, name varchar(1024), age int);
create table t2(id int, age int);
create PROCEDURE p1(id int, age int)
BEGIN
insert into t2 values(id, age);
END;
create PROCEDURE p1()
BEGIN
declare id int, age int;
set id = 1;
set age = 23;
call p1(id, age);
END;
create trigger test_0 after insert on t2
BEGIN ATOMIC
insert into t1 values(1, 'monetdb', 24);
call p1();
END;
drop trigger test_0;
drop procedure p1();
drop procedure p1(int,int);
drop table t1;
drop table t2;
## Comment 7466
Date: 2007-05-10 10:42:33 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
reassigned to romulo, seems dependency list related.
## Comment 7467
Date: 2007-05-10 12:15:38 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The new error for this bug after updating my M5 and SQL (cvs branch).
~ MAPI = monetdb@localhost:35709
=QUERY = insert into t2 values(0, 24);
ERROR = !TypeException:user.p1[21]:'user.p1' undefined in: _39:any := user.p1()
~ MAPI = monetdb@localhost:35709
=QUERY = insert into t2 values(0, 24);
ERROR = !ERROR: Syntax error in script
the original non-normalized output was: ERROR = !MALException:user.p1[0]:Syntax error in script
12:13:20 >
12:13:20 > Done.
12:13:20 >
I assign the bug to Niels....
Regards,
Romulo
## Comment 7468
Date: 2007-05-10 17:04:01 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
This bug is fixed for M4-SQL.
The test to cover it can be found:
sql/src/test/BugTracker/Tests/infinite_loop.SF-1714809.sql
However, in M5 the infinite loop continue.
with optimizer='off' I get:
~ MAPI = monetdb@localhost:37605
=QUERY = insert into t2 values(0, 24);
ERROR = !TypeException:user.p1[21]:'user.p1' undefined in: _39:any := user.p1()
~ MAPI = monetdb@localhost:37605
=QUERY = insert into t2 values(0, 24);
ERROR = !ERROR: Syntax error in script
the original non-normalized output was: ERROR = !MALException:user.p1[0]:Syntax error in script
Niels, to who I should assign this BUG? TO you or Martin?
Regards,
Romulo
## Comment 7469
Date: 2007-05-10 17:28:52 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: YES
The bug can be closed.
The resolution for for M5 was a simple compilation of SQL to M5.
Yes, I know stupid error once again.
The test to cover it can be found:
sql/src/test/BugTracker/Tests/infinite_loop.SF-1714809.sql
Regards,
Romulo
## Comment 7470
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714809 at http://sourceforge.net/support/tracker.php?aid=1714809
| Infinite loop.... | https://api.github.com/repos/MonetDB/MonetDB/issues/1331/comments | 0 | 2020-11-30T09:33:27Z | 2024-06-27T11:24:26Z | https://github.com/MonetDB/MonetDB/issues/1331 | 753,313,931 | 1,331 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-07 11:32:54 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-05-09 10:49:03 +0200
## Comment 7453
Date: 2007-05-07 23:32:54 +0200
From: @drstmane
Since 2007-03-10 (almost) all XRpc tests fail on MacOS (Darwin) with
!ERROR: during RPC call to "localhost:45639"
!ERROR: HTTP Error Code : 403 Verbidden
!ERROR: SOAP Fault Code : env:Sender
!ERROR: SOAP Fault Reason:
!ERROR: Called module not allowed.
!ERROR: CMDhttp_post: operation failed.
See, e.g.,
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XMark_XRpc/q01.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/benchmarks_XMark_XRpc/q20.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/tests_XRpc/add_dbl.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/tests_XRpc/insertNode.err.00.html
## Comment 7454
Date: 2007-05-08 08:20:21 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
[For those who care]
The problem seems to be that testing uses "/var/tmp/" as prefix / target location for testing.
On MacOs (Darwin), however, "/var" is a symlink to "/private/var";
hence, the logical path "/var/tmp/..." does not match the physical one "/private/var/tmp/...".
With XRpc, the logical one is listed in xrpc_trusted but the physical one is used as requested location.
See, e.g.,
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/tests_XRpc/add_dbl.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Darwin8.9.0/runtime/XRpc_BATs.out.00.html
(Hopefully) fixed by making testing use the physical prefix / target location "/private/var/tmp/".
## Comment 7455
Date: 2007-05-08 08:20:51 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Seems to be "my" bug report ...
## Comment 7456
Date: 2007-05-08 09:40:18 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
Is it possible to let Xrpc canonicalise the path before storing it? The next Darwin release is said to have no /private anymore.
## Comment 7457
Date: 2007-05-08 11:42:14 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
Hi Stefan,
Thanks for fixing this. I saw the problem earlier. Since the current branch is not that stable, I thought I can wait for a while to see if the problem will disappear. But then, I forget to check again :P
Canonicalise the path before storing it might be a good/flexible idea. I will have a look how this can be done.
Jennie
## Comment 7458
Date: 2007-05-09 10:30:00 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Assigned to Peter, as his checkin on Fri Mar 9 2007 seems to trigger (cause) the problem:
http://monetdb.cvs.sourceforge.net/monetdb/pathfinder/runtime/xrpc_server.mx?view=logrev1.21
http://monetdb.cvs.sourceforge.net/monetdb/pathfinder/runtime/xrpc_server.mx?r1=1.20&r2=1.21
## Comment 7459
Date: 2007-05-09 10:47:40 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed in CVS.
This turned out to be not a MacOS (Darwin) specific problem,
but two general BUGs in isAllowed():
(1)
In "BATloop(<bat>, p, q)", the loop/iterator variable is "p", not "q" !
(2)
BATs xrpc_trusted, xrpc_admin & xrpc_user are [str,void(nil)],
hence, in isAllowed() we need to use BUNhead instead of BUNtail
to read the BUNs'\'' (head!) values.
I have not idea, why the code apparently happened to work on all other platforms,
despite these quite foundamental BUGs;
might be little-/big-endial related.
Cetero Censeo:
For the umpteenth time,
this proves how important and "handy" multi-platform testing is to detect and eventually fix *foundamental* bugs in the code
--- provided anybody does care at all ...
## Comment 7460
Date: 2007-05-09 10:49:03 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
stmane did fix this one.
## Comment 7461
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714603 at http://sourceforge.net/support/tracker.php?aid=1714603
| PF: XRpc tests fail on MacOS (Darwin) since 2007-03-10 | https://api.github.com/repos/MonetDB/MonetDB/issues/1330/comments | 0 | 2020-11-30T09:33:24Z | 2024-06-27T11:24:25Z | https://github.com/MonetDB/MonetDB/issues/1330 | 753,313,896 | 1,330 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-07 05:15:01 +0200
From: Wouter Alink <<alink>>
To: @yzchang
Version: -- development
Last updated: 2007-05-18 12:47:49 +0200
## Comment 7439
Date: 2007-05-07 17:15:01 +0200
From: Wouter Alink <<alink>>
I was playing with Mserver 4.17.1 (monetdb_install.sh --nightly=current). I couldn't find any documentation, so I just played around a little. I noticed the following:
starting Mserver:
Mserver --dbinit="module(pathfinder);"
and then browsing to the website:
http://localhost:50001/admin/
works fine.
pressing Ctrl-R (refresh) a few times (3 or 4 is often enough) makes Mserver segfault.
it is not necessary to solve this bug instantly, but i just wanted you guys to know. i got another error when i pressed a button, but I might not have configured the system correctly. (i probably have to play some more with it)
p.s. if you would like me to wait with testing these new features, then please tell me.
## Comment 7440
Date: 2007-05-07 18:22:49 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
Hi Wouter,
Thanks for trying the admin interface and reporting the problem.
But I can not get Mserver crash in the way you describe (the code has just been checked out from CVS and rebuilt).
I can refresh the page several tens of times without problem.
Can you please provide some more system information?
- OS (Windows/Linux)?
- Bits (32/64)?
- Compile option (debug/optimize)?
Thanks in advance!
Jennie
PS> All admin/xrpc related problems can be assigned to me.
PSS> Those new features are made to be played with, so don't wait:)
## Comment 7441
Date: 2007-05-07 19:09:58 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
64-bit, fedora core 6
from the SR1 release tarball:
monetdb_install.sh --enable-optimise --enable-xquery --nightly=current
(i modified the script in that it adds the --enable-oid32 option)
note that the system is not able to communicate with the internet.
ps, jennie do you perhaps have a pointer to some documentation? i don't know how to correctly start the xrpc service (is module(pathfinder); enough?). i also have difficulty setting the correct values in MonetDB.conf. who is a user, who is a admin? do i need to add each host that is allowed to do xrpc, or can i use wildcards?
## Comment 7442
Date: 2007-05-07 19:26:37 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
To avoid confusion and redundant/extra/use-less bug chasing and debugging, please provide accurate information in your bug reports:
"SR1" is (IMHO) MonetDB-Venus-SuperBall-SR1.tar.bz2, and hence the leatest (bug-fix) *release* MonetDB 4.16.2 / pathfinder 0.16.2 not the CVS head version...
## Comment 7443
Date: 2007-05-07 19:31:28 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
... actually, Wouter's info is (still) confusing:
"from the SR1 release tarball" indicates the *release* tarball,
but "--nightly=current" suggests nightly builds.
Could you please clearify?
## Comment 7444
Date: 2007-05-07 19:35:56 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
the monetdb_install.sh script from the package SR1, and using it to install the nightly build (current).
there was no 'pathfinder-nightly-build' option, so i marked it 'pathfinder-head'
sorry for the confusion.
## Comment 7445
Date: 2007-05-07 19:37:36 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
oops
## Comment 7446
Date: 2007-05-07 19:38:13 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
He downloaded ssb-sr1 which contains a monetdb-install.sh script, and used --nightly=current. What is possibly an issue here, is how a machine without internet access can download the current nightlies. There may be a bug here that the script just installs using the packaged sources. Wouter, can you tell if the script downloaded sources for you?
## Comment 7447
Date: 2007-05-07 19:40:04 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
mid-air collision, sorry
## Comment 7448
Date: 2007-05-07 19:53:59 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: YES
ah, now i see the confusion...
the script installed fine (it downloaded the sources from the internet using a proxy ;).
the user under which the Mserver ran did not know how to reach the internet. (i thought this might be relevant information regarding the xrpc bug i desribed in the bugreport)
## Comment 7449
Date: 2007-05-08 11:29:49 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
Hi Wouter,
It should not matter if Mserver has access to the Internet or not, since the admin page is on your localhost. I cannot reproduce your problem, maybe because I don't use the nightly build, but CVS source?
Documentation about XRPC can be found here: http://monetdb.cwi.nl/projects/monetdb/XQuery/Documentation/XRPC.html
It was written for the last release, so I see that some information is out-dated. I will update the documentation a.s.a.p.
Regards,
Jennie
PS> Here is a quick update of the XRPC documentation:
- Now the XRPC modules are loaded by default, hence
$ Mserver --dbinit="module(pathfinder);"
also starts the XRPC server
- XRPC options in MonetDB.conf.in:
* xrpc_port: to which port the xrpc server should listen, default 50001, or mapi_port+1 if you changed mapi_port
* xrpc_open: works the same as mapi_open
* xrpc_options: the only implemented option is "timing", for performance tests, usually not used.
* xrpc_admin: usually you the admin web interface can only be access via localhost, if you want to access the page from another host than where Mserver with XRPC is running, you need to add the IP or the hostname of the other host to this list
* xrpc_user: the same idea as xrpc_admin, but for retrieving XML docs from MXQ.
* xrpc_trusted: this concerns a security check added after the last release. We now only execute functions in a module, which location (the 'at'-hint) has one of the prefixes in this list. For example, if this list contains the prefix "www.monetdb.nl", then we will execute functions in modules such as "www.monetdb.nl/foo.xq" and "www.monetdb.nl/bar/foo.xq".
Of course, the default values of these options can be set when you start Mserver with --set option=value
There is also a proc "add_xrpc_trusted" with which you can add more prefixes to the list "xrpc_trusted".
## Comment 7450
Date: 2007-05-18 00:33:29 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
I could reproduce a crash, sometimes in a free() [disconnect() in the shttpd code]
did you run a trace Wouter (i.e. experience this in gdb? typed 'bt' then?)
Valgrind did not report anything, and a made a changes (initializing the entire connection struct to 0).
since then, I have not been able to get this crash again -- but that says little.
Wouter: in the *latest* HEAD, can you still reproduce this problem?
## Comment 7451
Date: 2007-05-18 00:47:49 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
I think I found this bug. Plase confirm that it's gone
## Comment 7452
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714399 at http://sourceforge.net/support/tracker.php?aid=1714399
| XQuery/XRPC: refresh admin-page causes segmentation fault | https://api.github.com/repos/MonetDB/MonetDB/issues/1329/comments | 0 | 2020-11-30T09:33:21Z | 2024-06-27T11:24:24Z | https://github.com/MonetDB/MonetDB/issues/1329 | 753,313,856 | 1,329 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-07 04:31:57 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-07-20 06:35:07 +0200
## Comment 7431
Date: 2007-05-07 16:31:57 +0200
From: Jan Rittinger <<jan.rittinger>>
If I compile and run the query BugTracker/nested_element_construction.SF-1712578.xq with the algebra version (pf -A) I get a merged_union error which boils down to duplicate entries in the qname relations. The reason are the following two lines from find_qn_bulk() in pf_support.mx which both modify the same relation:
1.188 (stmane 10-Jan-07): qn_prefix_uri_loc.append(_pref_uri_loc, true);
1.195 (sjoerd 20-Feb-07): ws.fetch(QN_PREFIX_URI_LOC + UPDATE).find(cont).insert(_pref_uri_loc);
## Comment 7432
Date: 2007-05-07 17:43:06 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Just as a quick info:
===================================================================
2007/01/10 - stmane: runtime/pf_support.mx,1.188
propagated changes of Friday Jan 05 2007 - Wednesday Jan 10 2007
from the XQuery_0-14 branch to the development trunk
===================================================================
===================================================================
2007/02/20 - sjoerd: runtime/pf_support.mx,1.195
propagated changes of Sunday Feb 11 2007 - Tuesday Feb 20 2007
from the XQuery_0-16 branch to the development trunk
===================================================================
## Comment 7433
Date: 2007-05-17 23:50:51 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7434
Date: 2007-05-18 07:26:41 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
finishing Peter'\''s fix of find_qn_bulk():
the new code used "if (update) ...",
while the function'\''s signature read "... bit add ...";
this caused about 20 tests to fail with
"!ERROR: unknown variable '\''update'\''."
fixed by renaming the parameter "add" to "update"
(the latter indeed seems more intuitive/descriptive than the former).
this makes the ~20 failing tests work, again.
test added in pathfinder/tests/BugTracker/Tests/nested_element_construction.SF-1712578_alg.*
## Comment 7435
Date: 2007-06-26 11:12:00 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: YES
This bug seems to be open again.
module ("pathfinder");
var ws := ws_create ();
var a := bat (void,str).seqbase(0@0).append("foo").append("foo").append("foo").append("bar").append("bar").append("bar");
var b := add_qnames(a.project (""), a.project (""), a, ws);
print(ws.fetch(QN_LOC).fetch(0));
-----------------
h t name
oid str type
-----------------
[ 0@0, "foo" ]
[ 1@0, "bar" ]
[ 0@0, "foo" ]
[ 1@0, "bar" ]
Whereas the following snippet does the correct thing:
var b := [add_qname](a.project (""), a.project (""), a, CONST ws);
-----------------
h t name
void str type
-----------------
[ 0@0, "foo" ]
[ 1@0, "bar" ]
## Comment 7436
Date: 2007-07-20 17:36:30 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
If I run the test tests/BugTracker/Tests/nested_element_construction.SF-1712578_alg, I get no differences. So what is the problem then?
If the problem is what you describe in your comment about add_qnames behaving differently than [add_qname], then that could perhaps be a different bug. I don't immediately see how that relates to find_qn_bulk, which is what the bug was originally about.
In other words, if this is a new bug, please use a new bug report, if it is the same bug but in a different guise, please elaborate, and if there is no bug, then please close this report.
## Comment 7437
Date: 2007-07-20 18:35:07 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: YES
Sjoerd, as you correctly observed the test in the testweb runs ok. If I however slightly modify the input (as described in my previous comment) find_qn_bulk again shows the behavior described in the original bug report: it fills the qname relation with duplicate values.
I hope that was enough reason to re-open the bug (instead of creating a new one). I said 'was' because on my current version the below example works fine :D.
Thus I'll feel free to close the bug again till I (hopefully not) find a new counter example.
## Comment 7438
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714360 at http://sourceforge.net/support/tracker.php?aid=1714360
| PF/runtime: find_qn_bulk() adds new qnames twice | https://api.github.com/repos/MonetDB/MonetDB/issues/1328/comments | 0 | 2020-11-30T09:33:19Z | 2024-06-27T11:24:23Z | https://github.com/MonetDB/MonetDB/issues/1328 | 753,313,824 | 1,328 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-07 11:45:04 +0200
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2007-05-07 02:18:31 +0200
## Comment 7427
Date: 2007-05-07 11:45:04 +0200
From: @drstmane
from modules/pftijah/pftijah.mx:
========
define LOGGERFILE "/tmp/TIJAH_LOGGER"
/*
* the CMDtj_log is created to be able to print in very difficult IO areas
*/
int CMDtj_log(str format, int* i) {
FILE* f;
if ( (f = fopen(LOGGERFILE,"a")) ) {
^^^
fprintf(f,(const char*)format,*i);
fclose(f);
}
return GDK_SUCCEED;
}
========
apparently, "/tmp/TIJAH_LOGGER" is only appended to, but never cleaned or overwritten;
this eventually causes /tmp/ to filled if one uses pathfinder regularly --- even *without* using pftijah --- except from having it loaded via pathfinder...
$ l -h /tmp/TIJAH_LOGGER
-rw-rw-rw- 1 manegold ins 6.9G May 7 11:25 /tmp/TIJAH_LOGGER
^^^^
$ wc -l /tmp/TIJAH_LOGGER
423802741 /tmp/TIJAH_LOGGER
$ head -9 /tmp/TIJAH_LOGGER
+ mps(start-0)
+ mps(start-1)
+ mps(finish-1)
+ mps(start-2)
+ mps(start-3)
+ mps(start-4)
+ mps(finish-4)
+ mps(start-5)
+ mps(start-6)
$ tail -9 /tmp/TIJAH_LOGGER
+ mps(start-85)
+ mps(finish-85)
+ mps(finish-80)
+ mps(finish-79)
+ mps(finish-78)
+ mps(finish-77)
+ mps(start-77)
+ mps(start-78)
## Comment 7428
Date: 2007-05-07 12:56:34 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
side note: always use /var/tmp if on a unixish environment. e.g. on Solaris you're writing into swap when writing into /tmp
## Comment 7429
Date: 2007-05-07 14:18:31 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
Logger file is now initialized by using tj_setlog() and is never set hard in the code to prevent accidents like this. The file is also cleared every time tj_setlog() is called and cannot accidently grow this large.
## Comment 7430
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714133 at http://sourceforge.net/support/tracker.php?aid=1714133
| PFtijah: "/tmp/TIJAH_LOGGER" never cleaned => fills /tmp/ ! | https://api.github.com/repos/MonetDB/MonetDB/issues/1327/comments | 0 | 2020-11-30T09:33:17Z | 2024-06-27T11:24:22Z | https://github.com/MonetDB/MonetDB/issues/1327 | 753,313,799 | 1,327 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-07 11:34:44 +0200
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2007-05-07 02:09:43 +0200
## Comment 7424
Date: 2007-05-07 11:34:44 +0200
From: @drstmane
define LOGGERFILE "/tmp/TIJAH_LOGGER"
in modules/pftijah/pftijah.mx is obviously a Unix specific PATH and hence not (necessarily) portable.
## Comment 7425
Date: 2007-05-07 14:09:43 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Originator: NO
Removed hard wired "/tmp/TIJAH_LOGGER" from code. This file is now set by the application using tj_setlog().
## Comment 7426
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1714129 at http://sourceforge.net/support/tracker.php?aid=1714129
| PFtijah: hard-wired "/tmp/TIJAH_LOGGER" not portable | https://api.github.com/repos/MonetDB/MonetDB/issues/1326/comments | 0 | 2020-11-30T09:33:14Z | 2024-06-27T11:24:21Z | https://github.com/MonetDB/MonetDB/issues/1326 | 753,313,765 | 1,326 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-05 11:42:32 +0200
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2007-05-07 03:13:02 +0200
## Comment 7420
Date: 2007-05-05 23:42:32 +0200
From: @drstmane
See below for details.
This is actually the same problem as reported for test
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1204095-a.out.00.html
in
https://sourceforge.net/tracker/index.php?func=detail&aid=1712463&group_id=56967&atid=482468
hence, it could be related to checkins of 2007-04-25.
var b:=new(oid,void).rename("PD").persists(true);
var pd:=bat("PD");
pd.insert(0@0,nil);
pd.insert(1@0,nil);
pd.insert(2@0,nil);
commit();
pd.print();
-----------------
h t name
oid void type
-----------------
[ 0@0, nil ]
[ 1@0, nil ]
[ 2@0, nil ]
pd.info().reverse().like("sorted").print();
---------------------------------
h t name
str str type
---------------------------------
[ "65", "hsorted" ]
[ "0", "hnosorted" ]
[ "0", "hnosorted_rev" ]
[ "65", "tsorted" ]
[ "0", "tnosorted" ]
[ "0", "tnosorted_rev" ]
pd.info().reverse().like("dense").print();
-------------------------
h t name
str str type
-------------------------
[ "1", "hdense" ]
[ "0", "hnodense" ]
[ "0", "tdense" ]
[ "0", "tnodense" ]
pd.info().reverse().like("seqbase").print();
-------------------------
h t name
str str type
-------------------------
[ "0@0", "hseqbase" ]
[ "nil", "tseqbase" ]
pd.seqbase().print();
[ 0@0 ]
pd.delete(2@0);
!WARNING: BATpropcheck: BAT PD(26) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT PD(26) remains marked radix-clustered on 32 bits; not checked!
pd.print();
-----------------
h t name
oid void type
-----------------
[ 1@0, nil ]
[ 0@0, nil ]
pd.info().reverse().like("sorted").print();
---------------------------------
h t name
str str type
---------------------------------
[ "64", "hsorted" ]
[ "2", "hnosorted" ]
[ "0", "hnosorted_rev" ]
[ "65", "tsorted" ]
[ "2", "tnosorted" ]
[ "0", "tnosorted_rev" ]
pd.info().reverse().like("dense").print();
-------------------------
h t name
str str type
-------------------------
[ "0", "hdense" ]
[ "0", "hnodense" ]
[ "0", "tdense" ]
[ "0", "tnodense" ]
pd.info().reverse().like("seqbase").print();
-------------------------
h t name
str str type
-------------------------
[ "0@0", "hseqbase" ]
[ "nil", "tseqbase" ]
pd.seqbase().print();
[ 0@0 ]
## Comment 7421
Date: 2007-05-05 23:47:06 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Test added in MonetDB4/tests/BugTracker/Tests/delete_from_persistent_dense_BAT.SF-1713584.*
## Comment 7422
Date: 2007-05-07 15:13:02 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Originator: NO
Fixed in CVS.
## Comment 7423
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1713584 at http://sourceforge.net/support/tracker.php?aid=1713584
| delete on committed dense persistent BAT causes prop. error | https://api.github.com/repos/MonetDB/MonetDB/issues/1325/comments | 0 | 2020-11-30T09:33:12Z | 2024-06-27T11:24:20Z | https://github.com/MonetDB/MonetDB/issues/1325 | 753,313,722 | 1,325 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 11:59:07 +0200
From: Maurice van Keulen <<mvankeulen>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-17 02:50:22 +0200
## Comment 7415
Date: 2007-05-04 11:59:07 +0200
From: Maurice van Keulen <<mvankeulen>>
[milprint_summer backend]
Query below fails for attached document
declare function deep-equal($node1 as node(), $node2 as node())
as xs:boolean
{
let $e1 := $node1/descendant-or-self::element()
,$e2 := $node2/descendant-or-self::element()
return
if (count($e1) eq count($e2))
then let $e1s := for $e in $e1 return ($e/name(),$e/string())
,$e2s := for $e in $e2 return ($e/name(),$e/string())
return string-join($e1s,",") eq string-join($e2s,",")
else false()
};
let $doc := doc("imdb2.xml")
return deep-equal(<a>{$doc}</a>,<a>{$doc}</a>)
Error message is:
ERROR = !ERROR: mposjoin: missing matches. The result requires the same number of tuples as the first two input arguments: 2 < 20.
!ERROR: CMDmposjoin: operation failed.
Mserver -d10 does not provide any additional info.
## Comment 7416
Date: 2007-05-04 19:34:48 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
[ I assume, this has been discovered with the pathfinder CVS head, right? ]
Added test in
pathfinder/tests/BugTracker/Tests/mposjoin_missing_matches.SF-1712595.*
It does fail as reported.
! correct/expected output still needs to be provided/approved !
## Comment 7417
Date: 2007-05-17 02:50:22 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7418
Date: 2007-05-17 12:49:42 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
For completeness:
approved stable output after Peter's fix.
## Comment 7419
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712595 at http://sourceforge.net/support/tracker.php?aid=1712595
| Error mposjoin: missing matches | https://api.github.com/repos/MonetDB/MonetDB/issues/1324/comments | 0 | 2020-11-30T09:33:09Z | 2024-06-27T11:24:19Z | https://github.com/MonetDB/MonetDB/issues/1324 | 753,313,687 | 1,324 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 11:18:50 +0200
From: Maurice van Keulen <<mvankeulen>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-05-06 07:45:09 +0200
## Comment 7410
Date: 2007-05-04 11:18:50 +0200
From: Maurice van Keulen <<mvankeulen>>
Attached query uses two nested for loops to create a nested documents. The number of elements created is correct, but the element names of the innermost created elements is bogus. It gives
<result><tags>a b c d</tags><a><a/><a/><b/><b/></a><b><c/><c/><d/><d/></b><c><a/><a/><b/><b/></c><d><c/><c/><d/><d/></d></result>
while it should be
<result><tags>a b c d</tags><a><a/><b/><c/><d/></a><b><a/><b/><c/><d/></b><c><a/><b/><c/><d/></c><d><a/><b/><c/><d/></d></result>
## Comment 7411
Date: 2007-05-04 14:05:06 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
I can only say that it works for me (with mps.c version 1.355).
## Comment 7412
Date: 2007-05-04 19:23:40 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Added test in pathfinder/tests/BugTracker/Tests/nested_element_construction.SF-1712578.*
It does fail as reported with the latest CVS HEAD code base (mps r1.374).
## Comment 7413
Date: 2007-05-06 19:45:09 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Originator: NO
This problem was introduced in mps version 1.356 where the former namespace separator ':' was replaced by NS_ACCEL_SEP. A "::" string was accidentally replaced by a single NS_ACCEL_SEP character.
## Comment 7414
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712578 at http://sourceforge.net/support/tracker.php?aid=1712578
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Mix-up in nested element construction | https://api.github.com/repos/MonetDB/MonetDB/issues/1323/comments | 0 | 2020-11-30T09:33:07Z | 2024-06-27T11:24:18Z | https://github.com/MonetDB/MonetDB/issues/1323 | 753,313,651 | 1,323 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:44:14 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-05-12 12:57:30 +0200
## Comment 7401
Date: 2007-05-04 02:44:14 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/pf_add-doc_property_error.SF-1706438.err.00.html
MAPI = monetdb@localhost:36938
QUERY = pf:add-doc("/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/source/pathfinder/tests/BugTracker/Tests/pf_add-doc_property_error.SF-1706438.doc.xml","OneNowOneLater_read-only.2.xml","OneNowOneLater_read-only")
ERROR = !WARNING: BATpropcheck: BAT 1000000129_rid_level(-4036) was incorrectly marked keyed!
MAPI = monetdb@localhost:36938
QUERY = pf:add-doc("/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/source/pathfinder/tests/BugTracker/Tests/pf_add-doc_property_error.SF-1706438.doc.xml","OneNowOneLater_updateable.2.xml","OneNowOneLater_updateable",10)
ERROR = !WARNING: BATpropcheck: BAT 1000000135_rid_level(-4102) was incorrectly marked keyed!
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/corrupt_after_update.SF-1706640.err.00.html
ERROR = !WARNING: BATpropcheck: BAT tmp_10560(4464) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_10560(4464) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_10561(4465) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_10561(4465) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_10511(4425) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_10511(4425) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_10616(4494) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_10616(4494) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_10606(4486) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_10606(4486) remains marked radix-clustered on 32 bits; not checked!
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1204095-a.out.00.html
!WARNING: BATpropcheck: BAT doc_timestamp(54) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT doc_timestamp(54) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT collection_name(27) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT collection_name(27) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT collection_size(50) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT collection_size(50) remains marked radix-clustered on 32 bits; not checked!
## Comment 7402
Date: 2007-05-04 08:16:07 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The first one
(http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/pf_add-doc_property_error.SF-1706438.err.00.html)
seems to work fine, again, after Peter removed msync() from the pathfindercode, again.
## Comment 7403
Date: 2007-05-05 08:44:20 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The following checkins are possibly related:
===================================================================
2007/05/03 - boncz: MonetDB4/src/modules/plain/bat.mx,1.14
- make msync() work with [void,void] bats
===================================================================
2007/05/03 - sjoerd: pathfinder/runtime/pf_support.mx,1.223
Two more occurrences: use the new built-in function densebat instead of MIL loop.
===================================================================
2007/05/03 - sjoerd: pathfinder/runtime/pathfinder.mx,1.355
pathfinder/runtime/pf_support.mx,1.222
Use the new built-in function densebat instead of MIL loop.
===================================================================
2007/05/03 - sjoerd: pathfinder/runtime/pf_support.mx,1.224
Figure out whether the MAP_PID bat got changed by the update.
If it wasn't, don't log it and don't overwrite the master.
We need to pass this information to ws_precommit as well, but that
needs to be implemented still.
Peter, when you made your changes to ws_precommit, change the call to
ws_precommit here as well.
===================================================================
2007/05/03 - boncz: pathfinder/runtime/pathfinder.mx,1.356
pathfinder/runtime/pf_support.mx,1.225
pathfinder/runtime/xrpc_client.mx,1.33
pathfinder/runtime/xrpc_server.mx,1.39
pathfinder/runtime/xrpc/admin/admin.js,1.7
further changes to stabilize MXQ under tpox
- when two overlapping transactions modify map_pid, abort the second that tries to commit.
(this can be avoided at commit time by adapting the master map_pid from the committed image)
- fix bug in ws_precommit: access to page_ws/attr_ws should be short-locked
- allow again non .xml/.XML name (sorry Djoerd, I had forgotten!)
- HTTP-xrpc document access should now use the prefix doc/
(e.g. http://host:50001/xrpc/doc/foo.xml == xrpc://host/doc/foo.xml)
===================================================================
2007/05/03 - boncz: pathfinder/runtime/pathfinder.mx,1.357
- exit msync() again
regrattably, this cannot be used yet, here -- it makes no sense in a checkpoint
system because the checkpoints saves the entire bats anyway
msync could be used later to sync new pages in order to offload logging volume
===================================================================
## Comment 7404
Date: 2007-05-05 18:24:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
In fact, the latter two, i.e.,
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/corrupt_after_update.SF-1706640.err.00.html
and
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1204095-a.out.00.html
already fail since 2007-04-26.
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-25
## Comment 7405
Date: 2007-05-05 18:26:32 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
See also
https://sourceforge.net/tracker/index.php?func=detail&aid=1712458&group_id=56967&atid=482468
## Comment 7406
Date: 2007-05-05 23:50:14 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The last one
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1204095-a.out.00.html
only fails due to circumsatnce created by the previous tests in the same directory.
The actual problem is easily reproducible;
fro details, see its own bug report 1713584
"delete on committed dense persistent BAT causes prop. error"
at
https://sourceforge.net/tracker/index.php?func=detail&aid=1713584&group_id=56967&atid=482468
## Comment 7407
Date: 2007-05-07 18:13:30 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The last one
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1204095-a.out.00.html
has been fixed by Sjoerd's fix for 1713584 (http://sourceforge.net/tracker/index.php?func=detail&aid=1713584&group_id=56967&atid=482468).
Hence, only
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_BugTracker/corrupt_after_update.SF-1706640.err.
00.html
remains failing.
## Comment 7408
Date: 2007-05-12 12:57:30 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed BUNins() & BUNappend() in MonetDB/src/gdk/gdk_bat.mx the same way as Sjoerd recently did with BUNdelete_():
When a dense (and hence sorted) column becomes non-sorted due to an update, it (obviously) also become non-dense;
hence, we need to unset no only the sortedness property, but also the denseness property.
## Comment 7409
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712463 at http://sourceforge.net/support/tracker.php?aid=1712463
| PF: various tests fail since 2007.05.04 | https://api.github.com/repos/MonetDB/MonetDB/issues/1322/comments | 0 | 2020-11-30T09:33:04Z | 2024-06-27T11:24:17Z | https://github.com/MonetDB/MonetDB/issues/1322 | 753,313,621 | 1,322 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:42:58 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-17 12:44:27 +0200
## Comment 7392
Date: 2007-05-04 02:42:58 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q4x.out.00.html
xquery("xml","doc(\"users.xml\")").printf();
!ERROR: err:FORG0005: function fn:exactly-one expects exactly one value.
!ERROR: interpret: no matching MIL operator to 'printf(void)'.
!MAYBE YOU MEAN:
! printf(str, ..any..) : void
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q6a.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q6b.out.00.html
incomplete result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_WebSite/books1.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_WebSite/books2.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_WebSite/books3.out.00.html
incomplete result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/inserting_multiple_elements.SF-1590583.out.00.html
xquery("xml","let $batch := (<beer/>,<mies/>) return do insert $batch after exactly-one(doc(\""+TST+"\")//noot)").printf();
!ERROR: err:FORG0005: function fn:exactly-one expects exactly one value.
## Comment 7393
Date: 2007-05-05 09:02:31 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The follwoing checkins are possibly related:
===================================================================
2007/05/01 - sjoerd: MonetDB/src/gdk/gdk_setop.mx,1.54
Fix setting of keyness property in [ks]union.
===================================================================
2007/05/01 - sjoerd: pathfinder/runtime/pf_support.mx,1.220
Don't indicate that we're done reading our snapshot too early.
===================================================================
2007/05/01 - sjoerd: pathfinder/runtime/pathfinder.mx,1.353
Fixed various bugs that surfaced in TPoX testing.
- index maintenance must take into account that indexes for frequent
values are not stored;
- access to ws_overlaps_ws must happen with short lock held.
===================================================================
2007/05/01 - sjoerd: MonetDB4/src/modules/plain/bat.mx,1.12
Two new builtin commands:
- densebat - quickly create a new [void,void] bat with content;
- accdel - delete accelerator.
===================================================================
2007/05/01 - sjoerd:
MonetDB4/tests/Availability/Tests/00_Mserver.stable.out,1.13
MonetDB4/tests/Availability/Tests/00_Mserver.stable.out.STATIC,1.7
MonetDB4/tests/Availability/Tests/00_Mserver.stable.out.WindowsNT,1.11
MonetDB4/tests/Availability/Tests/01_Modules_static.stable.out,1.5
Approved after adding commands accdel and densebat.
===================================================================
2007/05/01 - sjoerd: pathfinder/compiler/mil/milprint_summer.c,1.368
pathfinder/compiler/mil/milprint_summer.c,1.318.2.33
When you want to use a library function, make sure the appropriate
include file is included. For alloca there is an extra twist: for the
way to include alloca.h see e.g. monet_utils.mx. Do not hide the lack
of a declaration with a cast. The compiler then thinks the function
returns an int which may well be smaller in space than the actual
value, and hence some bytes of the value may get lost.
Having said this, alloca should *not* be used in a loop. It allocates
memory in each iteration which is only freed at the end of the
function. If the loop loops many times, that can be a lot of memory
which is allocated on the stack (which is a very finite resource!).
Here ends today's lesson.
===================================================================
2007/05/02 - sjoerd: pathfinder/configure.ag,1.104
pathfinder/configure.ag,1.90.2.13
pathfinder/compiler/mil/milprint_summer.c,1.369
pathfinder/compiler/mil/milprint_summer.c,1.318.2.34
Include malloc.h.
===================================================================
2007/05/02 - sjoerd: pathfinder/configure.ag,1.105
pathfinder/configure.ag,1.90.2.14
pathfinder/compiler/mil/milprint_summer.c,1.370
pathfinder/compiler/mil/milprint_summer.c,1.318.2.35
Include stdlib.h instead of malloc.h.
===================================================================
2007/05/02 - boncz: MonetDB4/src/modules/plain/bat.mx,1.13
- msync(bat[void,T], oid lo, [oid hi], bit async) : bat[void,T]
if the BAT has memory mapped BUN and tail heaps, these are
(partially) synced to backing storage.
the portion to be synced is given by a head oid-range.
semantically, msync is a no-op: its bat parameter is returned as-is.
===================================================================
2007/05/02 - boncz: pathfinder/runtime/pathfinder.mx,1.354
pathfinder/runtime/pf_support.mx,1.221
- use msync for incremental shredding, rather than save
in a mixed update load with updates and shreds, saving the entire bat
also saves the dirty pages caused by the updates rather than just he
shred. performance-wise it is better to batch those until the next checkpoint.
- fix perf issue that caused the nsloc index to be ignored on small
context sequences (surfaced in xmark Q6/Q7)
===================================================================
2007/05/02 - boncz: MonetDB/src/gdk/gdk_posix.mx,1.140
MonetDB/src/gdk/gdk_storage.mx,1.127
forgot this one:
- MT_msync gets one more parameter (offset)
===================================================================
2007/05/02 - stmane: pathfinder/compiler/mil/milprint_summer.c,1.371
pathfinder/compiler/mil/milprint_summer.c,1.318.2.36
new day new lesson:
When replacing alloca in a loop be malloc/realloc,
make sure that you do not forget/overwrite
the origin of the (m|re)alloc'ed buf-fer;
this should keep MonetDB?XQuery from segfaulting
(e.g.) when loading an XQuery module
(
new day, still the old lesson:
testing (would have) found this problem
)
===================================================================
## Comment 7394
Date: 2007-05-08 18:53:49 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
For
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q4x.out.00.html
and
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/inserting_multiple_elements.SF-1590583.out.00.html
extended testing reveals that newly inserted nodes are present in the document and do show up when their parent is retrieved, but they cannot be retrieved themselves:
========
xquery("xml","doc(\""+TST+"\")").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><aap>
Hello World
</aap></XQueryResult>
--------
xquery("xml","doc(\""+TST+"\")/aap").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><aap>
Hello World
</aap></XQueryResult>
--------
xquery("xml","do insert <noot>Bes</noot> as first into exactly-one(doc(\""+TST+"\")/aap)").printf();
--------
xquery("xml","doc(\""+TST+"\")").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><aap><noot>Bes</noot>
Hello World
</aap></XQueryResult>
--------
xquery("xml","doc(\""+TST+"\")/aap").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult><aap><noot>Bes</noot>
^^^^^^^^^^^^^^^^ correct!
Hello World
</aap></XQueryResult>
--------
xquery("xml","doc(\""+TST+"\")/aap/noot").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult></XQueryResult>
^^ <noot>Bes</noot> MISSING!!
--------
xquery("xml","doc(\""+TST+"\")//noot").printf();
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult></XQueryResult>
^^ <noot>Bes</noot> MISSING!!
--------
xquery("xml","let $batch := (<beer/>,<mies/>) return do insert $batch after exactly-one(doc(\""+TST+"\")//noot)").printf();
!ERROR: err:FORG0005: function fn:exactly-one expects exactly one value.
!ERROR: interpret: no matching MIL operator to 'printf(void)'.
!MAYBE YOU MEAN:
! printf(str, ..any..) : void
========
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q6a.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q6b.out.00.html
fail only as
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_R/Q4x.out.00.html
fails first.
## Comment 7395
Date: 2007-05-14 14:59:16 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
See also
1718622 'PF: strange behaviour of "rename"'
http://sourceforge.net/tracker/index.php?func=detail&aid=1718622&group_id=56967&atid=482468
## Comment 7396
Date: 2007-05-14 15:55:48 +0200
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
and this:
[ monetdb-Bugs-1718635 ] PF: node's sibling not accessable after deletion of the node
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1718622&group_id=56967
## Comment 7397
Date: 2007-05-14 15:57:57 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
See also
1718635 "PF: node's sibling not accessable after deletion of the node"
http://sourceforge.net/tracker/index.php?func=detail&aid=1718635&group_id=56967&atid=482468
## Comment 7398
Date: 2007-05-17 12:44:27 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed with Peter's fix for
1718622 'PF: strange behaviour of "rename"'
http://sourceforge.net/tracker/index.php?func=detail&aid=1718622&group_id=56967&atid=482468
1718635 "PF: node's sibling not accessable after deletion of the node"
http://sourceforge.net/tracker/index.php?func=detail&aid=1718635&group_id=56967&atid=482468
1718709 "PF: select deleted node results in error"
https://sourceforge.net/tracker/index.php?func=detail&aid=1718709&group_id=56967&atid=482468
Thanks!
## Comment 7399
Date: 2007-05-20 10:52:50 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Apparently, I was too quick to close this one;
the WebSite tests book[123] still fail;
Since all other problems initially reported here
are indeed fixed,
let's continuing with new bug report
[ 1722078 ] PF: WebSite tests books[123] fail since 2007.05.03
http://sourceforge.net/tracker/index.php?func=detail&aid=1722078&group_id=56967&atid=482468
## Comment 7400
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712461 at http://sourceforge.net/support/tracker.php?aid=1712461
| PF: several tests fail since 2007.05.03 | https://api.github.com/repos/MonetDB/MonetDB/issues/1321/comments | 0 | 2020-11-30T09:33:01Z | 2024-06-27T11:24:16Z | https://github.com/MonetDB/MonetDB/issues/1321 | 753,313,574 | 1,321 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:41:01 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-18 06:10:45 +0200
## Comment 7387
Date: 2007-05-04 02:41:01 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_XMP/Q06.out.00.html
wrong/incomplete result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_STRING/Q5.out.00.html
wrong result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qs29.out.00.html
incomplete result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_MBench/qa02.err.00.html
ERROR = !ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_1030,tmp_1020) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_1030,tmp_736) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: CTrefine: both BATs must have the same cardinality and their heads must form a 1-1 match.
!ERROR: CTrefine: operation failed.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/complex-XQuery-batbat_lng_add_inplace-error-OK.SF-1562868.err.00.html
ERROR = !ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2664,tmp_2637) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2664,tmp_2632) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2657,tmp_2653) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2657,tmp_2517) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2655,tmp_2653) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2655,tmp_2517) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: BATcount: BAT required.
!ERROR: BATfetchjoin(tmp_2670,mapi_1_int_values) does not hit always (|bn|=0 != 1=|l|) => can't use fetchjoin.
!ERROR: mposjoin: first two input BATs (pre & cont) must be head-aligned.
!ERROR: CMDmposjoin: operation failed.
!ERROR: interpret_params: CTrefine(param 2): evaluation error.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/complex-XQuery-batbat_lng_add_inplace-error-KO.SF-1562868.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/complex-XQuery-batbat_lng_add_inplace-error-KO2.SF-1562868.err.00.html
ERROR = !ERROR: err:XPTY0004: cast to 'string' does not allow empty sequences to be casted.
## Comment 7388
Date: 2007-05-05 15:32:01 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-27
## Comment 7389
Date: 2007-05-08 14:53:49 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
For completeness, also find attached a list of possibly related checkins propagated from the release branch(es).
File Added: Checkins_2007-04-27.x
## Comment 7390
Date: 2007-05-18 18:10:45 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
This was also related to the optimization for loop-lifted join (Jan R.) that did not fully work.
Regrettably, it appears there there are now problems with the
tests/W3C_usas_cases/XMP/Q04
that tests ordering across two for-loops.
That is clearly anothour bug, though.
However, I consider complex ordering (other than a single non-for-loop nested order by over some sequence) in mps unfixable. Maybe Jan could do it, but I doubt he wants it.
## Comment 7391
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712459 at http://sourceforge.net/support/tracker.php?aid=1712459
| PF: several tests fail since 2007.04.28 | https://api.github.com/repos/MonetDB/MonetDB/issues/1320/comments | 0 | 2020-11-30T09:32:59Z | 2024-06-27T11:24:15Z | https://github.com/MonetDB/MonetDB/issues/1320 | 753,313,543 | 1,320 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:38:49 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-05-12 12:57:43 +0200
## Comment 7382
Date: 2007-05-04 02:38:49 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_Update/update.err.00.html
ERROR = !WARNING: BATpropcheck: BAT tmp_1627(919) was incorrectly marked keyed!
!WARNING: BATpropcheck: BAT tmp_1466(822) was incorrectly marked keyed!
(only with 64-bit OIDs)
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQUF_Parts/Q3b2.out.00.html
!WARNING: BATpropcheck: BAT tmp_1617(911) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1617(911) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1555(877) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1555(877) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1125(597) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1125(597) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1531(857) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1531(857) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1534(860) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1534(860) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1553(875) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1553(875) remains marked radix-clustered on 32 bits; not checked!
!WARNING: BATpropcheck: BAT tmp_1554(876) was incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_1554(876) remains marked radix-clustered on 32 bits; not checked!
(only with 64-bit OIDs)
## Comment 7383
Date: 2007-05-05 15:27:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-25
## Comment 7384
Date: 2007-05-05 18:26:15 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
See also
https://sourceforge.net/tracker/index.php?func=detail&aid=1712463&group_id=56967&atid=482468
## Comment 7385
Date: 2007-05-12 12:57:43 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Fixed BUNins() & BUNappend() in MonetDB/src/gdk/gdk_bat.mx the same way as Sjoerd recently did with BUNdelete_():
When a dense (and hence sorted) column becomes non-sorted due to an update, it (obviously) also become non-dense;
hence, we need to unset no only the sortedness property, but also the denseness property.
## Comment 7386
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712458 at http://sourceforge.net/support/tracker.php?aid=1712458
| PF: prop. err. in update tests since 2007.04.26 (64-bit OIDs | https://api.github.com/repos/MonetDB/MonetDB/issues/1319/comments | 0 | 2020-11-30T09:32:56Z | 2024-06-27T11:24:14Z | https://github.com/MonetDB/MonetDB/issues/1319 | 753,313,520 | 1,319 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:36:09 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-18 04:26:23 +0200
## Comment 7377
Date: 2007-05-04 02:36:09 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/benchmarks_XPathMark/A11.out.00.html
empty result
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_XQuery/order.out.00.html
wrong attribute name
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugDay_2005-12-19_0.9.3/zero-or-one_merged-union_2.SF-1221336.out.00.html
extra nodes in result
## Comment 7378
Date: 2007-05-05 15:24:38 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-03
## Comment 7379
Date: 2007-05-07 18:15:07 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
The latter two
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_XQuery/order.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/tests_BugDay_2005-12-19_0.9.3/zero-or-one_merged-unio
n_2.SF-1221336.out.00.html
have been fixed by JanR's fix for 1712578 (http://sourceforge.net/tracker/index.php?func=detail&aid=1712578&group_id=56967&atid=482468).
Hence, only
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests10
3/GNU.64.64.d-Fedora6/benchmarks_XPathMark/A11.out.00.html
remains failing.
## Comment 7380
Date: 2007-05-18 16:26:23 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
only the XPathmark A11 bug was still giving the wrong output at the time of the fix (so the other problems were different ones)
the cause was one substitue of ":" into NS_ACCELK_SEP too many, because fn:name should format the result separated by ":"
fixed
## Comment 7381
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712456 at http://sourceforge.net/support/tracker.php?aid=1712456
| PF: various tests fail since 2007.04.04 | https://api.github.com/repos/MonetDB/MonetDB/issues/1318/comments | 0 | 2020-11-30T09:32:54Z | 2024-06-27T11:24:14Z | https://github.com/MonetDB/MonetDB/issues/1318 | 753,313,487 | 1,318 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:34:51 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-18 04:13:02 +0200
## Comment 7372
Date: 2007-05-04 02:34:51 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1648854.out.00.html
incomplete result
## Comment 7373
Date: 2007-05-05 15:15:26 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-02
## Comment 7374
Date: 2007-05-18 16:12:23 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
test output was correct
modified the test so it now probably does what was intended
## Comment 7375
Date: 2007-05-18 17:12:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Just for info:
The original test ran fine from 11.03.2007 (when it was added by Peter) til 02.04.2007.
With neither the test itself nor its output being changed,
it started failing from 03.04.2007 onwards, only with the development head;
with the XQuery_0-16 release branch, the original unmodified test still works fine.
## Comment 7376
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712455 at http://sourceforge.net/support/tracker.php?aid=1712455
| PF: test fails since 2007.04.03 | https://api.github.com/repos/MonetDB/MonetDB/issues/1317/comments | 0 | 2020-11-30T09:32:51Z | 2024-06-27T11:24:13Z | https://github.com/MonetDB/MonetDB/issues/1317 | 753,313,460 | 1,317 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:33:14 +0200
From: @drstmane
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2007-05-07 08:13:16 +0200
## Comment 7368
Date: 2007-05-04 02:33:14 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_W3C_use_cases_XQ_XMP/Q04alg.err.00.html
fatal error: projection results in duplicate attribute tem1' (attributes 7 and 3)
halted in /ufs/manegold/_/scratch0/Monet/Testing/Current/source/pathfinder/compiler/algebra/logical.c (PFla_project_), line 783
## Comment 7369
Date: 2007-05-05 15:13:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-04-02
## Comment 7370
Date: 2007-05-07 20:13:16 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Seems to have been fixed ("by coincidence"?) by JanR's(?) today's checkins.
(see testweb tomorrow).
## Comment 7371
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712454 at http://sourceforge.net/support/tracker.php?aid=1712454
| PF: tests/W3C_use_cases/XQ/XMP/Q04alg fails since 2007.04.03 | https://api.github.com/repos/MonetDB/MonetDB/issues/1316/comments | 0 | 2020-11-30T09:32:49Z | 2024-06-27T11:24:12Z | https://github.com/MonetDB/MonetDB/issues/1316 | 753,313,427 | 1,316 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:28:09 +0200
From: @drstmane
To: Maurice van Keulen <<mvankeulen>>
Version: -- development
Last updated: 2007-05-25 11:57:17 +0200
## Comment 7358
Date: 2007-05-04 02:28:09 +0200
From: @drstmane
Mtest10..2007.03.31_09-02-49..out:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1652527.err.00.html
ERROR = !ERROR: assert_order(tmp_1657): tail was not ordered.
!ERROR: CMDassert_order: operation failed.
## Comment 7359
Date: 2007-05-05 15:08:26 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-03-30
## Comment 7360
Date: 2007-05-08 14:53:43 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
For completeness, also find attached a list of possibly related checkins propagated from the release branch(es).
File Added: Checkins_2007-03-30.x
## Comment 7361
Date: 2007-05-18 15:32:52 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
fixed
## Comment 7362
Date: 2007-05-20 11:00:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
re-opened:
the error is indeed gone, again;
however, the output (still) differes from the original one
(that was fine until 2007.03.30);
see
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugsViaSourgeforce/ID.1652527.out.00.html
## Comment 7363
Date: 2007-05-20 11:10:05 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
see also
[ 1721701 ] PF: "!ERROR: Division by zero" after latest checkins
http://sourceforge.net/tracker/index.php?func=detail&aid=1721701&group_id=56967&atid=482468
## Comment 7364
Date: 2007-05-25 02:25:36 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Maurice, could you please check whether the curent output is ok (same for 1721701).
if so, these bgs an be closed.
## Comment 7365
Date: 2007-05-25 09:27:54 +0200
From: Maurice van Keulen <<mvankeulen>>
Logged In: YES
user_id=654674
Originator: NO
Although I initially thought that the current output does not look okay, because there is an empty movie-element (<movie/>) in the current output, while there should be an entire subtree there, this is most probably an effect of illegal use of the pf:nid and id-functions in the query, which worked in the beginning, but does not work any longer. The query does give the right result if I substitute the "integrate_e" function with the current version in probxml-module which does not make illegal use of pf:nid and id anymore.
I will adapt the test to use the proper "integrate_e" function in the probxml-module. Leaving this bug open until then.
## Comment 7366
Date: 2007-05-25 11:57:16 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
approved new output
## Comment 7367
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712451 at http://sourceforge.net/support/tracker.php?aid=1712451
| PF: test fails on assert_order() since 2007.03.31 | https://api.github.com/repos/MonetDB/MonetDB/issues/1315/comments | 0 | 2020-11-30T09:32:47Z | 2024-06-27T11:24:11Z | https://github.com/MonetDB/MonetDB/issues/1315 | 753,313,386 | 1,315 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:24:58 +0200
From: @drstmane
To: Wouter Alink <<alink>>
Version: -- development
Last updated: 2007-06-07 12:10:37 +0200
## Comment 7344
Date: 2007-05-04 02:24:58 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_StandOff_basic/test2.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_StandOff_basic/lltest.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_StandOff_video/scenes.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_StandOff_video/songs.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_StandOff_video/artists.out.00.html
incomplete result
## Comment 7345
Date: 2007-05-08 11:53:09 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
StandOff it Wouter's "baby".
Feel free to re-assign in case you think someone else is "responsible" and/or should have a look.
## Comment 7346
Date: 2007-06-05 18:34:35 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Hi Wouter
Are you fixing this for the release?
Peter
## Comment 7347
Date: 2007-06-05 18:38:45 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Hi Wouter
Are you fixing this for the release?
Peter
## Comment 7348
Date: 2007-06-06 20:03:46 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
i tried to look into this last night.
i didn't know where to look for the bug so i analysed the difference between the code just before 2007-03-13 using:
cvs diff -D 20070306 -D 20070313
but there seem to be no changes in pathfinder that are directly related to standoff.
the bug seems to be inside the step itself, the correct mil seems to be generated. the only thing is that the primitive operator (in case of test 'lltest' this is 'loop_lifted_select_narrow_step_with_nsloc_test()') produces an empty result (which is not the expected result).
i don't know if the indices are generated correctly, but that might be the problem. i'll have to look into that. does somebody (peter, stefan) have an idea of where i should look? (for the record: i didn't check in anything myself around 2007-03-13)
(p.s. i'm going on holiday friday morning, so i don't know if i have enough time to fix it)
## Comment 7349
Date: 2007-06-06 20:19:32 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Attached you find the log messages of all checking between the (wrt. Standoff) successful testing run on 2007/03/12 and the testing run on 2007/03/13 that first showed the above tests failing.
Maybe this helps...
File Added: Checkins_2007-03-12
## Comment 7350
Date: 2007-06-06 21:05:13 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
if excluding changes to amdb, buildtools and changes to pathfinder/NT and pathfinder/xrpc/admin/*.(js|html) (is this a valid assumption?) the below diff can be made. from what i understand of these changes the only checkin that could in a very far-fetched way be associated to standoff steps is the pathfinder.mx diff. i guess i'll have to reverse that diff and see if that is the cause. any better ideas are very welcome.
Index: pathfinder.mx
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/runtime/pathfinder.mx,v
retrieving revision 1.302
retrieving revision 1.304
diff -r1.302 -r1.304
28c28
< @' $Id: pathfinder.mx,v 1.302 2007/03/11 23:26:36 boncz Exp $
---
> @' $Id: pathfinder.mx,v 1.304 2007/03/13 04:13:03 boncz Exp $
4109c4109,4124
< if (ATOMfromstr(xquery_types[t].monet_tpe, &ctx->val, &ctx->vallen, argval[l]) <= 0)
---
vim > if (xquery_types[t].monet_tpe == TYPE_str) {
> /* we no longer do fancy strTostr for strings; just copied as is */
> char *t, *s = argval[l];
> int len = strlen(s)*2;
> int sep = *s++;
> if (len <= ctx->vallen) {
> if (ctx->val) GDKfree(ctx->val);
> ctx->val = GDKmalloc(len+1);
> if (ctx->val == NULL) return "xquery_function_call: malloc failure.\n";
> }
> for(t=(char*) ctx->val; *s && *s != sep; s++) {
> if (*s == '\\') *t++ = '\\'; /* very usefule on windows for conserving paths */
> *t++ = *s;
> }
> *t = 0;
> } else if (ATOMfromstr(xquery_types[t].monet_tpe, &ctx->val, &ctx->vallen, argval[l]) <= 0)
Index: ../modules/pftijah/pftijah.mx
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/modules/pftijah/pftijah.mx,v
retrieving revision 1.106
retrieving revision 1.107
diff -r1.106 -r1.107
29a30,31
> .USE pathfinder;
>
Index: ../configure.ag
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/configure.ag,v
retrieving revision 1.101
retrieving revision 1.102
diff -r1.101 -r1.102
231a232,233
> AC_CHECK_FUNCS(inet_ntop)
>
Index: serialize.mx
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/runtime/serialize.mx,v
retrieving revision 1.90
retrieving revision 1.91
diff -r1.90 -r1.91
2037c2037,2041
< if (*file == DIR_SEP || *file == '/') {
---
> if (DIR_SEP != '/') {
> char *s = file; /* normalize path on windows to windows DIR_SEP */
> do { if (*s == '/') *s = DIR_SEP; } while(*(++s));
> }
> if (*file == DIR_SEP) {
Index: xrpc_server.mx
===================================================================
RCS file: /cvsroot/monetdb/pathfinder/runtime/xrpc_server.mx,v
retrieving revision 1.22
retrieving revision 1.24
diff -r1.22 -r1.24
24c24
< @' $Id: xrpc_server.mx,v 1.22 2007/03/10 10:37:45 boncz Exp $
---
> @' $Id: xrpc_server.mx,v 1.24 2007/03/13 03:37:07 boncz Exp $
256c256
< "charset=\"utf-8\"\r\n\r\n" \
---
> "Content-type: text/xml; charset=\"utf-8\"\r\n\r\n" \
935c935
< "Content-Type: application/soap+xml; "
---
> "Content-type: text/xml; "
1021c1021,1027
< (void) inet_ntop(AF_INET, (void*) &addr, dotted, 80);
---
> ifdef HAVE_INET_NTOP
> (void) inet_ntop(AF_INET, (void*) &addr, dotted, sizeof(dotted));
> else
> /* not normally thread safe, but on Windows where we need this it
> * actually *is* thread safe */
> strncpy(dotted, inet_ntoa(addr), sizeof(dotted));
> endif
## Comment 7351
Date: 2007-06-06 21:12:03 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
In fact, the prolems we see now might be different from those that occured on 2007/03/13.
There was a period in between, where the StandOff tests did not run properly at all, with apparently noone noticeing/caring, until:
===================================================================
2007/04/18 - stmane: tests/StandOff/StandOff.py,1.4
make StandOff tests run, again:
Mserver needs to be started with --dbinit="module(pathfinder);"
===================================================================
Hence, while changes 2007/03/12 originally made these test fail "somehow", any changes between 2007/03/13 and 2007/05/04 (when I filed this bug report) might have made tests fail the way they do now.
Sorry that I did not watch the StandOff testing behaviour and results more frequently/closely/accurately/enthusiastically/conscientiously ...
## Comment 7352
Date: 2007-06-06 22:27:39 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
Hi,
Like Stefan said, these changes are likely not related to the problem.
To refresh the memory: a lot has changed in the runtime by the introduction of:
(1) updates
(2) indices
and their interaction !!!
Wouter.. three days to go for the release: would you conside doing some actual debugging. I think that is in order here.
Peter
## Comment 7353
Date: 2007-06-06 22:47:08 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
For info: reverting runtime/pathfinder.mx 1.302->1.304 did not change anything for the StandOff tests in my case.
## Comment 7354
Date: 2007-06-06 22:58:09 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
got a little further:
found at least one problem (hopefully that is the only one): when 'loop_lifted_select_narrow_step_with_nsloc_test' arrives at line 824 of pf_standoff, the 'cands' bat is empty. it looks like the candidates are pruned while they shouldn't have been. the candidate-list should include all element-nodes having a region or optionally all element-nodes.
don't know how/where to fix it yet, still working on it.
p.s. stefan, you're not the one to blame... ;)
## Comment 7355
Date: 2007-06-06 23:39:09 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Originator: NO
hopefully fixed (":" was changed to NS_ACCEL_SP). please close if tests turn out OK.
## Comment 7356
Date: 2007-06-07 00:10:37 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Works for me.
Consider it fixed.
Thanks!
## Comment 7357
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712450 at http://sourceforge.net/support/tracker.php?aid=1712450
| PF: StandOff tests fail since 2007.03.13 | https://api.github.com/repos/MonetDB/MonetDB/issues/1314/comments | 0 | 2020-11-30T09:32:43Z | 2024-06-27T11:24:10Z | https://github.com/MonetDB/MonetDB/issues/1314 | 753,313,351 | 1,314 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-04 02:23:35 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2007-05-18 03:13:15 +0200
## Comment 7340
Date: 2007-05-04 02:23:35 +0200
From: @drstmane
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/bogus_text_step.SF-1537500.mps.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora6/tests_BugTracker/bogus_text_step.SF-1537500.mps.err.00.html
empty result instead of (expected!) error message
## Comment 7341
Date: 2007-05-05 15:02:51 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Find attached a list of possibly related checkins.
File Added: Checkins_2007-03-07
## Comment 7342
Date: 2007-05-18 15:13:15 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Originator: NO
as the attribute steps yields an empty sequence, it is diffucult to raise this error. Maybe previously it got raised if the attribute tep would have produced a constant representation of 'kind'.
To detect this kind of semantic errors, the test should be done at compile-time during type analysis.
In this case, though, the error is not mandated by the XQuery specs; it is our restriction not to support XPath on attribute nodes. So maybe we should be glad that in case of empty input sequences, we can still comply with the spec and refrain from giving the error.
## Comment 7343
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1712449 at http://sourceforge.net/support/tracker.php?aid=1712449
| PF: "bogus_text_step" test fails since 2007.03.08 | https://api.github.com/repos/MonetDB/MonetDB/issues/1313/comments | 0 | 2020-11-30T09:32:40Z | 2024-06-28T07:36:57Z | https://github.com/MonetDB/MonetDB/issues/1313 | 753,313,307 | 1,313 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-03 03:15:11 +0200
From: @grobian
To: Romulo Goncalves <<goncalve>>
Version: -- development
Last updated: 2007-05-03 04:16:36 +0200
## Comment 7332
Date: 2007-05-03 15:15:11 +0200
From: @grobian
Most if not all stable.out files are missing in this directory.
If you don't have them, either
1) handcraft them
2) disable the tests (with a comment) so they aren't run
This way nobody knows what's wrong with them, so please do either one of the above, preferably 1).
## Comment 7333
Date: 2007-05-03 16:16:36 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
The out outputs were added.
However, the tests need to be fixed and for that I need Niels help.
Regards,
Romulo
## Comment 7334
Date: 2007-05-03 16:34:30 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Why has this one been closed --- i.e., what is/was the "Resolution:"?
## Comment 7335
Date: 2007-05-03 16:37:49 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
Niels check in (02-05-2007 00:03)
Update of /cvsroot/monetdb/sql/src/server
In directory sc8-pr-cvs16:/tmp/cvs-serv11322/src/server
Modified Files:
sql_parser.mx sql_psm.mx sql_select.mx sql_semantic.mx
Log Message:
added DROP PROCEDURE cases
added CREATE PROCEDURE
Both untested
fixed bug (crash) in select * from 'variable';
I run the same query (the one in used to present this bug) and it works.
Regards,
Romulo
## Comment 7336
Date: 2007-05-08 22:25:49 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
From Romulo's first comment, I read that this bug has been closed, because the reported problem (missing stable output file(s)) has been "Fixed" by adding the respective file(s).
Please correct me in case I'm wrong.
## Comment 7337
Date: 2007-05-08 22:29:06 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
Yes I closed it, what was the problem?
Did I forget something?
Regards,
Romulo
## Comment 7338
Date: 2007-05-08 22:31:39 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
I got it. The resolution, right?
Sorry....
Regards,
Romulo
## Comment 7339
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1711947 at http://sourceforge.net/support/tracker.php?aid=1711947
| src/test/Update_Delete_action: missing stable output files | https://api.github.com/repos/MonetDB/MonetDB/issues/1312/comments | 0 | 2020-11-30T09:32:37Z | 2024-06-27T11:24:08Z | https://github.com/MonetDB/MonetDB/issues/1312 | 753,313,269 | 1,312 |
[
"MonetDB",
"MonetDB"
] | Date: 2007-05-03 02:54:19 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2007-11-19 05:45:29 +0100
## Comment 7324
Date: 2007-05-03 14:54:19 +0200
From: @drstmane
For all parts of the MonetDB code of both the latest "Stable" release branch as well as the "Current" development trunk, there are tests that fail but have no open bug reports acompanying them.
These tests did once work fine, and now fail after / due to "recent" code changes --- as reported by the daily testing result emails.
The respective developers are welcome to check these tests, fix their respective code and/or approve the new intended / acceptable output according to their code changes.
For details, consult the TestWeb at
http://monetdb.cwi.nl/projects/monetdb/Development/TestWeb/Stable/index.html
and/or
http://monetdb.cwi.nl/projects/monetdb/Development/TestWeb/Current/index.html
## Comment 7325
Date: 2007-05-03 15:08:47 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
It looks to me like most of the SQL tests just fail due to a timeout (in the server) not detected by TestWeb, which result in an (X) instead of a (T).
## Comment 7326
Date: 2007-05-03 15:21:06 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
ignore my last comment. I misread Mtimeout's output of telling its child received signal 11 (SIGSEGV)
## Comment 7327
Date: 2007-05-13 23:03:09 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
For sql4 I cannot (in a very quick scan see) any more problems, ie did the problem go away again?
## Comment 7328
Date: 2007-05-14 13:02:55 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Well, this bug report was no referring to any particular problem or failing test, but was meant to trigger general awareness of the effects and impact of coding and check-ins.
I don't have time any more to report each problem individually, but hope that the respective active developers will use the aids provided by automatic nightly testing to monitor the effects and impact of their code changes (check-ins) also on other parts of the project.
In particular, I haven't been watching MonetDB5 and SQL (both M4 & M5) for quite some time, any more, trusting the respective active developers.
While still watching MonetDB4 & pathfinder quite closely until recently, other obligations and time constraints don't allow me to continue this in the same extensive way; hence, also here, I count on the responsibility of the respective active developers.
## Comment 7329
Date: 2007-06-05 22:46:09 +0200
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
lowered priority or meta/default bug reports
## Comment 7330
Date: 2007-11-19 17:45:29 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: YES
Closing meta-bug report --- there will always be tests that fail after checkins --- these will be fixed, as long as / as soon as someone does care about them.
## Comment 7331
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1711927 at http://sourceforge.net/support/tracker.php?aid=1711927
| Tests failing after "recent" code changes | https://api.github.com/repos/MonetDB/MonetDB/issues/1311/comments | 0 | 2020-11-30T09:32:34Z | 2024-06-28T07:36:57Z | https://github.com/MonetDB/MonetDB/issues/1311 | 753,313,228 | 1,311 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.