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: 2005-04-03 05:47:36 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-04-05 07:39:48 +0200
## Comment 2784
Date: 2005-04-03 17:47:36 +0200
From: @drstmane
Since Thursday March 24 2005, all "bad-???" tests in
compiler// , i.e., all tests that do test error
handling and/or messages of the XQuery compiler fail.
It could be that the changes were intended like this,
but then the stable output should be approved accordingly!
Hence, who ever did change Pathfinder code on (or
after) Wednesday March 23 2005, *please* check, whether
your changes are involved in making the tests fail, and
either fix your code or update the stable output as
intended!
## Comment 2785
Date: 2005-04-04 10:53:44 +0200
From: @drstmane
Logged In: YES
user_id=572415
Apparently, all these tests result in a segmentation fault,
which is most probably not intended:
========
parse error: XQuery parsing failed
halted in
/var/tmp/monet.GNU_32-d.732/pathfinder/compiler/parser/parser.y
(PFparse), line 2238
!ERROR: Segmentation fault.
The Pathfinder compiler experienced an internal problem.
You may want to report this problem to the Pathfinder
development team (pathfinder@inf.uni-konstanz.de).
When reporting problems, please attach your XQuery input,
as well as the following information:
Invocation: command line
Compilation stage: 0
We apologize for the inconvenience...
./bad-001: line 6: 5375 Segmentation fault pf -s1 <<'EOT'
(: This must not be interpreted as /* :)
(/) *
EOT
========
(cf. e.g.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU_32-d_Fedora3/compiler_parser/bad-001.err.00.html
)
Note, the parsing error is intended; however, this obviously
should not result in a segfault...
## Comment 2786
Date: 2005-04-05 15:59:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
The segmentation faults reported here and in BUG #1176566 "correct error messages triggers segfault"
(http://sourceforge.net/tracker/index.php?func=detail&aid=1176566&group_id=56967&atid=482468)
were actually caused by the fact that the call of "setjmp()"
was hidden
in function "hidden_setjmp()" (which was done to keep some
compilers from
complaining about clobbered variables).
Fix:
remove "hidden_setjmp()", again;
instead, we call "setjmp()" before variables are declared.
also moved the "setjmp()" call for the stand-alone compiler
from "main()" to "pf_compiler()";
the "setjmp()" call for the in-lined server-side compiler
is in "pf_compiler_MonetDB()".
Still, the "bad-???" tests in compiler// show differences in
the output:
(1) The return status/value of "pf" in case of an error
seems to have changed.
(2) Some error messages do not appear any more.
In case either (or both) of these changes were intended,
*please* let me know and/or approve the new output!
In case either (or both) of these changes were NOT
intended, *please* check and fix the respective code!
(Unfortunately, I'm apparently not familiar enough with the
code to fix it myself. Sorry!)
## Comment 2787
Date: 2005-04-05 19:11:31 +0200
From: @drstmane
Logged In: YES
user_id=572415
Ok, I finally found the reason for part (2), the missing
error messages:
With the change from "quiet" to "debug", the default
*silently* changed
from "quiet==TRUE" to "debug==0", and thus most messages of
the compiler
where not printed any more.
Changed the default back to "debug=1".
Thus, we get back at least the error messages that cause the
compiler to stop;
other error messages/warning are (still) only appended to
"PFerrbuf",
but this is then never printed |-( ...
(This holds for the stand-alone compiler; I have no idea
what happens
with the in-lined server-side compiler --- there are no
tests for that one,
yet ...)
(
Minor fix on the way:
Basic rules of boolean algebra:
quiet <=> debug==0 => !quite <=> debug!=0
and not
quiet <=> debug==0 => !quite <=> debug==0
)
I haven't found, yet, why the return value/status of "pf"
has changed ...
## Comment 2788
Date: 2005-04-05 19:39:48 +0200
From: @drstmane
Logged In: YES
user_id=572415
Finally, I also found the cause for the changed return
value/status of "pf",
and restored the original behaviour
by passing the error codes through the setjmp/longjmp
up to the final exit()
(yes, this even works with the "ugly" setjmp/longjmp
solution ;-))
## Comment 2789
Date: 2005-10-06 17:08:17 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TEST exists
Test existed before the bug ;-)
## Comment 2790
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1175855 at http://sourceforge.net/support/tracker.php?aid=1175855
| PF: all "bad-???" tests in compiler// fail since Mar 24 2005 | https://api.github.com/repos/MonetDB/MonetDB/issues/510/comments | 0 | 2020-11-30T08:40:04Z | 2024-06-27T11:03:08Z | https://github.com/MonetDB/MonetDB/issues/510 | 753,275,996 | 510 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-04-01 01:02:20 +0200
From: Wouter Alink <<alink>>
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-04-01 02:32:32 +0200
## Comment 2779
Date: 2005-04-01 13:02:20 +0200
From: Wouter Alink <<alink>>
Depth is stored in 'chr'-field causing probably
problems with depths larger than 128 (which is the
NIL-value)
To example documents have been attached. doc01 has 130
node-levels and doc02 has 127 node-levels.
Doc02 shreds fine, but shredding doc01 crashes monet.
I guess setting XML_DEPTH_MAX in shredder.mx to 127
would solve the problem, but that is probably not up to
me to decide ;)
--------------------
[alink@volans src]$ Mserver
Monet Database Server V4.7.1
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(pathfinder);
MonetDB>shred_doc("/ufs/alink/shred_err02.xml","shred_err02.xml");
Shredded XML doc("shred_err02.xml"), total time after
commit=0.055s
MonetDB>shred_doc("/ufs/alink/shred_err01.xml","shred_err01.xml");
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
[alink@volans src]$
## Comment 2780
Date: 2005-04-01 13:03:12 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
included doc02
## Comment 2781
Date: 2005-04-01 14:32:31 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Improved error checking and handling for this case. Maximum
depth still is 256; this works fine with my examples. The
example shred_err01.xml had a depth of 258 (instead of
130!); this will never work indeed when PRE_LEVEL is (void,chr).
## Comment 2782
Date: 2005-10-06 16:58:59 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TEST / SUCCESS
Test added as shred_at_depth_258.SF-1174747
## Comment 2783
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1174747 at http://sourceforge.net/support/tracker.php?aid=1174747
| Mserver crashes when shredding at depth | https://api.github.com/repos/MonetDB/MonetDB/issues/509/comments | 0 | 2020-11-30T08:40:01Z | 2024-06-27T11:03:07Z | https://github.com/MonetDB/MonetDB/issues/509 | 753,275,962 | 509 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-26 09:16:36 +0100
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-29 10:15:57 +0200
## Comment 2774
Date: 2005-03-26 21:16:36 +0100
From: @drstmane
from (e.g.)
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.ntv_32-d/Fedora3/make-k.out :
icc -no-gcc -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.ntv_32-d.25322/pathfinder/compiler -I..
-I/var/tmp/monet.ntv_32-d.25322/pathfinder/compiler -Iinclude
-I/va
r/tmp/monet.ntv_32-d.25322/pathfinder/compiler/include
-I/var/tmp/monet.ntv_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB
-I/var/tmp/monet.ntv_32-d.2
5322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/common
-I/var/tmp/monet.ntv_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/gdk
-I/var/tmp/monet.ntv_32-
d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/monet
-I/var/tmp/monet.ntv_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/plain
-I/var/tmp/monet.ntv
_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/contrib
-I/var/tmp/monet.ntv_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/mapi
-I/var/tmp/mone
t.ntv_32-d.25322/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/C
-I/soft/local/include -we140 -wd1418 -c99 -Wall -w2 -mp1 -O3
-restrict -unroll
-tpp6 -axKWNPB -we266 -Werror
-wd1418,1419,279,310,981,810,444,193,111,177,171,181,764,269,108,188,1357,102,70,1572
-c /var/tmp/monet.ntv_32-d.25322/pathfi
nder/compiler/main.c
/var/tmp/monet.ntv_32-d.25322/pathfinder/compiler/main.c(700):
error 140: too many arguments in function call
if (hidden_setjmp(PFexitPoint) != 0 ) {
^
compilation aborted
for /var/tmp/monet.ntv_32-d.25322/pathfinder/compiler/main.c
(code 2)
make[3]: *** [main.o] Error 2
make[3]: Target `all-am' not remade because of errors.
make[3]: Leaving directory
`/var/tmp/monet.ntv_32-d.25322/pathfinder/Fedora3/compiler'
make[2]: *** [all-recursive] Error 1
make[2]: Target `all' not remade because of errors.
make[2]: Leaving directory
`/var/tmp/monet.ntv_32-d.25322/pathfinder/Fedora3/compiler'
## Comment 2775
Date: 2005-03-29 10:15:57 +0200
From: @peterboncz
Logged In: YES
user_id=591107
sorry about that. with gcc it compiled so I had not noticed.
## Comment 2776
Date: 2005-03-29 10:25:48 +0200
From: @drstmane
Logged In: YES
user_id=572415
No problem. That's what we have nightly testing for ;-)
Thanks for the fix!
## Comment 2777
Date: 2005-10-06 16:46:29 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TESTed by COMPILATION
## Comment 2778
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1171137 at http://sourceforge.net/support/tracker.php?aid=1171137
| PF: too many arguments in function call ("hidden_setjmp") | https://api.github.com/repos/MonetDB/MonetDB/issues/508/comments | 0 | 2020-11-30T08:39:58Z | 2024-06-27T11:03:06Z | https://github.com/MonetDB/MonetDB/issues/508 | 753,275,921 | 508 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-25 03:09:26 +0100
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-04-12 12:43:50 +0200
## Comment 2770
Date: 2005-03-25 15:09:26 +0100
From: @grobian
When trying to compile MonetDB on a Darwin 6.8 machine,
the build process stops at this point:
Making install in tools
g++ -Wall -W -g -Werror -Wno-format
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -D_REENTRANT -o .libs/Mserver
Mserver.o initmodules.o -L/sw/lib -L../monet/.libs
-lmonet
-L/Volumes/Scratch/monet-current/MonetDB/build/src/gdk/.libs
-L../modules/plain/.libs -l_algebra -l_arith -l_bat
-l_kernel -l_str -liconv -l_sys -l_trans -lm
-L../gdk/.libs -lbat
-L/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs
-ldl -lpthread -L../common/.libs -lstream -lmutils -lz
-lbz2
ld: warning multiple definitions of symbol _setenv
/Volumes/Scratch/monet-current/MonetDB/build/src/gdk/.libs/libbat.dylib(gdk_posix.lo)
definition of _setenv
/usr/lib/libpthread.dylib(setenv.So) definition of _setenv
ld: warning suggest use of -bind_at_load, as lazy
binding may result in errors or different symbols being
used
symbol _optarg used from dynamic library
/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs/libmutils.dylib(monet_options.lo)
not from earlier dynamic library
/usr/lib/libSystem.B.dylib(getopt.So)
symbol _optind used from dynamic library
/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs/libmutils.dylib(monet_options.lo)
not from earlier dynamic library
/usr/lib/libSystem.B.dylib(getopt.So)
symbol _optopt used from dynamic library
/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs/libmutils.dylib(monet_options.lo)
not from earlier dynamic library
/usr/lib/libSystem.B.dylib(getopt.So)
symbol _getopt used from dynamic library
/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs/libmutils.dylib(monet_options.lo)
not from earlier dynamic library
/usr/lib/libSystem.B.dylib(getopt.So)
symbol _opterr used from dynamic library
/Volumes/Scratch/monet-current/MonetDB/build/src/common/.libs/libmutils.dylib(monet_options.lo)
not from earlier dynamic library
/usr/lib/libSystem.B.dylib(getopt.So)
ld: Undefined symbols:
_alloca
make[2]: *** [Mserver] Error 1
make[1]: *** [install-recursive] Error 1
make: *** [install-recursive] Error 1
[orion:~] fabian%
(several warnings about redefinitions, and finally the
undefined symbol _alloca.
Relevant bootstrap output:
using /usr/bin/glibtool and /usr/bin/glibtoolize.
automake 1.6.1 is 1.5 or newer. Good.
autoconf 2.59 is 2.57 or newer. Good.
libtool 1.4.2 is 1.4 or newer. Good.
Python 2.2 is 2.0.0 or newer. Good.
## Comment 2771
Date: 2005-04-12 12:43:50 +0200
From: @grobian
Logged In: YES
user_id=963970
It would've been cool if it'd work, but it's just an 'old'
OS, so let's leave it like that and put no additional
efforts into it.
## Comment 2772
Date: 2005-10-06 16:45:28 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TESTed by COMPILATION
## Comment 2773
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1170533 at http://sourceforge.net/support/tracker.php?aid=1170533
| Compilation issues on Darwin 6.8 | https://api.github.com/repos/MonetDB/MonetDB/issues/507/comments | 0 | 2020-11-30T08:39:55Z | 2024-06-28T13:35:17Z | https://github.com/MonetDB/MonetDB/issues/507 | 753,275,891 | 507 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-24 02:32:24 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-24 04:42:36 +0100
## Comment 2766
Date: 2005-03-24 14:32:24 +0100
From: Jan Rittinger <<jan.rittinger>>
dbinit with pathfinder crashes Mserver while it stays alife
if load is done afterwards.
> Mserver --dbinit="module(pathfinder);"
Monet Database Server V4.6.1
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-suse-linux/32bit; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
> Mserver
Monet Database Server V4.6.1
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-suse-linux/32bit; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(pathfinder);
MonetDB>
## Comment 2767
Date: 2005-03-24 16:40:52 +0100
From: @peterboncz
Logged In: YES
user_id=591107
appears to be an old one; I only was the unlucky guy that added
some extra work to the initialization of the pathfinder module (it
now loads pf_support).
Thus, the MIL stack got deeper, calls THRhighwater() and
crashes, because the executing thread is not registered.
Previously, with a more shallow stack, THRhighwater was not
called, and thus it survived.
The problem is that the --dbinit option uses monet_eval() very
early; from within scheduleClient, not through the MIL job queue.
Only when monetInterpreter() is eneterd the thread registers as a
Monet ineterpreter.
fixed now
when the main thread enters serving the MIL job ques it calls i
## Comment 2768
Date: 2005-10-06 16:44:52 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TEST exists
Any runtime pathfinder test, tests this problem
## Comment 2769
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1169869 at http://sourceforge.net/support/tracker.php?aid=1169869
| Mserver --dbinit="module(pathfinder);" crashes | https://api.github.com/repos/MonetDB/MonetDB/issues/506/comments | 0 | 2020-11-30T08:39:52Z | 2024-06-28T13:35:16Z | https://github.com/MonetDB/MonetDB/issues/506 | 753,275,870 | 506 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-24 01:39:12 +0100
From: Jens Teubner <<teubner>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-24 04:45:14 +0100
## Comment 2762
Date: 2005-03-24 13:39:12 +0100
From: Jens Teubner <<teubner>>
The command line Pathfinder compiler no longer prints any
messages if an error occurs:
$ echo '&^$(*&$^*' | ./pf
$ echo $?
0
$
This seems to have something to do with Peter's changes.
## Comment 2763
Date: 2005-03-24 16:45:14 +0100
From: @peterboncz
Logged In: YES
user_id=591107
well, yes, it was Jan's hidden_setjmp() that actually caused it.
what actually happens is hard to comprehend, because after
falling into the failure branch it should go there, but something
appears to be messed up whet the setjmp return value is again
returned through a function call.
maybe even a compiler bug
the resolution (workaround?) was to print the PFerrbuf right
away in hidden_setjmp()
## Comment 2764
Date: 2005-10-06 16:43:33 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: DONE/ TEST EXISTS
## Comment 2765
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1169835 at http://sourceforge.net/support/tracker.php?aid=1169835
| pf no longer prints error messages | https://api.github.com/repos/MonetDB/MonetDB/issues/505/comments | 0 | 2020-11-30T08:39:50Z | 2024-06-28T13:35:16Z | https://github.com/MonetDB/MonetDB/issues/505 | 753,275,834 | 505 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-24 11:30:27 +0100
From: Jan Flokstra <<jflokstra>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-03-24 12:55:12 +0100
## Comment 2752
Date: 2005-03-24 11:30:27 +0100
From: Jan Flokstra <<jflokstra>>
I'm compiling Pathfinder on SuSe9.2 and I'm having
troubles with the HAVE_GC variable. It is not
automaitically set in my configuration. The function
CMDxquery() does not work when this variable is not
set! It always used to work perfect but since the
changes of wednesnight 23/3 it fails. I'm having 340
Segmentation violations in Mtest too. But I think this
is the same problem!
Config:
automake 1.9.1 is 1.5 or newer. Good.
autoconf 2.59 is 2.57 or newer. Good.
libtool 1.5.8 is 1.4 or newer. Good.
Python 2.3.4 is 2.0.0 or newer. Good.
## Comment 2753
Date: 2005-03-24 12:55:12 +0100
From: @peterboncz
Logged In: YES
user_id=591107
as described in HowToStart-PF, you must configure MonetDB
with --enable-gc=DIR
## Comment 2754
Date: 2005-03-24 13:10:24 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
I expect that should be --with-gc=yes or --with-gc=DIR.
## Comment 2755
Date: 2005-03-24 13:23:46 +0100
From: @drstmane
Logged In: YES
user_id=572415
Indeed the code did and does correctly say "--with-gc",
while the documentation both in the check-in messages and
the HowtoStart-PF wrongly says "--enable-gc" ... |-((
## Comment 2756
Date: 2005-03-24 13:38:47 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
This (again) triggers an old problem: Which MonetDB version do we
depend on?
I'm on the 4.6 (MonetDB_4-6) branch here. And I thought that this is what
we had agreed on. My ./configure neither has --enable-gc, nor --with-gc.
Another problem I see is that this GC stuff may raise some problems
when users want to compile MonetDB themselves. The Pathfinder
compilation procedure does not check for the presence of GC in MonetDB
and compiles fine without complaining.
## Comment 2757
Date: 2005-03-24 14:18:18 +0100
From: @peterboncz
Logged In: YES
user_id=591107
which MonetDB version? ==> we are working towards a new
*release* thus we are talking about HEAD.
It will still be possible to use a non-gc MonetDB with pathfinder,
only not in server-side xquery mode, so there is no reason to halt
compilation.
I will add an appropriate runtime error message in the non-gc
MonetDB if one attempts server-side xquery with it.
## Comment 2758
Date: 2005-03-24 14:25:40 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Ran configure with --with-gc=/usr/lib and it works.
## Comment 2759
Date: 2005-03-24 14:41:58 +0100
From: @peterboncz
Logged In: YES
user_id=591107
as described in HowToStart-PF, you must configure MonetDB
with --enable-gc=DIR
## Comment 2760
Date: 2005-10-06 16:39:10 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TESTed by COMPILATION
GC is gone...
## Comment 2761
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1169771 at http://sourceforge.net/support/tracker.php?aid=1169771
| HAVE_GC not properly set on SuSe | https://api.github.com/repos/MonetDB/MonetDB/issues/504/comments | 0 | 2020-11-30T08:39:46Z | 2024-06-28T13:35:15Z | https://github.com/MonetDB/MonetDB/issues/504 | 753,275,809 | 504 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-23 10:14:54 +0100
From: @njnes
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-24 10:39:09 +0100
## Comment 2749
Date: 2005-03-23 22:14:54 +0100
From: @njnes
MapiClient -lsql and MapiClient -lmil used to start an
interactive
client now an extra -i or --interactive is needed.
## Comment 2750
Date: 2005-10-06 16:38:32 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: Done / NO TEST
Cannot test interactively / automatically
## Comment 2751
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1169426 at http://sourceforge.net/support/tracker.php?aid=1169426
| MapiClient doesn't run interactively anymore | https://api.github.com/repos/MonetDB/MonetDB/issues/503/comments | 0 | 2020-11-30T08:39:43Z | 2024-06-27T11:03:00Z | https://github.com/MonetDB/MonetDB/issues/503 | 753,275,778 | 503 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-20 12:25:20 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2006-01-13 09:46:24 +0100
## Comment 2743
Date: 2005-03-20 12:25:20 +0100
From: @drstmane
Since Monday Mar 14 2005, MIL fork() (and hence o.a.,
all MAPI tests) fail when compiled with the PGI
compiler (pgcc) on Debian 3.0 on x86_64 (spin):
========
**ERROR: in routine alloca() there is a
stack overflow: thread 0, max 8180KB, used 8KB, request
200B
========
The only code changes on that day do not seem to be
related at all:
===================================================================
2005/03/14 - vzzzbx: src/modules/plain/str.mx,1.46.2.2
addition to http://sourceforge.net/support/tracker.php?aid=1161583 (infinite loop):
- documentation updated
- return of GDK_FAIL on empty 'sep' string
===================================================================
2005/03/14 - sgoeldner:
src/mapi/clients/perl/DBD/monetdb.pm,1.30.2.8
Unified the handling of subroutine parameters.
This is (mostly) a cosmetic change.
===================================================================
Moreover, the only documented system changes are that
Python-2.4 and vtk-4.4.2 have been installed in /soft/
Debugging "reveals"(?) the following:
========
11:30:18 manegold@spin:~/Monet/Stable/MonetDB $ gdb Mserver
GNU gdb 6.0-debian
Copyright 2003 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public
License, and you are
welcome to change it and/or distribute copies of it
under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show
warranty" for details.
This GDB was configured as "x86_64-linux"...
(gdb) r
Starting program:
/var/tmp/StM/_Stable_PREFIX_MONETDB_PGI_64_d_--enable-debug_/bin/Mserver
Program received signal SIG32, Real-time event 32.
0x0000002a967e2b3e in pthread_getconcurrency () from
/lib64/libpthread.so.0
(gdb) break CMDfork
Breakpoint 1 at 0x2a95f0e1e3: file
/data/home/manegold/Monet/Stable/MonetDB/src/modules/plain/kernel.mx,
line 329.
(gdb) cont
Continuing.
Monet Database Server V4.6.1
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-unknown-linux-gnu/64bit;
dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>fork(ls());
Breakpoint 1, CMDfork (stk=0, lt=0x5ae668,
res=0x7fbffff2b8)
at
/data/home/manegold/Monet/Stable/MonetDB/src/modules/plain/kernel.mx:329
329 if (lt->cnt != @1) {
(gdb) next
721 if (GDKembedded && monet_forks != 0) {
(gdb)
730 c = forkClient(stk, TRUE);
(gdb)
Program received signal SIG32, Real-time event 32.
0x0000002a967e2b3e in pthread_getconcurrency () from
/lib64/libpthread.so.0
(gdb) next
Single stepping until exit from function
pthread_getconcurrency,
which has no line number information.
0x0000002a967e2347 in pthread_create () from
/lib64/libpthread.so.0
(gdb)
Single stepping until exit from function pthread_create,
which has no line number information.
MT_create_thread (t=0x7fbfffec28, f=0x4032e8,
arg=0x7fbfffec24)
at
/data/home/manegold/Monet/Stable/MonetDB/src/gdk/gdk_system.mx:380
380 if (ret == 0)
(gdb)
381 *t = (MT_Id) (((int) newt) +
1); /* use pthread-id + 1 */
(gdb)
383 MT_thread_sigmask(&orig_mask, NULL);
(gdb)
385 return ret;
(gdb)
386 }
(gdb)
monetFork () at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_process.mx:174
174 monet_down_sema(monet_forksync,
"monetFork");
(gdb)
175 if (status)
(gdb)
178 monet_unset_lock(monet_critical,
"monetFork");
(gdb)
179 return p;
(gdb)
180 }
(gdb)
forkClient (stk=0, thread=1) at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_client.mx:299
299 return son;
(gdb)
300 }
(gdb)
CMDfork (stk=0, lt=0x5ae668, res=0x7fbffff2b8) at
/data/home/manegold/Monet/Stable/MonetDB/src/modules/plain/kernel.mx:731
731 if (c == NULL) {
(gdb)
735 res->vtype = TYPE_int;
(gdb)
736 res->val.ival = c->stk;
(gdb)
737 c->tree = lt->yysons[0];
(gdb)
738 QMwakeup(QMenqueue(c->stk, c->tree, 0,
(int) (c - monet_clients)));
(gdb) step
QMenqueue (stk=1, data=0x5d6008, barrier=0, kill=1) at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_queue.mx:586
586 p = newRequest(stk, data, barrier);
(gdb) next
587 p->ev_status = EVAWAIT;
(gdb)
588 p->ev_kill = -kill;
(gdb)
589 putRequest(p);
(gdb)
590 PARDEBUG THRprintf(GDKerr, "QMenqueue:
return key %d\n", p->ev_seqnr);
(gdb)
592 return p->ev_seqnr;
(gdb)
593 }
(gdb)
**ERROR: in routine alloca() there is a
stack overflow: thread 0, max 8180KB, used 9KB, request
200B
CMDfork (stk=0, lt=0x5ae668, res=0x7fbffff2b8) at
/data/home/manegold/Monet/Stable/MonetDB/src/modules/plain/kernel.mx:739
739 lt->yysons[0] = NULL;
(gdb) bt
0 CMDfork (stk=0, lt=0x5ae668, res=0x7fbffff2b8)
at
/data/home/manegold/Monet/Stable/MonetDB/src/modules/plain/kernel.mx:739
1 0x0000002a9567ce8d in .st12948 () at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_interpreter.mx:814
2 0x0000002a956862b5 in interpret_seqblock (stk=0,
lt=0x5b8be8, res=0x7fbffff2b8, scope=0)
at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_interpreter.mx:1737
3 0x0000002a9567c453 in .st12948 () at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_interpreter.mx:724
4 0x0000002a9569bebd in handleRequest
(t=0x2a95b4a7a0, q=0x5d8678, res=0x7fbffff2b8)
at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_queue.mx:534
5 0x0000002a9569c424 in doRequest (t=0x2a95b4a7a0,
preference=0x0)
at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_queue.mx:560
6 0x0000002a956efcb7 in monetInterpreter
(status=0x7fbffff354)
at
/data/home/manegold/Monet/Stable/MonetDB/src/monet/monet_process.mx:112
7 0x0000000000403f5a in main (argc=1,
av=0x7fbffff3e8) at Mserver.cc:443
(gdb)
========
Does anybody have any idea, what might go wrong, here?
## Comment 2744
Date: 2005-03-22 13:50:50 +0100
From: @drstmane
Logged In: YES
user_id=572415
For now, we consider the problem "un-solvable" and
discontinue testing with the PGI compiler.
## Comment 2745
Date: 2005-05-23 08:08:28 +0200
From: @drstmane
Logged In: YES
user_id=572415
Bug still occurs with PGI version 6.0.
## Comment 2746
Date: 2005-10-06 16:36:54 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: Done / tests exits
any mapi is okay for this bug. Currently no testing using
pgi is done.
## Comment 2747
Date: 2006-01-13 21:46:24 +0100
From: @drstmane
Logged In: YES
user_id=572415
We do not use/support the PGI compiler any more, hence, this
one has become obsolete.
## Comment 2748
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166912 at http://sourceforge.net/support/tracker.php?aid=1166912
| MIL fork() failes with PGI on Debian 3.0 on x86_64 | https://api.github.com/repos/MonetDB/MonetDB/issues/502/comments | 0 | 2020-11-30T08:39:40Z | 2024-06-28T13:35:14Z | https://github.com/MonetDB/MonetDB/issues/502 | 753,275,742 | 502 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-19 11:47:53 +0100
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-03-24 04:11:14 +0100
## Comment 2738
Date: 2005-03-19 11:47:53 +0100
From: @grobian
Currently the Windows installer uses for its DBFARM the
All Users profile directory. However, this is
confusing behaviour, allows other users to see each
other's data, and disallows separate users to use
Mserver concurrently (there is an For Everyone install
option).
The location of the DBFarm should be in the user home
path, instead of the all users home path.
Instead of %ALLUSERSPROFILE%, I would suggest to use
%HOMEPATH% to get the above described behaviour.
for more information on windwows variables, see
http://home.earthlink.net/~rlively/MANUALS/ENV/INDEX.HTM
## Comment 2739
Date: 2005-03-24 16:11:14 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
If MonetDB was installed for "Just me", we now put the
dbfarm into %APPDATA%\MonetDB\dbfarm. %APPDATA% translates
to the user's Application Data folder.
If another user on the system calls the Mserver.bat
(possibly indirectly), his (her) data will go to the All
Users' Application Data folder.
## Comment 2740
Date: 2005-03-24 17:10:27 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
For completeness:
Now the database is always in the user's Applicatin Data folder.
## Comment 2741
Date: 2005-10-06 16:35:20 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: Done / NO Test
Impossible currently to verify installations on windows
automatically
## Comment 2742
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166460 at http://sourceforge.net/support/tracker.php?aid=1166460
| Windows installer should user profile instead of all users | https://api.github.com/repos/MonetDB/MonetDB/issues/501/comments | 0 | 2020-11-30T08:39:37Z | 2024-06-27T11:02:58Z | https://github.com/MonetDB/MonetDB/issues/501 | 753,275,700 | 501 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-19 11:42:16 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-19 03:47:46 +0100
## Comment 2734
Date: 2005-03-19 11:42:16 +0100
From: @drstmane
The stable test outputs the following tests need to be
updated/approved according to the changes of Friday Mar
18 2005:
compiler/Tests/usageH
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU_32-d_Fedora3/compiler/usageH.out.00.html
runtime/Tests/load
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU_32-d_Fedora3/runtime/load.out.00.html
runtime/Tests/sigs
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/GNU_32-d_Fedora3/runtime/sigs.out.00.html
cf.
http://monetdb.cwi.nl/Development/TestWeb/Stable/approve/
## Comment 2735
Date: 2005-03-19 15:47:46 +0100
From: @drstmane
Logged In: YES
user_id=572415
Couldn't resist ;-) :
updated stable output according to Peter's changes of Friday
Mar 18 2005
## Comment 2736
Date: 2005-10-06 16:33:11 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TEST exists
## Comment 2737
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166456 at http://sourceforge.net/support/tracker.php?aid=1166456
| stable test outputs need to be updated/approved | https://api.github.com/repos/MonetDB/MonetDB/issues/500/comments | 0 | 2020-11-30T08:37:52Z | 2024-06-27T11:02:57Z | https://github.com/MonetDB/MonetDB/issues/500 | 753,274,636 | 500 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-19 11:38:03 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-23 12:40:58 +0100
## Comment 2729
Date: 2005-03-19 11:38:03 +0100
From: @drstmane
After the changes of Friday Mar 18 2005, we get:
========
[...]
icc -no-gcc -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.ntv_32-d.1754/pathfinder/runtime -I..
-I/var/tmp/monet.ntv_32-d.1754/pathfinder/runtime
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/common
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/gdk
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/monet
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/plain
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/contrib
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/mapi
-I/var/tmp/monet.ntv_32-d.1754/MonetDB/.ntv_32-d/Fedora3/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I/var/tmp/monet.ntv_32-d.1754/pathfinder/runtime/../compiler/include
-we140 -wd1418 -c99 -Wall -w2 -mp1 -O3 -restrict
-unroll -tpp6 -axKWNPB -we266 -Werror
-wd1418,1419,279,310,981,810,444,193,111,177,171,181,764,269,108,188,1357,102,70,1572
-c pathfinder.c -KPIC -DPIC -o .libs/pathfinder.o
/var/tmp/monet.ntv_32-d.1754/pathfinder/runtime/pathfinder.mx(2319):
error 593: variable "ret" was set but never used
int ret = GDK_FAIL;
^
compilation aborted for pathfinder.c (code 2)
make[3]: *** [pathfinder.lo] Error 1
[...]
========
cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.ntv_32-d/Fedora3/make.out
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.ntv_32-d/Fedora3/make-k.out
## Comment 2730
Date: 2005-03-21 19:25:34 +0100
From: @drstmane
Logged In: YES
user_id=572415
Re-opened, since the bug has not been fixed, yet!
(At least I cannot find any trace of a fix, neither in the
code, nor in any CVS check-in, nor in a comment in the bug
report...)
## Comment 2731
Date: 2005-03-23 12:40:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS by using the proper return value.
## Comment 2732
Date: 2005-10-06 19:35:00 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / COMPILATION
## Comment 2733
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166454 at http://sourceforge.net/support/tracker.php?aid=1166454
| runtime/pathfinder.mx fails to compile with Intel's icc | https://api.github.com/repos/MonetDB/MonetDB/issues/499/comments | 0 | 2020-11-30T08:37:49Z | 2024-06-27T11:01:27Z | https://github.com/MonetDB/MonetDB/issues/499 | 753,274,605 | 499 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-19 11:24:07 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-03-23 03:56:40 +0100
## Comment 2722
Date: 2005-03-19 11:24:07 +0100
From: @drstmane
After the check-ins of Friday Mar 18 2005, the
following tests fail:
src/gdk/Tests/reload
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB/.mTests103/GNU_32-d_Fedora3/src_gdk/reload.out.00.html
tests/BugReports/Tests/no.089
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB/.mTests103/GNU_32-d_Fedora3/tests_BugReports/no.089.out.00.html
tests/by_Peter/Tests/trim
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB/.mTests103/GNU_32-d_Fedora3/tests_by_Peter/trim.out.00.html
## Comment 2723
Date: 2005-03-21 10:45:42 +0100
From: @peterboncz
Logged In: YES
user_id=591107
I checked both Stable & Current, but just cannot see any
differences in the mentioned tests.
Has this been resolved alread(or gone away out of itself)?
## Comment 2724
Date: 2005-03-21 19:21:52 +0100
From: @drstmane
Logged In: YES
user_id=572415
The changes were only in the MonetDB_4-6 branch, hence they
occur only with the latest (non-released) stable CVS version
(MonetDB 4.6.1).
The testweb still shows the problems (differences & errors!)
under all three of the above links!
Hence, nothing has been solved, yet...
## Comment 2725
Date: 2005-03-22 19:01:12 +0100
From: @drstmane
Logged In: YES
user_id=572415
All non-administrative code changes on Friday Mar 18 2005
where made by Sjoerd, hence I assign this to him.
## Comment 2726
Date: 2005-03-23 15:56:40 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed by removing a DELTAsave in BATload_intern and a
DELTAload in BATloaddesc.
## Comment 2727
Date: 2005-10-06 19:34:35 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: ALREADY IN TESTWEB
## Comment 2728
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166451 at http://sourceforge.net/support/tracker.php?aid=1166451
| Changes of Friday Mar 18 2005 make tests fail | https://api.github.com/repos/MonetDB/MonetDB/issues/498/comments | 0 | 2020-11-30T08:37:47Z | 2024-06-27T11:01:26Z | https://github.com/MonetDB/MonetDB/issues/498 | 753,274,586 | 498 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-18 11:34:28 +0100
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-03-23 10:11:47 +0100
## Comment 2714
Date: 2005-03-18 23:34:28 +0100
From: @drstmane
The "setjmp" call in runtime/serialize.mx
"consistently" causes a BUS ERROR on Itanium (running
Red Hat Enterprise Linux WS release 4 (Nahant), Linux
2.6.9-5.EL) when compiled with Intel's icc 8.1.
In case we want to keep supporting pathfinder on that
platform, we need to check, whether we to abondon the
use of setjmp/longjmp; unless we find another solution...
## Comment 2715
Date: 2005-03-21 10:43:21 +0100
From: @peterboncz
Logged In: YES
user_id=591107
I am against the frivolous use of setjmp; just for error handling.
Even worse, this way of jumping out of a routine is simply wrong:
it will cause dangling BBP references; that is, BATs fixed by
BATdescriptor will never get unfixed anymore.
Please write proper error handling code; ie.. make each routine
return an error code, and check for it accordingly everywhere.
## Comment 2716
Date: 2005-03-21 11:07:13 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
OK, I will change the code and remove the setjmp/longjmp
stuff. My motivation for using this was that the code was
nuch clearer and faster.
I even asked for permission to do so! When dangling BBP refs
is a problem I can fix this also within the setjmp/longjmp
solution.
If Peter is finished with the Mapi stuff I will make the
adaptions beceause for this I think I need an exclusive lock
because every second line of code will change,
Jan.
## Comment 2717
Date: 2005-03-21 19:17:23 +0100
From: @drstmane
Logged In: YES
user_id=572415
The "permission" to use set/jmp/longjmp was "granted" under
the presime that the resulting code will be (1) correct and
(2) portable.
Apparently, both seem to be a problem (though I have no
idea, why it only fails with icc on Itanium, but not with
icc on x86 platforms...).
Hence, I'd appreciate a solution, what ever it might be.
## Comment 2718
Date: 2005-03-23 10:11:47 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
After a bit of grumbling I removed the setjmp/longjmp
construction and changed the error handling to the
old-fashioned if-statements. All should work now and I
consider the bug closed.
## Comment 2719
Date: 2005-03-23 10:18:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
For the "bookkeeping": documented the "Resolution", i.e.,
the reason why the report was closed.
## Comment 2720
Date: 2005-10-06 19:33:59 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / HARDWARE AND CODING STYLE
## Comment 2721
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1166283 at http://sourceforge.net/support/tracker.php?aid=1166283
| "setjmp" causes BUS ERROR with icc on Itanium | https://api.github.com/repos/MonetDB/MonetDB/issues/497/comments | 0 | 2020-11-30T08:37:44Z | 2024-06-27T11:01:25Z | https://github.com/MonetDB/MonetDB/issues/497 | 753,274,555 | 497 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-18 12:39:24 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-03-23 05:34:59 +0100
## Comment 2708
Date: 2005-03-18 12:39:24 +0100
From: @drstmane
On Fedora2/64-bit on x86_64/AMD64, we get:
[...]
icpc -no-gcc -DHAVE_CONFIG_H -I.
-I/tmp/monet.ntv_64-d.21295/MonetDB/src/mel -I../..
-I/tmp/monet.ntv_64-d.21295/MonetDB/src/mel -we140
-wd1418 -c99 -Wall -w2 -we266 -Werror
-wd1418,1419,279,310,981,810,444,193,111,177,171,181,764,269,108,188,1357,102,70,1572
-c -o mel.yy.o mel.yy.cc
/usr/include/c++/3.4.2/bits/postypes.h(72): error:
identifier "int64_t" is undefined
typedef int64_t streamoff;
^
compilation aborted for mel.yy.cc (code 2)
make[4]: *** [mel.yy.o] Error 2
[...]
## Comment 2709
Date: 2005-03-18 12:48:45 +0100
From: @drstmane
Logged In: YES
user_id=572415
... "Fedora2/64-bit on x86_64/AMD64" is db2/sn-1 ...
## Comment 2710
Date: 2005-03-23 17:32:54 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
This is a bug in the compiler / system include files.
int64_t is defined in stdint.h which is included somewhere
deep down in the include of iostream. The typedef is
surrounded by ifndef __int8_t_defined. The token
__int8_t_defined is defined in sys/types.h which defines
some of the same types as stdint.h and sets this token
because of that, but it does not define all the ones
stdint.h defines. One of the ones it *doesn't* define is
int64_t.
It could possibly be worked around by hacking an include of
stdint.h at the top of the generated mel.yy.cc, but that is,
of course, a hack.
## Comment 2711
Date: 2005-03-23 17:34:59 +0100
From: @njnes
Logged In: YES
user_id=43556
Seems the new gcc compiler requires a new type int64_t which
isn't defined
because the /usr/include/features.h file doesn't set
__GLIBC_HAVE_LONG_LONG. This because the maintainer of libc
didn't add the || defined(__x86_64__) for the intel and PGI
compilers jet.
So local patch of the /usr/include/features.h file helped!
## Comment 2712
Date: 2005-10-06 19:33:03 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / COMPILATION
## Comment 2713
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1165877 at http://sourceforge.net/support/tracker.php?aid=1165877
| Compiling MonetDB with Intel's icc fails with Fedora2/64-bit | https://api.github.com/repos/MonetDB/MonetDB/issues/496/comments | 0 | 2020-11-30T08:37:39Z | 2024-06-27T11:01:24Z | https://github.com/MonetDB/MonetDB/issues/496 | 753,274,505 | 496 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-16 06:21:19 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-03-16 09:04:18 +0100
## Comment 2704
Date: 2005-03-16 18:21:19 +0100
From: @grobian
consider the following dialogue:
CREATE TABLE "sys"."first" (
"id" int NOT NULL,
CONSTRAINT "first_id_pkey" PRIMARY KEY ("id")
);
CREATE TABLE "sys"."second" (
"id" int NOT NULL,
"ref" int,
CONSTRAINT "second_id_pkey" PRIMARY KEY ("id"),
CONSTRAINT "second_ref_fkey" FOREIGN KEY
("ref") REFERENCES "sys"."first" ("id")
);
monetdb-> insert into "second" values (1, null);
Error: ERROR: INSERT INTO: FOREIGN KEY constraint
second.second_ref_fkey violated
The NULL value should be allowed, as there is no NOT
NULL constraint on the ref column.
## Comment 2705
Date: 2005-03-16 21:04:17 +0100
From: @njnes
Logged In: YES
user_id=43556
there was a bug in the handling of foreign key checks for
single inserts, probably introduced in the last memory leak
cleanup.
## Comment 2706
Date: 2005-10-06 19:32:39 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/foreign_key_null.SF-1164646.*
## Comment 2707
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1164646 at http://sourceforge.net/support/tracker.php?aid=1164646
| SQL: FOREIGN KEY constraints also apply to NULL values | https://api.github.com/repos/MonetDB/MonetDB/issues/495/comments | 0 | 2020-11-30T08:37:37Z | 2024-06-27T11:01:23Z | https://github.com/MonetDB/MonetDB/issues/495 | 753,274,468 | 495 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-16 11:04:31 +0100
From: Wouter Alink <<alink>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-18 01:06:30 +0100
## Comment 2694
Date: 2005-03-16 11:04:31 +0100
From: Wouter Alink <<alink>>
When using "pf | MapiClient"
After an error is returned on the MIL-script generated
by pathfinder,
pathfinder will spit out garbage (and when handling
large documents, it can be a lot of garbage).
Another thing is when I try to stop pathfinder from
spitting garbage at me I type Ctrl-C and this results
in a crashed Mserver. I guess this is a bug as well.
## Comment 2695
Date: 2005-03-17 00:12:51 +0100
From: @drstmane
Logged In: YES
user_id=572415
What exactly is "garbage", here?
Could you please provide a (small) example to reproduce the
problem?
## Comment 2696
Date: 2005-03-17 10:02:51 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
for example the same case as with the previous bug:
<?xml version="1.0" encoding="utf-8"?>
<a> <a> <b/> </a> <a/> </a>
and query
doc("http://sourceforge.net/support/tracker.php?aid=01.xml")//a[.//b]
it will first report an error and afterwards some garbage
(it looks like xml, but it actually is a list of random strings)
## Comment 2697
Date: 2005-03-17 10:48:14 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
I think we already had this discussion with Sjoerd half a year
ago. At this point I probably wanted more error messages --
which is nonsense after the first thing went wrong.
The only thing I do is generate an ERROR() call in Monet.
Perhapse somebody knows how to stop evaluation or how to
ask if there was an error before starting to print the output.
## Comment 2698
Date: 2005-03-17 11:16:48 +0100
From: @drstmane
Logged In: YES
user_id=572415
1) The ERROR is actually not triggered by some generated
ERROR(),
but rather seems to originate from some BUG in
ll_descendant.
(Would have saved me some time, if Wouter would have
told us ;-)
2) When pipeing MIL into either Mserver or MapiClient, both
simply go one
trying to interpret the remainder of the MIL stream
even after an error.
There is IMHO no easy/feasable way to avoid this.
In case you give the respective MIL code as a file
parameter on the
command line (e.g., Mserver pf.mil / MapiClient
pf.mil), processing
stops after the first error.
3) Trying to fix the above mentioned bug in ll_descendant
(and not being
the most experienced XQuery interpreter myself):
What would be the expected/correct output of the given
example?
## Comment 2699
Date: 2005-03-17 11:24:59 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
I actually kept it general, because any (i'm not sure if it
is any, but is more than one) error generated by MIL (a
second example, if MIL reports a shortage of memory) it will
afterwards return garbage.
## Comment 2700
Date: 2005-03-17 11:30:28 +0100
From: @drstmane
Logged In: YES
user_id=572415
Well, as I said, I should have read your other bug report
more carefully.
Sorry.
And as I explained, the "garbage" problem is not PF
specific/related,
but a principle MonetDB "problem" ("feature"??).
## Comment 2701
Date: 2005-03-18 13:06:30 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed "by coincidence" with Peter's latest changes: the
PF-generated MIL is now wrapped in a MIL '{}" blokc, and
hence processing stops after the first error.
## Comment 2702
Date: 2005-10-06 19:27:08 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST SCRIPT PROVIDED
## Comment 2703
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1164341 at http://sourceforge.net/support/tracker.php?aid=1164341
| PF: ouput of garbage on error | https://api.github.com/repos/MonetDB/MonetDB/issues/494/comments | 0 | 2020-11-30T08:37:34Z | 2024-06-27T11:01:22Z | https://github.com/MonetDB/MonetDB/issues/494 | 753,274,434 | 494 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-16 10:58:02 +0100
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-17 02:09:02 +0100
## Comment 2684
Date: 2005-03-16 10:58:02 +0100
From: Wouter Alink <<alink>>
i think something goes wrong in shredding whitespace,
but i'm not sure. look at the results below. (i'm using
cvs-pathfinder)
--------------
given xml document (http://sourceforge.net/support/tracker.php?aid=01.xml): (with spaces between
the elements)
<?xml version="1.0" encoding="utf-8"?>
<a> <a> <b/> </a> <a/> </a>
and query
doc("http://sourceforge.net/support/tracker.php?aid=01.xml")//a[./b]
results in the correct answer
<a><b/></a>
but query:
doc("http://sourceforge.net/support/tracker.php?aid=01.xml")//a[.//b]
returns:
MAPI = anonymous@localhost:50000
QUERY = if (iter.count() != 0)
ERROR = !ERROR: ll_descendant: BATextend of BAT
'tmp_483' (483) failed: required 4294967295 buns
(4294967288 bytes), got only 52 buns (416 bytes).
!ERROR: ll_descendant: bunins failed.
!ERROR: PFll_descendant: operation failed.
-------------------------------------------
given xml document (http://sourceforge.net/support/tracker.php?aid=02.xml): (without spaces between
de elements)
<?xml version="1.0" encoding="utf-8"?>
<a><a><b/></a><a/></a>
and query:
doc("http://sourceforge.net/support/tracker.php?aid=02.xml")//a[.//b]
results in the correct answer:
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult>
<a><a><b/></a><a/></a><a><b/></a>
</XQueryResult>
----------------------------------------------
given the query:
<t><a> <a> <b/> </a> <a/> </a></t>//a[.//b]
the correct result is returned:
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult>
<a><a><b/></a><a/></a><a><b/></a>
</XQueryResult>
## Comment 2685
Date: 2005-03-16 17:38:00 +0100
From: @drstmane
Logged In: YES
user_id=572415
The shredder is JanF's "baby".
## Comment 2686
Date: 2005-03-17 09:58:52 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
I do not see why this is a shredding problem! The first
query works so the database representation seems correct.
I've seem some testsets fail which adress weird descendent
steps and Jan.R said he would take a look at these. Maybe
this problem is in the same class? I pass the bug to Jan.R.
## Comment 2687
Date: 2005-03-17 10:12:51 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
The reason I thought it would be the shredding was that the
last example (same xml as first example, but without use of
doc()-function) gave the correct output, so that's why I
thought the bug is located close to the doc()-function.
## Comment 2688
Date: 2005-03-17 11:16:27 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
The first query is correct and returns '<a> <b/> </a>'.
The second one crashes in the second descendant step ('b').
The third one is correct.
The fourth one is also correct, because whitespaces are
ignored in xquery. That means the input to query also contains
no whitespaces.
My conclusion: no problems with the shredder but with the
descendant step (like the error message suggests).
I guess that something in the nametest goes wrong, but this is
only a guess. As the best one to ask here is Stefan I will
close the round and assign it to Stefan :)
## Comment 2689
Date: 2005-03-17 11:27:19 +0100
From: @drstmane
Logged In: YES
user_id=572415
Oops, apparently, I didn't read this carefully enough (or
not at all?) ...
SORRY!
The bug in ll_descendant seem to indecate that
(1) the BAT representation of the documents with and without
whitespace
seem to differ (otherwise, there would not be
different behaviour)
--- could this be?
(2) this difference either trigers a bug in ll_descendant,
or the representation
in case of whitespace is inconsistent, i.e., ll_desc
works on wrong data,
and hence fails.
I'll need to analyse this in more detail...
... but I vhave the idea, that both, shredding and ll_desc
are involved ... ;-)
## Comment 2690
Date: 2005-03-17 14:09:02 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS:
The bug indeed seemed to have been in ll_descendant;
at least, I now get the correct answer with both documents,
i.e., with and without whitspace.
## Comment 2691
Date: 2005-03-21 12:40:55 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Just to round up the discussion (and clarify expected behavior):
- Whitespace *is* significant in XML documents. So documents
with and without whitespace are expected to result in different
BAT representations.
Whitespace behavior changes if a DTD is given in the XML
prolog. If no character data is allowed (according to a DTD)
at a certain point in the document, then whitespace will be ignored.
(The libxml2 library should actually handle this. So in a sense,
there "can't" be a problem with the shredder, I think.)
- For *constructed* elements (last example), whitespace behavior
is supposed to depend on the `declare xmlspace' declaration
in the XQuery query prolog. Pathfinder defaults to `strip' here, so
whitespace shall be ignored. (Haven't tested Pathfinder's behavior,
though.)
## Comment 2692
Date: 2005-10-06 19:26:08 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: REASSIGN vzzzbx
## Comment 2693
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1164338 at http://sourceforge.net/support/tracker.php?aid=1164338
| PF: desc step crashes - no incorrect shredding of whitespace | https://api.github.com/repos/MonetDB/MonetDB/issues/493/comments | 0 | 2020-11-30T08:37:31Z | 2024-06-27T11:01:21Z | https://github.com/MonetDB/MonetDB/issues/493 | 753,274,401 | 493 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-14 12:47:48 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-16 07:14:06 +0100
## Comment 2680
Date: 2005-03-14 00:47:48 +0100
From: Jan Rittinger <<jan.rittinger>>
abs(dbl(0)).print();
[ -0 ]
At least in my opinion abs should return only positive
values :)
## Comment 2681
Date: 2005-03-16 19:14:06 +0100
From: @drstmane
Logged In: YES
user_id=572415
Is "-0" negative?
Actually, the problem somewhere in the compiler and/or the
limitations of floating point representations, and hence
beyond our control:
========
$ cat abs.c
include <stdlib.h>
include <stdio.h>
int main() {
double n = 0.0, p = ((n>0)?n:-n);
printf("n=%g, p=%g\n", n, p);
return 0;
}
$ gcc abs.c -o abs
$ ./abs
n=0, p=-0
^ ^^
========
## Comment 2682
Date: 2005-10-06 19:25:06 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / INVALID
## Comment 2683
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1162659 at http://sourceforge.net/support/tracker.php?aid=1162659
| abs() returns negative value: -0 | https://api.github.com/repos/MonetDB/MonetDB/issues/492/comments | 0 | 2020-11-30T08:37:28Z | 2024-06-28T13:35:14Z | https://github.com/MonetDB/MonetDB/issues/492 | 753,274,376 | 492 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-11 08:02:26 +0100
From: Jan Rittinger <<jan.rittinger>>
To: @njnes
Version: -- development
Last updated: 2005-10-06 08:09:48 +0200
## Comment 2674
Date: 2005-03-11 20:02:26 +0100
From: Jan Rittinger <<jan.rittinger>>
The split function in str.mx (strSplit) gets into an infinite
loop if the seperator string is empty.
split("foo","");
The only reason why I submitted the bug report is, that I
wasn't sure what the expected output is (error message,
original string, empty string or a string for each character
:)).
## Comment 2675
Date: 2005-03-13 11:41:47 +0100
From: @njnes
Logged In: YES
user_id=43556
in case of empty string, just insert the whole string, ie
don't split.
## Comment 2676
Date: 2005-10-06 19:23:54 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: TEST ADDED / UNKNOWN
COMMAND: split(str, str) : BAT[void,str]
MODULE: str
COMPILED: by adm on Sat Oct 1 15:40:04 2005
Split the string s based on the seperator sep
note: the separator may not be an empty string.
ahem...
either the documentation is wrong, or the function is b0rken
right now, because the test appears to return nothing.
reopening the bug, not checking in output, just the test script.
## Comment 2677
Date: 2005-10-06 19:56:00 +0200
From: @njnes
Logged In: YES
user_id=43556
The strSplit apperantly simply refuses to split on empty
strings, ie gives an error.
## Comment 2678
Date: 2005-10-06 20:09:47 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/split_inf_loop.SF-1161583.*
(approved error message, all seems ok now)
## Comment 2679
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1161583 at http://sourceforge.net/support/tracker.php?aid=1161583
| infinite loop | https://api.github.com/repos/MonetDB/MonetDB/issues/491/comments | 0 | 2020-11-30T08:37:26Z | 2024-06-27T11:01:19Z | https://github.com/MonetDB/MonetDB/issues/491 | 753,274,342 | 491 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-11 02:22:18 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-03-21 10:58:19 +0100
## Comment 2667
Date: 2005-03-11 14:22:18 +0100
From: Jan Rittinger <<jan.rittinger>>
With the following MIL script Mserver crashes:
module("pathfinder");
shred_doc("/tmp/book.xml","blub");
delete_doc("blub");
quit();
!ERROR: HEAPfree: 04/263.buns was not mapped
!OS: Invalid argument
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
If I quit between the shred_doc and the delete_doc
everything works fine. (That's why I think the bug is not in
the shred_doc/delete_doc code :))
## Comment 2668
Date: 2005-03-17 14:36:59 +0100
From: @drstmane
Logged In: YES
user_id=572415
The problem is as follows:
To gain performance (with large documents), I recently
ensure that pathfinder BATs get memory mapped, by adding
".mmap(STORE_MMAP)" before making the respective BATs
persistent. MonetDB no work so that this mmap only gets
effective on the BATs have been unloaded and loaded again
(e.g., because the Mserver has been stopped and restarted).
In the given example, hence, this isn't the case; i.e., the
BATs are marked as (to be) memory mapped, but in fact they
are still malloced.
When "delete_doc()" executes "persists(FALSE)" on these
BATs, a "GDKmunmap()" is triggered, which obviously fails.
Possible solutions:
- add unload/load sequence in the shredder to make the
memory mapping effective
- initially create the BATs in the shredder as mmapped
- add a check in persists(FALSE), whether the to be mmaped
BAT is already mmapped
I don't know yet, which is the simplest/best one, but I'll
try to have a look at it...
## Comment 2669
Date: 2005-03-18 12:51:54 +0100
From: @drstmane
Logged In: YES
user_id=572415
The problems *seems* to have disappeared with the latest
changes to both MonetDB & pathfinder --- at least it now
works fine for me and with nightly testing.
Peter, if you still see this problem, could you please
update your MonetDB, and check again?
## Comment 2670
Date: 2005-03-21 10:48:11 +0100
From: @peterboncz
Logged In: YES
user_id=591107
Last week, Sjoerd checked in a fix in GDK, where on BAT load, a
all heap base pointers are set to NULL.
Can you check whether the newest MonetDB still exhibits this
bug? I think it should have gone away now.
## Comment 2671
Date: 2005-03-21 10:58:19 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
it seems to be gone
## Comment 2672
Date: 2005-10-06 19:16:33 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / probably too big, and no data file
## Comment 2673
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1161355 at http://sourceforge.net/support/tracker.php?aid=1161355
| memory mapping crashes Mserver | https://api.github.com/repos/MonetDB/MonetDB/issues/490/comments | 0 | 2020-11-30T08:37:22Z | 2024-06-27T11:01:18Z | https://github.com/MonetDB/MonetDB/issues/490 | 753,274,300 | 490 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-10 01:57:43 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-16 07:49:53 +0100
## Comment 2663
Date: 2005-03-10 01:57:43 +0100
From: Jan Rittinger <<jan.rittinger>>
The attached example produces the following warning
(only) with debugmask 10:
-------------------------
h tmp_309 name
oid int type
-------------------------
[ 3@0, 1 ]
[ 4@0, 3 ]
!WARNING: BATpropcheck: BAT tmp_309 was
incorrectly marked sorted!
!WARNING: BATpropcheck: BAT tmp_309 remains
marked radix-clustered on 32 bits; not checked!
Sorry -- I was to lazy to reproduce the problem without
the whole pathfinder module stuff :)
## Comment 2664
Date: 2005-03-16 19:49:53 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed property settings of outerjoin in CVS (MonetDB_4-6
branch):
The outerjoin result cannot be tail-sorted in case NILs are
produced.
## Comment 2665
Date: 2005-10-06 19:15:00 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: REASSIGN stmane (how do I set the debug
mask?)
## Comment 2666
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1160316 at http://sourceforge.net/support/tracker.php?aid=1160316
| wrong properties (says debugmask 10) | https://api.github.com/repos/MonetDB/MonetDB/issues/489/comments | 0 | 2020-11-30T08:37:20Z | 2024-06-27T11:01:17Z | https://github.com/MonetDB/MonetDB/issues/489 | 753,274,273 | 489 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-09 09:41:34 +0100
From: Dolf Smits <<nl115177>>
To: @njnes
Version: -- development
Last updated: 2005-03-23 10:15:49 +0100
## Comment 2642
Date: 2005-03-09 09:41:34 +0100
From: Dolf Smits <<nl115177>>
After creating a lot of tables, during development work
the server works fine.
After clearing all the created tables by dropping them,
the server crashes.
After startup the server crashes as soon as an
application is contacting the server.
I include the sql files which i loaded (using maven, ant
and the jdbc driver)
the script drop.sql is for dropping the tables and induces
the problem.
There are no SQL errors reported.
I also attach a bitnmap of the error
I also attach logfiles (sql logging level =1)
I'm running on windows 2000
## Comment 2643
Date: 2005-03-09 09:47:02 +0100
From: @grobian
Logged In: YES
user_id=963970
I think the uploading of the attach went wrong. SourceForge
requires you to tick the "Check to Upload and Attach a
File:" think, otherwise it won't upload.
## Comment 2644
Date: 2005-03-09 10:50:07 +0100
From: @grobian
Logged In: YES
user_id=963970
I loaded all tables as in your zip and used drop.sql to drop
them using JdbcClient. I ran them all in order, and there
seems to be no problem here, except for this:
Error on line 78: DROP TABLE: unknown table: a
Error on line 79: DROP TABLE: unknown table: b
I'm using auto-commit, do you use it as well?
I do notice this when trying to redump your schemas in one file:
% java -jar ~/develop/MonetJDBC/jars/MonetDB_JDBC.jar -f
all_tables.sql -p55123 -btest -d
Exception in thread "main" java.lang.Exception: Cyclic
dependancy graphs are not supported (foreign key relation
references self)
which is in this case nothing more than an indication that
there are cycles in the tables.
## Comment 2645
Date: 2005-03-09 10:54:09 +0100
From: @grobian
Logged In: YES
user_id=963970
seems I have forgotten one sql file to load. When using one
transaction to drop the tables (auto-commit = off) all goes
fine for me also.
## Comment 2646
Date: 2005-03-09 11:49:53 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
The scripts are run in the following order:
phase3ojb-schema.sql
phase1-schema.sql
phase2-schema.sql
registry-schema.sql
prefs-schema.sql
security-schema.sql
populate-db-default.sql
test-persistence-schema.sql
i dont know how the commit settings are but the log seems
to show that everyting is destroyed and after that commited.
it alse soeems that if i drop and build immideatly, the server
keeps running
if I just drop the tables, after some time the server crashes
## Comment 2647
Date: 2005-03-09 12:30:23 +0100
From: @grobian
Logged In: YES
user_id=963970
I don't have the populate-db-default.sql file. I have
executed the files in the order as you said. Does the
populate file issue some insert queries?
## Comment 2648
Date: 2005-03-09 12:35:11 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
Sorry,
forgot to include the populate file, it is attached now.
It does do all sorts of inserts
## Comment 2649
Date: 2005-03-14 11:32:39 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
It seems a bit difficult to reproduce, but it seems it has
something to do when the scripts generate errors.
If everything goes correctly, it wil be stable. But (too) often I
end up with a crashing server.
As also the JDBC driver is rather incomplete in some ways,
I'll stop testing with monet. It seemed a rather interesting
database, but at this moment it seems to be too immature.
Keep up the work!
## Comment 2650
Date: 2005-03-14 11:37:29 +0100
From: @grobian
Logged In: YES
user_id=963970
Can you please list the functionalities that the JDBC driver
lacks for you? Perhaps we can work on some of them. Thanks
for trying, at least we have learnt from your user
experience and have fixed some bugs.
## Comment 2651
Date: 2005-03-15 09:32:50 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
The problem started with the fact that the sql scripts are
generated using torque, and torque has no support for
monetDB.
So I needed to write some simple scripts for torque to get it
to genearate the scripts in such a way that monet accepts
them.
I than ran into the JDBC problem.
Problem is the perepared statement part.
Lots of it is not implemented, but the errorr's are all the
same. I changed the error reporting (operation not supported)
was patched to name which operation is not supported.
I found that the setObject (int, object, int) was not supported
but used in the application.
I patched that to find out wat datatypes were called, and
patching it for datatype 12 (and other string/char values)
helped.
I then needed to patch for sqltype 5, I did it the same way
and then got some class loader constraint violations from the
application.
I don't know for sure the are due to JAVA/JDBC problems, or
to incorrectly casting datatypes to SQL datatypes.
The problem is, I don't know which datatypes are supported
by MonetDB, and which SQL features are supported and
which are not yet supported. This makes it a long road to a
working implementation.
I think that when more information about what is supported
and more documentation is much needed for simple users
(and in my opinion, a database should "just work" as it is
some service you need for some application, and not a goal
in itself (as it might be for a database developer))
Maybe, we can track some things down and get it up and
running after all.
Dolf
## Comment 2652
Date: 2005-03-15 09:50:29 +0100
From: @grobian
Logged In: YES
user_id=963970
It's a pity you didn't report on the setObject lack, because
we already implemented this after the release, and could
have given it to you.
We are aware of our weak documentation side. I appreciate
your efforts to try and patch the JDBC driver to make it
work, but I would have appreciated it even more if you
reported your problem, as for me fixing the problem is
probably less difficult and as in this case, the problem was
already tackled.
We are eager to learn from real user experiences, as for our
cases, the database "just works".
Again we appreciate the patching of the sources, but please
let us know the issue first, even if only just to give you a
direction. Our priority list is heavily influenced by known
problems of people, as visible via the bug lists. Besides
that, we like to help you get going as far as we can.
## Comment 2653
Date: 2005-03-15 11:46:09 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
In this case, where to download the new jdbc driver.
looking at the download page, I can only find the 1.1.RISE
version, not any later versions.
can you tell me where to download the latest version?
Thanks
## Comment 2654
Date: 2005-03-15 13:52:18 +0100
From: @grobian
Logged In: YES
user_id=963970
You can find a freshly compiled version here:
http://homepages.cwi.nl/~fabian/monetjdbc/MonetDB_JDBC.jar
## Comment 2655
Date: 2005-03-16 11:39:52 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
It seems that the jdbc driver is working now. I corrected some
errors in the mappings from datatypes, and it seems to run
now.
Can you confirm the correct mappings, as stated below?
is there some way to run conditional statements with SQL,
like "drop table a if exists"?
I have a table, with severals coumns, if which two do give me
a problem.
columns 1 is the node_id, column2 is the parent_nodeid
it has a constraint that parent_nodeid should always refer to a
nodeID. Adding the first line always violates this constraint. In
other database using a NULL value for the parent_nodeid
works, with monet this is refused. How to add the first row?
(referencing it's own row also doesn't work)
Thanks
mapping dataypes:
-------------------------------------------------------------------------
BIT = boolean
TINYINT = integer
SMALLINT = integer
INTEGER = integer
BIGINT = integer
FLOAT = float
REAL = float
BOOLEANCHAR = char
BOOLEANINT = integer
DOUBLE = double precision
NUMERIC = numeric
DECIMAL = decimal
CHAR = char
VARCHAR = varchar
LONGVARCHAR = clob
DATE = date
TIME = time
TIMESTAMP = timestamp
BINARY = blob
VARBINARY = blob
LONGVARBINARY = blob
NULL =
OTHER =
JAVA_OBJECT =
DISTINCT =
STRUCT =
ARRAY =
BLOB =
CLOB =
REF =
## Comment 2656
Date: 2005-03-16 18:08:31 +0100
From: @grobian
Logged In: YES
user_id=963970
I don't know where exactly you 'corrected some errors', so I
cannot easily check if I made a few errors, but when looking
at your list, it doesn't seem to make much sense to me:
tinyint = smallint = int = integer?
float = float? -> in SQL float is double or real, so it
should be double
what is a booleanchar or booleanint?
numeric == decimal
longvarchar != clob (otherwise why the difference?)
*binary != blob
(blobs and clobs only allow stream operations)
blob = clob = nothing?
I think I just don't understand what you try to map here,
but please let me know if you can give me a specific case
where you think the performed mapping is wrong.
IF EXISTS clauses are not in the SQL standard, and thus not
supported. Simply use auto-commit or START
TRANSACTION/COMMIT statements around them.
I will try out your foreign key problem here.
## Comment 2657
Date: 2005-03-16 18:23:36 +0100
From: @grobian
Logged In: YES
user_id=963970
I reported the foreign key issue as a separate bug:
http://sourceforge.net/tracker/index.php?func=detail&aid=1164646&group_id=56967&atid=482468
thanks for reporting it.
## Comment 2658
Date: 2005-03-21 10:07:50 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
About the list.
It should read,
(the type we defined in the scripts) = (the type as defined in
this database)
Normally of it stand like
blob =
This means that this datatype does not need any translation.
now the question is, do I need to put anything, anywhere
behind the = . (in your answer you already say, float = double)
Do I need to give more mappings?, Or are all the datatypes
named already defined in Monet?
BIT =
TINYINT =
SMALLINT =
INTEGER =
BIGINT =
FLOAT = double
REAL =
BOOLEANCHAR = (something unknown)
BOOLEANINT = (something unknown)
DOUBLE =
NUMERIC =
DECIMAL =
CHAR =
VARCHAR =
LONGVARCHAR =
DATE =
TIME =
TIMESTAMP =
BINARY =
VARBINARY =
LONGVARBINARY =
NULL =
OTHER =
JAVA_OBJECT =
DISTINCT =
STRUCT =
ARRAY =
BLOB =
CLOB =
REF =
(i'll try to find out why and when the booleanchar and
booleanint are used, and find some usefull mapping)
I don't understand your idea about putting start
transaction/commit around failing commands.
I think the command still fails so the script generator will
abort.
Thanks, Dolf
## Comment 2659
Date: 2005-03-21 10:09:45 +0100
From: Dolf Smits <<nl115177>>
Logged In: YES
user_id=1235545
About the list.
It should read,
(the type we defined in the scripts) = (the type as defined in
this database)
Normally of it stand like
blob =
This means that this datatype does not need any translation.
now the question is, do I need to put anything, anywhere
behind the = . (in your answer you already say, float = double)
Do I need to give more mappings?, Or are all the datatypes
named already defined in Monet?
BIT =
TINYINT =
SMALLINT =
INTEGER =
BIGINT =
FLOAT = double
REAL =
BOOLEANCHAR = (something unknown)
BOOLEANINT = (something unknown)
DOUBLE =
NUMERIC =
DECIMAL =
CHAR =
VARCHAR =
LONGVARCHAR =
DATE =
TIME =
TIMESTAMP =
BINARY =
VARBINARY =
LONGVARBINARY =
NULL =
OTHER =
JAVA_OBJECT =
DISTINCT =
STRUCT =
ARRAY =
BLOB =
CLOB =
REF =
(i'll try to find out why and when the booleanchar and
booleanint are used, and find some usefull mapping)
I don't understand your idea about putting start
transaction/commit around failing commands.
I think the command still fails so the script generator will
abort.
Thanks, Dolf
## Comment 2660
Date: 2005-03-21 10:21:22 +0100
From: @grobian
Logged In: YES
user_id=963970
Okay! I completely misunderstood the meaning of your list.
I'm sorry.
So, we support everything according to the SQL99 standard,
if we support it at all. Things we currently don't support,
or don't support very well are:
BIT (we support boolean)
TINYINT (accepted, but mapped on a SMALLINT)
BOOLEANCHAR (not supported)
BOOLEANINT (not supported)
VARBINARY (not supported)
LONGVARBINARY (not supported)
OTHER (not supported)
JAVA_OBJECT (not supported)
DISTINCT (not supported)
STRUCT (not supported)
ARRAY (not supported through SQL)
BLOB (supported, but weak, no streams)
CLOB (supported, JDBC support without streams)
REF (not supported)
beware of float, depending on the precision you define it
will become either a real or double.
Does this help you a bit further?
## Comment 2661
Date: 2005-10-06 19:13:17 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / TOO BIG
## Comment 2662
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1159653 at http://sourceforge.net/support/tracker.php?aid=1159653
| Server crashes after creating and dropping tables | https://api.github.com/repos/MonetDB/MonetDB/issues/488/comments | 0 | 2020-11-30T08:37:16Z | 2024-06-27T11:01:16Z | https://github.com/MonetDB/MonetDB/issues/488 | 753,274,234 | 488 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-08 06:20:12 +0100
From: @peterboncz
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-16 10:16:11 +0100
## Comment 2638
Date: 2005-03-08 18:20:12 +0100
From: @peterboncz
It seems that aggregation aggrX3 style does not work
correctly if the tail column (the values to be aggregated)
happens to be a void column.
var v := bat(void,oid).insert(nil,1@0).insert(nil,1@0).insert
(nil,2@0).seqbase(0@0).reverse.access(BAT_READ);
var e := bat(oid,void).insert(1@0,nil).insert(2@0,nil).access
(BAT_READ);
{min}(v,e).print();
it came out in a larger TREC processing script:
module(aggrX3,xtables,ascii_io);
TRECDB@batloop(){
var cname := $t;
var dext := bat(cname + "_docno");
var did := bat(cname + "_did");
var tid := bat(cname + "_tid");
BUG: does not work without int-cast
var dpos := {min}([int](did.reverse()), dext);
dpos := did.join(dpos);
dpos := [-]([int](did.mirror()), dpos);
var tdid := [lng](tid);
tdid := [<<](tdid, const 32);
tdid := [or](tdid, [lng](did));
tdid := tdid.reverse().join(dpos);
tdid.access(BAT_WRITE);
tdid.order().reverse();
tdid.access(BAT_READ);
var td := tdid.mark(oid(0)).reverse();
var pos := tdid.reverse().mark(oid(0)).reverse();
tdid := CTrefine(td,pos).mark(oid(0)).reverse();
td := tdid.join(td);
pos := tdid.join(pos);
var t := [>>](td, const 32);
var d := [and](td, const lng(INT_MAX));
VAR bs := new(void,bat);
bs.insert(nil, t);
bs.insert(nil, d);
bs.insert(nil, pos);
VAR fmt := dump_format(bs, sprintf("%stmp/%
s/t_d_pos.fmt", datadir, cname));
dump_data(fmt, sprintf("%stmp/%s/t_d_pos.csv.bz2",
datadir, cname), -1); }
## Comment 2639
Date: 2005-03-16 22:16:10 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS (MonetDB_4-6 branch).
## Comment 2640
Date: 2005-10-06 19:11:51 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: REASSIGN to stmane for MIL knowledge
## Comment 2641
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1159198 at http://sourceforge.net/support/tracker.php?aid=1159198
| aggrX3: {min}([oid,void], [oid,any]) does not work | https://api.github.com/repos/MonetDB/MonetDB/issues/487/comments | 0 | 2020-11-30T08:37:13Z | 2024-06-27T11:01:15Z | https://github.com/MonetDB/MonetDB/issues/487 | 753,274,190 | 487 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-07 09:58:21 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-03-07 10:34:44 +0100
## Comment 2634
Date: 2005-03-07 21:58:21 +0100
From: @grobian
When trying to use a comparison with a timestamp in a
query, no tuples are returned:
START TRANSACTION;
CREATE TABLE "datetest" (
"low" timestamp,
"high" timestamp
);
INSERT INTO "datetest" VALUES ('2004-12-27
16:29:57.409', '9999-12-31 00:00:00. 000');
select * from datetest where now() > low and now() < high;
ROLLBACK;
The select query returns no tuples, also when using a
fixed date instead of now().
The query select * from datetest where low < high; does
yield in the expected output.
The given case is checked in CVS.
## Comment 2635
Date: 2005-03-07 22:34:44 +0100
From: @njnes
Logged In: YES
user_id=43556
the swapping l<r to r>l didn't work properly
## Comment 2636
Date: 2005-10-06 19:09:26 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: ALREADY IN TESTWEB
## Comment 2637
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1158581 at http://sourceforge.net/support/tracker.php?aid=1158581
| SQL: timestamp comparisons do not work | https://api.github.com/repos/MonetDB/MonetDB/issues/486/comments | 0 | 2020-11-30T08:37:10Z | 2024-06-27T11:01:14Z | https://github.com/MonetDB/MonetDB/issues/486 | 753,274,154 | 486 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-07 03:45:32 +0100
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-03-07 10:36:15 +0100
## Comment 2629
Date: 2005-03-07 15:45:32 +0100
From: @MarcinZukowski
Using += gives a very misleading error:
MonetDB>var x:=1; x+=x;
!ERROR: "var x:=1; x+=x;"
! ^
! new MIL syntax (>=4.4) requires parentheses for all operators.
!.
## Comment 2630
Date: 2005-03-07 17:21:48 +0100
From: @MarcinZukowski
Logged In: YES
user_id=607094
Just to compare, in nostalgic mode it also doesn't work, it says:
MonetDB>var x:=1; x+=x;
!ERROR: "var x:=1; x+=x;"
! ^
! can't help you here, sorry.
## Comment 2631
Date: 2005-03-07 22:45:22 +0100
From: @MarcinZukowski
Logged In: YES
user_id=607094
Now, that was informative :(
## Comment 2632
Date: 2005-10-06 19:08:22 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2633
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1158329 at http://sourceforge.net/support/tracker.php?aid=1158329
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.
| Misleading error message for += | https://api.github.com/repos/MonetDB/MonetDB/issues/485/comments | 0 | 2020-11-30T08:37:06Z | 2024-06-28T13:35:13Z | https://github.com/MonetDB/MonetDB/issues/485 | 753,274,123 | 485 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-06 06:39:23 +0100
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-03-17 12:29:13 +0100
## Comment 2621
Date: 2005-03-06 18:39:23 +0100
From: Wouter Alink <<alink>>
When a division-by-zero occurs inside a select
statement, a server-crash is the result. (on winXP)
select 3/0; works fine and returns: division by zero
create table a (x int, y int);
insert into a values (1,2);
insert into a values (3,0);
commit;
select x/y from a; results in a Mserver-crash
## Comment 2622
Date: 2005-03-07 22:40:21 +0100
From: @njnes
Logged In: YES
user_id=43556
bat_arith batbat_* has no zero checking.
## Comment 2623
Date: 2005-03-08 07:27:22 +0100
From: @drstmane
Logged In: YES
user_id=572415
Maybe, we need two versions of such operators, one with
checking, and one without (for pure speed) ...
I'll have a look...
## Comment 2624
Date: 2005-03-17 00:06:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
(Kind of) fixed in CVS (MonetDB_4-6 branch):
bat_arith's "[/]" and "[%]" now check division by zero,
and produce a NIL in the result in that case.
OPEN / TODO:
Should we rather stop and issue an error in case of division
by zero?
The generic multiplex does it this way.
Comments? Opinions?
I'll leave this open, until we have a final decision...
## Comment 2625
Date: 2005-03-17 00:12:44 +0100
From: @njnes
Logged In: YES
user_id=43556
yes errors would be better.
## Comment 2626
Date: 2005-03-17 00:29:13 +0100
From: @drstmane
Logged In: YES
user_id=572415
Ok:
In case of Division by Zero, bat_arith's "[/]" & "[%]"
now stop processing and issue an error
instead of inserting NILs.
## Comment 2627
Date: 2005-10-06 19:07:37 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / DUPLICATE
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/div_by_zero.SF-987304.*
## Comment 2628
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1157817 at http://sourceforge.net/support/tracker.php?aid=1157817
| SQL: Mserver crashes on division by 0 | https://api.github.com/repos/MonetDB/MonetDB/issues/484/comments | 0 | 2020-11-30T08:37:03Z | 2024-06-27T11:01:12Z | https://github.com/MonetDB/MonetDB/issues/484 | 753,274,088 | 484 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-05 02:17:35 +0100
From: Rene Vincent Jansen <<rvjansen>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-03-06 08:39:13 +0100
## Comment 2608
Date: 2005-03-05 14:17:35 +0100
From: Rene Vincent Jansen <<rvjansen>>
MacOSX 10.3.8., 2*2.5Ghz G5
Hi Guys, I am in an R&D group at ABN AMRO and have built a
metadata/reference data repository application entirely built on
binary relationships. So it seemed a natural fit to try its performance
on MonetDB. The thing is, our data loader uses the mentioned JDBC
call that does not seem to be implemented. What I would like to
know is whether there is some kind of restriction, or that it is
something just in the planning. I am afraid I have no time to look into
it or contribute to the solution myself right now, but that might
change if the performance shows to be much better than DB2, SQL
Server, MySQL or PostgreSQL, on which we can run out of the box.
## Comment 2609
Date: 2005-03-05 14:26:17 +0100
From: @grobian
Logged In: YES
user_id=963970
Rene,
Thanks for your interest in MonetDB/SQL. The setObject is
indeed not (yet) implemented. I will have a look at it to
see if I can do it. My first judgements are that it is just
an overlooked issue.
Can you tell which types you are trying to set using the
setObject method?
## Comment 2610
Date: 2005-03-05 14:45:24 +0100
From: Rene Vincent Jansen <<rvjansen>>
Logged In: YES
user_id=375088
These include JDBC Types 4 (Integer), 1 (char), 93 (TimeStamp), 12
(varchar), -1 (long varchar) and 8 (double).
## Comment 2611
Date: 2005-03-05 21:18:09 +0100
From: @grobian
Logged In: YES
user_id=963970
Ah... I already feared this somehow. The setObject method
IS implemented, but it just calls a set<Type> method, and
you probably sumbled upon one which is currently not supported.
What type do you use for the long varchar (which is not a
recognised type in MonetDB/SQL)?
Can you show a little piece of SQL that creates the table as
you use it, so I can reconstruct your problem here?
## Comment 2612
Date: 2005-03-05 21:48:45 +0100
From: Rene Vincent Jansen <<rvjansen>>
Logged In: YES
user_id=375088
It seems to fail on an Integer (type 4), with value -4, so it seems not to be
a long varchar problem (which is defined as varchar(255) anyway. The
unload is from DB2, the SQL DDL I used here is the PostgreSQL version)
This is a fragment of the trace.
304 *=* when sqlTp = java.sql.Types.TIMESTAMP
>>> "0"
314 *=* otherwise
315 *-*
*=* do
316 *=* this.ps.setObject(cn, s, sqlTp)
>>> "1"
>>> "-4"
>>> "4"
319 *=* catch sqlE = SQLException
320 *=* System.err.println(sqlE.getMessage())
>>> "Operation currently not supported!"
Operation currently not supported!
321 *=* say this.lastKey "columnnumber" cn "sqlTp" sqlTp "string" s
>>> "OID columnnumber 1 sqlTp 4 string -4"
OID columnnumber 1 sqlTp 4 string -4
322 *=* say "row tag number:" this.rowCount
>>> "row tag number: 1"
row tag number: 1
323 *=* sqlE.printStackTrace()
java.sql.SQLException: Operation currently not supported!
at nl.cwi.monetdb.jdbc.MonetPreparedStatement.setObject(Unknown
Source)
at
com.abnamro.midms.sqlxml.InsertXMLRowHandler.buildInsert(InsertXMLR
owHandler.nrx:618)
## Comment 2613
Date: 2005-03-06 09:44:31 +0100
From: @grobian
Logged In: YES
user_id=963970
Thanks for your trace.
The setObject method with signature
setObject(int parameterIndex, Object x, int targetSqlType)
Sets the value of the designated parameter with
the given object.
is not yet implemented. I will make this implementation as
far as that is possible.
## Comment 2614
Date: 2005-03-06 13:40:19 +0100
From: @grobian
Logged In: YES
user_id=963970
I checked in an implementation of the two missing setObject
methods on the current branch. Are you able to quickly
fetch the updated files, or do you prefer me to give you a
new JDBC jar-file?
I haven't tested the added code yet, I have to write a test
for that first. Please let me know your experiences.
## Comment 2615
Date: 2005-03-06 13:44:43 +0100
From: Rene Vincent Jansen <<rvjansen>>
Logged In: YES
user_id=375088
Aha, great. A new jar would be most convenient; I did not set up CVS
access yet.
## Comment 2616
Date: 2005-03-06 13:46:20 +0100
From: @grobian
Logged In: YES
user_id=963970
is it fine with you if I send it by email to your sf email
address?
## Comment 2617
Date: 2005-03-06 13:48:52 +0100
From: Rene Vincent Jansen <<rvjansen>>
Logged In: YES
user_id=375088
please send to rvjansen@xs4all.nl
## Comment 2618
Date: 2005-03-06 20:39:13 +0100
From: @grobian
Logged In: YES
user_id=963970
Fixed in current branch
## Comment 2619
Date: 2005-10-06 19:02:16 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: HANDLED BY meltdown
BugDay_2005-10-06: NO TEST / POSTPONED
needs a test JDBC app to do this.
## Comment 2620
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1157252 at http://sourceforge.net/support/tracker.php?aid=1157252
| preparedStatement.setObject not implemented | https://api.github.com/repos/MonetDB/MonetDB/issues/483/comments | 0 | 2020-11-30T08:37:00Z | 2024-06-27T11:01:11Z | https://github.com/MonetDB/MonetDB/issues/483 | 753,274,060 | 483 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-03 06:54:30 +0100
From: @swingbit
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-03-09 11:12:18 +0100
## Comment 2604
Date: 2005-03-03 18:54:30 +0100
From: @swingbit
.COMMAND kunion ( BAT[any::1,any::2] left,
BAT[any::1,any::2] right) :
BAT[any::1,any::2] = CMDkunion;
"Returns the union of two BATs; looking at
head-columns only.
Results in all BUNs of 'left' that are not in
'right', plus all BUNs
of 'right'. *no* double-elimination is done. If you
want this, do:
'kunion(left.kunique,right.kunique)' or:
'sunion(left,right).kunique'."
Either the implementation or the documentation is wrong.
It does exactly what it says here, if you swap every
'left' and 'right' occurrences.
Test:
var left := new(oid, int);
left.insert(oid(0), 0);
left.insert(oid(1), 1);
left.insert(oid(2), 2);
left.insert(oid(3), 3);
left.insert(oid(4), 4);
var right := new(oid, int);
right.insert(oid(1), 10);
right.insert(oid(2), 20);
right.insert(oid(4), 40);
kunion(left,right).print(); correct according to
doc, wrong result
kunion(right,left).print(); intended result
## Comment 2605
Date: 2005-03-09 11:12:18 +0100
From: @njnes
Logged In: YES
user_id=43556
indeed the implementation is the reversed of the
documentation. Fixed
documentation to fit the implementation
## Comment 2606
Date: 2005-10-06 18:59:20 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2607
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1156059 at http://sourceforge.net/support/tracker.php?aid=1156059
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.
| kunion, wrong implementation or wron documentation? | https://api.github.com/repos/MonetDB/MonetDB/issues/482/comments | 0 | 2020-11-30T08:36:57Z | 2024-06-27T11:01:10Z | https://github.com/MonetDB/MonetDB/issues/482 | 753,274,038 | 482 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-03-01 09:57:11 +0100
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-03-16 09:46:19 +0100
## Comment 2600
Date: 2005-03-01 21:57:11 +0100
From: @MarcinZukowski
Do this:
var b:=bat(void,int);
b.insert(0@0,1);
b.rename("xyz");
b.persists(true);
commit();
And now run such script twice:
proc int2lng(str intbat, str lngbat) : void
{
printf("Destroying %s\n", lngbat);
destroy(lngbat);
commit();
printf("Converting int %s to lng %s\n", intbat, lngbat);
var i := bat(intbat);
var l:=[lng](i);
l.rename(lngbat);
l.persists(true);
commit();
}
int2lng("xyz", "xyz_lng");
quit();
The second time I run it it crashes on destroy or first commit.
Then every second time it crashes.
## Comment 2601
Date: 2005-03-16 21:46:19 +0100
From: @njnes
Logged In: YES
user_id=43556
The BATstore read from disk contains the bun heap base
pointer. We now
reset this too zero. This solves the crash, ie the batstore
gets correctly freed
and the heap isn't freed (as its not there, its only the
BATdesc).
## Comment 2602
Date: 2005-10-06 18:58:30 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCEED
tests/BugDay_2005-10-06_4.9.3/Tests/crash_destroy_commit.*
## Comment 2603
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1154552 at http://sourceforge.net/support/tracker.php?aid=1154552
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.
| Crash on destroy/commit | https://api.github.com/repos/MonetDB/MonetDB/issues/481/comments | 0 | 2020-11-30T08:36:54Z | 2024-06-27T11:01:09Z | https://github.com/MonetDB/MonetDB/issues/481 | 753,274,013 | 481 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-28 06:51:41 +0100
From: Wouter Alink <<alink>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-11-20 03:28:45 +0100
## Comment 2594
Date: 2005-02-28 18:51:41 +0100
From: Wouter Alink <<alink>>
When a BAT starts with a nil value, multiplexing fails.
See example below:
MonetDB>var a := bat(int,int);
MonetDB>a.insert(1,2);
MonetDB>a.insert(2,3);
MonetDB>var b := bat(int,int);
MonetDB>b.insert(1,4);
MonetDB>b.insert(2,9);
MonetDB>print(a,b);
-------------------------
h tmp_61 tmp_50 name
int int int type
-------------------------
[ 1, 2, 4 ]
[ 2, 3, 9 ]
MonetDB>var b := bat(int,int);
MonetDB>b.insert(int(nil),int(nil));
MonetDB>b.insert(1,4);
MonetDB>b.insert(2,9);
MonetDB>print(a,b);
-------------------------
h tmp_61 tmp_50 name
int int int type
-------------------------
MonetDB>
## Comment 2595
Date: 2005-03-16 22:45:21 +0100
From: @njnes
Logged In: YES
user_id=43556
we now check on the second insert in a bat if the first is
nil and unset sorted if so. (later inserts of nil are no
problem as nil < any value).
## Comment 2596
Date: 2005-10-06 18:49:04 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/multiplex_nil.SF-1153664.*
## Comment 2597
Date: 2005-10-11 14:44:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
re-opened bug report, as the newly added test script
(tests/BugDay_2005-10-06_4.9.3/Tests/multiplex_nil.SF-1153664.*)
revelas that the bug still (again?) occurs, but only with
32-bit OIDs, not with 64-bit OIDs!??
cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.mTests103/index_short.html
or directly
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.mTests103/GNU.32.32.d-Fedora4/tests_BugDay_2005-10-06_4.9.3/multiplex_nil.SF-1153664.out.00.html
## Comment 2598
Date: 2005-11-20 15:28:45 +0100
From: @drstmane
Logged In: YES
user_id=572415
Apparently fixed by Peter.
## Comment 2599
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1153664 at http://sourceforge.net/support/tracker.php?aid=1153664
| Multiplexing fails... with a nil-value | https://api.github.com/repos/MonetDB/MonetDB/issues/480/comments | 0 | 2020-11-30T08:36:51Z | 2024-06-27T11:01:07Z | https://github.com/MonetDB/MonetDB/issues/480 | 753,273,984 | 480 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-26 03:29:20 +0100
From: Agustin Schapira <<schapirama>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-02-28 03:55:10 +0100
## Comment 2588
Date: 2005-02-26 15:29:20 +0100
From: Agustin Schapira <<schapirama>>
There is a mismatch between what rename(bat,str) accepts as a
valid BAT name, and what bat(str) can handle. If I include a hyphen
(-) when renaming a BAT, then using bat(str) to retrieve it doesn't
work. Example:
var a:=bat(int,int);
a.rename("x");
bat("x").print();
-----------------
h x name
int int type
-----------------
works correctly, but
var a:=bat(int,int);
a.rename("x-");
a.print();
-----------------
h x- name
int int type
-----------------
bat("x-").print();
!ERROR: interpret_params: print(param 1): invalid BAT.
fails.
Maybe rename should not accept the (-) as a valid character?
## Comment 2589
Date: 2005-02-28 12:43:56 +0100
From: @drstmane
Logged In: YES
user_id=572415
Agustin,
As of MonetDB 4.6.0, rename does issue an error with
non-supported characters:
========
Monet Database Server V4.6.0
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>var a:=bat(int,int);
MonetDB>a.rename("x-");
!ERROR: CMDrename: identifier expected: x-
!ERROR: CMDrename: operation failed.
!ERROR: interpret_params: not(param 1): evaluation error.
!ERROR: yytree: rename: x-.
MonetDB>a.print();
-----------------
h tmp_29 name
int int type
-----------------
MonetDB>
========
Actually, this fix has already been in CVS since Feb 10 2005.
Are you indeed using MonetDB 4.6.0 as the "Group" field of
this bug report suggests, or are you using an older version?
I'll leave this open for now, until I get your reply, but
iMHO, this bug is not "valid" any more.
Stefan
## Comment 2590
Date: 2005-02-28 14:09:32 +0100
From: Agustin Schapira <<schapirama>>
Logged In: YES
user_id=62434
Stefan,
I just tried this on 4.6.0_rc3, and it still fails:
Monet Database Server V4.6.0_rc3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for powerpc-apple-darwin7.7.0/32bit; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>var a:=bat(int,int);
var a:=bat(int,int);
MonetDB>a.rename("x-");
a.rename("x-");
MonetDB>a.print();
a.print();
-----------------
h x- name
int int type
-----------------
On 4.6.0 (final version) it does indeed work as you described (rename fails
when given an invalid char)...
I was still using a release candidate version.
Thanks,
-Agustin
## Comment 2591
Date: 2005-02-28 15:55:10 +0100
From: @drstmane
Logged In: YES
user_id=572415
Agustin,
the fix was in the release candidate 4.6.0_rc3, but only as
of Feb 10 2005; hence, your checkout was probably before
that date.
Assuming that the final 4.6.0 now works fine for you, I
close this bug report.
Stefan
## Comment 2592
Date: 2005-10-06 18:45:53 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/rename_bat_hyphen.*
(sorry, forgot sf bug number :/ )
## Comment 2593
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1152369 at http://sourceforge.net/support/tracker.php?aid=1152369
| Mismatch between rename(bat, str) and bat(str) | https://api.github.com/repos/MonetDB/MonetDB/issues/479/comments | 0 | 2020-11-30T08:36:49Z | 2024-06-27T11:01:06Z | https://github.com/MonetDB/MonetDB/issues/479 | 753,273,954 | 479 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-24 11:25:30 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-10-13 09:09:53 +0200
## Comment 2583
Date: 2005-02-24 11:25:30 +0100
From: @grobian
monetdb-> select cast(value as varchar(3)) from sessions;
+---------------+
| convert_value |
+===============+
| 0 |
| 250 |
| |
| true |
| sys |
| monetdb |
| monetdb |
| 0 |
| true |
| true |
+---------------+
10 rows
monetdb-> select cast(value as char(3)) from sessions;
+---------------+
| convert_value |
+===============+
| 0 |
| 250 |
| |
| true |
| sys |
| monetdb |
| monetdb |
| 0 |
| true |
| true |
+---------------+
10 rows
Either an error should have been raised, or the values
should have been truncated to reflect the maximum
allowed size (3).
Since it is a explicit cast, the values should be
truncated imho.
## Comment 2584
Date: 2005-03-09 13:02:42 +0100
From: @njnes
Logged In: YES
user_id=43556
Only the lookup of the truncate function failed (because of
changes to the
various int types).
## Comment 2585
Date: 2005-10-06 17:57:07 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / FAILURE
regression, the original bug is open again. I patched the
output to what I would expect to get.
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/var_char_cast_truncate.SF-1150907.*
## Comment 2586
Date: 2005-10-13 21:09:53 +0200
From: @drstmane
Logged In: YES
user_id=572415
Appears to be fixed in CVS, now.
## Comment 2587
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1150907 at http://sourceforge.net/support/tracker.php?aid=1150907
| SQL: cast to smaller (var)char does not truncate strings | https://api.github.com/repos/MonetDB/MonetDB/issues/478/comments | 0 | 2020-11-30T08:36:45Z | 2024-06-27T11:01:05Z | https://github.com/MonetDB/MonetDB/issues/478 | 753,273,902 | 478 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-23 12:55:30 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-03-09 11:54:45 +0100
## Comment 2579
Date: 2005-02-23 12:55:30 +0100
From: @grobian
Observe the following scenario:
monetdb-> create table t(id int);
Operation successful
monetdb-> insert into t values (1), (2);
1 affected row
monetdb-> select * from t;
+------------+
| id |
+============+
| 1 |
| 2 |
+------------+
2 rows
The two rows are correctly inserted, but the affected
row count does not reflect this correctly.
## Comment 2580
Date: 2005-03-09 11:54:45 +0100
From: @njnes
Logged In: YES
user_id=43556
now the inserts are first collected and then inserted in one
go. The count is now done on this single big insert.
## Comment 2581
Date: 2005-10-06 17:48:39 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/affected_rows_count.SF-1149757.*
## Comment 2582
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1149757 at http://sourceforge.net/support/tracker.php?aid=1149757
| SQL: inconsistant affected rows count on multi-valued insert | https://api.github.com/repos/MonetDB/MonetDB/issues/477/comments | 0 | 2020-11-30T08:36:41Z | 2024-06-27T11:01:04Z | https://github.com/MonetDB/MonetDB/issues/477 | 753,273,869 | 477 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-22 04:55:46 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-02-24 05:20:40 +0100
## Comment 2574
Date: 2005-02-22 16:55:46 +0100
From: Wouter Alink <<alink>>
See below, a bat created using a name of a non-existing
bat, results in a invalid bat, without warning.
MonetDB>var t := bat("frboierfwfr");
MonetDB>t.type().print();
[ 4 ]
MonetDB>t.info().print();
!ERROR: interpret_params: info(param 1): invalid BAT.
!ERROR: interpret_params: print(param 1): evaluation error.
## Comment 2575
Date: 2005-02-22 19:33:21 +0100
From: @drstmane
Logged In: YES
user_id=572415
Actually (and dispite my previous statement that acutally
tirggered this bug report) the error message is correct and
also the bat(<str>) cast works correctly:
- The bat(<str>) cast correctly detects that no BAT with the
given name exists,
and hence, it correctly returns "bat_nil".
- However, "bat_nil" cannot not be handled by any MIL
command, as already the interpreter checks for "valid" BATs,
considers "bat_nil" as invalid, and issues the above error.
- Thus, not even `isnil(bat("frboierfwfr"));' can work ...
Two questions arise:
- Should the "bat(<str>)" cast should (in addition to
returning "bat_nil") also echo a warning?
This is not handy, in case the "bat(<str>)" cast is used
internally, and "bat_nil" return are handled elsewhere.
- Should we make "isnil(bat(nil))" work, which basically
means patching the interpreter somehow?
IMHO, both solutions are not good, but I don't have a better
one, yet ...
## Comment 2576
Date: 2005-02-24 17:20:40 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
I just discovered that Marcin already mentioned this
behaviour in a feature request (607094). This topic will be
closed.
## Comment 2577
Date: 2005-12-21 11:34:20 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay: stmane: NO TEST / REDUNDANT / FEATURE REQUEST
## Comment 2578
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1146347 at http://sourceforge.net/support/tracker.php?aid=1146347
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.
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| Invalid BAT with type 4 | https://api.github.com/repos/MonetDB/MonetDB/issues/476/comments | 0 | 2020-11-30T08:36:38Z | 2024-06-28T13:35:55Z | https://github.com/MonetDB/MonetDB/issues/476 | 753,273,830 | 476 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-22 10:56:38 +0100
From: @yzchang
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2005-02-22 06:39:35 +0100
## Comment 2566
Date: 2005-02-22 10:56:38 +0100
From: @yzchang
After altering a table to add a foreign key in it, if I
then want to have a look at the definition of this
table by '\d <tablename>', Mserver crashes with the
following error message:
MonetDB>Mserver:
../../../../src/storage/bat/bat_store.mx:1689: key_dup:
Assertion `n' failed.
Aborted
If I restart Mserver with SQL server, error occurs in
the same file, but in another line:
MonetDB>Mserver:
../../../../src/storage/bat/bat_store.mx:1689: key_dup:
Assertion `n' failed.
Aborted
## Comment 2567
Date: 2005-02-22 13:00:04 +0100
From: @njnes
Logged In: YES
user_id=43556
can you include the full sql sequence into this bug report
(or attach the
sql files)?
## Comment 2568
Date: 2005-02-22 13:07:08 +0100
From: @yzchang
Logged In: YES
user_id=341633
Here is the file I used
## Comment 2569
Date: 2005-02-22 13:12:25 +0100
From: @yzchang
Logged In: YES
user_id=341633
Here is the file I used
## Comment 2570
Date: 2005-02-22 13:12:38 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
When you attach a file, don't forget to check the box,
otherwise the file will not be attached.
## Comment 2571
Date: 2005-02-22 13:18:21 +0100
From: @yzchang
Logged In: YES
user_id=341633
I have noticed that too. The file can now be downloaded at
sourforge.
## Comment 2572
Date: 2005-10-06 16:17:18 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/alter_table_describe.SF-1146092*
## Comment 2573
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1146092 at http://sourceforge.net/support/tracker.php?aid=1146092
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.
| Mserver crashes when view an altered table | https://api.github.com/repos/MonetDB/MonetDB/issues/475/comments | 0 | 2020-11-30T08:36:34Z | 2024-06-27T11:01:02Z | https://github.com/MonetDB/MonetDB/issues/475 | 753,273,790 | 475 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-22 10:34:31 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-02-22 06:39:56 +0100
## Comment 2563
Date: 2005-02-22 10:34:31 +0100
From: @grobian
consider the following scenario:
monetdb-> select class from range;
+------------+
| class |
+============+
| 17 |
| 17 |
| 17 |
| 2 |
| 17 |
| 16 |
| 16 |
| 16 |
| 16 |
| 16 |
| 11 |
| 16 |
| 16 |
| 18 |
| 18 |
| 16 |
+------------+
16 rows
monetdb-> select class from range except select 16;
Error: ERROR: interpret: no matching MIL operator to
'append(BAT[void,int], BAT[void,sht])'.
MAYBE YOU MEAN:
append(BAT[oid,void], BAT[oid,void]) :
BAT[oid,void]
append(BAT[void,any::1], BAT[oid,any::1]) :
BAT[void,any::1]
append(BAT[void,any::1], BAT[void,any::1]) :
BAT[void,any::1]
append(BAT[void,any::1], any::1) : BAT[void,any::1]
append(BAT[oid,any::1], BAT[oid,any::1]) :
BAT[oid,any::1]
append(BAT[oid,any::1], any::1) : BAT[oid,any::1]
monetdb-> select class from range except select cast
(16 as integer);
+------------+
| class |
+============+
| 17 |
| 2 |
| 16 |
| 11 |
| 18 |
+------------+
5 rows
monetdb-> select class from range except all select
cast (16 as integer);
+------------+
| class |
+============+
| 17 |
| 17 |
| 17 |
| 2 |
| 17 |
| 16 |
| 16 |
| 16 |
| 16 |
| 16 |
| 11 |
| 16 |
| 16 |
| 18 |
| 18 |
| 16 |
| 16 |
+------------+
17 rows
monetdb->
from the scenario above it is clear that a desirable
upcast isn't made automagically and that the EXCEPT is
treated like a UNION instead of the expected (k)diff
behaviour.
## Comment 2564
Date: 2005-10-06 15:24:23 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: ALREADY IN TESTWEB
sql/src/test/bugs/except-union-intersect-bug-sf-1146079.sql
## Comment 2565
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1146079 at http://sourceforge.net/support/tracker.php?aid=1146079
| SQL: EXCEPT is picky and doesn't work | https://api.github.com/repos/MonetDB/MonetDB/issues/474/comments | 0 | 2020-11-30T08:36:31Z | 2024-06-27T11:01:01Z | https://github.com/MonetDB/MonetDB/issues/474 | 753,273,770 | 474 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-21 11:12:42 +0100
From: @sjoerdmullender
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-21 11:49:48 +0100
## Comment 2559
Date: 2005-02-21 11:12:42 +0100
From: @sjoerdmullender
The Arjen_01* tests on Darwin fail. The problem is, a
number which should get printed as -30 gets printed as
4294967266 (there are a few others like this as well).
The problem seems to be that this particular number is
of type ssize_t and gets sprintf'ed using the SSZFMT
format. This is all correct. However, Darwin manages
to format the number as if it were unsigned.
This happens in local_itoa which is called from
CMDinfo, both in bat.mx.
## Comment 2560
Date: 2005-02-21 11:49:48 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Darwin treats %zd as an usigned format, so this is actually
a bug in Darwin. We now work around this by not using this
format on Darwin.
## Comment 2561
Date: 2005-10-06 15:22:33 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: ALREADY IN TESTWEB
## Comment 2562
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1145306 at http://sourceforge.net/support/tracker.php?aid=1145306
| test failure on Darwin | https://api.github.com/repos/MonetDB/MonetDB/issues/473/comments | 0 | 2020-11-30T08:36:29Z | 2024-06-27T11:01:00Z | https://github.com/MonetDB/MonetDB/issues/473 | 753,273,739 | 473 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-20 11:12:01 +0100
From: Jan Rittinger <<jan.rittinger>>
To: @njnes
Version: -- development
Last updated: 2005-02-22 03:27:46 +0100
## Comment 2544
Date: 2005-02-20 23:12:01 +0100
From: Jan Rittinger <<jan.rittinger>>
I have installed Suse9.2 without anything. That's
probably why configure_pathfinder prints the following
lines (@1). With this messages I expect pathfinder to
compile with the loader disabled. But I get an compile
error (@2).
@1:
checking for xml2-32-config... no
checking for xml2-config... no
checking libxml/parser.h usability... no
checking libxml/parser.h presence... no
checking for libxml/parser.h... no
configure: WARNING: The libxml2 XML parsing library was
not found or you have
requested to not use it.
This will disable XML Schema import
functionality in the
Pathfinder compiler. The XML document
loader does not work
without the libxml2 library and will
not be compiled at all.
checking for dlopen in -ldl... (cached) yes
checking for pthread_create in -lpthread... yes
checking gc.h usability... yes
checking gc.h presence... yes
checking for gc.h... yes
checking for GC_malloc in -lgc... yes
checking for u_long... no
checking for db.h... no
configure: WARNING: The BerkeleyDB library was not
found on your system or you
have requested not to use it. This will
disable the duplicate
removal feature in the XML document loader.
@2:
gcc -DHAVE_CONFIG_H -I.
-I/local_tmp/rittinge/pathfinder/loader -I..
-I/local_tmp/rittinge/pathfinder/loader
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/common
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/gdk
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/monet
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/plain
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/contrib
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/mapi
-I/local_tmp/rittinge/MonetDB/Linux/include/MonetDB/C-g
-O2 -std=c99 -Wall -W
-Werror-implicit-function-declaration -Werror
-Wno-format -Wno-unused-function -Wno-unused-label
-Wno-strict-aliasing -c
/local_tmp/rittinge/pathfinder/loader/pf-shred.c
/local_tmp/rittinge/pathfinder/loader/pf-shred.c:155:27:
libxml/parser.h: Datei oder Verzeichnis nicht gefunden
/local_tmp/rittinge/pathfinder/loader/pf-shred.c:156:36:
libxml/parserInternals.h: Datei oder Verzeichnis nicht
gefunden
/local_tmp/rittinge/pathfinder/loader/pf-shred.c:291:
error: parse error before "FILE"
## Comment 2545
Date: 2005-02-22 10:43:50 +0100
From: @drstmane
Logged In: YES
user_id=572415
Thanks Niels!
I'm actually not sure, though, whether this fix will make
the rest of PF comple without libxml2: The inlined shredder
(runtime/shredder.mx) also requires libxml2 and most
probably silently assumes that is it there (there are no
checks whether any HAVE_LIBXML or HAVE_LIBXML_PARSER_H
macros are defined).
Since my machines do have libxml2, I haven't tested it yet.
JanR, could you eventaully test this, and report?
In case it doesn't work, JanF might have to have a look at
the inlined shredder. Either, he finds a way to get it at
least compile without libxml2 (maybe with then with "empty"
functionality); or we'll have to make lbixml2 a prerequisite
in PF's configure, and stop immediately, if configure cannot
find it.
## Comment 2546
Date: 2005-02-22 10:46:28 +0100
From: @drstmane
Logged In: YES
user_id=572415
[Assigned to JanF to get him a copy of this.]
## Comment 2547
Date: 2005-02-22 11:04:13 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
I saw Niels has disabled the loader because of xml2 and this already
made me a bit worried. The shredder is just as tightly coupled to
libxml2 as the loader.
I think there is no point in running pathfinder without the loader or
shredder because we could never run a query because there is no
data!
But if persist and would like to do this the only way to make
pathfinder run is to disable the shredder too! I think a giant ifdef
checking for LIBXML2 would be sufficient. In the else we would
need to generate default error printing routines for the shred
functions.
(PS: I ran SuSe 9.2 default too also and even I needed 2 extra
packages to make bootstrap/configure work.)
## Comment 2548
Date: 2005-02-22 11:12:19 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Alternatively, issue an ERROR message from configure when
libxml2 isn't found.
## Comment 2549
Date: 2005-02-22 12:02:00 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
I reopened the bug report, because the shredder.mx also
'had' problems and is not fixed according to Stefans comment.
I can't test it anymore, because now libxml2-devel is
installed :(
## Comment 2550
Date: 2005-02-22 12:02:17 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
I reopened the bug report, because the shredder.mx also
'had' problems and is not fixed according to Stefans comment.
I can't test it anymore, because now libxml2-devel is
installed :(
## Comment 2551
Date: 2005-02-22 12:29:45 +0100
From: @njnes
Logged In: YES
user_id=43556
okay we simply bailout if lib xml is missing
## Comment 2552
Date: 2005-02-22 13:25:00 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
if you bail out please produce an error message, which fits
the problem.
In my case not libxml2 but libxml2-devel was missing.
## Comment 2553
Date: 2005-02-22 13:35:03 +0100
From: @njnes
Logged In: YES
user_id=43556
well libxml2 is a software package which suse and redhat
break atleast in
libxml2 and libxml2-devel. But other distribution may use
different names.
So probably we should just indicate what is missing ie libs
and/or include
files.
## Comment 2554
Date: 2005-02-22 15:25:35 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
if you bail out please produce an error message, which fits
the problem.
In my case not libxml2 but libxml2-devel was missing.
## Comment 2555
Date: 2005-02-22 15:27:46 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
sorry somehow the reload button opened it again
## Comment 2556
Date: 2005-02-22 16:20:24 +0100
From: @grobian
Logged In: YES
user_id=963970
don't use the reload button on sourceforge after posting.
just go back to the bug list page and click it again, or
select the url and press enter to force your browser
forgetting its POST cache.
## Comment 2557
Date: 2005-10-06 15:19:25 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / COMPILATION
## Comment 2558
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1145059 at http://sourceforge.net/support/tracker.php?aid=1145059
| pf doesn't compile without libxml | https://api.github.com/repos/MonetDB/MonetDB/issues/472/comments | 0 | 2020-11-30T08:36:26Z | 2024-06-27T11:00:59Z | https://github.com/MonetDB/MonetDB/issues/472 | 753,273,710 | 472 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-18 02:49:00 +0100
From: Jens Teubner <<teubner>>
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-04-15 08:57:26 +0200
## Comment 2536
Date: 2005-02-18 14:49:00 +0100
From: Jens Teubner <<teubner>>
shred_doc() appears to have problems with XML documents that use
namespaces:
The document
<foo xmlns:a='http://www.a.com/'>
<a:bar>Foobar</a:bar>
</foo>
results in
Mserver: Opening and ending tag mismatch: a line 4 and a:bar
: Unknown error 1135097575
The pf-shred loader loads the same XML document fine. Actually, I'm a bit
surprised by this problem, as it is be the libxml2 library that checks for tag
name (mis)matches. (The above error message does not come from the
code in shredder.mx.)
There are actually two more issues I discovered which concern document
loading:
1. When loading the above document, we currently store the prefix `a' in
the namespace field (or, the corresponding BAT). This prefix, however, can
freely be chosen by the author of the XML document, and the document
should behave identically, if we use any other prefix. So I'd suggest to not
store the namespace prefix in the pre|ns container, but rather the URI that
is assigned to this prefix. This may make the shredding code a bit more
tricky. Maybe there's some feature in the libxml2 library that helps us
here.
2. (This issue I'm not sure about myself.) Consider an XML document that
references a DTD. This DTD may contain entity definitions that are required
to parse the XML document (e.g. `ä'). In that case, shredding results
in error messages about undefined entities. DTDs can also affect the
behavior of SAX itself: SAX will not report ``ignorable whitespace''.
Whitespace is ignorable, if no text nodes can appear at that place in the
document, according to the DTD (no PCDATA or mixed content). So should
we try to respect DTD references in XML documents? Or make this an
option during document loading? (xmllint also complains in case of
undefined entities, except if --valid is turned on. xmllint may, by the way,
be used as a command line utility to simply expand all entities before
loading the document into MonetDB.)
## Comment 2537
Date: 2005-02-18 15:54:41 +0100
From: @drstmane
Logged In: YES
user_id=572415
Same bug occurs when loading the "document_2.xml" of the
XPathMark benchmark (cf. tests/XPathMark/Tests/document_2*
and
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU_32-d_Fedora3/tests_XPathMark/document_2.err.00.html
)
## Comment 2538
Date: 2005-02-18 17:36:50 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Problem (0). is solved. This was a case of 1 memory
optimization too far. Also improved error message for XML
errors. The current version erratically calls the Mserver error
() function! Will check-in the updates on monday (I cannot
check in from my notebook at home!)
I will try to make on option for problem 1. When there is no
support from libxml2 I think the biggest problem is the scope
of the xmlns definition. But the program already keeps track
of a nodestack so if we attach a symbol table there it may
even be easy!
And I'm also not very sure about problem 2:-)
## Comment 2539
Date: 2005-04-15 08:33:12 +0200
From: @drstmane
Logged In: YES
user_id=572415
Jens, Jan,
I haven't followed you recent changes in detail:
What's the status of this bug report?
Are all issues solved?
If so, we should close the bug report.
If only 2. still open, we shoudll consider making it a bug
report/feature request of its own, and close this one.
Thanks in advance,
Stefan
## Comment 2540
Date: 2005-04-15 08:57:26 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
I think all of the points are fixed now. This was a really
long one (bug/dtd load/id(refs)/xmlns! If a point is not
fixed we should make a new report for clarity!
## Comment 2541
Date: 2005-10-06 15:11:46 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: REASSIGN vzzbx
## Comment 2542
Date: 2005-10-06 15:21:27 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-10-06: Done.
ALREADY IN TESTWEB
## Comment 2543
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1143799 at http://sourceforge.net/support/tracker.php?aid=1143799
| shred_doc has problems with namespaces | https://api.github.com/repos/MonetDB/MonetDB/issues/471/comments | 0 | 2020-11-30T08:36:23Z | 2024-06-27T11:00:58Z | https://github.com/MonetDB/MonetDB/issues/471 | 753,273,685 | 471 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-17 04:29:43 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-02-17 05:02:10 +0100
## Comment 2532
Date: 2005-02-17 16:29:43 +0100
From: Wouter Alink <<alink>>
converting a string to a timestamp does not work
neither is there a test specified to test this
behaviour (no wonder it passes the test ;)
see code below:
MonetDB>module(monettime);
MonetDB>var d := timestamp("14-aug-2002 16:14");
MonetDB>d.print();
[ nil ]
MonetDB>var d := timestamp();
MonetDB>d.print();
[ 14 ]
## Comment 2533
Date: 2005-02-17 17:02:10 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
problem is only in website documentation (fixed)...
(monettime states the correct format: year-month-day)
## Comment 2534
Date: 2005-10-06 15:10:15 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / INVALID
## Comment 2535
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1124816 at http://sourceforge.net/support/tracker.php?aid=1124816
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.
| convert string -> timestamp | https://api.github.com/repos/MonetDB/MonetDB/issues/470/comments | 0 | 2020-11-30T08:36:21Z | 2024-06-27T11:00:57Z | https://github.com/MonetDB/MonetDB/issues/470 | 753,273,654 | 470 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-17 10:36:49 +0100
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-22 09:33:41 +0100
## Comment 2527
Date: 2005-02-17 10:36:49 +0100
From: Wouter Alink <<alink>>
On the website the section Documentation -> Core ->
Kernel Guide
is not rendered correctly. Seems like tex-statements
are not translated to HTML.
## Comment 2528
Date: 2005-02-17 14:53:54 +0100
From: @drstmane
Logged In: YES
user_id=572415
The problem stems from the fact, that the literal texi
directives that had [to] be[en] introduced in the gdk*.mx
files for texi extraction appear also literally in the tex &
html extraction, and hence "mess-up" the web site.
See also RFE 1122341 "Clean-up Mx: remove all html
extraction/generation feature"
(http://sourceforge.net/tracker/index.php?func=detail&aid=1122341&group_id=56967&atid=482471).
## Comment 2529
Date: 2005-02-22 09:33:41 +0100
From: @njnes
Logged In: YES
user_id=43556
we now use texinfo for the kernel guide. This is inline with
a complete conversion to texinfo (mx -> texinfo -> html).
## Comment 2530
Date: 2005-10-06 14:50:55 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2531
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1124595 at http://sourceforge.net/support/tracker.php?aid=1124595
| Kernel Guide not rendered correctly | https://api.github.com/repos/MonetDB/MonetDB/issues/469/comments | 0 | 2020-11-30T08:36:18Z | 2024-06-27T11:00:56Z | https://github.com/MonetDB/MonetDB/issues/469 | 753,273,627 | 469 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-16 01:50:46 +0100
From: @grobian
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2005-03-16 11:40:28 +0100
## Comment 2518
Date: 2005-02-16 13:50:46 +0100
From: @grobian
An attempt to stop a running Mserver using a TERM
signal (or Ctrl-C) in its running terminal results in a
fatal error message, indicating the kernel abruptly
stopped.
For running MonetDB as daemon, it is useful to be able
to signal Mserver in order to let it gracefully stop.
At the moment this seems to be impossible.
MonetDB>!FATAL: Interrupt received
## Comment 2519
Date: 2005-02-16 14:32:40 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
The message indicates that Mserver *does* handle interrupt
(i.e. SIGINT). If it didn't, you would just get a shell
prompt and no "FATAL" message.
But I agree, Mserver should exit gracefully when it receives
SIGTERM or SIGINT.
But isn't this really an RFE?
## Comment 2520
Date: 2005-02-16 15:26:34 +0100
From: @grobian
Logged In: YES
user_id=963970
no, it's a bug, because it makes it impossible to set up a
daemonised MonetDB without running the risk to break
something upon stop. The option monet_daemon however
suggests that this is possible, and under normal user driven
circumstances this is the way to start and stop a database.
(via a init.d start/stop script or perhaps a windows service)
## Comment 2521
Date: 2005-02-16 15:43:17 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
It's trivial to add
ifdef SIGTERM
(void) signal(SIGTERM, BATSIGinterrupt);
endif
to BATSIGinit in gdk_utils.mx. This will catch SIGTERM (in
addition to SIGINT which is already covered) and calls
BATSIGinterrupt when terminated.
The next step is to do something useful in BATSIGinterrupt.
Currently it calls GDKfatal which in turn calls GDKexit
which (surprise!) exits. If this isn't good enough,
somebody should define what *is* good enough.
## Comment 2522
Date: 2005-02-16 16:20:28 +0100
From: @grobian
Logged In: YES
user_id=963970
Ok, I messed up the signals a bit. The !FATAL message only
appears on a SIGINT signal (as by Ctrl-C). For KILL and
TERM the OS seems to handle the signal.
I think TERM should be caught and used to normally quit the
server, without a FATAL message.
% kill -INT 7865
MonetDB>!FATAL: Interrupt received
% kill -TERM 7884
MonetDB>Terminated
% kill -KILL 7902
MonetDB>Killed
## Comment 2523
Date: 2005-02-16 16:32:41 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
I agree.
I don't see the usefulness of saying "!FATAL: Interrupt
received" when the server is interrupted. I think SIGINT
and SIGTERM can be handled in the same way by doing a
graceful exit (i.e. cleanup + exit(1)).
Of course, SIGKILL can't be caught.
I'd suggest adding the lines
ifdef SIGTERM
(void) signal(SIGTERM, BATSIGinterrupt);
endif
to BATSIGinit, and to change the call to GDKfatal in
BATSIGinit to a call to GDKexit (i.e., skipping the printing
of the "!FATAL" message).
By the way, the default signal of the "kill" command is SIGTERM.
## Comment 2524
Date: 2005-03-16 23:40:28 +0100
From: @njnes
Logged In: YES
user_id=43556
okay CTRL-C and SIGTERM are now catched by BATSIGinterupt which
simply exits
## Comment 2525
Date: 2005-10-06 14:49:07 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: CAN NOT GET THE STUPID SCRIPT TO WORK /
TAKES FAR TOO LONG / I AM TOO STUPID TO SOLVE THE BUG
BugDay_2005-10-06: REASSIGN to someone who can script Mtest
Bleh. Attaching my attempt to do it.
## Comment 2526
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1123914 at http://sourceforge.net/support/tracker.php?aid=1123914
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.
| Mserver terminates "the hard way" on SIGTERM | https://api.github.com/repos/MonetDB/MonetDB/issues/468/comments | 0 | 2020-11-30T08:36:16Z | 2024-06-27T11:00:55Z | https://github.com/MonetDB/MonetDB/issues/468 | 753,273,601 | 468 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-15 02:38:56 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-03-23 04:05:20 +0100
## Comment 2511
Date: 2005-02-15 14:38:56 +0100
From: Wouter Alink <<alink>>
max(),min() and sum() include incorrectly nil-values in
their result
CREATE TABLE test (aap int);
INSERT INTO test VALUES (1),(2),(nil);
SELECT MIN(aap),MAX(aap),SUM(aap) FROM test
results in: null, null, null
instead of the expected: 1,2,3
The error lies in the underlying mil-functions (e.g.
.max())
## Comment 2512
Date: 2005-02-15 22:20:59 +0100
From: @yzchang
Logged In: YES
user_id=341633
I have read it in a database book that the NULL value is
treated according to the following two rules:
1. when we operate on a NULL and nay value, including
another NULL, using an arithmetic operator like x or +, the
result is NULL.
2. when we compare a NULL value and any value, including
another NULL, using a comparison operator like = or >, the
result is UNKNOWN.
Maybe MonetDB implements these two rules, which explains the
results you got.
## Comment 2513
Date: 2005-03-17 00:10:02 +0100
From: @drstmane
Logged In: YES
user_id=572415
There are two options:
Either, we change the default behaviour to ignore NILs,
or we add second "_no_nil" variants for all aggregates, na
dmake only these ignore NILs (just like with count)...
In any case, the case that all values are NIL needs to be
treated properly.
Comments? Opinions?
## Comment 2514
Date: 2005-03-21 10:55:45 +0100
From: @peterboncz
Logged In: YES
user_id=591107
"The error lies in the underlying mil-functions (e.g.
.max())"
Ha! Of course not!
In MIL, the semantics of 'nil' is 'value unknown', and all functions
respect that. In general, a MIL command thatgets at least one nil
parameter should return nil. Logical exceptions are e.g. or
(true,nil) = true and and(false,nil) = false.
This semantics for nil was mostly appreciated (and enforced) by
Data Distilleries. That does not mean it is great. I recognize that
other applications may have other wishes. Also, the current nil
implementation is quite inefficient (testing for a special value in all
inner loops is not very attractive -- it had better been kept
separate in attached nil vectors).
For SQL, this behavior is indeed incorrect. I think that for all
aggregates on NULLABLE columns, we should filter out the nils
first, or introduce some new aggregates ({max_nonnil}, etc) in
aggrX3.
## Comment 2515
Date: 2005-03-23 16:05:20 +0100
From: @njnes
Logged In: YES
user_id=43556
SQL indeed had inconsistent NULL definitions
any operation on NULL should result in NULL, ie NULL = UNKNOWN
In case of groups NULL get a different meaning 'NO VALUE',
so they are
ignored.
So in the new implementation we now filter out NULL's before
calculating
aggregations
## Comment 2516
Date: 2005-10-06 13:35:03 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/max_min_sum_null.*
fixed SQL
## Comment 2517
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1123132 at http://sourceforge.net/support/tracker.php?aid=1123132
| max(),min() and sum() include nil | https://api.github.com/repos/MonetDB/MonetDB/issues/467/comments | 0 | 2020-11-30T08:36:13Z | 2024-06-27T11:00:54Z | https://github.com/MonetDB/MonetDB/issues/467 | 753,273,566 | 467 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-15 02:04:27 +0100
From: @mlkersten
To: @njnes
Version: -- development
Last updated: 2005-02-15 10:27:09 +0100
## Comment 2503
Date: 2005-02-15 02:04:27 +0100
From: @mlkersten
Upon start of the SQL session:
sql> select count(*) from tables
[ 33 ]
sql> select count(*) from tables T, tables S
[ 729 ]
which ofcourse should have been 1089
## Comment 2504
Date: 2005-02-15 09:29:13 +0100
From: @drstmane
Logged In: YES
user_id=572415
This is actually not a problem/bug with the cross product,
but rather a "leak" / feature in the "tables" table:
temporary (result-?) tables are added during a session, but
"disappear", again, after a session
(see below). I guess, this does not block the release, does it?
sql>09:26:30 manegold@sn-1:~ $ MapiClient -lsql
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
1 id
[ 24 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
2 id
[ 27 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
3 id
[ 30 ]
sql>09:26:37 manegold@sn-1:~ $ MapiClient -lsql
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
1 id
[ 24 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
2 id
[ 27 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
3 id
[ 30 ]
sql>
## Comment 2505
Date: 2005-02-15 09:33:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
This is actually not a problem/bug with the cross product,
but rather a "leak" / feature in the "tables" table:
temporary (result-?) tables are added during a session, but
"disappear", again, after a session
(see below). I guess, this does not block the release, does it?
sql>09:26:30 manegold@sn-1:~ $ MapiClient -lsql
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
1 id
[ 24 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
2 id
[ 27 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
3 id
[ 30 ]
sql>09:26:37 manegold@sn-1:~ $ MapiClient -lsql
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
1 id
[ 24 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
2 id
[ 27 ]
sql>select count(*) from tables;
3 querytype
tables table_name
count_id name
int type
10 length
1 tuplecount
3 id
[ 30 ]
sql>
## Comment 2506
Date: 2005-02-15 09:38:02 +0100
From: @mlkersten
Logged In: YES
user_id=490798
I haven;t checked it on other tables, but....
I was sleepy tonight while filing it and forgot one essential
ingredient. It is M5/SQL code generation.
So for the time being it becomes invalid.
## Comment 2507
Date: 2005-02-15 09:38:11 +0100
From: @mlkersten
Logged In: YES
user_id=490798
I haven;t checked it on other tables, but....
I was sleepy tonight while filing it and forgot one essential
ingredient. It is M5/SQL code generation.
So for the time being it becomes invalid.
## Comment 2508
Date: 2005-02-15 09:39:04 +0100
From: @njnes
Logged In: YES
user_id=43556
Indeed the crossproduct it self works fine. As such the
release can continue.
Intermediates are stored in a ttables table. Somehow these
inserts do not mix well with the crossproduct/union's done
in this query (remeber tables is
a view). The fix seems to be to not store intermediates into
the temp table
anymore.
## Comment 2509
Date: 2005-10-06 13:27:42 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-10-06_2.9.3/Tests/cross_product.SF-1122802.*
## Comment 2510
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1122802 at http://sourceforge.net/support/tracker.php?aid=1122802
| Cross product problems | https://api.github.com/repos/MonetDB/MonetDB/issues/466/comments | 0 | 2020-11-30T08:36:10Z | 2024-06-27T11:00:53Z | https://github.com/MonetDB/MonetDB/issues/466 | 753,273,543 | 466 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-14 06:12:06 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-02-14 08:56:51 +0100
## Comment 2497
Date: 2005-02-14 18:12:06 +0100
From: Wouter Alink <<alink>>
Just compiled MonetDB with the latest cvs.
I could not start sql_server anymore... (see
description below)
Starting only mapi_server does not give any problems.
--------
[alink@volans MonetDB]$ Mserver
Monet Database Server V4.7.0
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(sql_server);
!ERROR: " var hi :=
oid(int(qgram_first.find(oid(qgram+1)))-1);"
!
^
! can't help you here, sorry.
MonetDB>module("sql_server");
!ERROR: interpret: unknown variable 'TIMEZONES'.
!ERROR: interpret_params: exist(param 1): evaluation error.
!ERROR: interpret_params: not(param 1): evaluation error.
!ERROR: yytree: TIMEZONE: GMT.
!ERROR: interpret_params: tzone_local(param 1):
evaluation error.
!ERROR: yytree: sql_server_start: GMT.
MonetDB>
## Comment 2498
Date: 2005-02-14 20:53:41 +0100
From: @drstmane
Logged In: YES
user_id=572415
I just tried it with the latest CVS versions, both the HEAD
and the release branches (MonetDB_4-6 & SQL_2-6), and it
works fine.
Could you please cvs update and try again?
Please run bootstrap before running make.
Btw, unless you really require any new features in the HEAD
version, you're highly recommended to use the release versions!
## Comment 2499
Date: 2005-02-14 20:56:51 +0100
From: @grobian
Logged In: YES
user_id=963970
confirming.
looks like your build is messed up somehow. Marking this
bug as invalid.
## Comment 2500
Date: 2005-02-14 21:00:23 +0100
From: @njnes
Logged In: YES
user_id=43556
I'm cannot reproduce this problem with the current version
## Comment 2501
Date: 2005-10-06 13:18:09 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / OUT OF DATE
## Comment 2502
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1122520 at http://sourceforge.net/support/tracker.php?aid=1122520
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.
| sql_server fails to start | https://api.github.com/repos/MonetDB/MonetDB/issues/465/comments | 0 | 2020-11-30T08:36:07Z | 2024-06-28T13:35:54Z | https://github.com/MonetDB/MonetDB/issues/465 | 753,273,520 | 465 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-14 03:36:26 +0100
From: @swingbit
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-14 05:43:48 +0100
## Comment 2493
Date: 2005-02-14 15:36:26 +0100
From: @swingbit
Take a mil script where the LAST line
a) is commented
and
b) does not end with new line
then:
$ Mserver < script.mil
produces
!WARNING: readClient: unexpected end of file;
discarding partial input
$ Mserver script.mil
hangs.
With MapiClient it hangs as well.
## Comment 2494
Date: 2005-02-14 17:43:48 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed. It appears the scanner function input() can return
EOF (i.e. -1) which wasn't catered for.
## Comment 2495
Date: 2005-10-06 13:15:50 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: ALREADY IN TESTWEB
tests/BugsViaSourgeforce/Tests/ID.1122410.*
## Comment 2496
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1122410 at http://sourceforge.net/support/tracker.php?aid=1122410
| Mserver hangs on last commented line with no NL | https://api.github.com/repos/MonetDB/MonetDB/issues/464/comments | 0 | 2020-11-30T08:36:05Z | 2024-06-27T11:00:51Z | https://github.com/MonetDB/MonetDB/issues/464 | 753,273,500 | 464 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-13 06:15:55 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-02-14 05:19:58 +0100
## Comment 2489
Date: 2005-02-13 18:15:55 +0100
From: @drstmane
18:15:09 manegold@draco:~/Monet/Stable/MonetDB $ Mx
Usage: Mx <flags> <file>.mx.
-t Produce LaTeX document (default).
-i Produce texi document .
-c Extract code.
-R <dir> Set target directory to <dir>).
-S <style> Set LaTeX documentstyle to 'style'.
-s Produce nroff -ms document.
-1 Single column (default) .
-2 Double column.
-H <n> Set hide level to 'n' (-H0
default).
-d Produce a draft document.
-x <extension> Extract <extension> labelled code.
-w Extract HTML code.
-W Extract HTML code for inclusion.
-D <id> Define macro 'id'.
-T <string> Define default hide text <string>.
-l No line and alike statements.
-n Non changed files won't be touched.
-+ Treat @c (C code) as @C (C++ code).
18:15:13 manegold@draco:~/Monet/Stable/MonetDB $ Mx -W
doc/mel.mx
Mx: invalid option -- W
Mx:Unknown flag:?.
Usage: Mx <flags> <file>.mx.
-t Produce LaTeX document (default).
-i Produce texi document .
-c Extract code.
-R <dir> Set target directory to <dir>).
-S <style> Set LaTeX documentstyle to 'style'.
-s Produce nroff -ms document.
-1 Single column (default) .
-2 Double column.
-H <n> Set hide level to 'n' (-H0
default).
-d Produce a draft document.
-x <extension> Extract <extension> labelled code.
-w Extract HTML code.
-W Extract HTML code for inclusion.
-D <id> Define macro 'id'.
-T <string> Define default hide text <string>.
-l No line and alike statements.
-n Non changed files won't be touched.
-+ Treat @c (C code) as @C (C++ code).
18:15:15 manegold@draco:~/Monet/Stable/MonetDB $
## Comment 2490
Date: 2005-02-14 17:19:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS:
removed non-existing option from usage message
TODO:
Mx seems to have option "-C", "_B", & "-g", which are not
listed in the usage message.
Since I don't know what they do, I cannot add them...
## Comment 2491
Date: 2005-10-06 13:08:50 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / OUT OF SCOPE
## Comment 2492
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1121868 at http://sourceforge.net/support/tracker.php?aid=1121868
| Mx usage message lists unknown(?) option "-W" | https://api.github.com/repos/MonetDB/MonetDB/issues/463/comments | 0 | 2020-11-30T08:36:02Z | 2024-06-27T11:00:50Z | https://github.com/MonetDB/MonetDB/issues/463 | 753,273,470 | 463 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-13 06:13:58 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2006-01-14 10:03:23 +0100
## Comment 2483
Date: 2005-02-13 18:13:58 +0100
From: @drstmane
When extracting TeX or HTML with "-t"/"-w", Mx seems to
ighnore the title as specified via the "@t" directive.
Thus, all Mx-extracted pages on the MonetDB web site
miss there title between the two horizontal lines at
the top of the page.
I vaguely recall that "some time ago" there used to be
a title, at least with TeX extraction...
## Comment 2484
Date: 2005-07-25 09:42:33 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Is this still relevant? Or can it be dropped.
## Comment 2485
Date: 2005-07-26 07:52:09 +0200
From: @drstmane
Logged In: YES
user_id=572415
This is no langer relevant wrt. the Mx-extracted pages on
the MonetDB web site, since they get the correct title now
using a fixed style sheet.
However, stand-alone extracted documentation still comes
without the title, which is at least "unexpected".
I recall, that the titles might have been removed
deliberately to facilitate the gathering of documentation
extracted from various Mx files into a single large
document, and with the new/fixed style sheets, explicite
titles in the extracted html documentation might even be
inconvenient for the web site generation.
Not being an Mx expert, I'm hence not sure, whether there is
a solution that suits the conflicting purposes:
Mx-extraction for stand-alone .tex file (and possibly .html
files) with titles vs. Mx-extraction for web site generation
(.html) and single large document (.tex / .texi)...
## Comment 2486
Date: 2005-10-06 13:07:23 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / OUT OF SCOPE
Bup, increasing priority, maybe it can be closed sometime
## Comment 2487
Date: 2006-01-14 22:03:23 +0100
From: @drstmane
Logged In: YES
user_id=572415
Since nobody else but me seemed to have ever hit this
problem, and my main concern --- building the MonetDB web
site --- is fixed / uses a work around, I close this one
with "Won't fix" to clean-up the open bug list...
## Comment 2488
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1121867 at http://sourceforge.net/support/tracker.php?aid=1121867
| Mx -t/-w ignores documen titles (@t) | https://api.github.com/repos/MonetDB/MonetDB/issues/462/comments | 0 | 2020-11-30T08:36:00Z | 2024-06-28T13:35:54Z | https://github.com/MonetDB/MonetDB/issues/462 | 753,273,444 | 462 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 11:37:07 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-14 09:54:52 +0100
## Comment 2478
Date: 2005-02-11 23:37:07 +0100
From: @drstmane
Sjoerd already know about this, but I still add it for
the bookkeeping:
After today's tcpip fixes (for Windows) test
tests/BugsViaSourgeforce/Tests/ID.1049942
now hangs (at least) on Linux.
The test belongs to BUG #1049942 "TCP/IP module RPC
crashes server when returning a bat"; see also
http://sourceforge.net/tracker/index.php?func=detail&aid=1049942&group_id=56967&atid=482468
## Comment 2479
Date: 2005-02-12 10:29:17 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
It also hangs on Windows.
## Comment 2480
Date: 2005-02-14 21:54:52 +0100
From: @drstmane
Logged In: YES
user_id=572415
This bug has become invalid, since the respective code
changes have been reverted.
## Comment 2481
Date: 2005-10-06 13:03:45 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / REDUNDANT
tests/BugsViaSourgeforce/Tests/ID.1049942.*
## Comment 2482
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1121099 at http://sourceforge.net/support/tracker.php?aid=1121099
| tests tests/BugsViaSourgeforce/Tests/ID.1049942 hangs | https://api.github.com/repos/MonetDB/MonetDB/issues/461/comments | 0 | 2020-11-30T08:35:57Z | 2024-06-28T13:35:53Z | https://github.com/MonetDB/MonetDB/issues/461 | 753,273,414 | 461 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 03:14:59 +0100
From: @grobian
To: @mlkersten
Version: -- development
Last updated: 2006-01-22 09:04:08 +0100
## Comment 2470
Date: 2005-02-11 15:14:59 +0100
From: @grobian
Currently the gold demo is not correctly installed on
'make install'. In the doc/MonetDB directory only the
load.mil and init.mil scripts are placed. One would
expect the whole gold directory to be there together
with the required directories that the load.mil script
uses.
Currently (as a work around) the MIL Gold Demo's html
tutorial says that you need a source distribution to
run the tutorial, but this is undesirable of course.
## Comment 2471
Date: 2005-02-22 10:01:44 +0100
From: @njnes
Logged In: YES
user_id=43556
We should make all these examples data files downloadable
and add
support to load/insert using data from the internet. This
means we
only install mil files, possibly format files.
## Comment 2472
Date: 2005-10-06 13:02:22 +0200
From: @grobian
Logged In: YES
user_id=963970
bump
maybe we can solve this error the next release
## Comment 2473
Date: 2005-12-21 11:35:04 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay: stmane: NO TEST / INSTALLATION
## Comment 2474
Date: 2006-01-15 16:36:02 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Added a tarfile to our assets directory and updated the
webpages. Now the user does not need the source anymore.
Double check doc and solution, then it can be closed.
## Comment 2475
Date: 2006-01-22 09:04:08 +0100
From: @mlkersten
Logged In: YES
user_id=490798
The data has been packaged and can be obtained from
both the source distribution and the website.
## Comment 2476
Date: 2006-01-22 09:25:53 +0100
From: @grobian
Logged In: YES
user_id=963970
given the description and diffs of Martin's checkins as well
as his comment on the bug here, this bug has not been fixed,
but just rejected, as the gold demo is still not installed.
The chosen solution now is ok with me.
## Comment 2477
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120791 at http://sourceforge.net/support/tracker.php?aid=1120791
| MIL Gold Demo is not properly installed | https://api.github.com/repos/MonetDB/MonetDB/issues/460/comments | 0 | 2020-11-30T08:35:54Z | 2024-06-28T13:35:52Z | https://github.com/MonetDB/MonetDB/issues/460 | 753,273,392 | 460 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 02:59:53 +0100
From: @swingbit
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-02-11 04:25:29 +0100
## Comment 2466
Date: 2005-02-11 14:59:53 +0100
From: @swingbit
I suppose this is another function that nobody ever
used (not that I did...)
rev_select() should perform a select with the reverse
of the function passed as parameter.
PROC rev_select(bat[any::1,any::2] b, str fcn, any::2
val) : bat[any::1,any::2]
{
if (fcn = "=") { return b.mil_select("=",val); }
else if (fcn = "!=") { return b.mil_select("!=",val); }
else if (fcn = ">") { return b.mil_select("<",val); }
...
}
rev_select(a, "=", v) calls rev_select(a, "=", v)
instead of rev_select(a, "!=", v)
The same holds for "!=".
## Comment 2467
Date: 2005-02-11 16:25:29 +0100
From: @drstmane
Logged In: YES
user_id=572415
"rev_select" is not "neg_select":
While "neg_select" negates the predicate,
"rev_select" only kind of "swaps" the operands or "mirrors"
the predicate.
For "=" and "!=", the current implementation is actually
correct.
However, I'm not quite sure, what the excat semantics for
"<", "<=", ">", ">=" are supposed to be:
The help message sais
"(value OP column) select rather than (column OP value)"
which I would read as
"the operands are simply swapped", i.e.,
"rev_select(b,OP,v)" does not do a "b OP v", but rather a "v
OP b",
which is equal to a "mil_select(b,PO,v)" with the following
OP->PO translation:
OP PO
< >=
<= >
> <=
>= <
However, the current implementation uses the following
OP->PO translation:
OP PO
< >
<= >=
> <
>= <=
This, I'd interpret as
"the predicate is mirror", i.e.,
"(value PO column) select rather than (column OP value)"
Peter,
could you please explain what you originally intended with
"rev_select", and
either show, why my interpretations is wrong,
or fix the semantic mismatch between help message and
implementation?
Thanks in advance!
Stefan
## Comment 2468
Date: 2005-10-06 13:01:22 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / INVALID
## Comment 2469
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120779 at http://sourceforge.net/support/tracker.php?aid=1120779
| wrong call in rev_select() | https://api.github.com/repos/MonetDB/MonetDB/issues/459/comments | 0 | 2020-11-30T08:35:51Z | 2024-06-28T13:35:52Z | https://github.com/MonetDB/MonetDB/issues/459 | 753,273,361 | 459 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 02:24:25 +0100
From: @yzchang
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-02-14 02:00:12 +0100
## Comment 2460
Date: 2005-02-11 14:24:25 +0100
From: @yzchang
Several pages under the title "Documentation" are
colored (partly) blue, and it becomes green on
"mouseOver". Is this the intention or is it a bug?
Regards,
Jennie
## Comment 2461
Date: 2005-02-11 14:28:15 +0100
From: @grobian
Logged In: YES
user_id=963970
a bug
## Comment 2462
Date: 2005-02-14 14:00:12 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed by closing <a name=...> tags properly in .mx files
## Comment 2463
Date: 2005-02-14 16:43:02 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed by closing <a name=...> tags properly in .mx files
## Comment 2464
Date: 2005-10-06 12:55:54 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2465
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120765 at http://sourceforge.net/support/tracker.php?aid=1120765
| annoying colored pages | https://api.github.com/repos/MonetDB/MonetDB/issues/458/comments | 0 | 2020-11-30T08:35:48Z | 2024-06-27T11:00:44Z | https://github.com/MonetDB/MonetDB/issues/458 | 753,273,329 | 458 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 02:20:14 +0100
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-11 04:22:14 +0100
## Comment 2456
Date: 2005-02-11 14:20:14 +0100
From: @grobian
Options are:
-b t/f | --blocked=true/false /* blocked mode */
-d | --debug=[level] /* set debug level */
-D | --dump /* dump sql database */
% MapiClient -lsql -d 1
1: cannot open
the --dedug=value variant however does work
% MapiClient -lsql --debug=1
sql>\q
I have no idea what it does or where to find which
value to use to actually debug something.
(from a user perspective)
## Comment 2457
Date: 2005-02-11 16:22:14 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
The --debug/-d flag didn't do anything, so I removed it.
## Comment 2458
Date: 2005-10-06 12:55:12 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / OBSOLETE
## Comment 2459
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120764 at http://sourceforge.net/support/tracker.php?aid=1120764
| MapiClient debug flag does not function | https://api.github.com/repos/MonetDB/MonetDB/issues/457/comments | 0 | 2020-11-30T08:35:45Z | 2024-06-27T11:00:43Z | https://github.com/MonetDB/MonetDB/issues/457 | 753,273,295 | 457 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 02:04:12 +0100
From: @yzchang
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-02-11 02:13:22 +0100
## Comment 2453
Date: 2005-02-11 14:04:12 +0100
From: @yzchang
A wrong path of the file "MonetDB_JDBC.jar" is used in
the SQL documentation. Fabian has already corrected the
first occurrence of the wrong path, but the wrong path
is used several times in the same page.
Regards,
Jennie
## Comment 2454
Date: 2005-10-06 12:48:36 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2455
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120752 at http://sourceforge.net/support/tracker.php?aid=1120752
| Wrong path in SQL documentation | https://api.github.com/repos/MonetDB/MonetDB/issues/456/comments | 0 | 2020-11-30T08:35:42Z | 2024-06-27T11:00:42Z | https://github.com/MonetDB/MonetDB/issues/456 | 753,273,267 | 456 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 02:01:14 +0100
From: @swingbit
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-02-11 02:58:08 +0100
## Comment 2448
Date: 2005-02-11 14:01:14 +0100
From: @swingbit
module(radix);
normal(0).print();
makes Mserver crash.
## Comment 2449
Date: 2005-02-11 14:22:45 +0100
From: @swingbit
Logged In: YES
user_id=883451
For completeness on radix module:
uniform(0) correctly returns an empty bat,
but
uniform(-1) makes Mserver crash
## Comment 2450
Date: 2005-02-11 14:58:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS:
normal(0) and uniform(0) both now correctly return and empty BAT
with negative size parameter, both issue an error.
## Comment 2451
Date: 2005-10-06 12:38:20 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/radix_normal_segfault.SF-1120748.*
## Comment 2452
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120748 at http://sourceforge.net/support/tracker.php?aid=1120748
| Segmentation fault on normal(0); | https://api.github.com/repos/MonetDB/MonetDB/issues/455/comments | 0 | 2020-11-30T08:35:39Z | 2024-06-27T11:00:41Z | https://github.com/MonetDB/MonetDB/issues/455 | 753,273,239 | 455 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 01:48:43 +0100
From: @yzchang
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-02-11 03:11:03 +0100
## Comment 2444
Date: 2005-02-11 13:48:43 +0100
From: @yzchang
(I'm not very sure if this is a bug, but) I have
several problems when I'm trying to load and use the
gold demo database as explained in the MonetDB site
"Got going > Totorials > Gold mining".
1. Executing the command "Mserver --dbname=gold <
load.mil" results in an error, because the file
load.mil is not located in the current working
directory. (Have I forgot any configuration before I
can execute this command?)
Since it is in the distribution directory scripts/gold,
I decided to use the absolute path of the file, and
then I got the next problem.
2. Executing the command " Mserver --dbname=gold <
(...)/scripts/gold/load.mil" results in some strenge
outputs:
$Mserver --dbname=gold (...)/scripts/gold/load.mil
Monet Database Server V4.7.0
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>MonetDB>MonetDB>MonetDB>MonetDB>!ERROR:
ASCIIload: could not open file pwd/format
!ERROR: ascii_io_load: operation failed.
!ERROR: interpret_params: load_data(param 1):
evaluation error.
!ERROR: interpret_params: make_persistent(param 1):
evaluation error.
MonetDB>MonetDB>MonetDB>MonetDB
Regards,
Jennie
## Comment 2445
Date: 2005-02-11 15:11:02 +0100
From: @grobian
Logged In: YES
user_id=963970
Updated the documentation to be more specific about where
`pwd` is supposed to be.
## Comment 2446
Date: 2005-10-06 11:47:20 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / DOCUMENTATION
## Comment 2447
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120741 at http://sourceforge.net/support/tracker.php?aid=1120741
| Problem with trying out the gold demo database | https://api.github.com/repos/MonetDB/MonetDB/issues/454/comments | 0 | 2020-11-30T08:35:36Z | 2024-06-27T11:00:40Z | https://github.com/MonetDB/MonetDB/issues/454 | 753,273,204 | 454 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-10 10:57:12 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-11 09:11:56 +0100
## Comment 2439
Date: 2005-02-10 22:57:12 +0100
From: @drstmane
Sorry, Sjoerd, yet another one:
MonetDB>print(1+1);
[ 2 ]
MonetDB>print(1-1);
[ 0 ]
MonetDB>print((1)+(1));
[ 2 ]
MonetDB>print((1)-(1));
[ 0 ]
MonetDB>print((1)+ 1);
[ 2 ]
MonetDB>print((1)- 1);
[ 0 ]
MonetDB>print((1) +1);
[ 2 ]
MonetDB>print((1) -1);
!ERROR: "print((1) -1);"
! ^
! can't help you here, sorry.
MonetDB>print((1)+1);
[ 2 ]
MonetDB>print((1)-1);
!ERROR: "print((1)-1);"
! ^
! can't help you here, sorry.
MonetDB>quit();
## Comment 2440
Date: 2005-02-10 23:01:00 +0100
From: @drstmane
Logged In: YES
user_id=572415
... only after a ')', that is ...
## Comment 2441
Date: 2005-02-11 09:11:56 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed:
MonetDB>print(1+1);
[ 2 ]
MonetDB>print(1-1);
[ 0 ]
MonetDB>print((1)+(1));
[ 2 ]
MonetDB>print((1)-(1));
[ 0 ]
MonetDB>print((1)+ 1);
[ 2 ]
MonetDB>print((1)- 1);
[ 0 ]
MonetDB>print((1) +1);
[ 2 ]
MonetDB>print((1) -1);
[ 0 ]
MonetDB>print((1)+1);
[ 2 ]
MonetDB>print((1)-1);
[ 0 ]
MonetDB>quit();
## Comment 2442
Date: 2005-10-06 11:43:57 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/operator_-_vs_sign_-.SF-1120397.*
## Comment 2443
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120397 at http://sourceforge.net/support/tracker.php?aid=1120397
| MIL parser fails to distiunguish operator "-" from sign "-" | https://api.github.com/repos/MonetDB/MonetDB/issues/453/comments | 0 | 2020-11-30T08:35:33Z | 2024-06-27T11:00:39Z | https://github.com/MonetDB/MonetDB/issues/453 | 753,273,159 | 453 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-10 08:57:06 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-10 09:29:59 +0100
## Comment 2435
Date: 2005-02-10 20:57:06 +0100
From: @drstmane
With the latest changes of the parser, we get:
MonetDB>print('c');
[ nil ]
MonetDB>print("s");
[ "s" ]
MonetDB>print(type("s"));
[ 11 ]
MonetDB>print(type('c'));
[ 11 ]
MonetDB>
i.e., character literals are obviously parsed as
str(nil) ...
Sorry, Sjoerd, still some work to do ... ;-)
## Comment 2436
Date: 2005-02-10 21:29:59 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed:
MonetDB>print('c');
[ c ]
MonetDB>print("s");
[ "s" ]
MonetDB>print(type("s"));
[ 11 ]
MonetDB>print(type('c'));
[ 2 ]
## Comment 2437
Date: 2005-10-06 11:11:46 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: TEST ADDED / SUCCESS
tests/BugDay_2005-10-06_4.9.3/Tests/character_string_constants.SF-1120297.*
## Comment 2438
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120297 at http://sourceforge.net/support/tracker.php?aid=1120297
| MonetDB parses character constats as str(nil) | https://api.github.com/repos/MonetDB/MonetDB/issues/452/comments | 0 | 2020-11-30T08:35:29Z | 2024-06-27T11:00:38Z | https://github.com/MonetDB/MonetDB/issues/452 | 753,273,121 | 452 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-10 03:33:26 +0100
From: @njnes
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-10 04:55:08 +0100
## Comment 2430
Date: 2005-02-10 15:33:26 +0100
From: @njnes
When running a mil script with large blobs, ie strings
the Mserver
exits with
input buffer overflow, can't enlarge buffer because
scanner uses REJECT
## Comment 2431
Date: 2005-02-10 16:55:08 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed by doing string matches ourselves (instead of by flex)
and by removing all REJECTs.
## Comment 2432
Date: 2005-10-06 10:54:59 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: RELEASED / REASSIGN need knowledge on
what kind/amount of data this was dealing with, ie. what was
problematic. Suggesting Niels or Sjoerd.
## Comment 2433
Date: 2005-10-06 16:30:15 +0200
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-10-06: Claimed by niels
BugDay_2005-10-06: TEST / SUCCESS
Test added as large_string.SF-1120061
## Comment 2434
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120061 at http://sourceforge.net/support/tracker.php?aid=1120061
| parser doesn't like large strings | https://api.github.com/repos/MonetDB/MonetDB/issues/451/comments | 0 | 2020-11-30T08:35:26Z | 2024-06-27T11:00:37Z | https://github.com/MonetDB/MonetDB/issues/451 | 753,273,094 | 451 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-09 11:14:46 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-02-09 12:34:19 +0100
## Comment 2424
Date: 2005-02-09 11:14:46 +0100
From: Wouter Alink <<alink>>
ORDER BY is not accepted in combination with a UNION
----
sql>SELECT id AS test FROM aap UNION SELECT id AS test
FROM aap ORDER BY test;
MAPI = monetdb@localhost:45123
QUERY = SELECT id AS test FROM aap UNION SELECT id AS
test FROM aap ORDER BY test;
ERROR = !ERROR syntax error at token (344) in
statement: select id as test from aap union select id
as test from aap order
## Comment 2425
Date: 2005-02-09 11:52:01 +0100
From: @grobian
Logged In: YES
user_id=963970
you need some extra parenthesis
as in:
select * from (select * from test union select * from test)
as a order by id;
SELECT * FROM (SELECT id AS test FROM aap UNION SELECT id AS
test FROM aap) AS myaap ORDER BY test;
## Comment 2426
Date: 2005-02-09 12:34:19 +0100
From: @grobian
Logged In: YES
user_id=963970
Although the obscure SQL standard seems to assume this is
correct,
it is hard to find out in the parser to which part the order
by applies. It seems not worth fixing it, as the statement
as such is not clear by itself.
Use an outer select and order on that one.
## Comment 2427
Date: 2005-10-06 10:51:39 +0200
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-10-06: CLAIMED BY meltdown
BugDay_2005-10-06: NO TEST / INVALID
BugDay_2005-10-06: NO TEST / REDUNDANT
sql/src/test/bugs/subselect_union-bug-sf-941826.sql
## Comment 2428
Date: 2005-11-09 16:00:56 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, ALREADY IN TESTWEB
sql/src/test/bugs/subselect_union-bug-sf-941826.sql
## Comment 2429
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1119199 at http://sourceforge.net/support/tracker.php?aid=1119199
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.
| ORDER BY is not accepted in combination with a UNION | https://api.github.com/repos/MonetDB/MonetDB/issues/450/comments | 0 | 2020-11-30T08:35:22Z | 2024-06-28T13:35:51Z | https://github.com/MonetDB/MonetDB/issues/450 | 753,273,067 | 450 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-09 10:47:26 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2010-05-04 11:18:15 +0200
## Comment 2420
Date: 2005-02-09 10:47:26 +0100
From: Wouter Alink <<alink>>
a UNION does not give the expected result, it discards
the duplicates.
see examples below.
------
sql>SELECT id FROM aap;
3 querytype
aap table_name
id name
int type
10 length
4 tuplecount
9 id
[ 1 ]
[ 10 ]
[ 9 ]
[ 6 ]
sql>SELECT * FROM aap UNION SELECT * FROM aap;
3 querytype
table_name
id name
int type
10 length
4 tuplecount
6 id
[ 1 ]
[ 10 ]
[ 9 ]
[ 6 ]
sql>SELECT 'a' AS test FROM aap;
3 querytype
table_name
test name
char type
1 length
4 tuplecount
7 id
[ "a" ]
[ "a" ]
[ "a" ]
[ "a" ]
sql>SELECT 'a' AS test FROM aap UNION SELECT 'b' AS
test FROM aap;
3 querytype
table_name
test name
char type
1 length
2 tuplecount
8 id
[ "a" ]
[ "b" ]
## Comment 2421
Date: 2005-02-09 10:51:45 +0100
From: @grobian
Logged In: YES
user_id=963970
this is corrrect. If you want duplicates, use UNION ALL
## Comment 2422
Date: 2005-11-09 15:57:15 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, TEST ADDED/SUCCES
sql/src/test/union-duplicates.SF-1119186.sql
## Comment 2423
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1119186 at http://sourceforge.net/support/tracker.php?aid=1119186
Bug has invalid status, setting status to "NEW".
Previous status was "DELETED".
| UNION | https://api.github.com/repos/MonetDB/MonetDB/issues/449/comments | 0 | 2020-11-30T08:35:20Z | 2024-06-28T13:35:50Z | https://github.com/MonetDB/MonetDB/issues/449 | 753,273,039 | 449 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-09 10:29:59 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-02-09 12:46:52 +0100
## Comment 2417
Date: 2005-02-09 10:29:59 +0100
From: Wouter Alink <<alink>>
Using two different clients, updates are lost when a
transactions are running in parallel.
Expected behaviour would be that the transaction on
terminal 2 would not be allowed to commit.
--------
Terminal 1 (MapiClient):
sql>CREATE TABLE aap (id int);
sql>INSERT INTO aap VALUES (1);
sql>INSERT INTO aap VALUES (10);
sql>INSERT INTO aap VALUES (8);
sql>INSERT INTO aap VALUES (6);
sql>COMMIT;
Step 1:
sql>SELECT * FROM aap;
3 querytype
aap table_name
id name
int type
10 length
4 tuplecount
5 id
[ 1 ]
[ 10 ]
[ 8 ]
[ 6 ]
Step 3:
sql>UPDATE aap SET id=2 WHERE id=1;
4 querytype
[ 1 ]
sql>COMMIT;
7 querytype
Step 5:
sql>SELECT * FROM aap;
3 querytype
aap table_name
id name
int type
10 length
4 tuplecount
7 id
[ 1 ]
[ 10 ]
[ 7 ]
[ 6 ]
sql>
------------------
Terminal 2 (Mapiclient):
Step 2:
sql>SELECT * FROM aap;
3 querytype
aap table_name
id name
int type
10 length
4 tuplecount
1 id
[ 1 ]
[ 10 ]
[ 8 ]
[ 6 ]
Step 4:
sql>UPDATE aap SET id=7 WHERE id=8;
4 querytype
[ 1 ]
sql>COMMIT;
7 querytype
## Comment 2418
Date: 2005-11-09 15:03:04 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, TEST ADDED/ SUCCESS(?)
src/sql/updates_got_lost.SF-1119178.sql
[needs parallel processing]
## Comment 2419
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1119178 at http://sourceforge.net/support/tracker.php?aid=1119178
| Updates are lost | https://api.github.com/repos/MonetDB/MonetDB/issues/448/comments | 0 | 2020-11-30T08:35:14Z | 2024-06-27T11:00:34Z | https://github.com/MonetDB/MonetDB/issues/448 | 753,272,988 | 448 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-08 05:52:45 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-02-08 11:43:27 +0100
## Comment 2414
Date: 2005-02-08 17:52:45 +0100
From: Wouter Alink <<alink>>
After deleting all files/folders in the 'dbfarm' (and
forgetting to delete all folders in the 'log'),
starting sql_server_start() will result in a segfault.
For detailed description see below.
-------
rm -r dbfarm/*
Mserver
...
MonetDB>module(sql_server);
MonetDB>sql_server_start();
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1208082304 (LWP 22307)]
0x001bb50d in load_trans (bm=0x1da0fa, stk=0) at
bat_store.mx:912
912 bat_store.mx: No such file or directory.
in bat_store.mx
Current language: auto; currently c
(gdb) bt
0 0x001bb50d in load_trans (bm=0x1da0fa, stk=0) at
bat_store.mx:912
1 0x001bce92 in store_init (debug=139525600,
logdir=0x831fd08 "/ufs/alink/local/var/MonetDB/log",
dbname=0x831fc24 "demo", stk=513) at bat_store.mx:1481
2 0x001b470b in mvc_init (debug=0, stk=513) at
sql_mvc.mx:187
3 0x0018d7b6 in mvc_init_wrap (stk=513, lt=0x0,
res=0x850fde0)
at sql_server.mx:274
4 0x00412d9c in interpret (stk=513, lt=0x84a8238,
res=0xbfe611f0)
at monet_interpreter.mx:814
5 0x0013298c in CMDcatch (stk=513, lt=0x84fb018,
res=0xbfe611f0)
at kernel.mx:1087
6 0x00412d9c in interpret (stk=513, lt=0x84fb018,
res=0xbfe611f0)
at monet_interpreter.mx:814
7 0x00418f9a in interpret_assignment (stk=513,
lt=0x84fb150, res=0xbfe611f0)
at monet_interpreter.mx:1689
8 0x00414bf9 in interpret_var (stk=513, lt=0x84fb0e8,
res=0xbfe611f0)
at monet_interpreter.mx:1270
9 0x00412ce5 in interpret (stk=513, lt=0x84fb0e8,
res=0xbfe611f0)
at monet_interpreter.mx:786
10 0x0041928c in interpret_seqblock (stk=513,
lt=0x839bfc0, res=0xbfe611f0,
scope=1) at monet_interpreter.mx:1737
11 0x004127dc in interpret (stk=0, lt=0x839bfc0,
res=0xbfe611f0)
at monet_interpreter.mx:724
-
## Comment 2415
Date: 2005-11-09 14:46:42 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, NO TEST/ DOCUMENTATION
One should not play around with the database store directly
## Comment 2416
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1118746 at http://sourceforge.net/support/tracker.php?aid=1118746
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.
| SEQFAULT deleting dbfarm without deleting log | https://api.github.com/repos/MonetDB/MonetDB/issues/447/comments | 0 | 2020-11-30T08:35:11Z | 2024-06-27T11:00:33Z | https://github.com/MonetDB/MonetDB/issues/447 | 753,272,947 | 447 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-04 11:00:45 +0100
From: Wouter Alink <<alink>>
To: @mlkersten
Version: -- development
Last updated: 2005-02-09 06:35:55 +0100
## Comment 2407
Date: 2005-02-04 11:00:45 +0100
From: Wouter Alink <<alink>>
used: Mknife 1.6.0rc0 (that came with the latest
Windows Binaries)
Mknife fails to connect (when pressing help: "Could not
locate Netscape/Mozilla") when Mozilla is not
installed. The error-message is not obvious and is not
documented either... (and I had FireFox installed).
note1: after having installed and later uninstalled
Mozilla 1.7.5 (and both trying Mknife rc0 and rc2), I
could not reproduce the error anymore. I guess there
might be some registry values not deleted during
uninstalling Mozilla.
note2: After updating to 1.6.0rc2 Mknife incorrectly
states that there is a new update available. (the whole
help-section actually is a mess, because for 'Manual'
and 'About MonetDB' it shows 'Access to
Mozilla/Netscape failed' (although installed and running)
note3: there is no bug-group for Mknife
## Comment 2408
Date: 2005-02-04 11:24:32 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Martin is responsible for Mknife.
## Comment 2409
Date: 2005-02-09 00:09:07 +0100
From: @mlkersten
Logged In: YES
user_id=490798
/*
* The help system is currently limited to switching to
* an external browser for access in the MonetDB Home page
* The problem is, ofcourse, which browser to start.
* this depends strongly on the host environment.
* We assume Mozilla availability for the time being
*/
It calls java.util.ViewInBrowser(<url>). We haven;t found an
easy way to resolve this issue.
## Comment 2410
Date: 2005-02-09 08:17:50 +0100
From: @grobian
Logged In: YES
user_id=963970
java.util is a package (so it has no members) and there
exists no ViewInBrowser class (regardless of the constructor
type) according to the web documentation.
## Comment 2411
Date: 2005-02-09 18:35:55 +0100
From: @mlkersten
Logged In: YES
user_id=490798
It now properly finds a browser, e.g. IE
Further improvements for documentation and functionality
are delayed to a next release.
## Comment 2412
Date: 2005-11-09 16:06:14 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, NO TEST/DOCUMENTATION
Mknife is not actively maintained.
## Comment 2413
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1116065 at http://sourceforge.net/support/tracker.php?aid=1116065
| Windows XP: Mknife 1.6.0rc2 | https://api.github.com/repos/MonetDB/MonetDB/issues/446/comments | 0 | 2020-11-30T08:35:07Z | 2024-06-28T13:35:49Z | https://github.com/MonetDB/MonetDB/issues/446 | 753,272,910 | 446 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-03 05:06:35 +0100
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-04 11:22:28 +0100
## Comment 2403
Date: 2005-02-03 17:06:35 +0100
From: Wouter Alink <<alink>>
When starting Mserver on Windows XP XP2:
!Error: moduleClient: module(pcre) load error
but apparently this is not critical because a
MonetDB>-prompt appears afterwards.
## Comment 2404
Date: 2005-02-04 11:22:28 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
This bug has already been fixed in CVS.
As yet, no new installer with the fix has been produced.
## Comment 2405
Date: 2005-11-09 16:03:51 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09,Martin Kersten, NO TEST/ COMPILATION
## Comment 2406
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1115520 at http://sourceforge.net/support/tracker.php?aid=1115520
| Windows: error starting Mserver | https://api.github.com/repos/MonetDB/MonetDB/issues/445/comments | 0 | 2020-11-30T08:35:04Z | 2024-06-27T11:00:31Z | https://github.com/MonetDB/MonetDB/issues/445 | 753,272,886 | 445 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-03 05:04:34 +0100
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-04 11:21:56 +0100
## Comment 2399
Date: 2005-02-03 17:04:34 +0100
From: Wouter Alink <<alink>>
When installing MonetDB on Windows XP SP2:
Installation went fine (final message was: installation
was succesful)
But when starting MonetDB SQL Server:
Mserver.exe - Unable To Locate Component
This application has failed to start because
lib_algebra.dll was not found. Re-installing the
application may fix the problem.
Adding \Program Files\CWI\lib\MonetDB to the path (in
Mserver.bat) will fix the problem
Other (minor) bug: Mserver version-information is not
available when choosing 'properties' after a
right-click on the .msi-file.
## Comment 2400
Date: 2005-02-04 11:21:56 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
This bug has already been fixed in CVS.
As yet, no new installer with the fix has been produced.
## Comment 2401
Date: 2005-11-09 16:02:50 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09,Martin Kersten, NO TEST ADDED/ COMPILATION
## Comment 2402
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1115518 at http://sourceforge.net/support/tracker.php?aid=1115518
| Windows XP installation: lib_algebra.dll not found | https://api.github.com/repos/MonetDB/MonetDB/issues/444/comments | 0 | 2020-11-30T08:35:01Z | 2024-06-27T11:00:30Z | https://github.com/MonetDB/MonetDB/issues/444 | 753,272,853 | 444 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-02 10:57:46 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-11-09 11:47:10 +0100
## Comment 2395
Date: 2005-02-02 10:57:46 +0100
From: Wouter Alink <<alink>>
When granting a privilege on an object to PUBLIC,
regardless of the object, or privilege the Monet Server
gives a segfault and stops.
Example:
CREATE TABLE test (id INT);
COMMIT;
GRANT SELECT ON test TO PUBLIC;
## Comment 2396
Date: 2005-11-09 14:42:25 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay 2005-11-09, Martin Kersten, TEST ADDED/ FAILURE
src/tests/grant_public.SF-1114580.sql
## Comment 2397
Date: 2005-11-09 23:47:10 +0100
From: @njnes
Logged In: YES
user_id=43556
fixed bug in revoke
## Comment 2398
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1114580 at http://sourceforge.net/support/tracker.php?aid=1114580
| Segfault: GRANT ... TO PUBLIC | https://api.github.com/repos/MonetDB/MonetDB/issues/443/comments | 0 | 2020-11-30T08:34:57Z | 2024-06-27T11:00:29Z | https://github.com/MonetDB/MonetDB/issues/443 | 753,272,826 | 443 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-02 10:30:28 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-02-02 01:35:29 +0100
## Comment 2391
Date: 2005-02-02 10:30:28 +0100
From: Wouter Alink <<alink>>
using the following table:
CREATE TABLE test (id INT, name VARCHAR(10));
COMMIT;
the query:
ALTER TABLE test DROP name;
results in error:
ERROR = !Unknown table element (159686096)->token =
unknown
(where the number '159686096' is different each time
the ALTER TABLE is executed)
## Comment 2392
Date: 2005-02-02 13:35:29 +0100
From: @njnes
Logged In: YES
user_id=43556
added missing implementation
## Comment 2393
Date: 2005-11-09 14:26:46 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay_2005-11-09, Martin Kersten, TEST ADDED/SUCCESS
sql/src/tests/drop_column_alter_table.SF-1114557.sql
## Comment 2394
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1114557 at http://sourceforge.net/support/tracker.php?aid=1114557
| error using DROP COLUMN in ALTER TABLE | https://api.github.com/repos/MonetDB/MonetDB/issues/442/comments | 0 | 2020-11-30T08:34:54Z | 2024-06-27T11:00:28Z | https://github.com/MonetDB/MonetDB/issues/442 | 753,272,794 | 442 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-02 09:58:23 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-02-02 03:51:46 +0100
## Comment 2381
Date: 2005-02-02 09:58:23 +0100
From: Wouter Alink <<alink>>
using the following example table:
CREATE TABLE test (id INT);
INSERT INTO test VALUES (10);
INSERT INTO test VALUES (11);
INSERT INTO test VALUES (12);
COMMIT;
the query:
SELECT * FROM test WHERE (id=10) OR (id=11)
should (to my best knowledge) return:
[10]
[11]
but does not (an empty result is returned).
## Comment 2382
Date: 2005-02-02 10:11:33 +0100
From: @grobian
Logged In: YES
user_id=963970
Uhm.... works for me:
Welcome to the MonetDB interactive JDBC terminal!
Database: MonetDB 4.6.0_rc1
Driver: MonetDB Native Driver 1.1 (Rise_rc2)
Type \q to quit, \h for a list of available commands
auto commit mode: on
monetdb-> CREATE TABLE test (id INT);
INSERT INTO test VALUES (10);
INSERT INTO test VALUES (11);
INSERT INTO test VALUES (12);Operation successful
monetdb-> 1 affected row
monetdb-> 1 affected row
monetdb->
1 affected row
monetdb-> SELECT * FROM test WHERE (id=10) OR (id=11) ;
+------------+
| id |
+============+
| 10 |
| 11 |
+------------+
2 rows
## Comment 2383
Date: 2005-02-02 10:15:00 +0100
From: @njnes
Logged In: YES
user_id=43556
same with mapiclient
## Comment 2384
Date: 2005-02-02 11:00:35 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
But now try:
SELECT * FROM test WHERE (id=11) OR (id=12);
I get an empty result for this one where I got two results
for the one with (id=10) OR (id=11) .
## Comment 2385
Date: 2005-02-02 11:15:41 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Sorry for the wrong example...
but there definitely is a problem, I quess there is
something wrong with the ordering of the clauses... look at
the example below:
sql>SELECT id FROM test WHERE id<11 OR id>11;
gives an empty result
sql>SELECT id FROM test WHERE id>11 OR id<11;
gives:
[ 12 ]
[ 10 ]
## Comment 2386
Date: 2005-02-02 11:31:42 +0100
From: @njnes
Logged In: YES
user_id=43556
something seems broken in the kunion mil function
## Comment 2387
Date: 2005-02-02 11:51:28 +0100
From: @njnes
Logged In: YES
user_id=43556
correction the bug is in fetchjoin
## Comment 2388
Date: 2005-02-02 15:51:46 +0100
From: @njnes
Logged In: YES
user_id=43556
It seems were back with kunion. There the properties weren't
set correctly,
ie hdense will the head was dense.
## Comment 2389
Date: 2005-11-09 14:27:57 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay_2005-11-09, Martin Kersten, TEST ADDED/SUCCESS
## Comment 2390
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1114540 at http://sourceforge.net/support/tracker.php?aid=1114540
| incorrect results when using OR in SELECT statement | https://api.github.com/repos/MonetDB/MonetDB/issues/441/comments | 0 | 2020-11-30T08:34:51Z | 2024-06-27T11:00:27Z | https://github.com/MonetDB/MonetDB/issues/441 | 753,272,758 | 441 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-01 03:17:49 +0100
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-01 04:33:09 +0100
## Comment 2376
Date: 2005-02-01 15:17:49 +0100
From: Wouter Alink <<alink>>
This is more of a request kind of thingy...
there are no 'groups' or 'categories' in the
bug-tracker to submit documentation bugs to.
## Comment 2377
Date: 2005-02-01 16:33:09 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
There was already a group "Website", and I have now added a
category "Documentation".
## Comment 2378
Date: 2005-11-09 13:44:52 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: DOCUMENTATION
## Comment 2379
Date: 2005-11-09 13:47:21 +0100
From: @mlkersten
Logged In: YES
user_id=490798
BugDay_2005-11-09, Martin Kersten,NO TEST/DOCUMENTATION
## Comment 2380
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1114000 at http://sourceforge.net/support/tracker.php?aid=1114000
| documentation bugs cannot be submitted | https://api.github.com/repos/MonetDB/MonetDB/issues/440/comments | 0 | 2020-11-30T08:34:48Z | 2024-06-27T11:00:26Z | https://github.com/MonetDB/MonetDB/issues/440 | 753,272,726 | 440 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-01 03:14:14 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-02-01 05:52:42 +0100
## Comment 2371
Date: 2005-02-01 15:14:14 +0100
From: Wouter Alink <<alink>>
While installing the SQL front-end for MonetDB, it was
nowhere mentioned that I had to give the
sql_server_start(); command.
On the 'Documentation' -> 'Front-end' -> 'SQL'
->'getting started' section of the website a misleading
example is presented.
## Comment 2372
Date: 2005-02-01 17:45:42 +0100
From: @grobian
Logged In: YES
user_id=963970
yup, error/bug. should use the mil script to do it.
will check that in, needs to roll forward afterwards though.
thanks!
## Comment 2373
Date: 2005-02-01 17:52:42 +0100
From: @grobian
Logged In: YES
user_id=963970
changing state to fixed, but not closed as it is not on live
on the website yet, but fixed in the html source.
## Comment 2374
Date: 2005-11-09 14:21:37 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: NO TEST / DOCUMENTATION
## Comment 2375
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1113994 at http://sourceforge.net/support/tracker.php?aid=1113994
| sql_server_start(); not documented | https://api.github.com/repos/MonetDB/MonetDB/issues/439/comments | 0 | 2020-11-30T08:34:45Z | 2024-06-27T11:00:25Z | https://github.com/MonetDB/MonetDB/issues/439 | 753,272,672 | 439 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-25 09:14:25 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-01-27 04:39:28 +0100
## Comment 2368
Date: 2005-01-25 21:14:25 +0100
From: @grobian
in (pictor:src/test/pg_regress) fabian% sh runtest.sh
sql99/varchar.sql an insert is made into a table with a
varchar(4). Inserting a string with 6 characters seems
to work fine.
monetdb-> CREATE TABLE "sys"."varchar_tbl" (
monetdb=( "f1" varchar(4)
monetdb=( );
Operation successful
monetdb-> insert into varchar_tbl values ('long string');
1 affected row
monetdb-> select * from varchar_tbl;
+-------------+
| f1 |
+=============+
| long string |
+-------------+
1 row
I would expect an error message such as:
INSERT INTO VARCHAR_TBL (f1) VALUES ('long string');
ERROR: value too long for type CHARACTER VARYING(4)
## Comment 2369
Date: 2005-11-09 14:20:55 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: ALREADY IN TESTWEB
src/test/BugDay_2005-10-06_2.8/string_truncate.SF-919542.sql
## Comment 2370
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1109345 at http://sourceforge.net/support/tracker.php?aid=1109345
| SQL: varchar length not checked on insert | https://api.github.com/repos/MonetDB/MonetDB/issues/438/comments | 0 | 2020-11-30T08:34:42Z | 2024-06-27T11:00:24Z | https://github.com/MonetDB/MonetDB/issues/438 | 753,272,643 | 438 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-25 10:59:53 +0100
From: Jens Teubner <<teubner>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-01-27 01:02:14 +0100
## Comment 2363
Date: 2005-01-25 10:59:53 +0100
From: Jens Teubner <<teubner>>
I've accidentally used C style comments in a MIL script. The result is a
server crash. Shouldn't MonetDB react a bit more gracefully?
The simple example works with the MapiClient, as well as on the Mserver
console:
/* foo bar baz */
print (42);
Mserver, by the way, reports `fatal flex scanner internal error--end of
buffer missed'
## Comment 2364
Date: 2005-01-25 16:00:55 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
This is quite a specific case. You get a crash with Jens's
sequence, but not with e.g.
/* foo bar */
print(42);
The problem is as follows.
In the lexer (@l part of monet_parse.mx), the "*/" is
matched (line 162 of current stable version). The code
dealing with this checks whether this sequence makes sense
(it doesn't) and is then supposed to call REJECT so that the
lexer tries just "*" and the just "/".
The check involves a call to TBLSEARCH which in turn calls
input(). The call to input() is an actual function call.
One of the things it does is shift the characters that have
already been accepted off the beginning of the buffer (i.e.
it shifts the buffer to the left). It then reads in one
more line of input (the "print(42);\n" bit).
Subsequently, TBLSEARCH decides "*/" can't be used, so it
calls REJECT.
However, the REJECT implementation uses some local variables
that point into the buffer that was shifted by the call to
input(). Those variables are not valid anymore because of
the call to input(). The lexer can't cope with this, so it
reports a fatal error and exits.
I don't know enough about the intricacies of the lexer code,
so I can't propose a fix. I assign this to Peter in the
hope he can shed some light over a possible fix.
## Comment 2365
Date: 2005-01-27 13:02:14 +0100
From: @peterboncz
Logged In: YES
user_id=591107
fix: do not use REJECT anymore (apparently input()/unput()
usage inside lexer actions makes it impossible to use
REJECT).
REJECT was used to back off from a sequence of chars that
are permitted as operator names; if that sequence did not
match any existing operator in a loaded MIL module.
instead; the sequence of operator-chars that do not match
an operator is given back to the lexical analyzer buffer with
unput() *except* the first character, which is emitted as a
single char token.
REJECT was not needed as the char sequence could not
match any other subsequent rule anyway.
## Comment 2366
Date: 2005-11-09 14:02:51 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: ALREADY IN TESTWEB
src/tests/BugsViaSourgeforce/Tests/ID.1108971.milS
## Comment 2367
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1108971 at http://sourceforge.net/support/tracker.php?aid=1108971
| Syntax errors crash Mserver | https://api.github.com/repos/MonetDB/MonetDB/issues/437/comments | 0 | 2020-11-30T08:34:39Z | 2024-06-27T11:00:23Z | https://github.com/MonetDB/MonetDB/issues/437 | 753,272,613 | 437 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-24 07:12:27 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-09 03:39:31 +0100
## Comment 2356
Date: 2005-01-24 19:12:27 +0100
From: @drstmane
Initiated by a "real-life experience":
Currently, Mserver does not check on startup, whether
the database is binary compatible. Hence, accessing
(e.g.) a 32bit database with a 64bit Mserver (or vice
versa) might result in a server crash and/or a
corrupted database!
To avoid this, Mserver shoudl check at startup, whether
the database is compatible with the Mserver:
- 32bit vs. 64bit address space
- 32bit vs. 64bit oids (only with 64bit address space)
- little vs. big-endian (unless this is already taken
care of...)
## Comment 2357
Date: 2005-01-25 16:50:02 +0100
From: @drstmane
Logged In: YES
user_id=572415
Actually, this IS a BUG...
## Comment 2358
Date: 2005-01-25 18:25:08 +0100
From: @mlkersten
Logged In: YES
user_id=490798
The easiest solution I can think of is to include a version
file in BBP, which contains a bit-pattern that, when read
upon initializatoin, leads to a failed match.
Such an initialization action could later on be extended to
include encryption information as well.
## Comment 2359
Date: 2005-02-07 10:31:48 +0100
From: @grobian
Logged In: YES
user_id=963970
According to chapter 13.4 and 13.5 of the MonetDB Kernel
Implementation Notes v4.4 this should already been taken
care of...
## Comment 2360
Date: 2005-02-09 15:39:31 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed by adding a line with two numbers to the top of the
BBP giving the sizes of pointers (size_t) and oids respectively.
## Comment 2361
Date: 2005-11-09 14:00:02 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: NO TEST
## Comment 2362
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1108525 at http://sourceforge.net/support/tracker.php?aid=1108525
| Mserver should check whether database is binary compatible | https://api.github.com/repos/MonetDB/MonetDB/issues/436/comments | 0 | 2020-11-30T08:34:36Z | 2024-06-27T11:00:22Z | https://github.com/MonetDB/MonetDB/issues/436 | 753,272,583 | 436 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-14 11:40:57 +0100
From: @mlkersten
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-02 12:49:45 +0100
## Comment 2352
Date: 2005-01-14 23:40:57 +0100
From: @mlkersten
When you call MapiClient -lsql in a command window
different from the MonetDB distribution, it may return
the message
Could not open
C:\Document and Settings\sjoerd\My
Documents\src\MonetDB\NT\etc\MonetDB.conf
A more informative, non-local error message would be
helpful.
## Comment 2353
Date: 2005-01-18 11:28:24 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixing this is a bit involved:
The message comes from monet_options.mx if the config file
could not be opened. If no config file is specified on the
MapiClient commandline, the default is used which is
compiled in. The function that produces the error message
does not know where the file name comes from, i.e. it has no
idea whether the name was the compiled-in default or a
user-supplied file name. In the latter case, the message is
reasonable, in the former you might want a different error
message.
## Comment 2354
Date: 2005-11-09 13:44:41 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: NO TEST / INSTALLATION
## Comment 2355
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1102683 at http://sourceforge.net/support/tracker.php?aid=1102683
| Windows settings | https://api.github.com/repos/MonetDB/MonetDB/issues/435/comments | 0 | 2020-11-30T08:34:33Z | 2024-06-28T13:35:49Z | https://github.com/MonetDB/MonetDB/issues/435 | 753,272,551 | 435 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-13 06:29:29 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-01-23 09:49:51 +0100
## Comment 2347
Date: 2005-01-13 18:29:29 +0100
From: Jan Rittinger <<jan.rittinger>>
With different properties (mark vs. insert) the refine
extension behaves differently and therefore causes wrong
results in pf. (A rewritten query would be also a solution :
))
var item := bat (void,oid).seqbase(0@0).insert(nil,1@0).
insert(nil,2@0).insert(nil,3@0).insert(nil,4@0).insert(nil,
5@0);
var iter := item.project(1@0);
var kind := iter.project(97);
var sorting := iter.reverse().sort().reverse();
print(sorting);
sorting := sorting.CTrefine(kind);
print(sorting);
sorting := sorting.CTrefine(item);
print(sorting);
sorting := sorting.reverse().{min}().reverse().mark(0@0).
reverse();
print(sorting);
iter := iter.project(1@0);
item := iter.mark(1@0);
kind := iter.project(97);
var sorting := iter.reverse().sort().reverse();
print(sorting);
sorting := sorting.CTrefine(kind);
print(sorting);
sorting := sorting.CTrefine(item);
print(sorting);
sorting := sorting.reverse().{min}().reverse().mark(0@0).
reverse();
print(sorting);
## Comment 2348
Date: 2005-01-16 01:46:37 +0100
From: @drstmane
Logged In: YES
user_id=572415
The reason seems to be that in the second case, the tail of
item is not materialized (void iso. oid), and then the last
CTrefine yileds a wrong result.
See the attached log for details.
CTrefine uses GDKqsort, which requires a materialized tail;
I once made CTrefine "work" (i.e., not crash) on void-tails
by avoiding the GDKqsort in that case, but that's apparently
not enough.
IMHO, there is no other solution than either dis-allowing
CTrefine with void-tails or silently materialize them (at
least temporary) ...
Maybe / Hopefully Peter knows better ...
## Comment 2349
Date: 2005-01-23 21:49:51 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS (MonetDB_4-6 release branch; propagation to
main trunk will follow later):
Made CTrefine cope with a not materialized (void iso. oid)
tail of the second BAT:
In this case, CTrefine now locally materializes each chunk
thus that "sort_flush" and "GDKqsort" do work properly.
As usually: also added the respective test script to CVS.
## Comment 2350
Date: 2005-11-09 13:42:57 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: ALREADY IN TESTWEB
src/tests/BugsViaSourgeforce/Tests/ID.1101798.milS
## Comment 2351
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1101798 at http://sourceforge.net/support/tracker.php?aid=1101798
| refine extension unreliable | https://api.github.com/repos/MonetDB/MonetDB/issues/434/comments | 0 | 2020-11-30T08:34:30Z | 2024-06-27T11:00:20Z | https://github.com/MonetDB/MonetDB/issues/434 | 753,272,520 | 434 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-13 12:48:37 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-01-13 11:06:30 +0100
## Comment 2341
Date: 2005-01-13 12:48:37 +0100
From: @grobian
monetdb-> \d a
CREATE TABLE "sys"."a" (
"id" int NOT NULL,
"val" int,
CONSTRAINT "a_id_pkey" PRIMARY KEY ("id"),
CONSTRAINT "a_val_unique" UNIQUE ("val")
);
monetdb-> insert into a values (1, 1);
1 affected row
monetdb-> insert into a values (1, 2);
Error: ERROR: INSERT INTO: PRIMARY KEY constraint
a.a_id_pkey violated
monetdb-> insert into a values (2, 1);
Error: ERROR: INSERT INTO: UNIQUE constraint
a.a_val_unique violated
monetdb-> insert into a values (2, 2);
1 affected row
monetdb-> update a set id = 1 where val = 2;
1 affected row
monetdb-> select * from a;
+------------+------------+
| id | val |
+============+============+
| 1 | 1 |
| 1 | 2 |
+------------+------------+
2 rows
monetdb-> update a set val = 1 where val = 2;
1 affected row
monetdb-> select * from a;
+------------+------------+
| id | val |
+============+============+
| 1 | 1 |
| 1 | 1 |
+------------+------------+
2 rows
monetdb->
Seems like on updates a keycheck is not performed.
## Comment 2342
Date: 2005-01-13 12:56:01 +0100
From: @grobian
Logged In: YES
user_id=963970
It might be of interest that the NOT NULL constraint *is*
checked:
monetdb-> insert into a values (null, 2);
Error: ERROR: INSERT INTO: NOT NULL constraint violated for
column a.id
monetdb-> update a set id = null where id = 1;
Error: ERROR: UPDATE: NOT NULL constraint violated for
column a.id
## Comment 2343
Date: 2005-01-13 12:57:03 +0100
From: @grobian
Logged In: YES
user_id=963970
sorry, this should be a bug...
## Comment 2344
Date: 2005-01-13 23:06:30 +0100
From: @njnes
Logged In: YES
user_id=43556
fixed the update_ukey code was not called for single column
keys because these have no index.
## Comment 2345
Date: 2005-11-09 13:30:06 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: ALREADY IN TESTWEB
src/test/bugs/primary_key-bug-sf-975308.sql
## Comment 2346
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1101569 at http://sourceforge.net/support/tracker.php?aid=1101569
| SQL: key constraints not checked on UPDATE | https://api.github.com/repos/MonetDB/MonetDB/issues/433/comments | 0 | 2020-11-30T08:34:27Z | 2024-06-27T11:00:19Z | https://github.com/MonetDB/MonetDB/issues/433 | 753,272,494 | 433 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-11 11:52:05 +0100
From: sandy ganz <<sganz>>
To: @njnes
Version: -- development
Last updated: 2005-01-14 09:13:19 +0100
## Comment 2334
Date: 2005-01-11 23:52:05 +0100
From: sandy ganz <<sganz>>
Created a table to that accepts the BCP data from
Sybase. Works ok on smaller file sizes (~200meg) but
when trying to load a 14G file the MapiClient seems to
hang and no server activity is seen (or MapiClient) .
Also, not sure, but the sql ascii loader may be overly
picky on floating point fields that require 0.0 instead
of just 0.
Running the i686-redhat-linux-gnu/32bit built with the
matching client.
If the either is not built with large file access this
could be the problem.
The file contains a dump of a database that has about
30 col's with about 29 million rows, and about 14gig in
size.
The box is a dual xeon 2.8 with 6gig of ram running RH
2.4 .19
It Should fail with warning or load. Any pointers on
hard limits to the db as well?
Thanks, still trying to untangeling all the info...
Sandy
## Comment 2335
Date: 2005-01-12 00:39:56 +0100
From: @njnes
Logged In: YES
user_id=43556
As Monet uses direct memory access using normal c-pointers,
the maximum memory size is therefor related to the pointer
size. In your
case this is 32bit, ie at most a db of 4GB. If this is a new
xeon emt
you could try again with a 64bit monet.
I still think there could be problem with the copy into
statement. Could
you find where Mapi breaks. Now we know 200MB works but
14GB does not. Does 1G still work?
## Comment 2336
Date: 2005-01-12 18:33:06 +0100
From: sandy ganz <<sganz>>
Logged In: YES
user_id=233880
Not sure what the limit is on the copy into, running into
problems where it is failing on some of the BCP records. It
looks like that it may be either single (') or double (") quotes
in the data that is causing a problem. For example The phrase
>> 3" 7' Height << seems to be the record that offends the
loader. What rules about the format can you tell so I can try
larger sets.
## Comment 2337
Date: 2005-01-12 20:03:29 +0100
From: sandy ganz <<sganz>>
Logged In: YES
user_id=233880
More...
Did some fancy vi work on a smaller split to remove all quotes,
seemed to load up the data, then on commit, the server
tossed an exception, internal error. Mapi Hangs.
The size of the file was about 463 Meg.
The interesting thing is that the server will not restart. It
tosses a couple of errors like -
!WARNING: BATdescriptor: range erro 624
!ERROR: BATSIGcrash: Mserver internal error (Segmentation
fault), please restart.
!ERROR: (One potential cause could be that your disk might
be full...)
Sized of data are not all that large, so not sure what might be
up. I had moved the default dbfarm off of /var/lib/MonetDB to
a 100G diskset so I don't think that is any problem.
Does MonetDB do any other disk access other the the farm
and the logs as specified in the MonetDB.conf?
I guess I have to blow the database off disk and recreate
unless their is another way to reset the problem.
I can say that the Loader is an very important part of the db,
since it is not praticle to do SQL to insert the data for larger
sets (TOO SLOW) are their other ways to load data that is
more robust and still have SQL access?
Also if you have a better place for me to post I can do that
as well, but no one responded in the USER mail list. I have a
few other questions, but will wait until I get the loader going
better.
Thanks
Sandy
## Comment 2338
Date: 2005-01-14 21:13:19 +0100
From: @njnes
Logged In: YES
user_id=43556
MapiClient was hanging on an incorrect error message and
therefor
the server was hanging because no next input was send by the
client.
Solved, ie fixed the error message and made sending error
save from
multiline error messages.
## Comment 2339
Date: 2005-11-09 13:14:38 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: TEST ADDED / SUCCESS
src/test/BugDay_2005-11-09_2.9.3/Tests/hang_on_copy_into.SF-1100504.sql
## Comment 2340
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1100504 at http://sourceforge.net/support/tracker.php?aid=1100504
| MapiClient Hangs on Large 'Copy Into' | https://api.github.com/repos/MonetDB/MonetDB/issues/432/comments | 0 | 2020-11-30T08:34:24Z | 2024-06-27T11:00:17Z | https://github.com/MonetDB/MonetDB/issues/432 | 753,272,468 | 432 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-11 06:06:28 +0100
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-01-13 08:26:43 +0100
## Comment 2327
Date: 2005-01-11 18:06:28 +0100
From: @drstmane
from
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.ntv_32-d/Fedora2/make.out:
========
[...]
/bin/sh ../libtool --mode=compile icc -no-gcc
-DHAVE_CONFIG_H -I.
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime -I..
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/common
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/gdk
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/monet
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/plain
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/contrib
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/mapi
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime/../compiler/include
-we140 -c99 -Wall -w2 -mp1 -O3 -restrict -unroll
-tpp6 -axKWNPB -we266 -Werror
-wd1418,1419,279,310,981,810,444,193,111,177,171,181,764,269,108,188,1357,102,70
-c -o xquery.lo xquery.c
icc -no-gcc -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime -I..
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/common
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/gdk
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/monet
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/plain
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/contrib
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/mapi
-I/var/tmp/monet.ntv_32-d.14271/MonetDB/.ntv_32-d/Fedora2/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime/../compiler/include
-we140 -c99 -Wall -w2 -mp1 -O3 -restrict -unroll -tpp6
-axKWNPB -we266 -Werror
-wd1418,1419,279,310,981,810,444,193,111,177,171,181,764,269,108,188,1357,102,70
-c xquery.c -DPIC
/var/tmp/monet.ntv_32-d.14271/pathfinder/runtime/xquery.mx(324):
error 175: subscript out of range
isspecial_tab[(int)'\377'] = 1; /* means
write as octal */
^
compilation aborted for xquery.c (code 2)
make[3]: *** [xquery.lo] Error 1
[...]
========
Indeed, looking at the code in runtime/xquery.mx
reveals the following:
[...]
static unsigned char isspecial_tab[CHAR_MAX]; /* the
main table */
! with unsigned char's, CHAR_MAX is 255;
! with signed char's (which appears to be the default for
! many compilers; e.g. gcc & icc), CHAR_MAX is "only"
127 ...
static char* isspecial_val[16]; /* the values
of the special chars */
static int isspecial_cnt = 0; /* 0 = tables
are not initialized */
static void init_isspecial() {
int i;
for(i=0; i<CHAR_MAX; i++)
isspecial_tab[i] = 0;
for(i=0; i<32; i++)
isspecial_tab[i] = 1; /* means write as
octal */
isspecial_tab[(int)'\377'] = 1; /* means write
as octal */
! with unsigned char's, (int)'\377' is 255, hence equal to
! CHAR_MAX, and thus out of range (since isspecial_tab
! has size CHAR_MAX, not CHAR_MAX+1 ...;
! with signed char's, (int)'\377' is -1, and hence
obviously
! out of range ...
! (only icc seems to be smart enough to detect this ...)
[...]
## Comment 2328
Date: 2005-01-12 10:13:57 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Should be repaired now. Increased array size to CHAR_MAX+1
and changed initializing iteration from < to <=
## Comment 2329
Date: 2005-01-12 10:46:50 +0100
From: @drstmane
Logged In: YES
user_id=572415
Jan,
I had to re-open the bug report (sorry ;-), since your
changes do not completely fix the problem:
(At least) with gcc & icc, signed char's are default
(__CHAR_UNSIGNED__ is not defined); hence, (int)'\377' is
-1, and thus obviously "subscript out of range" ...
(int)(unsigned char)'\377' would be 255 (hence, >=0), but
would not fit the CHAR_MAX+1 = 127 subscript range, either,
in case of the (default) signed char's.
Using UCHAR_MAX instead of CHAR_MAX would also solve this
problem.
Since I'm not sure, whether using "(int)(unsigned
char)'\377'" and "UCHAR_MAX" would keep the code
(semantically) correct, I bounce this back to you ;-)
## Comment 2330
Date: 2005-01-12 11:11:13 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Perhaps you can use 0377 instead of '\377'. The former *is*
a positive integer with value 255, whereas the latter may be
equal to 255, or it may be equal to -1 (as Stefan pointed out).
## Comment 2331
Date: 2005-01-13 08:26:43 +0100
From: @drstmane
Logged In: YES
user_id=572415
Jan's latest changes seem to solve the problem,
hence, I now close this report.
Thanks for the quick action, Jan!
## Comment 2332
Date: 2005-11-09 13:00:04 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: NO TEST / COMPILATION
## Comment 2333
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1100291 at http://sourceforge.net/support/tracker.php?aid=1100291
| icc detected "subscript out of range" in runtime/xquery.mx | https://api.github.com/repos/MonetDB/MonetDB/issues/431/comments | 0 | 2020-11-30T08:34:21Z | 2024-06-27T11:00:16Z | https://github.com/MonetDB/MonetDB/issues/431 | 753,272,440 | 431 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-10 10:58:00 +0100
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-01-11 10:14:46 +0100
## Comment 2323
Date: 2005-01-10 10:58:00 +0100
From: @grobian
The ODBC driver doesn't compile with the native
compilers on Debian3 (spin) and Fedora in the testweb.
## Comment 2324
Date: 2005-01-11 10:14:46 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Added casts.
## Comment 2325
Date: 2005-11-09 13:26:25 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: NO TEST / COMPILATION
## Comment 2326
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1099345 at http://sourceforge.net/support/tracker.php?aid=1099345
| ODBC driver doesn't compile with native compilers | https://api.github.com/repos/MonetDB/MonetDB/issues/430/comments | 0 | 2020-11-30T08:34:19Z | 2024-06-27T11:00:15Z | https://github.com/MonetDB/MonetDB/issues/430 | 753,272,407 | 430 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-07 03:19:32 +0100
From: @sjoerdmullender
To: @njnes
Version: -- development
Last updated: 2005-01-11 11:55:30 +0100
## Comment 2319
Date: 2005-01-07 15:19:32 +0100
From: @sjoerdmullender
The following sequence causes a crash in the third
query. Note that there is no table crash_me.
create table crash_q (a integer,b char(10));
insert into crash_q values(1,'c');
update crash_q set b= (select b from crash_me where
crash_q.a = crash_me.a);
## Comment 2320
Date: 2005-01-11 11:55:30 +0100
From: @njnes
Logged In: YES
user_id=43556
subquery was used without checking the result
## Comment 2321
Date: 2005-11-09 13:26:52 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-11-09_2.8/Tests/crash_with_wrong_nesting.SF-1097883.*
## Comment 2322
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1097883 at http://sourceforge.net/support/tracker.php?aid=1097883
| crash in erroneous nested query | https://api.github.com/repos/MonetDB/MonetDB/issues/429/comments | 0 | 2020-11-30T08:34:16Z | 2024-06-27T11:00:14Z | https://github.com/MonetDB/MonetDB/issues/429 | 753,272,375 | 429 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-07 11:14:04 +0100
From: @sjoerdmullender
To: @njnes
Version: -- development
Last updated: 2005-01-11 08:34:57 +0100
## Comment 2311
Date: 2005-01-07 11:14:04 +0100
From: @sjoerdmullender
The following statement results in a Segmentation Fault:
select case when 1 > 2 then 'false' when 2 > 1 then
'true' end
## Comment 2312
Date: 2005-01-07 11:19:53 +0100
From: @grobian
Logged In: YES
user_id=963970
apart from that it crashes (which is incorrect of course)
the statement itself is incorrect.
'false' and 'true' are unequal in type. The first is
varchar(5), the latter varchar(4). You should resolve this
by either using real booleans or casting the both two
strings to varchar(5).
## Comment 2313
Date: 2005-01-07 11:21:03 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
The following also causes a crash:
select case 2 when 1 then 'false' when 2 then 'true' end
## Comment 2314
Date: 2005-01-07 11:27:25 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
I don't care whether the statement is correct or incorrect
SQL: it should not crash.
The statements come from the mysql crash-me test.
## Comment 2315
Date: 2005-01-07 11:29:20 +0100
From: @grobian
Logged In: YES
user_id=963970
just trying to help.
you don't have to tell me that it shouldn't crash.
## Comment 2316
Date: 2005-01-11 20:34:57 +0100
From: @njnes
Logged In: YES
user_id=43556
in the single value case we incorrectly returned a bat of
atoms instead of
a single atom
## Comment 2317
Date: 2005-11-09 13:23:48 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-11-09_2.8/Tests/crash_in_case.SF-1097749.*
## Comment 2318
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1097749 at http://sourceforge.net/support/tracker.php?aid=1097749
| crash in case | https://api.github.com/repos/MonetDB/MonetDB/issues/428/comments | 0 | 2020-11-30T08:34:13Z | 2024-06-27T11:00:13Z | https://github.com/MonetDB/MonetDB/issues/428 | 753,272,339 | 428 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-05 10:55:21 +0100
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-01-11 03:07:51 +0100
## Comment 2306
Date: 2005-01-05 10:55:21 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Test case:
sql>set auto_commit=1;
7 querytype
affected table_name
rows name
int type
10 length
1 tuplecount
1 id
[ 1 ]
sql>create table test(i int);
6 querytype
sql>insert into test values(1);
4 querytype
affected table_name
rows name
int type
10 length
1 tuplecount
2 id
[ 1 ]
sql>delete from test;
4 querytype
affected table_name
rows name
bigint type
20 length
1 tuplecount
3 id
[ 1 ]
sql>select * from test;
3 querytype
test table_name
i name
int type
11 length
1 tuplecount
4 id
[ 1 ]
## Comment 2307
Date: 2005-01-05 22:24:15 +0100
From: @grobian
Logged In: YES
user_id=963970
the problem is related to the moment the commit is done.
in autocommit simulation mode the same behaviour is there:
sql>create table t (id int);
6 querytype
sql>commit;
7 querytype
sql>insert into t values (1);
4 querytype
[ 1 ]
sql>commit;
7 querytype
sql>delete from t;
4 querytype
[ 1 ]
sql>commit;
7 querytype
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
1 tuplecount
3 id
[ 1 ]
sql>
However, the problem seems to be related to a commit after
the insert statement, as it seems that as soon as a commit
is placed after the insert (before the delete) the deletes
do not disappear.
commit after delete:
sql>create table t (id int);
6 querytype
sql>insert into t values (1);
4 querytype
[ 1 ]
sql>delete from t;
4 querytype
[ 1 ]
sql>commit;
7 querytype
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
0 tuplecount
21 id
sql>commit
more>;
7 querytype
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
0 tuplecount
22 id
sql>
commit after create:
sql>create table t (id int);
6 querytype
sql>commit;
7 querytype
sql>insert into t values (1);
4 querytype
[ 1 ]
sql>delete from t;
4 querytype
[ 1 ]
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
0 tuplecount
17 id
sql>commit;
7 querytype
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
0 tuplecount
18 id
sql>
commit after insert:
sql>create table t (id int);
6 querytype
sql>insert into t values (1);
4 querytype
[ 1 ]
sql>commit;
7 querytype
sql>delete from t;
4 querytype
[ 1 ]
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
0 tuplecount
13 id
sql>commit;
7 querytype
sql>select * from t;
3 querytype
t table_name
id name
int type
11 length
1 tuplecount
14 id
[ 1 ]
sql>
## Comment 2308
Date: 2005-01-11 15:07:51 +0100
From: @njnes
Logged In: YES
user_id=43556
marked a table dirty on table clear (delete all)
## Comment 2309
Date: 2005-11-09 13:21:03 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-11-09_2.8/Tests/delete_with_auto_commit.SF-1096337.*
## Comment 2310
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1096337 at http://sourceforge.net/support/tracker.php?aid=1096337
| DELETE doesn't work if auto_commit=1 | https://api.github.com/repos/MonetDB/MonetDB/issues/427/comments | 0 | 2020-11-30T08:34:10Z | 2024-06-27T11:00:12Z | https://github.com/MonetDB/MonetDB/issues/427 | 753,272,300 | 427 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-23 10:06:20 +0100
From: @drstmane
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2004-12-23 02:56:31 +0100
## Comment 2301
Date: 2004-12-23 10:06:20 +0100
From: @drstmane
from
http://monetdb.cwi.nl/testing/intern/monetdb/Current/MonetDB/.GNU_64-d/Debian3.0/test10.out:
[...]
make[2]: Entering directory
`/var/tmp/monet.GNU_64-d.23228/MonetDB/Debian3.0'
python is /usr/bin/python
python is /usr/bin/python
Python 2.1.3
Traceback (most recent call last):
File
"/var/tmp/monet.GNU_64-d.23228/MonetDB/Debian3.0/src/testing/Mtest.py",
line 3084, in ?
main(sys.argv)
File
"/var/tmp/monet.GNU_64-d.23228/MonetDB/Debian3.0/src/testing/Mtest.py",
line 2595, in main
quiet = bool(env['QUIET'])
NameError: global name 'bool' is not defined
FAIL: RunMtest
===================
1 of 1 tests failed
===================
make[2]: *** [check-TESTS] Error 1
make[2]: Leaving directory
`/var/tmp/monet.GNU_64-d.23228/MonetDB/Debian3.0'
make[1]: *** [check-am] Error 2
make[1]: Leaving directory
`/var/tmp/monet.GNU_64-d.23228/MonetDB/Debian3.0'
make: *** [check-recursive] Error 1
make: Target `check' not remade because of errors.
[...]
## Comment 2302
Date: 2004-12-23 10:11:57 +0100
From: @drstmane
Logged In: YES
user_id=572415
apparently, there's also Python 2.2 on spin (see below);
if it's too difficult to fix Mtest.py, maybe, we could make
that the default and have bootstrap require Python >= 2.2.0
iso. >= 2.0.0 ...
10:09:15 manegold@spin:~ $ l /usr/bin/python*
lrwxrwxrwx 1 root root 9 Jun 25 16:51
/usr/bin/python -> python2.1*
-rwxr-xr-x 1 root root 522504 Sep 7 2002
/usr/bin/python2.1*
-rwxr-xr-x 1 root root 742568 Sep 26 05:07
/usr/bin/python2.2*
10:09:45 manegold@spin:~ $
## Comment 2303
Date: 2004-12-23 14:56:31 +0100
From: @grobian
Logged In: YES
user_id=963970
- added support for legacy platforms
- rewrote quiet option to be triggered by -q or --quiet
instead of --QUIET=true
## Comment 2304
Date: 2005-11-09 13:14:46 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: NO TEST / REDUNDANT
Mtest is test anyway on all our supported platforms
## Comment 2305
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1090275 at http://sourceforge.net/support/tracker.php?aid=1090275
| Mtest doesn't work anymore with python <2.2 (e.g. on spin) | https://api.github.com/repos/MonetDB/MonetDB/issues/426/comments | 0 | 2020-11-30T08:34:08Z | 2024-06-27T11:00:11Z | https://github.com/MonetDB/MonetDB/issues/426 | 753,272,280 | 426 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-21 09:13:58 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-12-22 12:46:05 +0100
## Comment 2297
Date: 2004-12-21 21:13:58 +0100
From: @grobian
monetdb-> \d rawtriples
CREATE TABLE "sys"."rawtriples" (
"id" int(19) NOT NULL,
"subjns" int(19) NOT NULL,
"subjlname" varchar(255) NOT NULL,
"predns" int(19) NOT NULL,
"predlname" varchar(255) NOT NULL,
"objns" int(19) NOT NULL,
"objlname" varchar(255) NOT NULL,
"objlabelhash" int(19),
"objlang" varchar(16),
"objlabel" varchar,
"objisliteral" boolean
);
monetdb-> INSERT INTO rawtriples VALUES(1001, 1,
'100168990', 2, 'glossaryEntry', 0, '',
-5342185907394955428, NULL, 'adorning with mosaic', true);
1 affected row
monetdb-> select * from rawtriples;
+----------+----------+-----------+----------+---------------+----------+----------+--------------+---------+----------------------+--------------+
| id | subjns | subjlname | predns |
predlname | objns | objlname | objlabelhash |
objlang | objlabel | objisliteral |
+==========+==========+===========+==========+===============+==========+==========+==============+=========+======================+==============+
| 1001 | 1 | 100168990 | 2 |
glossaryEntry | 0 | | 0 |
<NULL> | adorning with mosaic | true |
+----------+----------+-----------+----------+---------------+----------+----------+--------------+---------+----------------------+--------------+
1 row
here the quite large number -5342185907394955428 is
truncated to 0 without notice.
additional question is how to store this number as
Sesame quite likes numbers in this order of magnitudes.
## Comment 2298
Date: 2004-12-22 00:46:05 +0100
From: @njnes
Logged In: YES
user_id=43556
improved parser checks for overflow
## Comment 2299
Date: 2005-11-09 13:13:32 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-11-09_2.8/Tests/overflow_checking.SF-1089334.*
## Comment 2300
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1089334 at http://sourceforge.net/support/tracker.php?aid=1089334
| SQL: overflow values silently truncated to 0 | https://api.github.com/repos/MonetDB/MonetDB/issues/425/comments | 0 | 2020-11-30T08:34:05Z | 2024-06-27T11:00:10Z | https://github.com/MonetDB/MonetDB/issues/425 | 753,272,254 | 425 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-21 08:04:25 +0100
From: Jan Rittinger <<jan.rittinger>>
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2004-12-22 09:06:35 +0100
## Comment 2293
Date: 2004-12-21 20:04:25 +0100
From: Jan Rittinger <<jan.rittinger>>
boolean (1) and boolean(0) return confusing results:
<?xml version="1.0" encoding="utf-8"?>
<XqueryResult>
1149668178
</XqueryResult>
<?xml version="1.0" encoding="utf-8"?>
<XqueryResult>
1149668187
</XqueryResult>
I suggest "true" and "false" meaning e.g.,
<?xml version="1.0" encoding="utf-8"?>
<XqueryResult>
true
</XqueryResult>
(Galax returns true() and false())
## Comment 2294
Date: 2004-12-22 09:06:35 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Repaired the bug. Bad format string used %d instead of %s.
## Comment 2295
Date: 2005-11-09 13:04:24 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-11-09_0.8/Tests/boolean.SF-1089276-*
## Comment 2296
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1089276 at http://sourceforge.net/support/tracker.php?aid=1089276
| print routines misses boolean values | https://api.github.com/repos/MonetDB/MonetDB/issues/424/comments | 0 | 2020-11-30T08:34:02Z | 2024-06-27T11:00:09Z | https://github.com/MonetDB/MonetDB/issues/424 | 753,272,215 | 424 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-16 04:29:40 +0100
From: @dkwakkel
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-01-11 01:23:42 +0100
## Comment 2289
Date: 2004-12-16 16:29:40 +0100
From: @dkwakkel
There is a BATleak in BATins:
b = BATmaterialize(b, BATcount(b)+ BATcount(n));
This new bat is not used by CMDinsert_bat:
if (BATins(*r=b,s) == NULL) ...
In addition the result of a oid-void insert is discutable:
a := new (void, int).seqbase (0@0);
a.insert (0@0, 1);
b := new (oid, int);
b.insert (1@0, 10);
b.insert (4@0, 20);
a.insert (b);
a.print;
What is the expected result?:
h tmp_69 name
void int type
-----------------
[ 0@0, 1 ]
[ 1@0, 10 ]
[ 2@0, 20 ]
Or need the last tuple be 4@0 iso 2@0?
## Comment 2290
Date: 2004-12-16 17:24:25 +0100
From: @njnes
Logged In: YES
user_id=43556
Donald I expect the 4@0 and get this with the current
version, ie work for
me. Which version are you using (you mil seems old).
## Comment 2291
Date: 2005-11-09 12:56:54 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: TEST ADDED / SUCCESS
MonetDB/tests/BugDay_2005-11-09_4.8/Tests/oid_insert_in_void.SF-1086508.*
## Comment 2292
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1086508 at http://sourceforge.net/support/tracker.php?aid=1086508
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.
| insert oid bat in void bat leaks bat / does not work | https://api.github.com/repos/MonetDB/MonetDB/issues/423/comments | 0 | 2020-11-30T08:33:59Z | 2024-06-28T13:35:48Z | https://github.com/MonetDB/MonetDB/issues/423 | 753,272,184 | 423 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-14 09:45:20 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-01-14 11:06:49 +0100
## Comment 2283
Date: 2004-12-14 21:45:20 +0100
From: @drstmane
When compiled with 64-bit (and hence be deault with
64-bit oid's), all XQuery tests in the current test set
in CVS do fail when handled by "pf -M | Mserver
--dbinit='module(pathfinder);'.
E.g., XMark q01 results in:
--------
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit;
dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
!ERROR: BATpropcheck: BAT tmp_111 is dense but not key!?
!ERROR: BATpropcheck: BAT tmp_127 is dense but not key!?
<?xml version="1.0" encoding="utf-8"?>
<XqueryResult>
NOT FOUND BAT[4][0] in WS!!!!
NIL ctx->curBAT[4]
NOT FOUND BAT[3][0] in WS!!!!
NIL ctx->curBAT[3]
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
--------
Others produce similar output.
When compiled with 64-bit, but 32-bit oid's
(--enable-odi32), everything work just fine.
Also, 32-bit Pathfinder on 64-bit Linux works fine.
(I did use a working 32-bit pf for these tests; cf. bug
report 1085355 "'pf -M' seg-fault's on 64-bit with all
XMark queries".)
## Comment 2284
Date: 2004-12-15 17:37:45 +0100
From: @drstmane
Logged In: YES
user_id=572415
Assigned to Jan Flokstra, since he should have a look into
the problems that occur in "result_print".
I'll try to chase the BATpropcheck errors that occur before
"result_print" is called...
## Comment 2285
Date: 2004-12-17 15:28:57 +0100
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Repaired the crash of the result print. Rest of the bugs are
outside my reach so I send the bug back to Stefan
## Comment 2286
Date: 2005-01-14 11:06:49 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS
## Comment 2287
Date: 2005-11-09 12:50:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: NO TEST / REDUNDANT
pf is tested in hundreds of tests on all supported platforms
and OID/BIT settings
## Comment 2288
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1085384 at http://sourceforge.net/support/tracker.php?aid=1085384
| 64-bit Pathfinder fails (crashes Mserver) with 64-bit oid's | https://api.github.com/repos/MonetDB/MonetDB/issues/422/comments | 0 | 2020-11-30T08:33:56Z | 2024-06-27T11:00:07Z | https://github.com/MonetDB/MonetDB/issues/422 | 753,272,143 | 422 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-14 08:57:41 +0100
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2004-12-15 06:16:19 +0100
## Comment 2278
Date: 2004-12-14 20:57:41 +0100
From: @drstmane
on 64-bit systems (I tested on Fedora Core 2 / Linux64
2.6.9), compiling XMark quieres with "pf -M" results in
a segfault for all but Q06.
The segfault seems to occur in some twig-generated
code, not matter whether a 64-bit or a 32-bit twig is
used. See attached log for details.
The nightly testing results suggest, that this is the
case on all 64bit systems:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/index.html
With a 32-bit compiled pf, things work fine; hence, I
suspect that twig resp. the twig generated code is not
made for 64-bit... |-(
## Comment 2279
Date: 2004-12-15 17:29:35 +0100
From: @drstmane
Logged In: YES
user_id=572415
Thanks to Peter's help, we just found out the the problem is
actually not twig, but gc (the garbage collector): while
"my" gc-6.3 causes the crashes, "Peter's" gc-6.2 woks fine.
More investigation will follow...
## Comment 2280
Date: 2004-12-15 18:16:19 +0100
From: @drstmane
Logged In: YES
user_id=572415
config:
Fedora Core 2 / Linux 2.6.9
64-bit
gcc 3.3.3
After recompiling both gc-6.2 & gc-6.3 myself, the problem
is still the same: pf works fine when linked with gc-6.2,
but produces segfaults with gc-6.3.
Hence, we stick to gc-6.2, for now.
## Comment 2281
Date: 2005-11-09 12:48:16 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-11-09, Stefan: NO TEST / REDUNDANT
pf is tested in hundreds of tests...
## Comment 2282
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1085355 at http://sourceforge.net/support/tracker.php?aid=1085355
| "pf -M" seg-fault's on 64-bit with all XMark queries | https://api.github.com/repos/MonetDB/MonetDB/issues/421/comments | 0 | 2020-11-30T08:33:52Z | 2024-06-28T13:35:47Z | https://github.com/MonetDB/MonetDB/issues/421 | 753,272,095 | 421 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-14 02:49:53 +0100
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-12-20 07:37:18 +0100
## Comment 2274
Date: 2004-12-14 14:49:53 +0100
From: @MarcinZukowski
When loading a module fails (e.g. due to missing
symbols), it seems it stays 'partly loaded', e.g. it is
visible in loaded() and you can do sigs() (returns
empty), but functionality is naturally missing.
Next load of the same (possibly fixed) module doesn't
do anything, except for increasing usage_count of it
module.
It's not really crucial, but it seems an error
nevertheless.
## Comment 2275
Date: 2004-12-20 19:37:18 +0100
From: @njnes
Logged In: YES
user_id=43556
This only happens while developing. Developers can easily
fix this with a
Mserver restart.
The real fix would be to correctly remove the mil parts of
the module, but this
is to much work for an no crucial bug.
## Comment 2276
Date: 2005-11-09 13:42:08 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / TOO HARD TO FIX AND TO TEST
## Comment 2277
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1085120 at http://sourceforge.net/support/tracker.php?aid=1085120
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.
| module loading problem | https://api.github.com/repos/MonetDB/MonetDB/issues/420/comments | 0 | 2020-11-30T08:33:49Z | 2024-06-28T13:35:46Z | https://github.com/MonetDB/MonetDB/issues/420 | 753,272,051 | 420 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-11 06:57:06 +0100
From: @drstmane
To: @njnes
Version: -- development
Last updated: 2005-01-12 12:28:15 +0100
## Comment 2270
Date: 2004-12-11 18:57:06 +0100
From: @drstmane
from
http://monetdb.cwi.nl/testing/intern/monetdb/Current/MonetDB/.mTests103/GNU_64-d_IRIX646.5/src_gdk/qsort.out.00.html
:
========
v_i.sort_rev().print();
!ERROR: BATSIGcrash: Mserver internal error (Bus
error), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
========
The bug seems to have sneaked-in with changes made
between Nov 10 2004 & Nov 26 2004: before Nov 10, the
qsort test (src/gdk/Tests/qsort.milM) worked fine on
all architectures; then it failed on all architectures
until Nov 26 2004; since Nov 27 2004, it work on all
architectures except IRIX64 & SunOS (64bit).
Since it works in Linux64, it *might* be endian related...
(Sorry, that I cannot locate the bug any better, but I
didn't monitor testing while being on vacation in
Thailand... ;-)
## Comment 2271
Date: 2005-01-12 00:28:15 +0100
From: @njnes
Logged In: YES
user_id=43556
the BATordered mx macro was also called with void bats. With
sort_rev
this lead to accessing the bat using lngCmp (while the tail
was int), leading
to an alignment error on sun/irix. We can avoid calling
BATordered (
improving the performance) as we know TYPE_void is ordered
(on direction
afcouse).
## Comment 2272
Date: 2005-11-09 13:31:42 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: ALREADY IN TESTWEB
src/gdk/Tests/qsort.milM
## Comment 2273
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1083533 at http://sourceforge.net/support/tracker.php?aid=1083533
| sort_rev crashes Mserver (Bus error) on IRIX & SunOS (64bit) | https://api.github.com/repos/MonetDB/MonetDB/issues/419/comments | 0 | 2020-11-30T08:33:46Z | 2024-06-27T11:00:04Z | https://github.com/MonetDB/MonetDB/issues/419 | 753,272,029 | 419 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-10 02:50:18 +0100
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2004-12-16 10:18:00 +0100
## Comment 2266
Date: 2004-12-10 14:50:18 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Test:
sql>create table test(a integer, b int);
6 querytype
sql>insert into test(a) values(NULL);
4 querytype
affected table_name
rows name
int type
0 length
1 tuplecount
1 id
[ 1 ]
sql>insert into test(b) values(NULL);
4 querytype
affected table_name
rows name
int type
0 length
1 tuplecount
2 id
[ 1 ]
sql>commit;
7 querytype
sql>delete from test;
4 querytype
affected table_name
rows name
int type
0 length
1 tuplecount
3 id
[ 2 ]
sql>delete from test;
4 querytype
affected table_name
rows name
int type
0 length
1 tuplecount
4 id
[ 2 ]
I'd expect [ 0 ] from the second delete.
## Comment 2267
Date: 2004-12-16 10:12:15 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
As of 2004-12-15 (after cvs update), I can no longer
reproduce this bug.
## Comment 2268
Date: 2005-11-09 13:29:24 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-11-09_2.9.3/Tests/delete_x_2.SF-1082915.sql
## Comment 2269
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082915 at http://sourceforge.net/support/tracker.php?aid=1082915
| SQL: DELETE (2x) | https://api.github.com/repos/MonetDB/MonetDB/issues/418/comments | 0 | 2020-11-30T08:33:43Z | 2024-06-27T11:00:03Z | https://github.com/MonetDB/MonetDB/issues/418 | 753,271,995 | 418 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-10 12:05:01 +0100
From: Steffen Goeldner <<sgoeldner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2006-12-04 09:08:22 +0100
## Comment 2259
Date: 2004-12-10 12:05:01 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Env: set DBI_DSN=dbi:ODBC:driver=MonetDB ODBC Driver
Running perl -MDBI _c.pl, with _c.pl:
DBI->connect->column_info( undef,'sys','env', undef )-
>dump_results;
shows
'', 'sys', 'env', 'name' , '12', 'varchar', '0', '0', '0', '0', '1', ''
, '', '12', '-1', '0', '0', 'yes'
'', 'sys', 'env', 'value', '12', 'varchar', '0', '0', '0', '0', '1', '',
'', '12', '-1', '0', '1', 'yes'
2 rows
The ordinal_position starts with 0, but it should start
with 1.
The attached patch adds 1 to sys.columns.number.
## Comment 2260
Date: 2004-12-10 22:46:44 +0100
From: @njnes
Logged In: YES
user_id=43556
patch applied
## Comment 2261
Date: 2005-11-09 13:23:37 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / POSTPONED
We need tests for *all* ODBC functions.
## Comment 2262
Date: 2006-11-10 04:20:01 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 2263
Date: 2006-11-29 10:04:45 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
over-ruling "sf-robot":
re-set to "Pending" since adding a proper test has neither been done nor finally been discarded, yet.
User (submitter) & developer (assignee),
please check, again.
## Comment 2264
Date: 2006-12-04 21:08:22 +0100
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Close this as there are other open bugs about the same missing ODBC-tests issue.
## Comment 2265
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082833 at http://sourceforge.net/support/tracker.php?aid=1082833
| SQLColumns.c: ordinal_position starts by 0 | https://api.github.com/repos/MonetDB/MonetDB/issues/417/comments | 0 | 2020-11-30T08:33:40Z | 2024-06-27T11:00:01Z | https://github.com/MonetDB/MonetDB/issues/417 | 753,271,969 | 417 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-10 11:42:18 +0100
From: Steffen Goeldner <<sgoeldner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2007-11-30 04:20:04 +0100
## Comment 2252
Date: 2004-12-10 11:42:18 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Running perl -MDBI _c.pl, with _c.pl:
DBI->connect->column_info( undef,'sys','env', undef )-
>dump_results;
shows
DBD::ODBC::db column_info failed: [Microsoft][ODBC
Driver Manager] Invalid string or buffer length (SQL-
HY090)(DBD: describe/SQLBindCol err=-1) at _c.pl line 1.
The ODBC log shows:
perl cols c2c-c10 EXIT
SQLDescribeCol with return code 0 (SQL_SUCCESS)
HSTMT 014A29B0
UWORD 7
UCHAR * 0x013FEC0D [
11] "column_size"
SWORD 255
SWORD * 0x01400B38 (11)
SWORD * 0x01400B42 (0)
SQLULEN * 0x01400B3C
(4294967292)
SWORD * 0x01400B40 (0)
SWORD * 0x01400B44 (2)
perl cols c2c-c10 ENTER SQLBindCol
HSTMT 014A29B0
UWORD 7
SWORD 1
<SQL_C_CHAR>
PTR 0x013FEB7C
SQLLEN -3
SQLLEN * 0x01400B58
perl cols c2c-c10 EXIT SQLBindCol
with return code -1 (SQL_ERROR)
HSTMT 014A29B0
UWORD 7
SWORD 1
<SQL_C_CHAR>
PTR 0x013FEB7C
SQLLEN -3
SQLLEN * 0x01400B58
DIAG [HY090] [Microsoft][ODBC
Driver Manager] Invalid string or buffer length (0)
The DataType (0) and ColumnSize (4294967292) are
wrong.
The attached patch changes
"integer" to "int"
in SQLColumns.c.
## Comment 2253
Date: 2004-12-10 22:47:19 +0100
From: @njnes
Logged In: YES
user_id=43556
applied patch
## Comment 2254
Date: 2005-11-09 13:22:17 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / POSTPONED
We need tests for *all* ODBC functions.
## Comment 2255
Date: 2006-11-10 04:20:01 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 2256
Date: 2006-11-29 10:04:38 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
over-ruling "sf-robot":
re-set to "Pending" since adding a proper test has neither been done nor finally been discarded, yet.
User (submitter) & developer (assignee),
please check, again.
## Comment 2257
Date: 2007-11-30 04:20:04 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 2258
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082816 at http://sourceforge.net/support/tracker.php?aid=1082816
| SQLColumns: integer vs. int | https://api.github.com/repos/MonetDB/MonetDB/issues/416/comments | 0 | 2020-11-30T08:33:37Z | 2024-06-27T11:00:00Z | https://github.com/MonetDB/MonetDB/issues/416 | 753,271,935 | 416 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-09 06:24:41 +0100
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-09 02:02:39 +0100
## Comment 2248
Date: 2004-12-09 18:24:41 +0100
From: @drstmane
When preparing bug report 1082284 "module ref-counting
tests fail since recent gdk-changes", I also noticed
the following:
The test (src/monet/Tests/modrefcnt1.milM) also checks,
whether the type
"lock" that is provided by module "lock" is correctly
removed only after the
second drop. This works fine when the test script is
read from stdin
(piped/redirected), but fails when the test script is
directly given on the command line. In the second case,
the type "lock" seems to be gone already after the
first drop, though the module still seems to be loaded
(as expected).
This holds for both MonetDB 4.4.3 & MonetDB 4.5.0 .
See logs below for details.
========
cat src/monet/Tests/modrefcnt1.milM
--------
module("lock");
module("lock");
VAR test := new(int,lock);
loaded();
test.print();
test.count().print();
drop("lock");
loaded();
test.print();
test.count().print();
drop("lock");
loaded();
test.print();
test.count().print();
quit();
========
========
Mserver --trace < src/monet/Tests/modrefcnt1.milM
--------
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module("lock");
MonetDB>module("lock");
MonetDB>
MonetDB>VAR test := new(int,lock);
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
-----------------
h tmp_28 name
int lock type
-----------------
MonetDB>test.count().print();
[ 0 ]
MonetDB>
MonetDB>drop("lock");
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
-----------------
h tmp_28 name
int lock type
-----------------
MonetDB>test.count().print();
[ 0 ]
MonetDB>
MonetDB>drop("lock");
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
!ERROR: DESCload: atom 'lock' unknown, in BAT '28'.
!ERROR: interpret: MIL operator invocation of
'print(BAT)' with illegal types.
MonetDB>test.count().print();
!ERROR: DESCload: atom 'lock' unknown, in BAT '28'.
!ERROR: interpret: MIL operator invocation of
'count(BAT)' with illegal types.
!ERROR: interpret_params: print(param 1): evaluation error.
[
the above errors are correct, as the type "lock" is
correctly gone after the second drop of module "lock".
]
MonetDB>
MonetDB>quit();
========
Mserver src/monet/Tests/modrefcnt1.milM
--------
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
-----------------
h tmp_28 name
int lock type
-----------------
[ 0 ]
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
!ERROR: interpret: MIL operator invocation of
'print(BAT[int,lock])' with illegal types.
[
After the first drop, module "lock" is correctly still
listed with ref-count 1, but type "lock" seems to be
gone...
]
MonetDB>
========
## Comment 2249
Date: 2005-02-09 14:02:39 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed. The reference counting in TBL_unloadmod which
decides whether a logical drop needs to be done if no
physical drop was done was using the wrong client id
(monet_clients entry).
## Comment 2250
Date: 2005-11-09 13:10:46 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: TEST ADDED / SUCCESS
MonetDB/tests/BugDay_2005-11-09_4.9.3/Tests/double_load_double_drop.SF-1082293.{prelude,milM}
## Comment 2251
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082293 at http://sourceforge.net/support/tracker.php?aid=1082293
| Atom of "double-loaded" module disappears after first drop | https://api.github.com/repos/MonetDB/MonetDB/issues/415/comments | 0 | 2020-11-30T08:33:33Z | 2024-06-27T10:59:59Z | https://github.com/MonetDB/MonetDB/issues/415 | 753,271,898 | 415 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-09 06:12:55 +0100
From: @drstmane
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-12-09 09:59:17 +0100
## Comment 2245
Date: 2004-12-09 18:12:55 +0100
From: @drstmane
Since Niels' changes to gdk_align.mx gdk_bbp.mx
gdk_system.mx on Monday, Dec 6 2004
("
small valgrind fixes
1) copy where dst == src
2) unlock mutex by a thread not owning the lock
Also introduced MT_{create,destroy,wait,signal}_cond
functions needed to
have BBPlock correctly wait for all unloads to finish
")
the module reference counting tests in src/monet do fail.
(Probabaly, failures of some "Availability" tests are
related to the same changes.)
Apparently, after loading module "lock" twice, the
first drop still seems to work correctly, while the
second drop now hangs the server.
This only occurs, when the test script (e.g.,
src/monet/Tests/modrefcnt1.milM) is piped into the
Mserver (or read via input redirection).
If the test script is given on the commandline, the
second drop triggers the following error:
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault), please restart.
!ERROR: (One potential cause could be that your disk
might be full...)
See logs below for details.
The "test from stdin" version works fine in MonetDB
4.5.0 where the respective gdk changes haven't been
applied, yet.
But the "test from file" also fails with 4.5.0, though
with a different error:
Mserver:
/ufs/manegold/Monet/4.4/MonetDB/src/monet/monet_tbl.mx:1788:
TBL_unloadmod: Assertion `r' failed.
Aborted (core dumped)
See below for details.
========
cat src/monet/Tests/modrefcnt1.milM
--------
module("lock");
module("lock");
VAR test := new(int,lock);
loaded();
test.print();
test.count().print();
drop("lock");
loaded();
drop("lock");
loaded();
quit();
========
Mserver --trace < src/monet/Tests/modrefcnt1.milM
--------
Monet Database Server V4.4.3
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module("lock");
MonetDB>module("lock");
MonetDB>
MonetDB>VAR test := new(int,lock);
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
-----------------
h tmp_28 name
int lock type
-----------------
MonetDB>test.count().print();
[ 0 ]
MonetDB>
MonetDB>drop("lock");
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>
MonetDB>drop("lock");
!FATAL: BATSIGabort: signal 2 caught by thread b7fe4aa1
[ Ctrl-C'ed since Mserver hangs ]
========
Mserver src/monet/Tests/modrefcnt1.milM
--------
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
-----------------
h tmp_28 name
int lock type
-----------------
[ 0 ]
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
Mserver:
/ufs/manegold/Monet/4.4/MonetDB/src/monet/monet_tbl.mx:1788:
TBL_unloadmod: Assertion `r' failed.
Aborted (core dumped)
========
========
Mserver --trace < src/monet/Tests/modrefcnt1.milM
--------
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module("lock");
MonetDB>module("lock");
MonetDB>
MonetDB>VAR test := new(int,lock);
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
-----------------
h tmp_28 name
int lock type
-----------------
MonetDB>test.count().print();
[ 0 ]
MonetDB>
MonetDB>drop("lock");
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
MonetDB>test.count().print();
MonetDB>
MonetDB>drop("lock");
MonetDB>loaded();
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
MonetDB>test.print();
MonetDB>test.count().print();
MonetDB>
MonetDB>quit();
========
Mserver src/monet/Tests/modrefcnt1.milM
Monet Database Server V4.5.0
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically
linked.
Visit http://monetdb.cwi.nl/ for further information.
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 2 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
-----------------
h tmp_28 name
int lock type
-----------------
[ 0 ]
---------------------------------
module usage_count name
str int type
---------------------------------
[ "algebra", 1 ]
[ "arith", 1 ]
[ "bat", 1 ]
[ "kernel", 1 ]
[ "lock", 1 ]
[ "main", 1 ]
[ "str", 1 ]
[ "sys", 1 ]
[ "trans", 1 ]
Mserver:
/ufs/manegold/Monet/4.4/MonetDB/src/monet/monet_tbl.mx:1788:
TBL_unloadmod: Assertion `r' failed.
Aborted (core dumped)
========
## Comment 2246
Date: 2005-11-09 12:55:45 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / REDUNDANT
## Comment 2247
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082284 at http://sourceforge.net/support/tracker.php?aid=1082284
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.
| module ref-counting tests fail since recent gdk-changes | https://api.github.com/repos/MonetDB/MonetDB/issues/414/comments | 0 | 2020-11-30T08:33:29Z | 2024-06-27T10:59:58Z | https://github.com/MonetDB/MonetDB/issues/414 | 753,271,854 | 414 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-09 03:51:19 +0100
From: Jens Teubner <<teubner>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2004-12-10 10:04:14 +0100
## Comment 2241
Date: 2004-12-09 15:51:19 +0100
From: Jens Teubner <<teubner>>
Current Pathfinder CVS does not compile on GCC 3.4.3:
array.c: In function `PFarray_vprintf':
array.c:223: error: implicit declaration of function `va_copy'
va_copy() was introduced with C99. GCC 3.4 is invoked with `-ansi
-std=c99' and does thus not support C99. This problem is similar to
other C99 issues we had seen in the past (e.g., initialization of
variables in for(;;) loops).
conf/monet.m4 sets `-ansi -std=c99' for GCC 3.x, except for GCC 3.
3, where (correctly) only `-std=c99' is set. I've patched monet.m4
on my system as below, which solves the problem for me.
However, I have no clue, why `-ansi' is set at all for GCC 3.x.
Seems like it doesn't make trouble on GCC 3.[0-2], but demanding
two standards at once (ISO C90 and C99) doesn't really make sense,
does it? So should we even omit `-ansi' for all the GCC 3.x versions,
or does that break anything on other systems?
*** /opt/MonetDB-4.4.2/share/MonetDB/conf/monet.m4 Thu Dec
9 15:34:28 2004
--- /home/teubner/work/monet/MonetDB/conf/monet.m4 Mon
Nov 15 15:38:45 2004
***************
*** 237,247 ****
*-solaris*)
AC_DEFINE(__EXTENSIONS__, 1, [Compiler flags])
;;
*-irix*|*-cygwin*|*-darwin*|2.*-*)
;;
! 3.[[3-4]]*-*)
AC_DEFINE(_POSIX_C_SOURCE, 199506L, [Compiler
flag])
AC_DEFINE(_POSIX_SOURCE, 1, [Compiler flag])
AC_DEFINE(_XOPEN_SOURCE, 600, [Compiler flag])
CFLAGS="$CFLAGS -std=c99"
CXXFLAGS="$CXXFLAGS -ansi"
--- 237,247 ----
*-solaris*)
AC_DEFINE(__EXTENSIONS__, 1, [Compiler flags])
;;
*-irix*|*-cygwin*|*-darwin*|2.*-*)
;;
! 3.3*-*)
AC_DEFINE(_POSIX_C_SOURCE, 199506L, [Compiler
flag])
AC_DEFINE(_POSIX_SOURCE, 1, [Compiler flag])
AC_DEFINE(_XOPEN_SOURCE, 600, [Compiler flag])
CFLAGS="$CFLAGS -std=c99"
CXXFLAGS="$CXXFLAGS -ansi"
## Comment 2242
Date: 2004-12-10 22:04:14 +0100
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS.
## Comment 2243
Date: 2005-11-09 12:51:11 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / COMPILATION
## Comment 2244
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1082161 at http://sourceforge.net/support/tracker.php?aid=1082161
| GCC 3.4 should not be invoked `-ansi -std=c99' | https://api.github.com/repos/MonetDB/MonetDB/issues/413/comments | 0 | 2020-11-30T08:33:27Z | 2024-06-27T10:59:57Z | https://github.com/MonetDB/MonetDB/issues/413 | 753,271,838 | 413 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-08 02:18:46 +0100
From: Steffen Goeldner <<sgoeldner>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2004-12-08 07:14:37 +0100
## Comment 2237
Date: 2004-12-08 14:18:46 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
RunMserver.bat.in is missing from MonetDB-4.4.2.tar.gz:
NMAKE : fatal error U1073: don't know how to
make '".\..\RunMserver.bat.in"'
## Comment 2238
Date: 2004-12-08 19:14:36 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS:
Next to RunMserver.bat.in, also src/utils/Mx/getopt.c &
src/utils/Mx/mygetopt.h were missing in the source
distribution (MonetDB-4.4.2.tar.gz).
Until a new source distribution will be made available for
download on SF (probably only with the next release), please
find the missing files in the attached MonetDB-4.4.2-a.tar.gz .
## Comment 2239
Date: 2005-11-09 12:49:28 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / COMPILATION
## Comment 2240
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1081341 at http://sourceforge.net/support/tracker.php?aid=1081341
| NMAKE: don't know how to make RunMserver.bat.in | https://api.github.com/repos/MonetDB/MonetDB/issues/412/comments | 0 | 2020-11-30T08:33:25Z | 2024-06-27T10:59:56Z | https://github.com/MonetDB/MonetDB/issues/412 | 753,271,814 | 412 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-08 12:04:01 +0100
From: Steffen Goeldner <<sgoeldner>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-12-15 12:05:17 +0100
## Comment 2232
Date: 2004-12-08 12:04:01 +0100
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Running the server with
--set sql_debug=1 > logs\1 2>&1
--set sql_debug=1 > logs\2 2>&1
--set sql_debug=1 > logs\3 2>&1
creates the logs in logs.tgz (attached).
During run 1 and 2, I had client activity (perl_dbd_test).
Logs 2 and 3 show
!ERROR: BATrename: name is in use
All logs are surprisingly different.
## Comment 2233
Date: 2004-12-09 11:27:07 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
To create client activity (perl_dbd_test):
- unpack t.tar.gz in MonetDB/src/mapi/clients/perl/DBD
- run
perl Makefile.PL
make
make test
## Comment 2234
Date: 2004-12-13 15:28:36 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
As of 2004-12-13 (after cvs update), I can no longer
reproduce this bug.
## Comment 2235
Date: 2005-11-09 12:48:05 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-11-09, sjoerd: NO TEST / OBSOLETE
The relevant rename has been removed from the code, so there
is no way to test this.
## Comment 2236
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1081274 at http://sourceforge.net/support/tracker.php?aid=1081274
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.
| !ERROR: BATrename: name is in use | https://api.github.com/repos/MonetDB/MonetDB/issues/411/comments | 0 | 2020-11-30T08:33:22Z | 2024-06-27T10:59:55Z | https://github.com/MonetDB/MonetDB/issues/411 | 753,271,786 | 411 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.