Date: 2012-05-29 10:53:59 +0200
From: Inho Kim <<georgios.kim>>
To: SQL devs <>
Version: 11.11.5 (Jul2012)
CC: @njnes, @drstmane

Last updated: 2013-05-22 13:36:32 +0200

Comment 17296

Date: 2012-05-29 10:53:59 +0200
From: Inho Kim <<georgios.kim>>

User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/536.5 (KHTML, like Gecko) Chrome/19.0.1084.52 Safari/536.5
Build Identifier: MonetDB v11.9.5

If where condition has single identifier expression, mserver5 crash.

for example..

select * from sys.tables where 1; ==> OK
select * from sys.tables where name ; ==> Crashed

Reproducible: Always

Comment 17300

Date: 2012-05-29 18:57:55 +0200
From: @njnes

fixed a bug in the optimizer which pushes functions down select/joins.

Comment 17463

Date: 2012-07-14 08:57:29 +0200
From: @drstmane

Re-opened and assigned to Jul2012 as test has being failing with Jul2012 & default for quite some time with

QUERY = select * from tables where name;
ERROR = !conversion from string to type bit failed.
!conversion from string to type bit failed.
!conversion from string to type bit failed.

cf.,

http://monetdb.cwi.nl/testweb/web/testgrid.php?serial=44603:a04c99a2b437&module=sql&targets=&order=platform,arch,compiler

http://monetdb.cwi.nl/testweb/tests/44603:a04c99a2b437/GNU-Fedora-x86_64-propcheck-assert/sql/mTests/test/BugTracker-2012/predicate_select.Bug-3090.err.diff.html

Comment 17464

Date: 2012-07-14 15:10:46 +0200
From: @sjoerdmullender

This should be easy enough to fix, however the question is, would that fix be correct? The problem is that non-Boolean values are being used as Booleans. Is that actually allowed?

Comment 17465

Date: 2012-07-15 10:38:39 +0200
From: @njnes

I think approving the error is much clearer. We have a strict rule for converting strings to bit (bool), ie the string should be true,false or NULL.

Comment 17516

Date: 2012-07-19 14:54:47 +0200
From: @grobian

Changeset b6b2bd626749 made by Fabian Groffen fabian@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=b6b2bd626749

Changeset description:

predicate_select: approve per comment 4 in bug #3090

Comment 17524

Date: 2012-07-20 10:53:06 +0200
From: @drstmane

Test proves to produce too variale output to be approved easily:
output (number of error messages) appears to depend on number of threads:
http://monetdb.cwi.nl/testweb/web/testgrid.php?serial=44688:b6b2bd626749&order=platform,arch,compiler&targets=&module=sql&tstlimit=test/BugTracker-2012

Maybe, we need to forcefully run it single-threaded, e.g., by using optimizer pipeline "sequential_pipe"?

Comment 17525

Date: 2012-07-20 12:32:16 +0200
From: @grobian

Perhaps it would make more sense to only return one error message, or an unique set of error messages encountered by the workers.

Comment 17533

Date: 2012-07-20 19:01:58 +0200
From: @drstmane

Changeset 807573750262 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=807573750262

Changeset description:

predicate_select.Bug-3090: approved output after recent changes

Comment 17534

Date: 2012-07-20 19:39:32 +0200
From: @drstmane

Sjoerd checked in Fabian's suggestion:

http://dev.monetdb.org/hg/MonetDB/rev/a528928726b6
http://dev.monetdb.org/hg/MonetDB/rev/5e08a7eba004

Comment 17535

Date: 2012-07-20 20:15:43 +0200
From: @drstmane

Changeset a13583ea7077 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=a13583ea7077

Changeset description:

predicate_select.Bug-3090: approved new output

Comment 17555

Date: 2012-07-22 13:33:28 +0200
From: @drstmane

Changeset 3121867aed83 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=3121867aed83

Changeset description:

rel_optimizer.c: fixed typo ("ne" -> "e")
found by segfault in test sql/test/BugTracker-2012/Tests/predicate_select.Bug-3090.sql

Comment 17556

Date: 2012-07-22 13:33:30 +0200
From: @drstmane

Changeset d4f17b9d4fc7 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=d4f17b9d4fc7

Changeset description:

predicate_select.Bug-3090: approved SciQL output

Comment 17637

Date: 2012-08-23 10:14:37 +0200
From: @sjoerdmullender

Jul2012-SP1 has been released.

Comment 18583

Date: 2013-03-02 12:58:39 +0100
From: @drstmane

Changeset 1d5a855e4e87 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=1d5a855e4e87

Changeset description:

predicate_select.Bug-3090: approved stable output after changeset [04ad0c9e2637](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=04ad0c9e2637) 

Comment 18744

Date: 2013-05-22 13:36:32 +0200
From: MonetDB Mercurial Repository <>

Changeset 6f4a9ef59b50 made by Stefan Manegold Stefan.Manegold@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=6f4a9ef59b50

Changeset description:

predicate_select.Bug-3090: (re-)approved nogeom output