Beta4 postgres.exe crash "...terminated by signal 5"

Greetings.
I am very new to PostgreSQL so apologies in advance if this bug report is a
waste of time for some reason.

Summary:
I have been experimenting with 8.0 Beta4 on Windows XP and the attached SQL
resulted in a crash of the database (not the postmaster as far as I can
tell). My code creates a function, then another function to test it. Then
the test function is executed. During the execution of the test function the
process crashes.

Version:
PostgreSQL 8.0.0beta4 on i686-pc-mingw32, compiled by GCC gcc.exe (GCC)
3.3.1 (mingw special 20030804-1)
OS: Windows XP SP1

Recreating the problem:
Create a database and run the attached script against it using psql.

Observed Outcome:
postgres.exe crashes. See log file extract below.

Expected Outcome:
Either the tests should run to completion or an exception should be raised
(in the likely event that my syntax or logic is flawed).

FYI I have implemented my logic in a new way, so I don't need a fix, but
thought I'd pass on the details all the same.

Best Regards,
- John.

pg_log Extract
=========

2004-11-16 08:58:41 NOTICE: 00000: PASS: / did not raise an exception

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:41 NOTICE: 00000: PASS: /topdir did not raise an exception

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:41 NOTICE: 00000: PASS: /topdir/subdir did not raise an
exception

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:41 NOTICE: 00000: PASS: /topdir/subdir/subsubdir did not
raise an exception

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:41 NOTICE: 00000: FAIL: should have been identified as an
invalid path

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:41 NOTICE: 00000: PASS: // did raise an exception

2004-11-16 08:58:41 LOCATION: exec_stmt_raise, pl_exec.c:1968

2004-11-16 08:58:47 LOG: 00000: server process (PID 3652) was terminated by
signal 5

2004-11-16 08:58:47 LOCATION: LogChildExit, postmaster.c:2282

2004-11-16 08:58:47 LOG: 00000: terminating any other active server
processes

2004-11-16 08:58:47 LOCATION: HandleChildCrash, postmaster.c:2175

2004-11-16 08:58:47 LOG: 00000: all server processes terminated;
reinitializing

2004-11-16 08:58:47 LOCATION: reaper, postmaster.c:2074

2004-11-16 08:58:47 LOG: 00000: database system was interrupted at
2004-11-15 17:43:17 GMT Standard Time

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4026

2004-11-16 08:58:47 LOG: 00000: checkpoint record is at 0/4EA3158

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4095

2004-11-16 08:58:47 LOG: 00000: redo record is at 0/4EA3158; undo record is
at 0/0; shutdown FALSE

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4123

2004-11-16 08:58:47 LOG: 00000: next transaction ID: 26681; next OID:
135172

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4126

2004-11-16 08:58:47 LOG: 00000: database system was not properly shut down;
automatic recovery in progress

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4182

2004-11-16 08:58:47 LOG: 00000: redo starts at 0/4EA3198

2004-11-16 08:58:47 LOCATION: StartupXLOG, xlog.c:4219

2004-11-16 08:58:47 WARNING: 01000: could not remove database directory
"C:/Program Files/PostgreSQL/8.0.0-beta4/data/base/128288"

2004-11-16 08:58:47 LOCATION: dbase_redo, dbcommands.c:1214

2004-11-16 08:58:48 LOG: 00000: record with zero length at 0/4EB9F38

2004-11-16 08:58:48 LOCATION: ReadRecord, xlog.c:2459

2004-11-16 08:58:48 LOG: 00000: redo done at 0/4EB9F08

2004-11-16 08:58:48 LOCATION: StartupXLOG, xlog.c:4277

2004-11-16 08:58:49 LOG: 00000: database system is ready

2004-11-16 08:58:49 LOCATION: StartupXLOG, xlog.c:4489

_________________________________________________________________
Don't just search. Find. Check out the new MSN Search!
http://search.msn.com/

"John Kennedy" <jmkennedyie(at)hotmail(dot)com> writes:
> I am very new to PostgreSQL so apologies in advance if this bug report is a
> waste of time for some reason.

Not at all. Looks like you've found a mistake in where variables of
pass-by-reference datatypes get stored. Thanks for the report!

regards, tom lane

Tom Lane wrote:
> "John Kennedy" <jmkennedyie(at)hotmail(dot)com> writes:
> > I am very new to PostgreSQL so apologies in advance if this bug report is a
> > waste of time for some reason.
>
> Not at all. Looks like you've found a mistake in where variables of
> pass-by-reference datatypes get stored. Thanks for the report!

Tom, I assume this was fixed. CVS is down now so I can't check.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> Tom Lane wrote:
>> "John Kennedy" <jmkennedyie(at)hotmail(dot)com> writes:
>>> I am very new to PostgreSQL so apologies in advance if this bug report is a
>>> waste of time for some reason.
>>
>> Not at all. Looks like you've found a mistake in where variables of
>> pass-by-reference datatypes get stored. Thanks for the report!

> Tom, I assume this was fixed.

Yes.

> CVS is down now so I can't check.

Works for me ...

regards, tom lane

Hi again.
Was incommunicado for a while...but just to close, FYI I have verified the
bug is fixed in both Beta 5 and RC1.

Another bug I spotted in Beta4 (variables in functions having their values
changing for no apparent reason) has also been fixed in beta 5. You fixed it
before I had time to report it: impressive!

It seems to me you have developed a fantastic platform: I wish you every
success with the release.
- John.

>From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
>To: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
>CC: John Kennedy <jmkennedyie(at)hotmail(dot)com>, pgsql-bugs(at)postgresql(dot)org
>Subject: Re: [BUGS] Beta4 postgres.exe crash "...terminated by signal 5"
>Date: Wed, 01 Dec 2004 17:01:17 -0500
>
>Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > Tom Lane wrote:
> >> "John Kennedy" <jmkennedyie(at)hotmail(dot)com> writes:
> >>> I am very new to PostgreSQL so apologies in advance if this bug report
>is a
> >>> waste of time for some reason.
> >>
> >> Not at all. Looks like you've found a mistake in where variables of
> >> pass-by-reference datatypes get stored. Thanks for the report!
>
> > Tom, I assume this was fixed.
>
>Yes.
>
> > CVS is down now so I can't check.
>
>Works for me ...
>
> regards, tom lane
>
>---------------------------(end of broadcast)---------------------------
>TIP 4: Don't 'kill -9' the postmaster

_________________________________________________________________
Don't just search. Find. Check out the new MSN Search!
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

"John Kennedy" <jmkennedyie(at)hotmail(dot)com> writes:
> Was incommunicado for a while...but just to close, FYI I have verified the
> bug is fixed in both Beta 5 and RC1.

> Another bug I spotted in Beta4 (variables in functions having their values
> changing for no apparent reason) has also been fixed in beta 5.

Probably the same bug, actually --- the basic problem was storage
allocated to variables getting recycled for other purposes ...

regards, tom lane