Re: BUG #7955: argv=0x0 on program start

From: Damian Weber <dweber(at)htw-saarland(dot)de>
To: Michael Meskes <meskes(at)postgresql(dot)org>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #7955: argv=0x0 on program start
Date: 2013-03-17 19:55:29
Message-ID: alpine.BSF.2.00.1303172033370.163@magritte.htw-saarland.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

On Sun, 17 Mar 2013, Michael Meskes wrote:

> Date: Sun, 17 Mar 2013 11:04:26 +0100
> From: Michael Meskes <meskes(at)postgresql(dot)org>
> To: dweber(at)htw-saarland(dot)de
> Cc: pgsql-bugs(at)postgresql(dot)org
> Subject: Re: [BUGS] BUG #7955: argv=0x0 on program start
>
> On Sat, Mar 16, 2013 at 12:12:49PM +0000, dweber(at)htw-saarland(dot)de wrote:
> > My program compiled with system compiler gcc-4.2.1 and ecpg from PostgreSQL
> > version 9.2.3 terminates by receiving SIGSEGV. The gdb debugger shows
> > argv=0x0 (see below). On PostgreSQL version 8.4.16 it runs without trouble.
> >
> > Debugger output:
> > b$ gdb bin/prog prog.core
> > GNU gdb 6.1.1 [FreeBSD]
> > ...
> > Program terminated with signal 11, Segmentation fault.
> > ...
> > #0 0x08048eec in main (argc=-1077949560, argv=0x0) at prog.pgc:87
> > 87 {
>
> Hmm, this looks like the program didn't even start. Honestly I fail to see how
> this could be a bug in PostgreSQL. Are the libraries and the program compiled
> with threading enabled? What happens if you use a very small test case instead
> of your program?

while narrowing things down, I installed PostgresQL 8.4.16, 9.0.12,
9.1.8 and 9.2.23 and I got several different symptoms
along the way, a) being a case where it shows something unrelated
to PostgresQL, so you may close the issue on the PostgresQL bug tracking
system

a) some non-postgres function (print_usage()) commented out -> program works
b) some non-postgres function (print_usage()) active -> program crashes with argv=0x0
c) postgres function (select cursor, open, fetch, close) commented out -> program works
d) postgres function (select cursor, open, fetch, close) active -> program crashes with argv=0x0

mysteriously no crashes using PostgresQL 8.4.16, so perhaps a change
in ecpg from 8.4 to 9.0 triggered that kind of bug, maybe some kind
of size limit which prevented the program to execute in the first place

after a reboot of the machine I cannot reproduce the symptom anymore

anyway, sorry for the noise

Best regards

Damian

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message navin.pro 2013-03-18 05:31:23 BUG #7956: Installation fails saying database cluster initialization failed
Previous Message Michael Meskes 2013-03-17 10:04:26 Re: BUG #7955: argv=0x0 on program start