[omniORB] SIGSEGV, Linux and omniORBPy1.1

[Gary Pennington]

Hi,

garyp@python:~> gcc -v
Reading specs from /usr/lib/gcc-lib/i486-linux/egcs-2.91.66/specs
gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)

I might be missing something here, but how does the version of gcc/egcs
impact the execution of  a python server ? I didn't build my
omniORB distribution, I'm using the pre-built linux distribution from your
website so I haven't done any compilation.

I have access to a SPARC machine (as you might have guessed) , so I'll set
up my SPARC environment tomorrow and see if I can reproduce the problem.

Yours,

Gary


Duncan Grisby wrote:

> On Tuesday 29 August, Gary Pennington wrote:
>
> > The problem isn't obviously deterministic, so I can't pin it down, but
> > gdb always reports the same problem.
>
> You shouldn't trust core files from multi-threaded Linux programs. I'm
> surprised you get anything usable at all.
>
> What compiler are you using?  The egcs which comes with some (possibly
> all) versions of Suse is not configured with --enable-threads, so it
> leads to random crashes under high load. That sounds like your
> problem. You should compile gcc 2.95 or egcs 1.1.2 for yourself,
> making sure that you use the --enable-threads switch to configure.
>
> If you are already using a suitably configured compiler, please run
> your server under gdb. That way, when it segfaults, you'll get a
> reliable indication of where it died.
>
> Cheers,
>
> Duncan.
>
> --
>  -- Duncan Grisby  \  Research Engineer  --
>   -- AT&T Laboratories Cambridge          --
>    -- http://www.uk.research.att.com/~dpg1 --