From: Gatewood, Pat (GatewoodWP++at++navair.navy.mil)
Date: 03/31/2003 13:14:38
I am using gcc-2.96. I have recompiled the ACE wrappers and TAO without
exceptions and without optimization. My API still works without the
Performer libs and segraults with them (still not actually calling any
Performer code.
When I tried compiling under RedHat 8.0 and the gcc3.2 compiler I ran into
problems with template instantiation. It seems like I am going to have a
lot of problems going that way.
I have the performer part compiled under gcc2.96 without the API and it
seems to work ok. Unfortunately it is necessary for the API to be linked
with the Performer application.
-----Original Message-----
From: Tom Flynn [mailto:flynnt++at++engr.sgi.com]
Sent: Wednesday, March 26, 2003 17:09
To: Gatewood, Pat
Cc: 'Jean-Christophe Lombardo'; 'info-performer++at++sgi.com'
Subject: RE: [info-performer] Performer 3.0 and pthreads
pthread vs. sproc isn't an issue here...it's probably a compiler related
problem. if you're using redhat 7.2, i'm guessing you're using gcc-2.96.
the 2.96 compiler is pretty buggy unto itself (as opposed to 2.91 or 2.95
which were pretty stable). in addition, the gcc2 compilers had problems
with exception handling, optimizing inline code, and a few other things.
either code would segfault where it shouldn't just by linking in libpf or
it would give a mess load of undefined references or other problems.
the problem you're seeing sounds like the same kind of thing that would
happen if one was using c++ exceptions with gcc2.
if you can, i recommend upgrading to redhat 8 and/or the gcc-3.2 compilers
(you'll need the gcc3 version of Performer). my gut instinct is that the
problem you're seeing will go away.
-tom
On Wed, 26 Mar 2003, Gatewood, Pat wrote:
> My OS is linux (RedHat 7.2).
>
> -----Original Message-----
> From: Jean-Christophe Lombardo [mailto:Jean-Christophe.Lombardo++at++cstb.fr]
> Sent: Wednesday, March 26, 2003 10:25
> To: Gatewood, Pat
> Subject: Re: [info-performer] Performer 3.0 and pthreads
>
>
> What is your OS ?
> There is no way of using pthread and performer in the same code on IRIX
> machine. You'll have to use sproc() instead of pthread.
>
> jcl
>
> Gatewood, Pat wrote:
> > I have an application with a CORBA API, intended to drive some Performer
> > code. The Performer code is currently commented out.
> >
> > When I link the CORBA API portion using "-lpthread" it runs as expected.
> > Adding "-lpf" causes the CORBA API code to get a segmentation fault.
> > Bear in mind that it is not currently calling anything related to
> Performer.
> >
> > I was under the impression that Performer 3.0 was thread safe. What's
> > happening here? How can I fix this?
> >
> > Thanks for the help.
> >
> > -----------------------------------------------------------------------
> > List Archives, Info, FAQ: http://www.sgi.com/software/performer/
> > Open Development Project: http://oss.sgi.com/projects/performer/
> > Submissions: info-performer++at++sgi.com
> > Admin. requests: info-performer-request++at++sgi.com
> > -----------------------------------------------------------------------
> >
>
> --
> Jean-Christophe Lombardo (mailto:Jean-Christophe.Lombardo++at++cstb.fr)
> Environnements Virtuels Enrichis / Enriched Virtual Environments
> CSTB | tel: +33 4 93 95 67 12 / fax: +33 4 93 95 67 33
> 290 route des Lucioles - BP 209 - 06904 Sophia Antipolis - France
>
> -----------------------------------------------------------------------
> List Archives, Info, FAQ: http://www.sgi.com/software/performer/
> Open Development Project: http://oss.sgi.com/projects/performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com
> -----------------------------------------------------------------------
>
-- "Mongooses are famous for their snake-fighting ability, and are almost always victorious because of their speed, agility, and timing and also because of their thick coat."
This archive was generated by hypermail 2b29 : Mon Mar 31 2003 - 13:15:11 PST