Re: [Boost-users] [thread] Segfault in Boost

From: Mat Marcus (mat-lists_at_[hidden])
Date: 2005-08-19 14:37:16

This sounds like the gcc vs. shared libs problem
again.
<http://aspn.activestate.com/ASPN/Mail/Message/boost/2786704>

 - Mat

--- David Abrahams <dave_at_[hidden]> wrote:

> 
> Jeff, it's always a good idea to put the name of the
> library in
> question in the subject line of your posts.
>
http://www.boost.org/more/discussion_policy.htm#effective
> 
> 
> "Jeff Lounsberry" <jeff_at_[hidden]> writes:
> 
> > I'm getting a segfault in Boost when I go to use
> mutexes. I built the
> > Boost unit test for threads with the following:
> >
> > bjam "-sTOOLS=gcc" "-sBUILD debug
> <runtime-link>static <threading>multi"
> >
> > *** errors detected in test suite "Boost.Threads:
> mutex test suite"; see
> > standard output for details
> >
> > [---]
> >
> >
> > *** errors detected in test suite "Boost.Threads:
> xtime test suite"; see
> > standard output for details
> > Segmentation fault
> >
> > [---]
> >
> > In Boost 1.33 there are errors with
> test_mutex_lib:
> >
>
root_at_unstable:~/scratch/boost_1_33_0/bin/boost/libs/thread/test/test_mutex_lib.test/gcc/debug/runtime-link-static/threading-multi>
> ./test_mutex_lib
> > Running 6 test cases...
> > unknown location(0): fatal error in
> "test_try_mutex": signal: memory
> > access violation
> >
> > *** errors detected in test suite "Boost.Threads:
> mutex test suite"; see
> > standard output for details
> >
> > [--]
> > In Boost 1.33 there are no errors with
> test_xtime_lib: 
> >
>
root_at_unstable:~/scratch/boost_1_33_0/bin/boost/libs/thread/test/test_xtime_lib.test/gcc/debug/runtime-link-static/threading-multi>
> ./test_xtime_lib
> > Running 2 test cases...
> >
> > *** No errors detected
> >
> > Well then, i'm thoroughly confused. It seems the
> unit test Boost has are
> > making calls to libthreaddb.so (NPTL).
> >
> > Anyone have any ideas? I've reproduced this on 2
> machines (opterons
> > 246), 2 distros now. What really boggles me is
> that the one xtime test
> > works on 1.33 but the mutex test doesn't. Simple
> mutex calls work fine
> > however. Are these specific problems with the test
> or something deeper?
> > I've seen no discussion on the mailing list.
> 
> -- 
> Dave Abrahams
> Boost Consulting
> www.boost-consulting.com
> 
> _______________________________________________
> Boost-users mailing list
> Boost-users_at_[hidden]
>
http://lists.boost.org/mailman/listinfo.cgi/boost-users
>