Re: [Boost-testing] Python + Solaris + -lrt

From: Caleb Epstein (caleb.epstein_at_[hidden])
Date: 2007-03-31 10:38:09

On 3/30/07, David Abrahams <dave_at_[hidden]> wrote:
>
> on Fri Mar 30 2007, "Caleb Epstein" <caleb.epstein-AT-gmail.com> wrote:

> >  (maybe that was yesterday?), but I have
> > never seen this error from today's run before:
> >
> > /openpkg/bin/ld:
> > /home/cae/boost-regression/RC_1_34_0/results/boost/bin.v2/libs/python/build/gcc-4.1.1_sunos_i86pc/debug/debug-symbols-off/libboost_python-gcc41-d-1_34.so.1.34.0:
> > pthread_getspecific: invalid version 2 (max 0)
> > /home/cae/boost-regression/RC_1_34_0/results/boost/bin.v2/libs/python/build/gcc-4.1.1_sunos_i86pc/debug/debug-symbols-off/libboost_python-gcc41-d-1_34.so.1.34.0:
> > could not read symbols: Bad value
>
> My guess is that pthreads are already statically linked into the
> executable, and the library search picked up a different version.

No, that can't be it.  As with many of the system libraries on
Solaris, there is no static version of the pthread library.  That same
link command without the -lpthread produces an identical error
message.

Or at least it did last night when I tried.  This morning the test
appears to have compiled and run cleanly.  My results aren't up on the
web page yet (31-Mar 10:34AM EDT) but it looks like this has now been
fixed.  Did you change something?

-- 
Caleb Epstein