The __vdso_clock_gettime problem seems to exist in the .a version. It doesn't show up when I build a shared library (libxfem.so) that is then loaded on runtime by a driver. This shared library has a class that contains boost threads. The error occurs when I construct a local thread (code snippet below). I do not know what the etiquette for attaching code files on boost mailing list is, so I'm not attaching my codes. But if you wish, I can send you the code files. boost::thread st( &XFE_Scene::run, _scene ); // error occurs here _threads[ 0 ] = boost::move( st ); When I try to load libxfem.so at run time, this is the error that I get. /home/kish1/Projects/Ashwini2.0/trunk/Engine/SafeDriver/src/CoreManager.cpp@initialize,82: fatal error: could not open libxfem.so /home/kish1/Projects/lib/libxfem.so: undefined symbol: _ZTIN5boost6detail16thread_data_baseE I greatly appreciate if you can decipher what exactly the problem is. I have a version running on an Intel Mac that has boost 1.37 installed, and the same code runs perfectly. It was also ran properly on a previous Fedora Core 7 version. On Thu, Jul 16, 2009 at 9:35 AM, Christoph Gysin<christoph.gysin@fr33z3.org> wrote:
Now, I write a small test program (given below). It gives me linker errors with undefined references to pretty much every pthread symbol.
Yes, boost::thread uses pthreads so you'll need to link against it.
If anyone wants I can paste all the errors in the next message. When I add the -lpthread flag to the compile, it gives me an undefined reference error with regards to __vdso_clock_gettime from /lib64/libpthread.a.
Looks like a known bug: http://sourceware.org/bugzilla/show_bug.cgi?id=5531
Chris -- echo mailto: NOSPAM !#$.'<*>'|sed 's. ..'|tr "<*> !#:2" org@fr33z3 _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users