[PATCHES][RFC] Fix a deadlock bug in static variable initialization

[Doug Kwan (關振德)]

That's why __GTHREAD_HAS_COND is only defined in gthr-posix.h. The
other thread systems still use the old broken code. I don't have
access to system running the other thread systems, I could not even
find documentations. If someone give me access to machines running
platforms, I am more than happy to change the code and test.

-Doug

2007/6/8, Andrew Pinski <pinskia@gmail.com>:

> A couple more, you only updated gthr-posix.h and gthr-single.h even
> though gthr-posix95.h exists and is used for some solaris based
> targets (older ones) and guard.cc is used there also and
> pthread_cond_wait exists for POSIX 95 :).  So the tests will fail on
> older solaris machines and maybe even on alpha OSF as that uses
> gthr-dce.h.  And also gthr-tpf.h.