On Tue, 2002-09-17 at 19:47, Thomas Pfaff wrote: > Agreed, I missed the point where magic is set to 0 if TlsAlloc has failed. > But i would really appreciate if you would apply the patch for > pthread_key::get that removes the set_errno(0) and preserves the WIN32 > LastError. > If you configure gcc to use the the pthreads interface for exceptions then > it makes heavy use of pthread_getspecific to read the actual eh context. > Some people might be very surprised when errno and Win32 LastError is > cleared behind her back (It took me 2 days to find the reason for this > about 2 years ago on mingw).
ok. Rob
signature.asc
Description: This is a digitally signed message part
