select should be a cancellation point

[Cedric Hombourger]

Hi,

I have looked at PR #1013 (http://bugs.uclibc.org/view.php?id=1013) and
concluded that the thread got never cancelled because select() is no longer
a cancellation point while specs say that it should. Anyone knows why it got
removed from libpthread/linuxthreads.old/wrapsyscall.c?

I wanted to compare with linuxthreads (instead of linuxthreads.old) but even
simple programs segfault. I haven't investigated why at this stage. Is it a
known issue?

What's the plan wrt. to the pthread library:
   - everybody will use NPTL as soon as merged into trunk?
   - stabilize linuxthreads and then remove linuxthreads.old?

Regards,
Cedric Hombourger
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.busybox.net/pipermail/uclibc/attachments/20070705/610534d9/attachment-0002.htm