finishing cancellation support

[Ulrich Drepper <drepper@redhat.com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm on the home stretch for the cancellation handling.  The secondary
cancellation points need to be handled and other functions not intended
to be cancellation points need to be modified appropriately.

The biggest problem (beside time) are the cancellation handlers.
Actually, it's not too much of a problem, not with the C cleanup
attribute handling.  This is why I'll bump the gcc requirement to gccs
with this cleanup handling.  It's nothing really new, any deployed glibc
already had to be compiled with such a gcc.  But it'll now be a
configure test.  gcc 3.3.1 has the code (and will probably released
soon).  The Red Hat backport to gcc 3.2 also has the changes.  So there
should be a compiler for everybody.

I'll be checking in code in small chunks.  The amount of work is too
large to do it in one scoop.  Hopefully everything remains usable.

The biggest problem is testing.  Especially with the exception based
cleanup.  I could really use some help.  While doing the work I'll
document which of the secondary cancellation points (see the second
table in 2.9.5.2 POSIX 2003) are actual cancellation points in NPTL.  If
I could get some help with writing test cases similar to the ones in
tst-cancel4.c in NPTL this would speed up things quite a bit.

- -- 
- --------------.                        ,-.            444 Castro Street
Ulrich Drepper \    ,-----------------'   \ Mountain View, CA 94041 USA
Red Hat         `--' drepper at redhat.com `---------------------------
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+/y7a2ijCOnn/RHQRApYnAKDGcbiUEErdhDMDAsXuILP/EqK4zgCfRy0E
NqwyTXxf8moXfwCo7LSeUno=
=F9lp
-----END PGP SIGNATURE-----