cvs commit: projects/validate/testcases/open_posix_testsuite/conformance/interfaces/pthread_mutex_timedlock 5-1.c 5-2.c 5-3.c

[Mike Makonnen]

On Wed, Dec 31, 2003 at 11:21:52AM -0800, Kris Kennaway wrote:
> On Wed, Dec 31, 2003 at 07:54:48AM -0800, Mike Makonnen wrote:
> > mtm         2003/12/31 07:54:48 PST
> > 
> >   FreeBSD projects repository
> > 
> >   Modified files:
> >     validate/testcases/open_posix_testsuite/conformance/interfaces/pthread_mutex_timedlock 
> >                                                                                            5-1.c 
> >                                                                                            5-2.c 
> >                                                                                            5-3.c 
> >   Log:
> >   Correct incorrect assertion about the behaviour of pthread_mutex_timedlock()
> >           o If the thread has the mutex already locked it will not block. It
> >             will return EDEADLK. Therefore, according to the standard,
> >             returning EINVAL if the timeout is invalid does not apply to
> >             such a thread.
> >   
> >   Better explanation for the case testing for immediate return
> >   with ETIMEDOUT when the timeout is set to the current time.
> 
> Isn't something a bit wrong with the testsuite if you have to make a
> lot of commits to port it to freebsd?

Heh. I knew someone was bound to bring this up.

The testsuite was developed on a Linux box. So, most of the patches have to
do with header inclusions and unimplemented or differing function signatures.
Also keep in mind that the commits were/are being made to three separate
parts: the testharness (Linux centric), tests developed for a Linux box, and
Posix conformance tests.  The only ones that should work right out of the box
with FreeBSD are the POSIX tests. And most of those tests do work out of the box.
Except for this one case most of the patching in this area is header
inclusions. In this regard I am tring to be careful to fix the header inclusions in
areas that don't affect the assertion that is being tested.

Cheers.
-- 
Mike Makonnen  | GPG-KEY: http://www.identd.net/~mtm/mtm.asc
mtm at identd.net | Fingerprint: 00E8 61BC 0D75 7FFB E4D3  6BF1 B239 D010 3215 D418
mtm at FreeBSD.Org| FreeBSD - Unleash the Daemon !