On Fri, 4 Nov 2011, Kai Tietz wrote:

2011/11/4 Prof Brian Ripley <rip...@stats.ox.ac.uk>:
On Fri, 4 Nov 2011, Kai Tietz wrote:

2011/11/4 Prof Brian Ripley <rip...@stats.ox.ac.uk>:

Thanks for picking up on this.  I am seeing about 100 examples of header
conflicts with R packages (after patching R itself to avoid the problem).

I've now read the source code and I do not see anything in
winpthreads/src/clock.c that is specific to pthreads: even the
thread-specific timer is standard Windows API calls.

Kai?

Well, the plan was (and is for now) to provide this API only in
combination with pthread-POSIX layer.  The cause for this is, that
timers share API to be breakable in terms of pthread's kill.
The header hick-up here is that we need to define within unistd.h - if
this API is present - the proper defines for it.

If you can think of a better way to provide this functionality in a
POSIX compatible way, then I will adopt this approach.

I mentioned one earlier in the thread:  E.g. Solaris guards within

#if (!defined(_STRICT_STDC) && !defined(__XOPEN_OR_POSIX)) || \
       (_POSIX_C_SOURCE > 2) || defined(__EXTENSIONS__)
#endif

I don't know what you would feel appropriate in your setup: _POSIX_C_SOURCE
is most natural to someone who does a lot of cross-platform work.

Brian

Brian,

I see that pthread_time.h is including here pthread.h, which is indeed
unintended and IMHO even wrong.  The cause might be the for the

Yes, including pthread.h was my original complaint.

timespec structure, which lives in timeb.h headers.  So by replacing
here the pthread.h by include for timeb.h header, and guard
functions-definition by your suggested pattern is for sure doable (and
even correct AFAICS).

Would this be ok for you?

Yes, thanks, it would help a lot.

Brian


Regards,
Kai

------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
Mingw-w64-public mailing list
Mingw-w64-public@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public


--
Brian D. Ripley,                  rip...@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595
------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
Mingw-w64-public mailing list
Mingw-w64-public@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public

Reply via email to