i see, that in 9434/9435 the check and the insque/remque issue is done. great.

i've just checked it back and it seems to work :)

thanks!

-piet

-------- Original-Nachricht --------
> Datum: Fri, 25 Jan 2008 10:50:37 -0500
> Von: Jeff Johnson <n3npq@mac.com>
> An: rpm-devel@rpm5.org
> Betreff: Re: patch-aw (1/2) PTHREAD_MUTEX_DEFAULT != PTHREAD_MUTEX_NORMAL

> 
> On Jan 24, 2008, at 5:38 PM, Peter Kalbus wrote:
> 
> > in qnx the macros are defined like:
> >
> > 	PTHREAD_MUTEX_DEFAULT	0	/* ((int)(_NTO_SYNC_NONRECURSIVE)) */
> > 	PTHREAD_MUTEX_NORMAL		3	/* ()int)(_NTO_SYNC_NONRECUESIVE |  
> > _NTO_SYNC_NOERRORCHECK))
> >
> > what is the problem, if they are not equal, so i can post that to  
> > the foundry27 to get clearification.
> >
> 
> The only context I have for the issue is here:
> 
>      https://bugzilla.redhat.com/show_bug.cgi?id=159024
> 
> The underlying problem is pthreads implementation specific,
> and there is a difference between the glibc and the cygwin  
> implementation.
> 
> Note that the issue is not critical for rpm porting anymore.
> 
> Originally pthread_mutex() was essential to the signal handler,
> and SIGCHLD was not handled correctly on Cygwin because
> of the implementation difference.
> 
> The current rpm signal handler avoids pthread_mutex() entirely,
> using a mutex from pipe(2) instead.
> 
> 
> > -piet
> >
> >
> >
> > note, that the atached patch contains another topic and that that  
> > patch is not yet cross platform capable.
> >
> >
> 
> I've added the insque/remque patch, will be in rpm-5.0.1.
> 
> 73 de Jeff
> 
> ______________________________________________________________________
> RPM Package Manager                                    http://rpm5.org
> Developer Communication List                        rpm-devel@rpm5.org