pthread_kill prototype

[Robert Collins]

> -----Original Message-----
> From: Norman Vine [ mailto:nhv@cape.com ]
> Robert Collins writes:
> 
>  Norman Vine [ mailto:nhv@cape.com ] wrote:
> >> >>  
> >> >
> >
> >Are you up to rebuilding those pacakges? I will be breaking 
> the ABI by
> >fixing this - and it will be fixed this week.
> >_POSIX_THREADS is referred to by
> > http://www.opengroup.org/onlinepubs/007908799/xsh/threads.htm
> l as being
> >present on XSI systems. Now as I don't know what an XSI system is, I
> >wonder why _POSIX_THREADS showed up for you?
> 
> Python uses _POSIX_THREADS to distinguish pthreads from
> other possible thread implementations

Yes I realise that - what I don't realise is _why_ that symbol was
defined when you were building! Cygwin doesn't define it, and python
shouldn't define it.
 
> >I will be defining some of those symbols as I go along, but 
> until I get
> >the time to find a full blown pthreads test suite that'll run under
> >cygwin... (Norman if you've got some interest.. hint hint)
> 
> It doesn't have a full blown test suite but GNU PTH has a fully
> POXIX compilent pthread package that compiles and seems to
> run just fine with Cygwin

PTH doesn't integrate in with cygwin behind the scenes, so things like
pthread_atfork won't work, and shared memory access to mutex's will be
broken as well. Or did you mean that their test suite works? 

Also are you talking the redhat hosted GNU portable threads for win32?
Or a more general library based threads implementation (because these
are often userland and as such not actually concurrent.

Perhaps a url for the one you are talking about?

> 
> Cheers
> 
> Norman Vine
>