Re: [help-serveez] Thread safety

[Andreas Rottmann]

stefan <address@hidden> writes:

> On 11 Dec 2002, Andreas Rottmann wrote:
> 
> > Raimund 'Raimi' Jacob <address@hidden> writes:
> >
> > > i tend more to --enable-some-threadsafe-functions (definitly
> > > --enable-threads, because that's not true) and then use
> > > pthreads. pthreads is the way to go for multithreading and
> > > multithreaded systems without pthreads are not supported. i
> > > agree with rotty that there are probably none of them, anyway.
> > >
> > That was the way I thought to go.
> 
> I also agree with that.
> 
You seem to contradict yourself. I read the above as "it doesn't
really make sense to layer ontop of pthreads", which contradicts your
following statements:

<quote stefan>
On 10 Dec 2002, Andreas Rottmann wrote:

> > > > As I already stated:  If ever -> then a wrapper API, because this what
> > > > Serveez is mostly about: system independence.
> > > >
> > > Hmm, one could argue POSIX is also about system independence ;-)
> >
> > But POSIX isn't implemented for all of them...  Of course for the most
> > common ones it actually is.
> >
> Are there any real systems _with_ _thread_ _support_ that don't have a
> pthread implementation?
>
> But I guess your point is that with a wrapper, it would be possible to
> use *exactly* the same code, even if there are no pthreads (the mutex
> operations, then would become simply null code). ACK?

Correct.  Also I would implement the necessary changes for windoze without
using a 3rd party library (as Raimi proposed).
</quote>

Could you clear this up?

Regards, Andy
-- 
Andreas Rottmann         | address@hidden        | address@hidden | 
address@hidden
http://www.8ung.at/rotty | GnuPG Key: http://www.8ung.at/rotty/gpg.asc
Fingerprint              | DFB4 4EB4 78A4 5EEE 6219  F228 F92F CFC5 01FD 5B62