[linux-audio-dev] GThread vs. pthread

iainduncan at telus.net iainduncan at telus.net
Sat Jun 12 19:15:41 UTC 2004


Another advantage if you're new to it as I am, is that there are some very 
good books on pthreads. And you can look at the Ardour and Softwerk code. ; )  
 
The O'Reilly book is old ( '96 ) but very clearly written and readable. Of 
course because it's old you can probably find it at the library. 
 
Iain 
 
 
> At 12 Jun 2004 00:09:05 -0500, 
> Jack O'Quin wrote: 
> >  
> > Pete Bessman <ninjadroid at gazuga.net> writes: 
> >  
> > > One more point in favor of GThread I just figured out is config 
> > > testing for it.  All you have to do is at a pkg-config check for 
> > > gthread-2.0 and your set.  Having glanced through other programs, 
> > > checking for pthreads seems a bit more involved than I care to be 
> > > (read: not at all, I hate the fscking autotools).  So, I think 
> > > I'll roll with GThread. 
> >  
> > Not a very strong reason.  Checking for pthreads is a little tricky, 
> > but hardly rocket science.  These checks (from JACK) work for quite a 
> > few platforms, though I'm sure there are others one could add... 
> >  
> > AC_CHECK_HEADER(/usr/include/nptl/pthread.h, 
> > 	[CFLAGS="$CFLAGS -I/usr/include/nptl"]) 
> >  
> > AC_CHECK_FUNC(pthread_create, [], 
> >     AC_CHECK_LIB(pthread, pthread_create, [], 
> > 	AC_MSG_ERROR([*** JACK requires POSIX threads support]))) 
>  
> Touche.  Since you did all the hard work for me, and pthreads isn't 
> OOPy like GThreads, I'll stick with pthreads :-) 
>  
> > There is also a conditional check for pthread_barrier_init(). 
>  
> Well, I don't even know wtf that is, so I think I'll be ok without it. 
>  
> [pb] 
>  
 
 




More information about the Linux-audio-dev mailing list