[88] in Pthreads mailing list archive
Re: pthreads porting help!
daemon@ATHENA.MIT.EDU (Christopher Provenzano)
Wed Aug 16 18:47:42 1995
To: Ken Raeburn <raeburn@cygnus.com>
Cc: neal@ctd.comsat.com (Neal Becker), pthreads@MIT.EDU, proven@MIT.EDU
In-Reply-To: Your message of "Wed, 16 Aug 1995 11:19:06 EDT."
<199508161519.LAA12575@kr-laptop.cygnus.com>
Date: Wed, 16 Aug 1995 17:18:52 EDT
From: Christopher Provenzano <proven@MIT.EDU>
>
> From: neal@ctd.comsat.com (Neal Becker)
> Date: Wed, 16 Aug 95 10:10:34 -0400 (EDT)
>
> Still trying to port pthreads to hpux9. The basic problem now is,
> what should be used for sigset_t? If we use the real hpux definition,
>
> Include the hpux system header file that defines the type.
sigset_t should be the same as what the system defines it to be.
>
> Then we need a lot of fixing up, for example assignments from
> __SIGEMTPYSET.
>
> Yup.
Take a look at machdep/sunos-5.3/__signal.h. Solaris uses a non integral
sigset_t too.
>
> But in pthreads/sig.c: we have definitions of the POSIX sigemptyset,
> etc. HPUX already has these. Does pthreads require that sigset_t be
> an int?
>
> Rewrite those functions to declare "static const sigset_t" variables
> initialized with __SIGEMPTYSET and __SIGFILLSET, and assign from
> those. Then they can be defined as structure or array initializers.
>
> (I don't know what other places might be assuming an integral type
> though. Try it and see what else doesn't compile...)
There shouldn't be any in the pthread code. By setting up __SIGEMPTYSET,
__SIGFILLSET, ... correctly everything will just work.
CAP
