From: Alain Magloire <alainm AT rcsm DOT ece DOT mcgill DOT ca>
Message-Id: <199907131721.NAA00083@mccoy2.ECE.McGill.CA>
Subject: Re: setitimer and getitimer
To: djgpp-workers AT delorie DOT com
Date: Tue, 13 Jul 1999 13:21:17 -0400 (EDT)
In-Reply-To: <Pine.SUN.3.91.990713094015.25972B-100000@is> from "Eli Zaretskii" at Jul 13, 99 09:41:06 am
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

Bonjour M. Eli Zaretskii

> 
> > According to DEC, this is standardized not by POSIX, but by XPG4-UNIX. 
> > That's the X/Open group definition of Unix.
> 
> That was my mistake: I thought setitimer was Posix, but it isn't.
> 

Posix.4(The realtime extensions, and later incorporate in Posix.1) defines a
set of interval Timers : timer_create, timer_gettime(), timer_delete(),
clock_getime(), etc ..

They are very flexible, some implementations actually use them as the
basis and define {g,s}etitimer() around it.

If the {g,s}etitimer() is the default facility in DJGPP, it is possible
to use it and implement the POSIX counterpart on top.  The problem is that
you'll also need to revisit the signal API, and implement the full Posix
signal semantics (SA_SIGINFO).

For example when you create a timer interval, you are not restricted
to SIG{ALRM,VTALRM,PROF}.  You can choose the signal to be delivered
SIGRTMIN .... SIGRTMAX.
etc ..


-- 
au revoir, alain
----
Aussi haut que l'on soit assis, on est toujours assis que sur son cul !!!