Message-Id:	<5.1.0.14.0.20010802213106.00aca020@pop.dorsai.org>
X-Sender:	pjfarley AT pop DOT dorsai DOT org
X-Mailer:	QUALCOMM Windows Eudora Version 5.1
Date:	Thu, 02 Aug 2001 21:48:41 -0400
To:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
From:	"Peter J. Farley III" <pjfarley AT dorsai DOT org>
Subject:	Re: libc question: Why no setpgrp()?
Cc:	djgpp-workers AT delorie DOT com
In-Reply-To:	<Pine.SUN.3.91.1010802100200.28510Q-100000@is>
References:	<3b68c7f6 DOT 8131322 AT news DOT escape DOT com>
Mime-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

At 10:02 AM 8/2/01 +0300, Eli Zaretskii wrote:
<Snipped>
 >IMHO, that configure script has a bug which should be fixed.  It
 >simply didn't consider that setpgrp might be absent.

Well, the script in question is from the gawk-3.1.0.tar.gz released by 
Arnold Robbins, and the test program is also commented as being 
authored by him.

However, there *is* a workaround.  The configure.in already has 
specific bypasses of this test for VMS and BeOS, and it was fairly 
trivial to add "MS-DOS" as another bypass case, which is the "right 
thing" for DJGPP (void argument to getpgrp).

<Snipped>
 >It's better to fix the script and send the patches to the package
 >maintainer.  setpgrp isn't a Posix function, so packages shouldn't
 >assume it exists without checking.

I will do that.  I also found another error in the M4 test he wrote for 
the strtod function, which I will also send to him.

BTW, even though the gawk-3.1.0 package has a pre-built "makefile" and 
other files for "pc" use, including DJGPP, I am trying to see how few 
"special purpose" DJGPP exceptions need to be used to build the package 
starting with configure and the all the latest DJGPP ports.

It's a "configure" learning exercise for me, but I'll pass on anything 
else I find to Arnold as I find them.

Thanks for the advice.

Peter

- Raw text -