Date: Sun, 3 May 1998 17:59:33 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: HANRIGOU Philippe <HANRIGOU AT cgste DOT mq>
cc: DJGPP <djgpp AT delorie DOT com>
Subject: Re: Problem with bash
In-Reply-To: <3548A39B.6CE0@CGSTE.MQ>
Message-ID: <Pine.SUN.3.91.980503175732.19706T-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Thu, 30 Apr 1998, HANRIGOU Philippe wrote:

> And I compiled tminit setting -D__DPMI_YIELD_BUG. Unfortunately, screen
> output does not repair the bug. Everything's the same (freezing on a
> __dpmi_yield call)!

I give up.  It bothers the heck out of me that __dpmi_yield is hanging
on some machines but not on others, since it means that some programs
might stop working on those machines (__dpmi_yield is called
internally by some library functions which use busy-waiting loops, and
by some popular programs like Info, Emacs and RHIDE).  But if you have
no other way, I guess disabling __dpmi_yield in Bash is the way to go.

I honestly hope that this problem only happens in Bash due to its active 
messing with selectors and such when it runs other DJGPP programs.  (The 
fact that it only hangs after running a DJGPP program seems to point that 
way.)  If it is true, it at least won't affect other programs which use 
__dpmi_yield.

I suggest to make the modified Bash binary available somewhere, for
those who bump across this problem in the future.

Thanks for your efforts in investigating this riddle.