delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp-workers/1999/03/07/06:29:55

Date: Sun, 7 Mar 1999 13:27:17 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: "Mark E." <snowball3 AT usa DOT net>
cc: djgpp-workers AT delorie DOT com
Subject: Re: Bash 2.03 updated
In-Reply-To: <199903041628.QAA94472@out1.ibm.net>
Message-ID: <Pine.SUN.3.91.990307132653.26544K-100000@is>
MIME-Version: 1.0
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

On Thu, 4 Mar 1999, Mark E. wrote:

> I've update the packages to fix the crash that was introduced in the 
> 2.02.1 -> 2.03 upgrade.

Almost there, I think.

I have two gripes and two observations.

Gripe no.1: When I set PATH_SEPERATOR=:, it seems like the behavior is
unaffected by the value of PATH_EXPAND.  To wit: on a DOS machine, the
subsidiary programs invoked by Bash, including Bash itself and Make,
get the PATH in the /dev/c/foo:/dev/d/bar form, which of course causes
many failures (Make fails to run commands, Bash doesn't find external
programs, etc.).  I tried the same on Windows, and there PATH is
*always* expanded, no matter how PATH_EXPAND is set (even if I set
LFN=n).  I'm not sure what's going on here.

Gripe no.2: Ctrl-C causes Bash to abort with a traceback, as if I
pressed Ctrl-BREAK.  This is in contrast to usual operation of DJGPP
programs, where Ctrl-C just exits with no traceback.  One problem with
this, besides the aesthetics, is that the traceback causes important
information to scroll off the screen.  Could we have the usual
behavior back, please?

Observation no.1: When Bash is aborted with Ctrl-C, the registers'
dump prints strange data for the FS selector: it has a very low base
address (e.g. 00004110) and a 64KB limit (0000ffff).  Is this normal,
and if so, how does this happen?

Observation no.2: This version of Bash is considerably slower than
1.14.7 for the same script.  I measured the time used by both to run a
typical configure script, and the old version is almost twice as fast
as the new one.  In particular, the new version sometimes ``thinks''
for prolonged periods of time (mostly about 1 second, but I saw it
stop for 5 seconds a couple of times, which is A LOT for a P166 in
plain DOS mode).  These pauses seem to be not entirely deterministic,
so maybe they have some kind of garbage collection going on (I don't
have the sources to look).  Pretty annoying, if you ask me.  It would
be swell, if you could look into this problem.

Can somebody compare these two versions on Unix, so we could at least
know if the problem is specific to the DJGPP port or not?

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019