Date: Tue, 31 Mar 1998 21:55:26 -0800 (PST)
Message-Id: <199804010555.VAA22797@adit.ap.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
From: Nate Eldredge <eldredge AT ap DOT net>
Subject: Re: Profiling code
Cc: djgpp-workers AT delorie DOT com
Precedence: bulk

At 11:01  3/31/1998 +0300, Eli Zaretskii wrote:
>
>On Mon, 30 Mar 1998, gareth Bushell wrote:
>
>> I am trying to optimize some code so I thought I'd use gprof.
>> the problem is that when I run the code the table output by 
>> gprof contains an entry _mono_putc. I don't use this function but it is 
>> taking up almost 70% of the execution time.
>
>This is a FAQ.  Section 13.2 of the DJGPP FAQ list explains that there
>is a bug in a library function used by programs compiled with -pg
>which causes this (and can even crash profiled programs in some
>configurations).  Get a patched version of the library from Tom
>Demmer's site and relink your program against it.

I'm just wondering: Is there a reason why we can't (or don't) update the
DJDEV package between releases? Other packages (like, say, Fileutils) are
patched to fix egregious bugs and re-released without changing the version
number. Tom Demmer's patch site is great, but it seems silly to have to tell
people who come up against known bugs to go and fetch a non-standard library
from a non-standard place. I assume this is a matter of software philosophy,
but IMHO known bugs should be *fixed*.

Just my $0.02

Nate Eldredge
eldredge AT ap DOT net