From: Nate Eldredge <neldredge AT hmc DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Newbie: HELP! It won't work!!!!
Date: Mon, 04 Oct 1999 16:17:48 -0700
Organization: Harvey Mudd College
Lines: 42
Message-ID: <37F9359C.DDC426F2@hmc.edu>
References: <37f92508 DOT 181091 AT news DOT image DOT dk>
NNTP-Posting-Host: mercury.st.hmc.edu
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: nntp1.interworld.net 939079109 16063 134.173.45.219 (4 Oct 1999 23:18:29 GMT)
X-Complaints-To: usenet AT nntp1 DOT interworld DOT net
NNTP-Posting-Date: 4 Oct 1999 23:18:29 GMT
X-Mailer: Mozilla 4.61 [en] (X11; U; Linux 2.2.13pre12 i586)
X-Accept-Language: en
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Sune Nielsen wrote:
> 
> Please don't rip my head off for posting this rather longish email,
> but I seriously need help :(
> 
> I've recently upgraded DJGPP from version 2.01 to 2.02, and now, for
> some reason of which I have absolutely no idea, it compiles my code
> into completely rubbish. Well... it actually compiles rather fine, but
> when I try to run the program it tells me something like
> 
> <snip>----------------------
> Call frame traceback:
> in function malloc+99
> in function _flsbuf+76
> etc...
> <snip>---------------------
> 
> I mean "malloc"... I donīt even use that function (itīs supposed to be
> C++ ;)
> Have I totally missed some point or... is my code really that buggy? I
> am positive that it worked under v2.01!

It's very likely to be a bug in your code.  v2.02 introduced a new
implementation of malloc (and C++'s `new' and `delete' are built on top
of the `malloc' family) that happens to be much less tolerant of
erroneous behavior like overrunning arrays and multiple freeing.

Note that your traceback doesn't necessarily point to the buggy code;
often something like an overrun will corrupt malloc's internal state and
cause it to crash on a later, valid operation.

Anyway, you should start hunting the bug.  An automated malloc checking
tool is often helpful in cases like these.  I've written one called YAMD
available from

http://www3.hmc.edu/~neldredge/yamd/

There are others on Simtelnet if you prefer.
-- 

Nate Eldredge
neldredge AT hmc DOT edu