Date: Thu, 10 Apr 1997 19:59:28 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Schuster <Schuster AT eev DOT e-technik DOT uni-erlangen DOT de>
cc: djgpp AT delorie DOT com
Subject: RE: Re: Compiler causes program crash
In-Reply-To: <334BB282.3B03@eev.e-technik.uni-erlangen.de>
Message-ID: <Pine.SUN.3.91.970410195906.3026D-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Wed, 9 Apr 1997, Schuster wrote:

> All my pointers are allocated with the new[] operation. It's a c++
> programm.

Using new[] is not a panacea against bugs.  You can still have
uninitialized pointer, or you might use the templates incorrectly.  A
debugging session should allow you to find the cause for crashes.

> I've also tried this and saw that it mostly crashed after compiling 
> when envoking a container-template.

Then you should inspect the variables and code that is connected with
these templates, and look for a pointer whose value is garbage, or for
an index that is too large.

> My solution (after "some" hours) : I've changed the Paramters of cwsdpmi
> 
> These are now:
> 
> Full name of paging file ("" to disable) ? [c:\cwsdpmi.swp]
> Number of page tables to initially allocate (0=auto) ? [0]
> Minimum application memory desired before 640K paging ? [432Kb]
> Paragraphs of DOS memory to reserve when 640K paging ? [3840]
> Paragraphs of memory for extra CWSDPMI internal heap ? [16192]

I fail to see how these changes could solve your problem.  I think
that you have caused the problem to move to another place (since on
Windows it still crashes).  The 16192 figure is too large: you waste
too much DOS memory without any serious reason (I never saw a case
where you would need to bump that number to more than 256).

Once again, I suggest you debug your program instead of trying to make
the bug go away by fiddling with irrelevant parameters.