X-Authentication-Warning:	delorie.com: mailnull set sender to djgpp-workers-bounces using -f
Date:	Tue, 1 Jan 2002 13:22:50 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	Andrew Cottrell <acottrel AT ihug DOT com DOT au>
cc:	Charles Sandmann <sandmann AT clio DOT rice DOT edu>,
	Andris Pavenis <pavenis AT lanet DOT lv>, Robert Hoehne <robert DOT hoehne AT gmx DOT net>,
	"Salvador Eduardo Tropea (SET)" <salvador AT inti DOT gov DOT ar>,
	djgpp-workers AT delorie DOT com
Subject:	Re: NTVDM Error and TVision.
In-Reply-To:	<001301c1929e$5f4e9c80$0102a8c0@acceleron>
Message-ID:	<Pine.SUN.3.91.1020101131542.22865B-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Tue, 1 Jan 2002, Andrew Cottrell wrote:

> I am trying to find why the Rhide CVS source causes a NTVDM error if I exit
> Rhide when it is in a window under XP.

What exactly is the NTVDM error?  Can you post all the details about what 
NTVDM does or prints?

> 1) In the source above if I uncomment the "destroy(App);" line of change the
> #if 0 to #if 1 to go back to the original code I get a NTVDM error during
> the exit stage if I exit the demo if it is in a window.

Can you trace into the code of `destroy' (perhaps after rebuilding 
TVision with debug info)?  It's hard to reason about possible problems in 
a C++ method one knows nothing about.

> 4) In tvision\include\tv\object.h the definition of destroy is:
>     inline void destroy( TObject *o )
>     {
>       if (o)
>       {
>         o->shutDown();
>         delete o;
>       }
>     }

Does the crash happen in the shutDown method ir when the object is 
deleted in "delete o;"?

> I am stumped as to where to look next as it now looks like a needle in a
> haystack problem and I am not familiar with this haystack.

The only method I know of is to run this under a debugger or add 
debugging printf statements, until you find the offending piece of code.  
When the problem is isolated to a few lines that don't involve overloaded 
operators, it will be possible to try to guess why does that code crash 
on XP but not elsewhere.

- Raw text -