Sender:	<solyom AT bmeik DOT eik DOT bme DOT hu>
Date:	Thu, 23 Dec 1993 10:49:20 +0100
From:	solyom AT bmeik DOT eik DOT bme DOT hu
Reply-To:	SOLYOM AT HUBME51 DOT bitnet
To:	DJGPP AT SUN DOT SOE DOT CLARKSON DOT EDU
Subject:	Re: Re: to SERIOUS problem

>You have just altered the arrangement of the stack and caused the problem to
>go elsewhere.

No. The program code has changed. I verified it before I run it. Besides
when I run the program, the original code (before removing the unused 
arguments) was perfectly sane, just at the stat() call it used 
	push bx
and at the Fopen call it used
	push [bp-120] 	(or similar)
and the bx register is not changed between the two calls.

>I wonder if you have a function of the same name as one that is used by stat
>or calloc? rather a wild guess.  Try calling Fopen on the first line
>of the function.  Are you sure there is not a subtle bug in Fopen()
>which is only appearing now?

I do not have anything whose name starts with a lowercase letter as
any function stat()  uses do.

>We had a problem with double precision arrays on the stack which was
>due to incorrect code being generated by the optimiser. I suggest you
>reproduce the bug and then try with and without optimisation.

I never tried to compile it without optimization, so perhaps it can help.
Thanks.

But the most mystical thing is that with 1.10 I could eliminate the
problem by changing the linker!

				Andras S˘lyom
				solyom AT bmeik DOT eik DOT bme DOT hu

- Raw text -