Mail Archives: djgpp/2004/11/24/10:49:52
> Date: Wed, 24 Nov 2004 09:59:15 -0500
> From: Pityocamptes AT aol DOT com
>
> I am not very familiar with these procedures (taking assembly for a class - using NASM). Could you explain in some detail what you mean by "In a nutshell, you need to build NASM with
> debugging information and post here more info about where the program crashes." Thanks.
I described that in more detail a day or two ago, but here it is
again, more detailed:
Step 1: Download NASM sources.
Step 2: Find the way to build it with debugging information, i.e. with
the -g switch to GCC. It could be that the Makefile that builds NASM
has a special target for that, or perhaps you need to edit the
Makefile a bit. (I don't know exactly how to do this, but perhaps
someone else will tell you the specifics.)
Step 3: Run the NASM you built and make it crash.
Step 4: With the crash message still on the screen, run symify.exe or
bfdsymify.exe to get the source file names and line numbers appended
to the addresses. (Section 9.3 of the DJGPP FAQ has more about symify
and its usage.) That will produce a traceback of function calls that
led to the crash. Post here all that information, exactly as it is
displayed on the screen.
Step 5: Look in the NASM sources for the code fragment near the place
where it crashed. The file name and line number where it crashed are
on the first (the topmost) line of the traceback. Post the code
fragment where it crashed here.
Step 6: Wait for someone to come up with ideas why it could crash,
given the info you posted. When someone responds, they will have some
ideas and some things for you to check or to do. Follow their advice.
HTH
- Raw text -