Mail Archives: djgpp/1997/10/27/05:34:19
At 11.17 1997-10-27 +0200, you wrote:
>
>On Mon, 27 Oct 1997, Peter Palotas wrote:
>
>> Stripping debug information and whatever, using 'strip' often makes my
>> programs give me a SIGSEGV. Is this understandable, or is it a bug that
>> should be looked into?
>
>It's a bug. Stripping debug info should make no difference for how does
>a program work (or doesn't work). Please describe the cases where it
>happens in more detail. In particular, what version of Binutils is that
>and how do you strip it (are you stripping the .exe or the raw COFF file)?
Okay, now I have reproduced an error that only occurs using strip, before
using djp on a file. I think the bugs I've encountered always came when
using strip in a combination with djp, however I'm not sure of this.
Anyway, just a simple file:
test.cc:
int main(void)
{
return 0;
}
Compiling it with no extra flags or anything, just:
gcc -c test.cc -o test.o
gcc -o test.exe test.o
And the program test.exe runs fine (although it doesn't do anything but
anyway).
Now if run djp on it, the program runs fine, and running strip on the
program after that, and the program still runs fine (although the
executable got larger so that made no sense).
However, running strip on the compiled .exe file, and THEN running djp on
the program and the program gives me a GPF.
I have version 2.8.1 of the binutils, version 1.04 of djp. I don't know
where the error lies, but I'm sure there is one!
(Sorry about the last similar post. The allegro things weren't even
neccessary to reproduce the error, and the commandlines I wrote to compile
the file would have resulted in an error since I didn't include -lalleg.)
-- Peter Palotas alias Blizzar -- blizzar AT hem1 DOT passagen DOT se --
*****************************************************
* A brief description of DJGPP: *
* NEVER BEFORE HAS SO FEW DONE SO MUCH FOR SO MANY! *
*****************************************************
- Raw text -