Mail Archives: djgpp-workers/2001/12/26/07:45:16
Hello.
Eli Zaretskii wrote:
> I don't think GDB is the culprit here: it supported stabs since v4.18,
> perhaps even before that. I'm quite sure stabs work for me with GCC
> 3.0.2 as well, but I cannot test that at the moment.
>
> Do -ggdb and -gstabs give you the same results?
No. I tried the following and they produced debug info that makes 'l main'
work in gdb:
-gstabs
-gstabs+
-gstabs+3
I tried the following and they produced debug info that makes 'l main'
fail in gdb:
-ggdb
-ggdb3
> If nothing else helps, could you make the offending program (source
> and unstripped binary) available for download somewhere, so I could
> look into it?
The sources that originally showed the problem are my current zippo work
area and it's not ready for a check-in and I don't have enough web space
to upload an archive of it. Anyway, I built some binaries of it that you
can download here:
http://www.phekda.freeserve.co.uk/richdawe/djgpp/gdb/gdb-problem.zip
The binaries are in zippo/ in the archive, with a log of the commands used
to build the sources.
I also tried to reproduce the problem using a simple case. The archive
contains this simple case in simple/. I built objects with -gstabs+ and
-ggdb3 and linked as plain objects and as libraries. I thought perhaps
binutils had stomped on the objects, when constructing libzippo, as part
of the zippo build process. Unfortunately the simple case shows no
difference between -gstabs+ and -ggdb3.
The stock zippo 0.1.5 sources show the same problem as my zippo work area.
You can get these here:
http://www.phekda.freeserve.co.uk/richdawe/zippo/#dl
Please let me know if you need anything else.
Thanks, bye, Rich =]
--
Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]
- Raw text -