This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Re: Stripped files
- From: Daniel Jacobowitz <drow at false dot org>
- To: Stephen & Linda Smith <ischis2 at cox dot net>
- Cc: gdb at sourceware dot org
- Date: Tue, 9 Jan 2007 21:52:42 -0500
- Subject: Re: Stripped files
- References: <45A44256.1070004@cox.net>
On Tue, Jan 09, 2007 at 06:33:10PM -0700, Stephen & Linda Smith wrote:
> Am I breaking any assumptions in the GDB code if the code running on a
> remote target is stripped and the code that GDB loads on the host has
> the symbolic information? In other words is GDB likely to get confused
> by this situation?
No, this is normal.
> If the situation in the above paragraph isn't a problem, would it be a
> good idea for GDB to sniff for the file with the symbolic information
> via an default extension. For example the file that would be on the
> remote target would be something like libfoo.so and the file that that
> gdb would look for first on the host system would be
> libfoo.so-withsymbols and then fi that didn't exist it would look for
> libfoo.so.
>
> What do you think of that?
The solution we've used for this before is .gnu_debuglink (see the GDB
and binutils manuals for more about that). You can also use
sysroot / solib-search-path if the debug symbol file has the same name as the
target file.
--
Daniel Jacobowitz
CodeSourcery