info sources regression with dwarf2/stabs

Bob Rossi bob@brasko.net
Tue Jun 15 13:51:00 GMT 2004


On Fri, Jun 11, 2004 at 10:14:45PM -0400, Bob Rossi wrote:
> On Fri, Jun 11, 2004 at 02:35:10PM -0400, Bob Rossi wrote:
> > On Fri, Jun 11, 2004 at 02:29:16PM -0400, Bob Rossi wrote:
> > > Hi,
> > > 
> > > A user of -file-list-exec-source-files noticed that this command does
> > > not initially list the files that are in a library. For example,
> > > 
> > > I have
> > >    $ ls
> > >    Makefile  lib.c  lib.h  lib.o  libmilib.a  main  main.c
> > > 
> > > And
> > >    $ cat Makefile
> > >    all:
> > >       gcc -c -g -Wall lib.c
> > >       ar cru libmilib.a lib.o
> > >       gcc -g -Wall main.c -o main -L. -lmilib -I.
> > > 
> > > When I do -file-list-exec-source-files originally, it doesn't show
> > > lib.c, after I do 'list lib.c:1' it will then show the filename. Also, I
> > > believe it shows all of the filenames in the library. 
> > > 
> > > My question is, is there any way to get this funcionality from the get
> > > go? Load the psymtabs for all the libraries?
> > > 
> > > Or is there a GDB command that forces loading all symbols into the
> > > partial symbol table at start?
> > 
> > Unfortunatly, I forgot to add one thing. The files from all libraries do
> > show up when the stabs debugging format is used. It seems like this is a
> > dwarf2 problem. Anyone know if this is a bug?
> > 
> > Or something maybe I could look into?
> 
> Finally, this is also a problem with the 'info sources' command. Does
> that make the problem a little more interesting to anyone? :)

Ping, any ideas?

Thanks,
Bob Rossi



More information about the Gdb mailing list