Re: whodunnit? info sources now returns duplicates

[Jim Blandy <jimb at redhat dot com>]

Daniel Jacobowitz <drow@false.org> writes:
> On Wed, Sep 15, 2004 at 11:14:56AM +0100, David Lecomber wrote:
> > This is a regression bug -- the version I have tested this are 
> > from CVS on 2004-08-17, and 2004-06-22.  
> > 
> > Any test program:
> > 
> > foo/test.c
> > 
> > int main() {
> > 	int x;
> > }
> > 
> > now compile - but in the directory above foo.
> > 
> > gcc -g foo/test.c
> > 
> > run gdb ./a.out 
> > 
> > For 20040622, only foo/test.c is listed in info sources (correct)
> > For 20040817 (and recent 20040913) foo/test.c and test.c are listed..
> > (incorrect)
> > 
> > When you add a breakpoint to main, and do info sources, the duplicate
> > vanishes..
> > 
> > Any takers?
> 
> Joel's patch to dwarf2read to create symtabs based on the line table
> also creates symtabs if the name of the compilation unit and the name
> in the line table don't match.  I noticed this last week and don't know
> what to do about it.
> 
> One workaround is a patch I whipped together to display the fullname in
> info sources output (arguably more useful anyway).  The two will have
> the same fullnames and we already have code to suppress printing the
> duplicates.  Attached.
> 
> What do you think?

Looks good to me.  If you've tested it, go ahead and commit.