[PATCH] Eliminate duplicate dwarf2 data

[Jason Merrill]

>>>>> Daniel Berlin <dberlin@redhat.com> writes:

 >>  > 2. Something i can't quite remember at the moment, but it made me
 >>  > give up trying to do it in the linker entirely. I really wish i
 >>  > could remember what it was. I want to say we were getting ahold of
 >>  > it at the wrong time, like after the labels were gone and it had
 >>  > relocated everything, so that we'd need to rewrite all the info. But
 >>  > i don't remember if that was really the issue or not.

 >> That would be problematic, but I'd be surprised if there wasn't a way to do
 >> it sooner.

 > There isn't, because gas is what is removing our labels, and generating
 > relocs, and then we don't get a hold of it until after relocs are
 > generated.

You mean after they've been relocated?  Don't the stabs optimizations
happen before then?

 > Almost. We can't easily convince bfd to remove pieces of a section, it
 > requires hacks because resetting the section size doesn't work.

How does the stabs optimization do it?

 > In addition, you also still have to fix up all the pubnames

True.  Ideally, we would want to find and remove the pubname chunk that
corresponds to the CU we're removing and remove it as well, though if we
don't we're only wasting space.

BTW, are you planning to use pubnames and aranges in GDB?

Jason