This is the mail archive of the
mailing list for the GCC project.
Re: eliminating debug info for symbols/types in unexpanded inlines
> I know Cary Coutant is working on a new duplicate header file
> elimination scheme, but as far as I know it isn't available yet. ?This
> patch is a much simpler solution we can use in the meantime. ?This patch
> has been in use for about a year inside Cisco without any problems so
> far, though Cisco is using an older gcc release, so it isn't a complete
> proof that this is safe with gcc mainline.
I hope to have the dwarf-4 branch ready to integrate into mainline in
a week or two. The algorithm for computing a signature was still in
flux in the DWARF workgroup discussions, but we finally nailed it down
and the proposal was accepted for DWARF 4 this week. Right now, I'm
investigating a bug that I think appeared with my last merge from
mainline; with a fix for that, with a patch to make the final
signature calculation match the DWARF spec, and with confirmation from
the editor of the DWARF spec for the new tag, attribute, and form
codes, it should be ready for review.
Offhand, though, I don't see any reason why what you're doing in this
patch would interfere in any way with comdat type info. It looks to me
like your patch is eliminating debug info that isn't at all needed,
while mine is only eliminating duplicate info. It certainly looks
worthwhile to me.