This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH 2/2] DWARF: make it possible to emit debug info for declarations only
- From: Pierre-Marie de Rodat <derodat at adacore dot com>
- To: Richard Biener <richard dot guenther at gmail dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Tue, 20 Jun 2017 16:34:20 +0200
- Subject: Re: [PATCH 2/2] DWARF: make it possible to emit debug info for declarations only
- Authentication-results: sourceware.org; auth=none
- References: <email@example.com> <firstname.lastname@example.org> <CAFiYyc3NfKn+3BshKjhwH-U31TkyUEeJ62BezkF0rkbTvshQbQ@mail.gmail.com> <email@example.com> <CAFiYyc3A+NU6d-ZtGqO8NKHa7nwJ0S-Ov4JT2TwMc=6PErf0xQ@mail.gmail.com> <firstname.lastname@example.org> <email@example.com> <CAFiYyc1LS0pps+RURi0OpkgoT9Ha-ToP9JXf6acN3mj3UsRkKw@mail.gmail.com>
On 06/20/2017 02:16 PM, Richard Biener wrote:
Nice. This looks ok.
Great, thank you!
I'm mildy curious about the deecrease of debuginfo size for cc1 -- did
you spot anything obvious there?
Well, the benchmark I exposed was for the whole file size, not just
.debug_info section size. Just to be sure, I compared object files for
both trunk and my patched tree: outside of Ada units, I only get the
gcc/dwarf2out.o: -10168 bytes
gcc/godump.o: +272 bytes
gcc/passes.o: +880 bytes
This diff comes from my changes themselves. I had a quick look at the
same for cc1’s .debug_info: there is the expected evolution, too,
I suspect Fortran wants to do sth similar as Ada for imported modules.
Maybe. I have zero Fortran knowledge, so I’ll let a Fortran expert
decide, if that is fine for you. :-) In any case, the back-end is ready
Pierre-Marie de Rodat