https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85640
--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Anyway, the problem discussed in #c2 to #c4 doesn't seem to exist in 10/trunk anymore, since: r10-2806-gdf7d46d925c7baca7bf9961aee900876d8aef225 And at least some of the changes that grew size are desirable for speed (and -O3 is not size optimization), so is there anything we want to do about this PR?
