https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124909
Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[16 regression] |gcc.dg/match-shift-cmp-3
|gcc.dg/match-shift-cmp-3 |fails with -fshort-enums
|fails with -fshort-enums |
--- Comment #2 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
I think I messed up my testing when looking at the gcc-15 output. The
optimization was only added in November 2025, so it could not be happening on
the release branch.
Looking deeper, it looks like we only handle a limited set of types, so for
example, 'signed char' doesn't get optimized either, nor does 'short'.