https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123212
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I don't see why you throw -O0 into the same bucket as -Og.
For -O0 compile time is really important. For -Og, the debuggability of code
is much more important, not really short compile time (after all, e.g.
var-tracing is enabled at -Og and that is sometimes very expensive). As
whether using jump tables or series of if switches or their combination doesn't
affect debuggability (while optimization of switches into table lookups does,
because one can't then step into the stmts which initialize those), I think we
should be able to choose whatever results in reasonable code generation for
-Og.
