This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] Enable autoinlining for size at -O1+
Daniel Jacobowitz <drow@false.org> writes:
> Why at -O1? Even with a somewhat smarter debugger than GDB, this will
> interfere with debugging; I'd have expected -O2.
I would like to also argue against it at -O1. I often use -O1 with -g
to debug; the -O0 code is way too stupid but the -O2 code tends
towards obscurity; -O1 is a good compromise when debugging. Inlining
functions at such a low optimization level means that you can no
longer break on those trivial functions.
Enabling it at -O2 (and -Os) is fine with me, though.
Perhaps we could further define -O1 as "optimizations appropriate for
the development/debug cycle"? That implies fast turnaround with
reasonable code, without resulting in obscure code or unexpected
side-effects in gdb.