Re: Live range splitting in new allocator

[Richard Henderson <rth at redhat dot com>]

On Mon, Jan 29, 2001 at 12:41:10PM -0800, Linus Torvalds wrote:
> That's my personal opinion too, but gcc has long had the rule that the
> presense or absense of "-g" should not affect code generation, only
> debugging information.

Indeed.  If -g subtly changes the code generated by
"-O99 -fdiscombobulate", it can be a real nightmare
finding a compiler bug.

> And too many people seem to think that "-O2 -g" is a valid
> combination and should be debuggable.

Too many people work on x86, where this happens to work most
of the time.  I find -O2 undebuggable more often than not on
Alpha and IA-64, and that just from the scheduler.

Using -O1 -g is what I prefer when I don't want to wait too
long on an overly slow program.



r~