Date:	Sat, 14 Jan 1995 12:01:31 +0900
From:	Stephen Turnbull <turnbull AT shako DOT sk DOT tsukuba DOT ac DOT jp>
To:	drupp AT cs DOT washington DOT edu
Cc:	djgpp AT sun DOT soe DOT clarkson DOT edu
Subject:	gcc = gcc -O2 ?

DRUPP:

   I think the whole reason for asking this question of the PC folks is that
   the Gnu people want to change the default for all platforms to -O2.

Oh.  Well, as I said (not particularly clearly) in another message,
consistency is highest on my list.  Another good reason for
consistency just occurred to me: how about people who use cross
compilers?  Yes, they only have to get it right once in the
configuration, but suppose that they publish their work to the 'net
*before* getting it right?

Also, I definitely vote for orthogonality of -g and -O; Bob Babcock's
reasoning is telling.

I don't know if it's still true, but the installation docs for GNU
Emacs used to say that -O2 generates more robust code than -O0, as
that was the most common compilation mode, especially for production
executables.  Thus -O2 compiled code is more heavily field tested.
    I'm not sure I believe the logic this implies (given GCC's
multipass algorithm), but that's what RMS said.
    --Steve

- Raw text -