This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Patch checked in for K&R stage1 cc on obtabs.c
- To: oliva at dcc dot unicamp dot br
- Subject: Re: Patch checked in for K&R stage1 cc on obtabs.c
- From: craig at jcb-sc dot com
- Date: 25 May 1999 20:07:33 -0000
- Cc: law at cygnus dot com, tromey at cygnus dot com, egcs-patches at egcs dot cygnus dot com
- Cc: craig at jcb-sc dot com
- References: <19622.927650475@upchuck.cygnus.com> <oryaidhy9i.fsf@lua.lbi.dcc.unicamp.br>
>BTW, isn't it time to make `make bootstrap' the default target, and
>require `make all' (or maybe `make onestage', so that `make all' could
>be equivalent to `make bootstrap') for a one-stage build? This would
>allow `configure;make;make install' to ``just work''. Of course this
>may break a lot of developers' scripts, but developers can easily
>handle this change, while users in general don't always figure out
>they need a different installation procedure for gcc.
Huzzah. (Which means, I agree. I've been thinking about the same
thing for awhile.)
In fact, as part of the configury/build rewrite, it'd be nice if
we could somehow have "make all" be the default as usual, and have
its dependencies look something like (pseudo-code):
gcc: stage2/gcc
...
stage2/gcc: stage1/gcc
...
stage1/gcc: ...
...
Then, for people who want to short-circuit that, we could provide
them a configure-time option (*not* a build-time one) that'd do the
trick...somehow:
CC=/usr/bin/cc ./configure --no-bootstrap ...
Or whatever.
The idea here is that restarting an aborted/crashed/interrupted build
shouldn't require special knowledge of where it stopped (no
"make bootstrap2" nonsense), among other things.
tq vm, (burley)