This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Analysis of remaining xm-host headers
- To: Zack Weinberg <zackw at stanford dot edu>
- Subject: Re: Analysis of remaining xm-host headers
- From: Stan Shebs <shebs at apple dot com>
- Date: Wed, 14 Mar 2001 16:57:58 -0800
- CC: gcc at gcc dot gnu dot org
- References: <20010314004354.E29383@stanford.edu>
Zack Weinberg wrote:
>
> i386/xm-next.h, m68k/xm-next.h, [...]
These are top candidates to be evaporated; I seriously doubt that
the current compiler could be gotten to work with old NeXT systems.
> rs6000/xm-darwin.h: Shuts off HAVE_DESIGNATED_INITIALIZERS,
> complaining of bugs in Apple's version of GCC. Could we maybe be more
> conservative about acceptable version ranges instead?
I agonized over this before adding it. The problem is, believe it
or not, a longstanding limitation in the ObjC frontend - we only
notice because OS X only has cc1obj preinstalled, and uses it for
both C and ObjC compilation (OK, stop laughing). We could write
an autoconf test for this, but I would rather fix the limitation.
(It's not done yet because I need to verify that it doesn't break
any of the ObjC code in OS X and GNUstep.)
BTW, it's really cool that you've done all this work! For software,
cleanliness really is next to godliness.
Stan