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: N8TM at aol dot com
- Subject: Re: Patch checked in for K&R stage1 cc on obtabs.c
- From: craig at jcb-sc dot com
- Date: 26 May 1999 03:17:35 -0000
- Cc: ghazi at caip dot rutgers dot edu, oliva at dcc dot unicamp dot br, egcs-patches at egcs dot cygnus dot com, law at cygnus dot com
- Cc: craig at jcb-sc dot com
- References: <914f57fd.247cbcef@aol.com>
>In a message dated 5/25/99 8:56:49 AM Pacific Daylight Time,
>ghazi@caip.rutgers.edu writes:
>
>> Why would we want g77 in K&R style? I feel its reasonable to
>> not only assume ANSI C but also GNU C for the language front ends
>> other than C. It would certainly make writing and supporting them
>> easier.
>It doesn't seem like there's much chance of writing g77 so that it will build
>directly on sub-ANSI/ISO compilers, nor would extra effort be justified.
The last part might be right, but I don't know of any reason why I can't
fairly easily ensure that g77 will build on K&R C compilers as part of
the rewrite.
As I said, the back end already provides most, if not all, of the important
facilities.
I believe it's the case that, right now, to "fix" g77's K&R bugs, all
that is needed is to PROTO-ize a whole bunch of definitions and
declarations (presumably a Perl or Guile script could do that right quick)
*and* -- the harder, and less-worthwhile part -- move all the
statement-expressions in target.h into target.c as proper functions.
(target.h and target.c happen to be the primary "targets" -- naming
something of a coincidence -- of the rewrite. If the rewrite accomplished
nothing else, rewriting those, and the related code in the front end,
would solve most/all of the cross-compiling problems, make it much easier
to support intrinsics in PARAMETER statements, and probably make `-r8'
and `-i8' easy to support as well.)
tq vm, (burley)