Alright, makes sense. Since I don't know what you mean by "all compilers," I added a configure test and I will make sure __sparc__ is used everywhere.
In regards to the second point (@function vs. 2), something has been bugging me for a while... Why do the .spec.c files have so much assembly code in them? To me, it looks like most of it could be replaced with an array and some creative use of global variables.
I'm willing to tackle the task of converting the .spec.c files to use as much C as possible, but I can only test it on sparc. Someone else would have to test it on i386 and with native .dlls
Gregg
To: Gregg Mattinson gm138242@opcom-mail.canada.sun.com Cc: wine-devel@winehq.com Subject: Re: __sparc__ not defined with Forte C From: Alexandre Julliard julliard@winehq.com Date: 15 Jul 2002 18:49:04 -0700 User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Common Lisp) MIME-Version: 1.0
Gregg Mattinson gm138242@scot.canada.sun.com writes:
ChangeLog: dlls/ntdll/signal_sparc.c include/basetsd.h include/wine/port.h server/context_sparc.c
- __sparc__ is not defined with Forte C, but __sparc is defined with both
Forte
and GCC. Same goes for __sun__
If __sparc and __sun work for all compilers we should use them everywhere, no need to try to support other forms. Otherwise it should be done with a configure test like we do for __i386__.
- Forte C doesn't like @function. Replaced it with 2.
This needs a configure test. A #ifdef SUNPRO_C is almost always the wrong thing to do.
-- Alexandre Julliard julliard@winehq.com
Gregg Mattinson gm138242@scot.canada.sun.com writes:
In regards to the second point (@function vs. 2), something has been bugging me for a while... Why do the .spec.c files have so much assembly code in them? To me, it looks like most of it could be replaced with an array and some creative use of global variables.
I doubt it. Many things need to have a very specific layout that cannot be done in plain C. If things are done in assembly there's usually a good reason for that.