This is the mail archive of the cygwin-apps@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Want to be a gcc/binutils maintainer?


 --- egor duda <deo@logos-m.ru> wrote: > Hi!
> 
> Wednesday, 07 November, 2001 Christopher Faylor cgf@redhat.com wrote:
> 
> CF> I got a couple of nibbles about this before.
> 
> CF> Anyone interested in this exciting and challenging responsibility?
> 
> well, i'm interested. that doesn't mean i'm sure i'm capable, though
> :)
> 
> i'm now in process of debugging and fixing regressions in -3_0 and
> -current branches, but, i still haven't looked at -mno-cygwin and
> dll-specific stuff.
> 
> i wonder if we can create a branch in gcc cvs tree to make
> cygwin-specific gcc development process a bit more "public"?
> 
> 

Egor, I think we do need something like that.  I have just uploaded a beta
(binaries, src, and diffs) of mingw gcc with fastcall support for mingw at
SF file release site.
https://sourceforge.net/project/showfiles.php?group_id=2435&release_id=60167
It is for 2.95.3 but the changes have also been tested successfully on gcc
trunk (and 3.0.2) with mingw bootstrap. My problem has been trying to make
the diffs public within the w32 crowd and getting some feedback before
going to gcc-patches. 


> 2001-11-07  Egor Duda  <deo@logos-m.ru>
> 
>         * config/i386/cygwin.h: Strip leading '*' from name, if any.
> 
> Index: gcc/config/i386/cygwin.h
> ===================================================================
> RCS file:
>
/usr/home/deo/cvsroot/cvs-mirrors/sourceware/gcc/gcc/config/i386/cygwin.h,v
> retrieving revision 1.33.2.15
> diff -u -p -2 -r1.33.2.15 cygwin.h
> --- gcc/config/i386/cygwin.h    2001/10/26 04:25:00     1.33.2.15
> +++ gcc/config/i386/cygwin.h    2001/11/06 21:10:16
> @@ -356,4 +356,5 @@ do {                                                 
>                       \
>    const char *_p;                                                      \
>    const char *_name = I386_PE_STRIP_ENCODING (SYMBOL_NAME);            \
> +  if (*_name == '*') _name++;                                          \
>    for (_p = _name; *_p && *_p != '@'; ++_p)                            \
>      ;                                                                  \
> 
> fixes one of testsuite regressions in 3_0 and make it possible to
> build -current natively. I don't know, though, if it's correct way to
> deal with this particular problem.
> 

The mingw patchset referred to above addresses a similar isssue (leading
'+'  in names for fastcall symbols, that have to be converted to
'@' before passing to assembler)  


Danny

http://briefcase.yahoo.com.au - Yahoo! Briefcase
- Manage your files online.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]