Re: reverting the gcc 3.2.2 Ada regression

[Joel Sherrill <joel dot sherrill at OARcorp dot com>]

Jakub Jelinek wrote:
> 
> On Fri, Feb 14, 2003 at 12:07:03AM +0100, Gabriel Dos Reis wrote:
> > | What I mentioned as alternative to this patch was adding
> > | $(LIBGCC2_INCLUDES) to one of {,ADA_}INCLUDES{,_FOR_SUBDIR}
> > | variables in ada/Makefile.in (depends on which files actually
> > | needed those includes).
> >
> > I seem to recall you did something to the redhat branch. I would
> > really like we have something consistent.
> 
> I just backed out those 4 TARGET_LIBGCC2_CFLAGS -> LIBGCC2_CFLAGS
> changes in Makefile.in. For gcc-3_2-rhl8-branch rtems is not a priority,
> but I needed something which builds on x86_64 immediately.
> On gcc-3_2-branch, I think those 4 lines should be backed out too,
> plus it needs something to make rtems working again.

I have reverted the patches in question to ada/Makefile.in and 
ada/Make-lang.in.

Double check me but I think adding LIBGCC2_INCLUDES to 
INCLUDES_FOR_SUBDIR is what you are looking for.  That appears
to be the constant used for the .h include path.  
ADA_INCLUDES_FOR_SUBDIR appears to be an Ada include path.

I have started a build to see if I can find an alternative
solution but I will be out of town for the next few days
so might now be able to report back until Tuesday.

>         Jakub

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel@OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985