Re: _ELIX_LEVEL problems

[Jeff Johnston <jjohnstn at redhat dot com>]

Joern RENNECKE wrote:
> Jeff Johnston wrote:
>
> > Aha.  I found the problem.  The $(ELIX_SOURCES) is missing from 
> > lib_a_SOURCES.  This was due to my change recently to fix the problem 
> > with the object prefix being added properly to all objects.  I'll 
> > start checking the other libc directories to ensure that I didn't mess 
> > those up as well.
> >
> > Patch checked in.
>
>
> Yes, the results are for sh-elf look sane now - the only new regressions 
> since Feburary are for new tests, and a timeout in g++.dg/eh/cleanup1.C 
> - it does that semi-randomly depending on load.
> That was also using this patch: 
> http://sourceware.org/ml/newlib/2006/msg00419.html .
> I take it that http://sourceware.org/ml/newlib/2006/msg00426.html 
> basically an approval, so I'm checking the patch in now.

Thanks.

> How did you re-generate libc/ctype/Makefile.in ?

The Makefile.in files without configure.in files (e.g. libc/ctype, 
libc/stdio, ...) are generated in the parent directory that has the 
configure.in.  Thus, from newlib/libc:

  automake --cygnus ctype/Makefile

-- Jeff J.