[Bug bootstrap/51094] [4.7 Regression] Bootstrap failure at revision 181279 on non-ELF targets

[iains at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51094

--- Comment #20 from Iain Sandoe <iains at gcc dot gnu.org> 2011-11-17 17:50:09 UTC ---
(In reply to comment #19)
> > --- Comment #18 from Iain Sandoe <iains at gcc dot gnu.org> 2011-11-15 09:34:00 UTC ---
> > (In reply to comment #16)
> >> > --- Comment #15 from Iain Sandoe <iains at gcc dot gnu.org> 2011-11-14 08:47:16 UTC ---
> >> > Created attachment 25814 [details]
> >> >   --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=25814
> >> > adjustment to final.c
> >> >
> >> > (just a heads up)
> >> > although not a bootstrap issue, this revision also caused a fair number of
> >> > regressions at m64 on i686/powerpc-darwin9.  This is where the m32 target has
> >> > an m64 multilib.
> >> >
> >> > I am testing the patch on *-darwin9 and x86-64-darwin10.
> >
> > The patch fixes those particular fails for me on *-darwin9, 
> > if it works for you then should it be posted - or does Jason want to amend?
> 
> Indeed: the patch is still necessary to get sort-of decent testsuite
> results on i386-pc-solaris2.1[01].

I think the patch needs a tweak -
- the case of largest negative int in the case where HOST_WIDE_INT is long
long.

I wonder if it's simply OK to bail to the original printf for that corner-case,
or whether it's worth handling specially.