Re: Why does sparc build for you? or, why are make check results reported when it hasn't bootstrapped since May 5?

[Eric Botcazou <ebotcazou at libertysurf dot fr>]

> > This is the faulty patch:
> >
> > 2003-04-16  J"orn Rennecke <joern.rennecke@superh.com>
> >
> >         Re-apply this patch:
> >
> >         2002-05-16  Dale Johannesen  <dalej@apple.com>
> >           * combine.c (cant_combine_insn_p):  Reenable combinations
> >           involving hard regs unless CLASS_LIKELY_SPILLED_P.
>
> What makes you think this patch is faulty?

It seems that the combiner aggressively optimizes the code without taking 
into account or updating the liveness informations (see attached files). The 
problem was first reported on 04/17 by Kaveh so it wasn't difficult to 
pinpoint your patch. Reverting it fixes the bootstrap failure.

> It's a fairly straightforward patch.  it was originally pulled because
> of an x86 Ada bootstrap failure, but this patch is quite important for
> processors that pass arguments in registers to get good code, and further
> investigation of the Ada failure showed that it was due to a bug in a
> splitter in i386.md , which was easily fixed.

It may have uncovered a bug elsewhere then.

-- 
Eric Botcazou

Attachment: natFloat.cc.20.life
Description: Text document

Attachment: natFloat.cc.21.combine
Description: Text document