Re: [pph] Support for references to symbols in other PPH images (issue4873051)

[Diego Novillo <dnovillo at google dot com>]

On 08/16/2011 02:04 PM, Gabriel Charette wrote:

> Setting *cache_ix to (unsigned) -1 used to be a "hack" (with a comment
> explaining it which was removed just below), to avoid a warning about
> it being unset in a branch, but that branch was only PPH_RECORD_END,
> in which case it didn't matter.

Yeah, I remember, but I set them to -1 now as an actual return value now.

> Now that we actually check in pph_cache_get that include_ix ==
> (unsigned) -1, I'm not so sure this is good... while (unsigned) -1 is
> a very large unsigned int, it's not impossible to have an actual cache
> entry with that number...

Oh, no.  If we ever need 4B entries in the cache, we will die much 
sooner than this.

> > @@ -457,13 +478,13 @@ pph_in_cxx_binding_1 (pph_stream *stream)
> >    cxx_binding *cb;
> >    tree value, type;
> >    enum pph_record_marker marker;
> > -  unsigned ix;
> > +  unsigned ix, include_ix;
>
> Sometimes you call the local variable "include_ix"...
>
> > @@ -505,13 +526,13 @@ pph_in_class_binding (pph_stream *stream)
> >   {
> >    cp_class_binding *cb;
> >    enum pph_record_marker marker;
> > -  unsigned ix;
> > +  unsigned image_ix, ix;
>
> ... and sometimes image_ix: consistency would be nice, although not necessary...

Thanks, I'll fix.  I'm horribly inconsistent with naming schemes (blame 
my evil twin).


Diego.