[m-rev.] for review: build HLDS by item kind

Zoltan Somogyi zoltan.somogyi at runbox.com
Thu Oct 29 15:00:32 AEDT 2015



On Thu, 29 Oct 2015 14:38:28 +1100 (AEDT), Julien Fischer <jfischer at opturion.com> wrote:
> > In fact, the immediate motivation for this diff is to make such a change.
> > This is a fix for Mantis bug 318, which requires adding abstract and/or
> > non-foreign definitions for types because foreign definitions for those types
> > *regardless* of their relative order in the source code.
> 
> That last sentence doesn't make sense.

It does with s/because/before/. Done.

> > diff --git a/compiler/add_pragma.m b/compiler/add_pragma.m
> > index e871b09..1c23133 100644
> > --- a/compiler/add_pragma.m
> > +++ b/compiler/add_pragma.m
> > @@ -20,13 +20,59 @@
> >
> >  %-----------------------------------------------------------------------------%
> > 
> > -:- pred add_pass_2_pragma(item_pragma_info::in,
> > -    item_mercury_status::in, module_info::in, module_info::out,
> > +:- inst item_pragma_info(I) for item_pragma_info/0
> > +    --->    item_pragma_info(I, ground, ground, ground).
> > +
> > +:- inst pragma_pass_2 for pragma_type/0
> 
> If you are going to start using typed insts in the compiler, then you should
> update the configure script's up-to-date check to ensure that they are supported
> by a bootstrap compiler.

Right. Since I think it is too early for that, I commented out
the "for pragma_type/0" part for now, both here and in make_hlds.m.

> The change looks fine otherwise.

Thanks for that. I made the other changes you recommended,
and ran a spellcheck on make_hlds_passes.m for good measure :-(

Zoltan.





More information about the reviews mailing list