Re: [PATCH] Add value_range_base (w/o equivs)

[Aldy Hernandez <aldyh at redhat dot com>]

On 11/11/18 3:53 AM, Richard Biener wrote:
> On Fri, 9 Nov 2018, Aldy Hernandez wrote:
>
> > On 11/9/18 9:19 AM, Richard Biener wrote:
> > > This adds value_range_base, a base class of class value_range
> > > with all members but the m_equiv one.
> >
> > First of all, thanks so much for doing this!
> >
> > > I have looked into the sole GC user, IPA propagation, and replaced
> > > the value_range use there with value_range_base since it also
> > > asserted the equiv member is always NULL.
> > >
> > > This in turn means I have written down that GC users only can
> > > use value_range_base (and fixed the accessability issue with
> > > adding a bunch of friends).
> >
> > > +
> > >   /* Range of values that can be associated with an SSA_NAME after VRP
> > > -   has executed.  */
> > > -class GTY((for_user)) value_range
> > > +   has executed.  Note you may only use value_range_base with GC memory.
> > > */
> > > +class GTY((for_user)) value_range_base
> > > +{
> >
> > GC users cannot use the derived value_range?  Either way could you document
> > the "why" this is the case above?
>
> I've changed the comment as it was said to be confusing.  The reason is
> the marking isn't implemented.

Ah, I see.  In which case, shouldn't you then remove the GTY() markers 
from the derived class?

/* Note value_range cannot currently be used with GC memory, only
   value_range_base is fully set up for this.  */
class GTY((user)) value_range : public value_range_base