coq-club AT inria.fr
Subject: The Coq mailing list
List archive
- From: Tom Prince <tom.prince AT ualberta.net>
- To: Michael Shulman <mshulman AT ucsd.edu>
- Cc: Coq Club <coq-club AT inria.fr>
- Subject: Re: [Coq-Club] typeclass methods in hints
- Date: Sat, 24 Dec 2011 12:47:02 -0700
On Fri, 23 Dec 2011 10:53:28 -0800, Michael Shulman
<mshulman AT ucsd.edu>
wrote:
> >> Hint Resolve @some_method.
> > You probably want to add ': typeclass_instances' here.
> Why?
Typeclass resolution uses the typeclass_instances hint db.
> >> I am told "some_method cannot be used as a hint". Is there a way to
> >> do this?
> >
> > Is is that some_method cannot be used as a hint by *auto*?
>
> I'm not sure -- are you asking me? This shouldn't cause the "Hint
> Resolve" command to fail, should it? The manual says that if a hint
> will only be used by eauto, then "Hint Resolve" just prints a warning.
I wasn't sure what you were seeing. It looks like this caused by having
a bad head contatnt. Either a forall, or bound variable, it looks like,
from the source.
> > There are a number of other things one can do. The simplest method is to
> > use
> > ':>' instead of ':' when declaring the projection ... this automatically
> > declares a hint for the projection.
>
> The method in question isn't a projection declared as part of the
> typeclass definition, but a function defined later which happens to
> take an argument belonging to a typeclass. (Is the word "method"
> wrong in this context?)
I don't know if there is standard terminology for this.
Tom
- Re: [Coq-Club] typeclass methods in hints, Tom Prince
- Re: [Coq-Club] typeclass methods in hints,
Michael Shulman
- Re: [Coq-Club] typeclass methods in hints, Tom Prince
- Re: [Coq-Club] typeclass methods in hints,
Michael Shulman
Archive powered by MhonArc 2.6.16.