[m-rev.] for review: don't allow nondefault mode functions in terms

[Mark Brown]

On Sat, Oct 31, 2015 at 8:33 PM, Mark Brown <mark at mercurylang.org> wrote:
> On Sat, Oct 31, 2015 at 6:10 PM, Zoltan Somogyi <zoltan.somogyi at runbox.com> wrote:
>> I am not sure about a flat "can't be resolved", but I certainly agree with
>> "can't be resolved without massive pain".

Hrm, I see what you mean.  :-(

>
> I haven't bootchecked the following diff yet, which I will be doing
> shortly. Does anyone see any problem in principle?

Wrong fix. See Fergus' comments about inst_is_ground here:

http://www.mercurylang.org/list-archives/developers/2001-August/012128.html

The same point applies to inst_is_unique, and I think some other
similar predicates as well. I haven't reached my pain threshold yet,
though...

Cheers,
Mark.