Re: master 12578d6: Change how (declare (modes store the data

[Lars Ingebrigtsen]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> BTW, maybe the generalization is to replace the "completion-predicate"
> and the "command-modes" into a function that returns a *score*, with
> some convention where a score lower than X means the command can't
> possibly work, and score lower than Y means the command is too unlikely
> to be useful to be worth listing, a score higher than Y means that the
> command has a special affinity with the current situation (like "written
> for the current mode"), ...

That sounds useful, but it sounds like a really difficult thing to
relate to.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no