Re: Generalizing find-definition

[Helmut Eller]

On Tue, Dec 16 2014, Dmitry Gutov wrote:

> Ok, sorry. I misunderstood. So, only strings? That's not bad, the
> string can be that "imprecise" value, and a property will signal that
> the backend will have to find out what the identifier is, maybe using
> the value of that property (likely marker), or the current value of
> point.

Yes, that's the idea.

>> There's still the problem of how to put text properties on the result of
>> completing-read.
>
> Do we really need to? I think the identifier-completion-table can
> manage to contain only unambiguous entries.
>
> Even if we could carry text properties through `completing-read', the
> user won't see them when choosing, so all strings in the table will
> have to be different anyway.

What I'm saying is that my original proposal of having a
read-identifier-form-minibuffer make sense as it allows backends to add
text properties more easily than a completion table.  And calling
completing-read isn't difficult for the backends if they already have
the completion table.

Helmut