Re: ELPA policy

[Dmitry Gutov <dgutov@yandex.ru>]

On 11/10/2015 10:45 PM, David Engster wrote:

> As you know, we will drop our upstream repository. During the last
> merge, I will also have to discuss with Eric what we can do to narrow
> CEDET's scope.
> I agree with you that semantic-symref is probably one of
> the packages that should be replaced with something more general.

In the end, I would leave only the "backend" part of semantic-symref 
(calling the tools + piping the results through Semantic), and use it in 
the Semantic xref backend. So the results will be displayed by xref.

Alas, xref is not up to feature parity yet (cannot group hits by 
containing functions, no checkboxes, simplistic "rename" command), and 
I'm having difficulties deciding on the changes to the xref data 
structures that would make that possible, as well as keep allowing 
certain other uses of xref, such as project-find-regexp.

I'll write that up in a separate thread later this week.

> But as
> long as it's there, I'd appreciate if you could leave whatever support
> you have coded for it.

Just to be clear, there's no "support" for semantic-symref in xref yet. 
We just use some low-level pieces of it as a "fast Grep".

> I'd also be happy to drop the completion GUI stuff, BTW. I've repeatedly
> said that company should be in core for that.

Also not exactly necessary. Semantic should make its 
completion-at-point-functions element the main entry point for 
completion, and Company will pick it up. You can even have argument list 
annotations, if you add the :annotation-function property.

I do want to have Company more widely available, but probably not in 
25.1. There are quite a few things I wanted to do with it first, and 
most are still untouched.