Re: RFC: Moving CEDET development to Emacs

David Engster <[email protected]> Wed, 21 Oct 2015 08:09:51 +0200
Newsgroups gmane.emacs.cedet
Message-ID <[email protected]>
Dmitry Gutov writes:
> Hi David,
>
> On 10/20/2015 10:21 PM, David Engster wrote:
>
>> So here's my proposal:
>>
>>    - We move CEDET development to Emacs proper, maybe on its own branch
>>      (might be necessary for times where Eric does not have a copyright
>>      lease from his employer).
>
> Forgive me for repeating myself, but I'd recommend moving to ELPA,
> either the whole CEDET, or most of it (maybe using the "external
> branch" approach).

I have considered that, but decided against it. The main reason is that
ELPA packages are second citizen and cannot be used as dependencies by
other core Emacs packages.

> While there's no hurry in that move itself, you wouldn't lose
> repository history, unlike the case when you move to Emacs, and then
> later move to ELPA.

The problem is that we have two histories: the one in Emacs, and the one
in CEDET. By moving to Emacs, we would keep the one in Emacs and loose
the one in CEDET. If we would like it the other way round, we'd have to
delete the files in Emacs and use subtree merges of our files.

> Depending on the upcoming conventions (which haven't settled), if
> CEDET is one of the ELPA packages that's officially bundled with
> Emacs, it could be a part of the new APIs anyway.

There's still no infrastructure for that bundling, AFAIK. Also, CEDET
has a rather involved build due to the parser generation from the
grammars, which is not supported by ELPA.

-David

------------------------------------------------------------------------------