Re: CEDET Merge

[Edward John Steere <edward.steere@gmail.com>]

>> From: Edward John Steere <edward.steere@gmail.com>
>> Date: Thu, 12 Jan 2017 21:32:36 +0200
>> 
>> Hello Everyone,
>> 
>> I've been working on a branch in upstream CEDET which brings it up to
>> date with the changes which have been made in Emacs core.  I started
>> this work with the intention of simplifying the merge process from CEDET
>> upstream into core so that users can benefit from improvements to CEDET
>> more easily.  I found that the divergence between the two versions of
>> CEDET has grown quite a bit and all of my work has been dedicated to
>> getting it to run on Emacs master (26?).
>> 
>> My branch now passes all of CEDETs tests when built against Emacs
>> @512e988.  I'd like to get these changes synchronised with Emacs core
>> and have my branch merged to upstream (i.e. CEDET on source forge).
>> 
>> This could be the first step with the second being a repackaging of
>> upstream so that it only contains the parts which are under active
>> development and can be installed via ELPA.
>> 
>> May I contribute my changes to a branch on the Savannah repository for
>> review/commentary?  (If yes, then It'll take a little bit of time to
>> merge the changes in my branch of CEDET back into Emacs core before I
>> actually push them up.)
>> 
>> Kind regards,
>> 
>> Edward Steere
>
> Does anyone know if the changes were brought into Emacs?  It looks
> like the merge branch got reasonably far:
>
> http://git.savannah.gnu.org/cgit/emacs.git/log/?h=scratch/last-cedet-merge

Hi Charles,

To my knowledge these changes never made it into Emacs.  In fact, a few
things changed since my first email:
 - David Engster got involved (he's been responsible for merges in the
   past if I remember correctly).
 - I was informed that my branch included elements of CEDET which were
   not intended to reach core (COGRE being a good example).
 - We divided up the work of merging tests and merging CEDET and since
   David had experience with this work he went ahead with the core
   elements.

At the same time as we were working on the merge we also started
discussing the possibility of changing the way that CEDET is
distributed.  In particular, we discussed the use of ELPA as a
distribution channel.  I originally argued in favour of this move
because I think that it's widely regarded that upstream CEDET is
difficult to install and setup with Emacs.  I also believe that these
difficulties in installing and setting it up have hampered it's adoption
and development (although this is mostly a suspicion based on various
blog posts which I've read about CEDET and my own experience in getting
started with, and then using it).

There were objections to making CEDET a package.  These related to the
difficulties which the maintainers have experienced in the past with
regards to supporting multiple versions of Emacs.  After some discussion
I deferred to their wisdom.  I agreed to assist with bringing the parts
of CEDET which are currently in Emacs core up to date and with the idea
that remaining pieces should probably live on in ELPA as packages.
However, the discussion continued.  There was renewed talk of developing
CEDET (as well as other larger packages which are distributed with
Emacs) separately to Emacs and that these packages should be brought in
as part of the tarball preparation step during releases (and for what
it's worth I think that this would be a better situation if we could
work through the problems raised.)

I believe that where these developments ended was with David's objection
to eventually pulling CEDET out of core (to be developed in the
aforementioned way) because of the problems which he raised in
developing it separately to Emacs core.

The problem which I face, without the help of those more experienced
than I, is that I don't have enough context on how to resolve some of
the problems which arise as part of doing a partial merge of CEDET into
core.  I'm not an expert wrt to the CEDET project -- I'm a user and an
occasional hacker who has an understanding of the project and wants to
give back, help it grow and enjoy the benefits of this growth and
improvement to Emacs.

I would be happy to discuss possibilities for continuing this
development.

I think that CEDET has more potential than that which has been exploited
thus far.  I know that there are performance problems when working with
larger projects and I know that these aren't intractable, they would
require that we continue to develop it, reconsider some of the
mechanisms by which it operates and make better use of the tools
available to us in Emacs.

Kind regards,

Edward Steere