Re: Summary and next steps for (package-initialize)

[Eli Zaretskii <eliz@gnu.org>]

> From: Radon Rosborough <radon.neon@gmail.com>
> Date: Sat, 19 Aug 2017 19:38:05 -0700
> 
> If you object to this proposal, please tell me what the concrete
> disadvantages to it would be, or explain why the advantages I have
> listed are not valid. Otherwise, let's start the process of figuring
> out the best implementation.

My take on this, after reading the discussions kindly pointed to by
Mark, is that your proposal will replace one problematic situation
with another.  Whether the new situation will be better than the
current one is arguable, but in any case I don't think it will be
significantly better.  Already a few people said they were unhappy
with such a solution.  I myself cannot say I like the idea of Emacs
creating an init file in the user's home directory.

I think we should instead explore the possibility that
package-initialize will be called only in startup.el.  That solution
came up during the discussions, but AFAICT was dismissed almost
without any serious consideration.  The issues raised against it could
probably be solved by splitting the package initialization in two, one
part before, the other after the user's init file is read.  We already
do something similar with initializing the frame appearance and its
faces.  However, such a possibility was not considered, unless I
missed something.

Admittedly, I know too little about package.el and the related
problems, so perhaps what I propose is hard to make work in practice,
or it even might be unworkable.  However, a couple of people proposed
the same idea back when this was discussed, and those people
definitely know much more about package.el than I do.  So I think we
should give this possibility one more chance, and this time we
shouldn't reject it unless some serious disadvantages of it are
uncovered and described in detail here, so we could all agree they are
serious obstacles.

Thanks.