all messages for Emacs-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: Radon Rosborough <radon.neon@gmail.com>
To: Eli Zaretskii <eliz@gnu.org>
Cc: Stefan Monnier <monnier@iro.umontreal.ca>, emacs-devel@gnu.org
Subject: Re: Friendly discussion about (package-initialize)
Date: Thu, 10 Aug 2017 10:06:34 -0700	[thread overview]
Message-ID: <CADB4rJFSfTim5jTfySuRpgoyt0JkUDU03B5ZpUocbtr0v_fVfQ@mail.gmail.com> (raw)
In-Reply-To: <83h8xfsx5m.fsf@gnu.org>

> This sounds like a contradiction. [...] Now I'm utterly confused.

Sorry, it looks like I failed to provide enough context in the name of
being concise. Let me try again.

The issue is complicated because there are actually a number of
different issues at play, which all interact in different ways.

Firstly, let me describe my ideal world. This would be a world in
which `package-initialize' was not called automatically anywhere, or
inserted into the init-file, or anything. However, new users could
still benefit from package.el because Emacs would generate a template
init-file which would contain a call to `package-initialize', if no
init-file already existed.

Now there is also the question of where it is appropriate to call
`package-initialize'. IMO, the only appropriate place to call it is in
the user's init-file. Doing it anywhere else smells like unnecessary
magic, and limits customizability. For example, calling it in
startup.el after loading the init-file means that package
customizations cannot be put in the init-file (unless you use
`after-init-hook', an advanced and rather nonstandard approach), but
the package management system still works after init. Can you possibly
think of any setup that would be *more* confusing to new users?

Then there is the issue of Emacs automatically modifying the init-file
to stick a call to `package-initialize' in it. I think this behavior
is disgusting because of how fragile and nonstandard it is.

Now let me address your specific concerns.

> If we call package-initialize from startup.el, why does it have to
> also be called from the init files?

IMO, it should not be called anywhere. I think having it only called
in startup.el would be a reasonable compromise. Having Emacs insist on
putting the call into the init-file, but then *also* calling it in
startup.el, makes no sense.

> since you are opposed to Emacs putting tyhis call to init files, why
> do you disagree that calling package-initialize in startup.el is
> wrong?

I do think that the correct place to call `package-initialize' is in
the init-file, but I object to Emacs taking care of this
automatically. It might seem like the best solution at first glance,
but doing it this way has a laundry list of horrible side effects (see
my original email).

> that's where Emacs performs all the initialization for the upcoming
> session, and that should include initializing the packages which
> need initialization.

Like I said, having the call be only in startup.el would be a
perfectly reasonable compromise IMO. My main issue is with Emacs
modifying the init-file automatically.

Does this make more sense? I can elaborate, but briefly:

1. I want `package--ensure-init-file' to go away ASAP.
2. I think Emacs should generate a template init-file; in addition to
   making `package--ensure-init-file' mostly superfluous, this would
   have other advantages in terms of improving default user
   experience.
3. If possible, I'd like startup.el to not call `package-initialize'
   either, but this is an order of magnitude less important than (1)
   and (2).

Best,
Radon



  reply	other threads:[~2017-08-10 17:06 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-07  0:37 Friendly discussion about (package-initialize) Radon Rosborough
2017-08-07  1:39 ` Stefan Monnier
2017-08-07  2:16   ` Radon Rosborough
2017-08-07  2:44     ` Stefan Monnier
2017-08-07  4:12       ` Radon Rosborough
2017-08-09 20:24         ` Stefan Monnier
2017-08-10  3:32           ` Radon Rosborough
2017-08-10  4:25             ` Eli Zaretskii
2017-08-10  4:39               ` Radon Rosborough
2017-08-10  7:24                 ` Eli Zaretskii
2017-08-10 17:06                   ` Radon Rosborough [this message]
2017-08-10 19:08                     ` Eli Zaretskii
2017-08-10 19:31                       ` Radon Rosborough
2017-08-10 20:00                       ` Mark Oteiza
2017-08-11  6:14                         ` Eli Zaretskii
2017-08-11  1:25                     ` Nick Helm
2017-08-11 21:43                       ` Stefan Monnier
2017-08-09 20:35         ` Interoperation between package managers (was: Friendly discussion about (package-initialize)) Stefan Monnier
2017-08-10  3:54           ` Radon Rosborough
2017-08-10 21:34             ` Interoperation between package managers Stefan Monnier
2017-08-11  2:14               ` Radon Rosborough
2017-08-11 22:05                 ` Stefan Monnier
2017-08-12 17:54                   ` Radon Rosborough
2017-08-12 20:53                     ` Jonas Bernoulli
2017-08-13 21:43                       ` Stefan Monnier
2017-08-13 21:25                     ` Stefan Monnier
2017-08-13 22:42                       ` Radon Rosborough
2017-08-13 23:32                         ` Stefan Monnier
2017-08-14  0:29                           ` Radon Rosborough
2017-08-14  8:02                             ` Stefan Monnier
2017-08-23 19:39             ` Nikolay Kudryavtsev
2017-08-23 20:58               ` Radon Rosborough
2017-08-24 12:36                 ` Nikolay Kudryavtsev
2017-08-24 20:11                   ` Radon Rosborough
2017-08-25 14:31                     ` Nikolay Kudryavtsev
2017-08-24 15:04               ` Ted Zlatanov
2017-08-24 15:46                 ` Nikolay Kudryavtsev
2017-08-07  3:20   ` Friendly discussion about (package-initialize) Noam Postavsky
2017-08-07  4:14   ` Mark Oteiza
2017-08-08  0:47     ` Stefan Monnier
2017-08-10 20:15       ` Mark Oteiza
2017-08-10 21:29         ` Stefan Monnier
2017-08-11  1:14           ` Mark Oteiza
2017-08-11  8:03             ` Clément Pit-Claudel
2017-08-07  6:52 ` Colin Baxter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CADB4rJFSfTim5jTfySuRpgoyt0JkUDU03B5ZpUocbtr0v_fVfQ@mail.gmail.com \
    --to=radon.neon@gmail.com \
    --cc=eliz@gnu.org \
    --cc=emacs-devel@gnu.org \
    --cc=monnier@iro.umontreal.ca \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/emacs.git
	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.