unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
From: "Clément Pit-Claudel" <cpitclaudel@gmail.com>
To: Yoni Rabkin <yoni@rabkins.net>
Cc: emacs-devel@gnu.org
Subject: Re: "Write a new package" culture instead of patches?
Date: Mon, 18 May 2020 15:35:53 -0400	[thread overview]
Message-ID: <5de873b7-0157-05ef-8aea-0635ad8a2276@gmail.com> (raw)
In-Reply-To: <87imgtb17b.fsf@rabkins.net>

On 18/05/2020 13.30, Yoni Rabkin wrote:
> I agree that it is their right to distribute Emms as they wish as long
> as they abide by the terms of the license, but I do not agree that their
> particular form of distribution is good for Emms (no quality control for
> those "needed by" projects; do they even work?) or if it is good for the
> people who enjoy Emms (maybe they steer people to use proprietary
> services.)

There is a bit of quality control, at package submission time (things regarding proper API documentation, proper namespacing, etc. — but nothing like tests, indeed).
I think the way they display thing isn't much different from what you get with "apt-cache rdepends" on a Debian system (it's not the same as the `Recommends: mpg321` or `Suggests: mp3info` lines that `apt` shows when I ask it `apt show emms`, for example).

>>>     * Not even linking to the Emms home page
>>>       (https://www.gnu.org/software/emms/).
>>
>> I think it does: I see this when I open the package in M-x list-packages:
>>
>>    Homepage: https://www.gnu.org/software/emms/
>>
>> The MELPA website links to the git repository instead.
> 
> Yes, that was what I was referring to.

Good point; I opened a ticket about this.

>>>     * Find a way of packaging a project as-is. For instance, Emms could
>>>       be distributed as is, and the M/ELPA software could simply point
>>>       at where Emms keeps its .el files for Emacs to find. This is
>>>       instead of how I see ELPA working now, which is to force the
>>>       software through a kind of a sieve (I think ELPA calls it a
>>>       recipe) where only a select few files come out the other end.
>>
>> It's trivial to make a recipe that includes all files, so I wouldn't
>> worry about this.
> 
> The Emms distribution already contains all of the files by defintion;
> none needed to be remove to begin with. I feel like we looking at the
> issue from two different viewpoints.

The package manager that comes by default in Emacs 24+ is able to produce ELC and info files automatically, so packages typically don't ship Makefiles.  Additionally, it makes certain assumptions about archive layout that EMMS' releases currently don't abide by; that's why MELPA has recipes.
Distributing through ELPA would require the same modifications: this is just the way package.el works. 

>> It will be great to have an improved EMMS recipe in MELPA!  If you run
>> into trouble, you should ask on the bug tracker; the MELPA folks are
>> great.
> 
> Why does Emms need to be offered through three different channels at the
> same time?
> 
> Ideally, I would contact the MELPA bug tracker and have Emms removed
> from MELPA, since it can be trivially downloaded from a GNU server

Downloading it from a GNU server is very complicated, compared to installing it through MELPA.

> and will hopefully soon be installable via ELPA.

I hope you can put it in ELPA; that would be even better.

> However, since nobody asked last time they installed Emms there, nothing
> would stop them from installing it on MELPA again, or modifying the
> recipe to exclude files again. Since MELPA offers the Emms project no
> control over distribution, I don't have much incentive to work on how
> Emms is distributed there, or to fix it and then schedule a weekly
> excursion to MELPA to see if someone has broken it.

This is not how it will work: EMMS was one of the earlier packages to be included in there, before there was a policy to keep maintainers in the loop.  Today, it wouldn't be included without asking.

> Please forgive me if I'm misunderstanding something fundamental about
> how MELPA works. As I've mentioned before, I don't use it or ELPA.

No worries. The short summary is that MELPA doesn't take an adversarial approach, so if you ask for your package to be removed, it will be removed.
But please don't, not before putting it on ELPA — it will break many users' configurations, since emms is rather popular there.

Do you keep statistics for your web server?  It would be useful to compare how many people install through MELPA and how many download releases directly.




  parent reply	other threads:[~2020-05-18 19:35 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-17 19:25 "Write a new package" culture instead of patches? ndame
2020-05-17 19:33 ` Eli Zaretskii
2020-05-17 19:43   ` ndame
2020-05-18  2:22     ` Eli Zaretskii
2020-05-17 19:48   ` Arthur Miller
2020-05-17 19:58   ` ndame
2020-05-18  5:41     ` Philippe Vaucher
2020-05-18 14:49       ` Eli Zaretskii
2020-05-18 15:22         ` Yoni Rabkin
2020-05-18 16:33           ` Clément Pit-Claudel
2020-05-18 17:30             ` Yoni Rabkin
2020-05-18 17:50               ` Dmitry Gutov
2020-05-18 19:17                 ` Clément Pit-Claudel
2020-05-18 19:31                   ` Dmitry Gutov
2020-05-18 20:13                 ` Yoni Rabkin
2020-05-18 21:23                   ` Dmitry Gutov
2020-05-18 19:35               ` Clément Pit-Claudel [this message]
2020-05-18 20:17                 ` Yoni Rabkin
2020-05-18 20:38                   ` Clément Pit-Claudel
2020-05-20  4:01                   ` Clément Pit-Claudel
2020-05-18 21:12                 ` Stefan Monnier
2020-05-18 15:57         ` Clément Pit-Claudel
2020-05-18 16:22           ` Stefan Kangas
  -- strict thread matches above, loose matches on Subject: below --
2020-05-11 16:41 dash.el [was: Re: Imports / inclusion of s.el into Emacs] Alfred M. Szmidt
2020-05-11 17:12 ` 조성빈
2020-05-12  3:16   ` Richard Stallman
2020-05-12  3:55     ` Stefan Monnier
2020-05-13  3:57       ` Richard Stallman
2020-05-13 12:27         ` Stefan Monnier
2020-05-14  5:10           ` Richard Stallman
2020-05-14 13:44             ` Stefan Monnier
2020-05-17  2:53               ` Richard Stallman
2020-05-17 13:01                 ` Eli Zaretskii
2020-05-17 13:38                   ` Dmitry Gutov
2020-05-17 14:24                     ` Eli Zaretskii
2020-05-17 18:27                       ` Dmitry Gutov
2020-05-17 18:52                         ` "Write a new package" culture instead of patches? Stefan Kangas
2020-05-17 19:42                           ` Dmitry Gutov
2020-05-17 22:14                             ` Yuan Fu
2020-05-17 22:44                               ` Arthur Miller
2020-05-17 23:13                               ` chad
2020-05-17 23:22                                 ` Stefan Monnier
2020-05-18  1:31                                   ` João Távora
2020-05-18  1:55                                   ` Tim Cross
2020-05-19  3:51                                   ` Richard Stallman
2020-05-19  3:51                               ` Richard Stallman
2020-05-19  4:33                                 ` Stefan Kangas
2020-05-17 21:14                           ` Alan Third
2020-05-17 22:02                             ` Arthur Miller
2020-05-18  7:58                               ` tomas
2020-05-18 12:08                                 ` Arthur Miller
2020-05-18 12:26                                   ` tomas
2020-05-18 23:07                                     ` arthur miller
2020-05-19  7:27                                       ` tomas
2020-05-17 21:51                           ` Matthias Meulien

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

  List information: https://www.gnu.org/software/emacs/

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

  git send-email \
    --in-reply-to=5de873b7-0157-05ef-8aea-0635ad8a2276@gmail.com \
    --to=cpitclaudel@gmail.com \
    --cc=emacs-devel@gnu.org \
    --cc=yoni@rabkins.net \
    /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 public inbox

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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).