GNU bug report logs - #79601
[PATCH] Indent startup.el and provider startup feature

Previous Next

To reply to this bug, email your comments to 79601 AT debbugs.gnu.org.
There is no need to reopen the bug first.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Yauhen Artsiukhou <jsirex <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: [PATCH] Indent startup.el and provider startup feature
Date: Wed, 8 Oct 2025 14:39:02 +0200

First patch does only indent of startup.el:
https://github.com/jsirex/emacs/commit/6e8bccc0816e3751d128161143debd6933335166.patch

Second patch provides  'startup feature, I believe it is missing:
https://github.com/jsirex/emacs/commit/dcb0aa1b74698bc60a21b60ce94bd642898f5c38.patch

PS. This is my first contribution
PPS. Attached as links because first patch is too big
--
Yauhen

Message #8 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Yauhen Artsiukhou <jsirex <at> gmail.com>, Andrea Corallo <acorallo <at> gnu.org>,
 Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 79601 <at> debbugs.gnu.org
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Wed, 08 Oct 2025 17:30:29 +0300

> From: Yauhen Artsiukhou <jsirex <at> gmail.com>
> Date: Wed, 8 Oct 2025 14:39:02 +0200
> 
> First patch does only indent of startup.el:
> https://github.com/jsirex/emacs/commit/6e8bccc0816e3751d128161143debd6933335166.patch

Thanks.  We don't like to make whitespace-only changes.  Instead, we
fix the whitespace around the places where we make real changes.

> Second patch provides  'startup feature, I believe it is missing:
> https://github.com/jsirex/emacs/commit/dcb0aa1b74698bc60a21b60ce94bd642898f5c38.patch

startup.el is preloaded, so such a provide doesn't really help
anything.  Why did you think we need it?

Andrea and Stefan, do we have any policy to have 'provide' in all the
preloaded packages?

Message #11 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Yauhen Artsiukhou <jsirex <at> gmail.com>, Andrea Corallo <acorallo <at> gnu.org>,
 79601 <at> debbugs.gnu.org
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Wed, 08 Oct 2025 10:53:45 -0400

> Andrea and Stefan, do we have any policy to have 'provide' in all the
> preloaded packages?

AFAIK, we don't.  We don't have a policy against them either, so
I wouldn't reject a patch just because it adds such a unnecessary
`provide`, but the patch should do something else that's useful.


        Stefan

Message #14 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Yauhen Artsiukhou <jsirex <at> gmail.com>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: Eli Zaretskii <eliz <at> gnu.org>, Andrea Corallo <acorallo <at> gnu.org>,
 79601 <at> debbugs.gnu.org
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Wed, 8 Oct 2025 17:14:28 +0200

I thought that having whitespace-only changes separately makes the
process easier.
But ok, next time I will not provide such patches.

Regarding `provide`, as an emacs user,  I did my best to write
`init.el` and sort things out. And I want to follow some structure.
I see many files have their own `defcustom`s and I decided to follow
each package with `use-package` to group settings.
So I did:
(use-package startup :ensure nil
  :init
  (setopt inhibit-startup-screen t))

But if demand is true or defer is false it tries to require package
and fails with:
Debugger entered--Lisp error: (error "Loading file
/usr/local/share/emacs/31.0.50/lisp/startup.elc failed to provide
feature ‘startup’")

Try `(require 'startup)`.
As a new user who tries to understand "how to write emacs" such
inconsistency is really confusing. When should I use `provide` and
when I should not.

Feel free to accept or reject any patch. At least I figured out how to
contribute :)

ср, 8 окт. 2025 г. в 16:53, Stefan Monnier <monnier <at> iro.umontreal.ca>:
>
> > Andrea and Stefan, do we have any policy to have 'provide' in all the
> > preloaded packages?
>
> AFAIK, we don't.  We don't have a policy against them either, so
> I wouldn't reject a patch just because it adds such a unnecessary
> `provide`, but the patch should do something else that's useful.
>
>
>         Stefan
>

Message #17 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Andrea Corallo <acorallo <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: Yauhen Artsiukhou <jsirex <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 79601 <at> debbugs.gnu.org
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Wed, 08 Oct 2025 15:10:32 -0400

Stefan Monnier <monnier <at> iro.umontreal.ca> writes:

>> Andrea and Stefan, do we have any policy to have 'provide' in all the
>> preloaded packages?
>
> AFAIK, we don't.  We don't have a policy against them either, so
> I wouldn't reject a patch just because it adds such a unnecessary
> `provide`, but the patch should do something else that's useful.

+1

Message #20 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Arash Esbati <arash <at> gnu.org>
To: Yauhen Artsiukhou <jsirex <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, Andrea Corallo <acorallo <at> gnu.org>,
 79601 <at> debbugs.gnu.org, Stefan Monnier <monnier <at> iro.umontreal.ca>
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Thu, 09 Oct 2025 10:48:49 +0200

Yauhen Artsiukhou <jsirex <at> gmail.com> writes:

> Regarding `provide`, as an emacs user,  I did my best to write
> `init.el` and sort things out. And I want to follow some structure.
> I see many files have their own `defcustom`s and I decided to follow
> each package with `use-package` to group settings.
> So I did:
> (use-package startup :ensure nil
>   :init
>   (setopt inhibit-startup-screen t))

Does the following form help?

  (use-package emacs
    :custom
    (inhibit-startup-screen t))

Best, Arash

Message #23 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Yauhen Artsiukhou <jsirex <at> gmail.com>
To: Arash Esbati <arash <at> gnu.org>
Cc: Eli Zaretskii <eliz <at> gnu.org>, Andrea Corallo <acorallo <at> gnu.org>,
 79601 <at> debbugs.gnu.org, Stefan Monnier <monnier <at> iro.umontreal.ca>
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Thu, 9 Oct 2025 12:31:48 +0200

Arash, yes I know about this.

Crucial difference here is (featurep 'emacs) is t.

So basically (require 'emacs) works as expected, but (require 'startup) fails.
inhibit-startup-screen is defined in startup.el.

So, if we abstract away a little bit, I see the following common pattern:
1. emacs has "something" implemented in lisp/foo/something.el.
2. I want to use it and perform standard routine: (progn (require
'something) (setopt something-enabled t))
    Same with use-package: (use-package something :ensure nil :init
(setopt this 0) :config (setopt that 1))

And it is a bit confusing that most files (provide 'feature-x) but
some of them don't. For example, (require 'indent) fails exactly the
same.

чт, 9 окт. 2025 г. в 10:48, Arash Esbati <arash <at> gnu.org>:
>
> Yauhen Artsiukhou <jsirex <at> gmail.com> writes:
>
> > Regarding `provide`, as an emacs user,  I did my best to write
> > `init.el` and sort things out. And I want to follow some structure.
> > I see many files have their own `defcustom`s and I decided to follow
> > each package with `use-package` to group settings.
> > So I did:
> > (use-package startup :ensure nil
> >   :init
> >   (setopt inhibit-startup-screen t))
>
> Does the following form help?
>
>   (use-package emacs
>     :custom
>     (inhibit-startup-screen t))
>
> Best, Arash

Message #28 received at 79601-done <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Yauhen Artsiukhou <jsirex <at> gmail.com>
Cc: arash <at> gnu.org, acorallo <at> gnu.org, 79601-done <at> debbugs.gnu.org,
 monnier <at> iro.umontreal.ca
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Sat, 25 Oct 2025 13:17:22 +0300

> From: Yauhen Artsiukhou <jsirex <at> gmail.com>
> Date: Thu, 9 Oct 2025 12:31:48 +0200
> Cc: Stefan Monnier <monnier <at> iro.umontreal.ca>, Eli Zaretskii <eliz <at> gnu.org>, 
> 	Andrea Corallo <acorallo <at> gnu.org>, 79601 <at> debbugs.gnu.org
> 
> Arash, yes I know about this.
> 
> Crucial difference here is (featurep 'emacs) is t.
> 
> So basically (require 'emacs) works as expected, but (require 'startup) fails.
> inhibit-startup-screen is defined in startup.el.
> 
> So, if we abstract away a little bit, I see the following common pattern:
> 1. emacs has "something" implemented in lisp/foo/something.el.
> 2. I want to use it and perform standard routine: (progn (require
> 'something) (setopt something-enabled t))
>     Same with use-package: (use-package something :ensure nil :init
> (setopt this 0) :config (setopt that 1))
> 
> And it is a bit confusing that most files (provide 'feature-x) but
> some of them don't. For example, (require 'indent) fails exactly the
> same.

No further comments, so I'm now closing this bug.

Message #31 received at 79601 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Yauhen Artsiukhou <jsirex <at> gmail.com>
Cc: Arash Esbati <arash <at> gnu.org>, Andrea Corallo <acorallo <at> gnu.org>,
 79601 <at> debbugs.gnu.org, Eli Zaretskii <eliz <at> gnu.org>
Subject: Re: bug#79601: [PATCH] Indent startup.el and provider startup feature
Date: Sun, 26 Oct 2025 20:33:06 -0400

> So, if we abstract away a little bit, I see the following common pattern:
> 1. emacs has "something" implemented in lisp/foo/something.el.
> 2. I want to use it and perform standard routine: (progn (require
> 'something) (setopt something-enabled t))
>     Same with use-package: (use-package something :ensure nil :init
> (setopt this 0) :config (setopt that 1))
>
> And it is a bit confusing that most files (provide 'feature-x) but
> some of them don't. For example, (require 'indent) fails exactly the
> same.

Indeed the `provide` thingies are still fundamentally designed for
(require <FEATURE>), i.e. for use by other packages, and not for
(use-package <FEATURE> ...) which is for init.el customization.

I consider any `require` in an init file to be a sign of a bug somewhere
(usually either a bug in the init file or in the required file), so I'm
very skeptical of this part of the `use-package` functionality.
But admittedly, `use-package` also uses <FEATURE> for
`with-eval-after-load` which is definitely legitimate in an init file.

The use of (use-package <FEATURE> ...) that you propose has a few other
problems for the preloaded files: we, Emacs maintainers, have not been
super careful about what to put where, and furthermore, we sometimes
move some definitions around, notably for reasons of ordering in
`loadup.el` and related bootstrap issues.

I tend to think it would be good to be more careful about what we put
where and treat those files more like actual packages (which just happen
to be preloaded), including obeying the usual "namespace prefix"
convention, but I don't think there's a general agreement about it, nor
much desire to implement such a change, or even much evidence that it
would bring much concrete benefit.

All in all, I think similar things tend to happen within multi-file
packages, so it makes sense to treat preloaded files as all members of
the `emacs` multi-file package.


        Stefan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.