GNU bug report logs - #73932
[PATCH] Add noconfirm to 'package-autoremove'

Previous Next

Full log

Message #14 received at 73932 <at> debbugs.gnu.org (full text, mbox):

From: Philip Kaludercic <philipk <at> posteo.net>
To: Sean Devlin <spd <at> toadstyle.org>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 73932 <at> debbugs.gnu.org
Subject: Re: bug#73932: [PATCH] Add noconfirm to 'package-autoremove'
Date: Sat, 26 Oct 2024 07:31:16 +0000

Sean Devlin <spd <at> toadstyle.org> writes:

> Hello,
>
>> On Oct 21, 2024, at 12:35 PM, Eli Zaretskii <eliz <at> gnu.org> wrote:
>> 
>>> From: Sean Devlin <spd <at> toadstyle.org>
>>> Date: Mon, 21 Oct 2024 11:58:34 -0400
>>> 
>>> Attached is a patch adding an optional NOCONFIRM argument to package-autoremove to skip user confirmation before removing packages.
>>> 
>>> I want to call package-autoremove in my init file to ensure there
>>> are no packages installed except those I configured explicitly via
>>> package-selected-packages.
>>> 
>>> For example, in one session I might install package XYZ
>>> interactively via the package menu to try it out. If I decide I
>>> like it, I’ll add it to package-selected-packages in my init file
>>> to keep it around. Otherwise, I want to make sure it gets blown
>>> away the next time I start Emacs.
>>> 
>>> This workflow is cumbersome in the current implementation of
>>> package-autoremove, which prompts the user before removing any
>>> packages. This patch adds an optional NOCONFIRM argument similar to
>>> what is implemented in package-install-selected-packages.
>>> 
>>> Please let me know if any changes are needed.
>> 
>> Thanks.  I'll let our package.el experts review and comment, but if
>> this is accepted, please add a NEWS entry for the new optional
>> behavior.
>
> Sounds good. Here is a patch with an attempt at a NEWS entry:

The change looks pretty uncontroversial to me.  Is the variation
interesting, where NOCONFIRM is a predicate so that we can selectively
remove packages?

> Please let me know if any changes are needed.
>
>> 
>> Also, would it make sense to turn NOCONFIRM on if the command is
>> invoked with a prefix argument?
>
> Sure, that might make sense. If we do that, should we do the same in package-install-selected-packages? It uses its NOCONFIRM argument in a similar way.

We can do that in a subsequent patch.

-- 
	Philip Kaludercic on siskin

Previous Next

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