[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Removing packages from archive-contents if removed from elpa-package
From: |
Philip Kaludercic |
Subject: |
Re: Removing packages from archive-contents if removed from elpa-packages |
Date: |
Mon, 31 Oct 2022 14:25:39 +0000 |
Stefan Monnier <monnier@iro.umontreal.ca> writes:
> Philip Kaludercic [2022-10-31 09:40:36] wrote:
>> there is at least one broken package on NonGNU ELPA
>> (color-theme-tangotango) that was added, then removed as dependencies
>> were missing, but it remained in the archive because it was added to the
>> archive-contents file. The following patch would make sure that any
>> package not listed in elpa-packages is also remove from
>> archive-contents:
>
> Thanks, but I don't think `elpaa--update-archive-contents` is the right
> place for that. This can be called N times for a given invocation of
> `elpaa-batch-make-all-packages`.
> Why not call a dedicated function from `elpaa-batch-make-all-packages`
> (i.e. alongside the call to `elpaa--publish-package-specs`)?
That can be done.
>> It might be a good idea to cache the interpretation of elpa-packages, so
>> I would also suggest the following:
>
> Not sure it's worth the trouble: the way the code is written, it should
> be called only once per Emacs session in the majority of cases, no?
Right, ideally somewhen around the beginning.
>> 1. This does not remove any tarballs. It wouldn't be hard to do, but I
>> am not sure if there would be any unintended side effects of doing
>> so -- at least immediately.
>
> I think we should keep the old HTML page and tarballs.
I agree.
>> 2. There are functioning packages on NonGNU ELPA like haskell-mode that
>> were also added to elpa-packages then removed, and still linger in
>> the archive without being updated. This would get rid of those too,
>> which might not be a good idea.
>
> Indeed, `haskell-mode` should be re-added to `elpa-packages`
> because IIRC it's still actively depended-upon (by GNU ELPA's `hcel`).
The issue here is that upstream isn't responsive and has asked for
haskell-mode to not be added until further notice :/