bug#69528: 30.0.50; [BUG] transient.el is not a member of package--builtin-versions

[Philip Kaludercic <philipk@posteo.net>]

Stefan Kangas <stefankangas@gmail.com> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> > What about making `lm-version' handle the "package-version" header then
>>> > using `lm-version' in loaddefs-generate--parse-file?  See patches.
>>>
>>> My main concern was if we want to have Package-Version always override
>>> Version, but if my patch modified loaddefs-gen, then I don't think there
>>> is much of a difference if we change lisp-mnt instead (in terms of the
>>> generality of the change).
>>>
>>> So I am fine with the change, and think we can merge it.  Eli: Is master
>>> still fine for these kinds of changes?
>>
>> I think so, yes.  But maybe I don't fully understand the effect of
>> this change?  Can you describe it?

Sorry for the late response, but Stefan summaries the situation well
below.

>> I also added the other maintainers, in case they have opinions on
>> this.
>
> I think the first patch is right, i.e. to use
>
>     (lm-version)
>
> instead of
>
>     (lm-header "version")
>
> So let's install that one, I think.

I agree.

> The effect of the second patch is to change `lm-version` to look for a
> "Package-Version" header if there is no "Version" header.
>
> This has two problems:
>
> 1. We didn't do that until now, and it's not clear to me what is the
>    issue that is prompting this change.  The transient.el issue seems to
>    have been fixed already.
>
> 2. The way I read the manual, it seems like "Package-Version" should be
>    preferred over "Version", if it exists:
>
>         ‘Package-Version’
>              If ‘Version’ is not suitable for use by the package manager, then a
>              package can define ‘Package-Version’; it will be used instead.
>              This is handy if ‘Version’ is an RCS id or something else that
>              cannot be parsed by ‘version-to-list’.
>
>    I'm also not sure we need to support packages with unusual versions
>    like RCS id's these days.  Is that use case still relevant?  Perhaps
>    we should simply deprecate the "Package-Version" header?

FWIW I use this for some of my own scripts that I version using RCS, so
I'd appreciate it if that functionality would stay.

-- 
	Philip Kaludercic on peregrine