Re: Stability of core packages (was: Not easy at all to upgrade :core packages like Eglot)

[João Távora]

On Tue, Apr 18, 2023 at 7:57 PM Jim Porter <jporterbugs@gmail.com> wrote:
>
> On 4/18/2023 5:57 AM, Eli Zaretskii wrote:
> > So this again goes back to the main issue: how should the stability
> > considerations affect development of core packages and their
> > "stability gradation"?
>
> It sounds to me like there are 3 or 4 levels (depending on how you count):
>
> * Stable: the version of a package included in the latest Emacs tarball
> * Latest: the latest version on GNU ELPA (etc)
> * Devel: the latest version on GNU-devel ELPA (etc)
>
> You could possibly add:
>
> * Core(?): the version of a package included in the tarball of the
> user's *current* Emacs installation

These are interesting levels, but I was under the impression
that the goal is to partition the set of :core packages according
to some kind of gradation, what Eli called "stability gradation".
The set can be found in the variable package--builtins.

In contrast, you seem to be proposing to partition the much larger
set of pairs (CORE-PACKAGE . VERSION) (where VERSION is the version
of CORE-PACKAGE, not Emacs).

These two exercises are potentially interesting, but merely as
academic exercises they are not particularly useful.

For the first exercise, the useful thing that can be extracted
from it is the optimal behaviour w.r.t to stability of our existing
tools when confronted with an argument from that set.

Such existing tools include package-install, use-package,
package-update, etc. For example.

A. The effects of the form (package-install 'eglot) moving from
Emacs 26,27,28 to Emacs 29 are drastically different.
That's bad stability.

B. The effects of, say, the form (package-install 'xref) moving
from Emacs 28 to Emacs 29 are the same.  That's good stability.

What I'd like to know if anyone else thinks, like me, that this is
a problem.  We can discuss how bad of a problem (I think it's
moderately serious).

But, more importantly, am I the only one that sees an easy
fix to this we can fix case A and keep case B untouched?

João