Re: emacs-29 9b775ddc057 1/2: ; * etc/EGLOT-NEWS: Fix wording of last change.

[João Távora]

On Sat, May 6, 2023 at 2:21 PM Eli Zaretskii <eliz@gnu.org> wrote:
>
> > From: João Távora <joaotavora@gmail.com>
> > Date: Sat, 6 May 2023 14:03:02 +0100
> > Cc: Dmitry Gutov <dmitry@gutov.dev>, emacs-devel@gnu.org
> >
> > On Sat, May 6, 2023 at 11:52 AM Eli Zaretskii <eliz@gnu.org> wrote:
> >
> > > I'm surprised, to say the least.  To recap:
> > >
> > >   . the entire discussion of bug#62720, of which this is fallout, was
> > >     about allowing package-install to upgrade Eglot, and according to
> > >     João, that's what package-install did before Emacs 29.  Was this
> > >     all based on a mistake? if so, why no one has said so in all that
> > >     long argument??
> >
> >
> > One more time: in Emacs 28 package-install doesn't upgrade,
> > but it installs the latest, which is incompatible behaviour
> > if you move to Emacs 29, where that won't happen.
>
> Is this for Eglot (and use-package), or is this for any package
> installed from ELPA?  IOW, does Emacs 29's package-install still
> install the latest version of a package from ELPA?  And does that
> happen even if some older version of a package is already installed?

Yes to the first, no to the second.

> > So if you're used to setting up a brand new Emacs 28 and
> > package-install Eglot to get versions with nice features and
> > bugfixes, you may be dismayed to find that doing the very
> > same thing in Emacs 29 results in what will probably be a
> > old version.
>
> Are you talking about users who didn't update their Eglot, except when
> a new Emacs version was released?  Or are you talking about users who
> updated Eglot from ELPA (using package-install) even between Emacs
> releases?  Or are you talking about something else entirely?

Just spend that minute, Eli.   Call it an investment.  And then
re-read what I wrote in that paragraph.  It will make sense.

> > And this problem will persist _and_ worsen during the lifetime of
> > Emacs 29
>
> Why do you insist on reiterating past disagreements, especially when
> they are tangents, is beyond me.

You brought it up!!! You asked "was this all based on a mistake?"
I just clarified and summarized what happens!

> > Anyway, experimenting and experiencing this yourself will
> > take at most a minute if you have Emacs 28 and Emacs 29
> > handy, so there's no shroud of mystery whatsoever.
>
> Why do you think I even have a minute to spare?

Clearly, unless you're some kind of superhuman being, you're
wasting much more time writing these messages. So I figured
a minute of experimentation would probably save a lot of
your time and our collective time.

You're still asking me questions, and then when I answer
them you say I'm bringing up old past disagreements.

> I'm asking the experts on package.el to describe what happens
> precisely _because_ I don't have that time.  For example, today, I've
> been at the keyboard, working on various Emacs issues for the last 7
> hours without any breaks, and this is my weekend!

For your health, take a break Eli.

> > But we already went through all this, and all these decisions
> > have been made already, and they are ultimately all legitimate
> > even if we all disagree.
>
> It sounds like what we went through was based on a misunderstanding,
> or at least Dmitry says it was.

Again, please invest that minute (after taking a break, of course).

> > I don't think the NEWS entry needed rewording: it didn't
> > say anything inaccurate or misleading.  You made it more
> > complete, and perfectly accurate, but also more inescrutable,
> > which is not really what NEWS should be about IMO.
>
> Clearly, if I thought the original wording was acceptable, I wouldn't
> have touched it

There was nothing wrong or misleading about it.  What possibly
can be wrong or unacceptable in announcing "Eglot can upgrade
itself to the latest version." to Eglot users?  But let's just let it go, your
reworking isn't that horrible either.

João