bug#38993: 26.3; EBDB saves the database on Emacs exit even when told not to

[Eric Abrahamsen <eric@ericabrahamsen.net>]

Jorge P. de Morais Neto <jorge+list@disroot.org> writes:

> Em [2020-01-08 qua 10:54:35-0800], Eric Abrahamsen escreveu:
>
>> Okay, I guess that would work: always add a hook, but only actually do
>> the save if `ebdb-save-on-exit' is t.  So if you have open *EBDB*
>> buffers, and unsaved changes, but you set `ebdb-save-on-exit' to nil,
>> then when you kill Emacs you'll still get prompted to save EBDB, but if
>> you say "no" then nothing further will happen.
>
> That would be a step in the right direction.  In addition to this you
> could improve the weird (and arguably harmful) EBDB idiosyncrasies that
> we previously talked about:
>
> 1. misleading the user by asking whether to save but then, if
>    `ebdb-save-on-exit' is t, saving the database anyway even if the user
>    answered in the negative; and
> 2. saving the "*EBDB*" buffer itself, which I believe is not useful.
>
> And the `ebdb-save-on-exit' option could be renamed to
> `ebdb-silently-save-on-exit', or, if that is too verbose, then at least
> its documentation should emphasize that Emacs will save with no
> questions asked.

I've already fixed #2 above, I just haven't made a new release yet (I'll
try to get the rest of the bugs you raised fixed first).

I still don't have a good solution for the first problem. I guess what
I'm leaning toward is 1: adding a note to the documentation/manual about
this weirdness, and 2: defaulting `ebdb-save-on-exit' to nil. The
average user will be using EBDB interactively, meaning that they'll get
the prompt both at "C-x s" and when killing Emacs. The
`ebdb-save-on-exit' option could be reserved for hackers who are using
EBDB in some more peculiar fashion.

>> I have an account at SourceHut, which is very FOSS-friendly, but so
>> far haven't put any repos there, maybe I could consider moving.
>
> You might want to consult
> <https://libreplanet.org/wiki/Fsf_2019_forge_evaluation>, as well as
> discussing the options with the free software community on IRC or XMPP.
> I myself have chosen GitLab to host my very simple static website.  I
> based the decision on GitLab's features, ease of use, and the C ethics
> grade from the FSF, a grade which seemed acceptable to me.  However I
> later learned that, at least according to the aforementioned LibrePlanet
> page, the GitLab ethics evaluation regressed.  I guess I should move
> away from it.

It looks to me like SourceHut is licensed under the GPL
(https://git.sr.ht/~sircmpwn/git.sr.ht/tree/master/LICENSE), so I'm
hoping that's an A! Or at least a B.

> By the way, if you forgive my impertinence, I suggest saying
> "free/libre" rather than "FLOSS", or, even worse, "FOSS".  Please see
> <https://www.gnu.org/philosophy/floss-and-foss.en.html>.

Thanks for that! No impertinence, I'll admit it's something I haven't
read or really thought much about before. I'll be aware of it from here
on out.

Thanks,
Eric