bug#68213: 30.0.50; completion-preview-tests failure in --without-x build

[Po Lu]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Po Lu <luangruo@yahoo.com>
>> Cc: Eshel Yaron <me@eshelyaron.com>,  Eli Zaretskii <eliz@gnu.org>,
>>   jm@pub.pink,  68213@debbugs.gnu.org
>> Date: Mon, 08 Jan 2024 14:16:54 +0800
>> 
>> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> 
>> > But why is `mouse-4` sometimes on `mouse-wheel-down-event` and sometimes
>> > on `mouse-wheel-down-alternate-event`?  What difference does it make?
>> >
>> > But, AFAICT the `wheel-up/down` events always mean the same, so I think
>> > `mwheel.el` should just always bind `wheel-DIR` events and then the
>> > `mouse-wheel-DIR-event` vars are used only for those cases where other
>> > events also need to be bound, like in the patch below, WDYT?
>> 
>> I think we should let sleeping dogs lie, because the last time changes
>> were made to these options in mwheel.el, several obscure problems
>> emerged under text terminals that required another round of changes,
>> yielding the current values.
>
> There's one dog we cannot let lie: the one that fails the test in
> completion-preview-tests to fail to compile -- this is what started
> this bug report, and I would like to solve that problem.

>From a cursory reading of this bug report, I gather that m-w-*-event and
m-w-*-alternate-event are at times defined to identical values, which
causes defvar-keymap to signal?  In that situation the correct course of
action is using [remap ...] to bind the mwheel commands to whatever
completion-preview needs it set to, right?

HTH.