bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Drew Adams]

If you pop up a menu (e.g. using `x-popup-menu') and one of its items
calls `minibuffer-message', the user will never see the message,
presumably because the call to `sit-for' in `minibuffer-message' sees
the mouse-up event (from choosing the menu item) as user input,
canceling the `sit-for' timeout.
 
Seems like we should be able to make this work somehow.  Perhaps
`sit-for' could accept an optional arg listing a set of events to
ignore, and `minibuffer-message' could then call it with mouse-up
in that list?
 
But should _all_ uses of `minibuffer-message' ignore `mouse-up' events
wrt the timeout?  Dunno.  Sounds doubtful.
 
Still, this seems like something we should be able to handle, so that
users can see a message associated with a menu item.
 
Example use: A user chooses a menu item to cycle some variable/behavior
(e.g. a sort order) to the next possible value, and the action ends with
a message echoing what the new value is.  Currently, the user can do
this over and over without ever seeing what the new value is each time.
In some contexts it might not be convenient for the user to stop the
overall interaction just to interrogate the value.  That is, use of
the popup menu might be only one link in a chain of user interactions.
 
One possibility might be to move the `sit-for' out of
`minibuffer-message', making the calling code be responsible for it
instead.  That would be analogous to the way `message' is used.  But I'm
not sure how that would affect other things - perhaps it is important
that `minibuffer-message' calls `sit-for' itself.  Dunno.
 

In GNU Emacs 24.0.50.1 (i386-mingw-nt5.1.2600)
 of 2011-02-14 on 3249CTO
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (4.4) --no-opt --cflags
-Ic:/imagesupport/include'
 

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Stefan Monnier]

> If you pop up a menu (e.g. using `x-popup-menu') and one of its items
> calls `minibuffer-message', the user will never see the message,
> presumably because the call to `sit-for' in `minibuffer-message' sees
> the mouse-up event (from choosing the menu item) as user input,
> canceling the `sit-for' timeout.
 
> Seems like we should be able to make this work somehow.  Perhaps

Agreed.  I think the way to attack this is elsewhere, tho: the mouse-up
should be eaten up by the menu code.


        Stefan

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Drew Adams]

> > Seems like we should be able to make this work somehow.
> 
> Agreed.  I think the way to attack this is elsewhere, tho: 
> the mouse-up should be eaten up by the menu code.

Yes, that sounds good.

Dunno whether that would ever be an inconvenience - I can't think of a negative
use case.  Seems like every menu selection action would always involve mouse
down + up.

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Lars Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

> If you pop up a menu (e.g. using `x-popup-menu') and one of its items
> calls `minibuffer-message', the user will never see the message,
> presumably because the call to `sit-for' in `minibuffer-message' sees
> the mouse-up event (from choosing the menu item) as user input,
> canceling the `sit-for' timeout.

Do you have a test case that demonstrates the problem?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#8116: [External] : Re: bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Drew Adams]

> Do you have a test case that demonstrates the problem?

The problem is specified in the bug report.

Whether the situation is problematic in the same way as
when the problem was reported, 10 years ago, someone
who uses Emacs 27 or the incipient 28 can investigate.

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Sun, 18 Jul 2021 02:30:08 +0200
> Cc: 8116@debbugs.gnu.org
> 
> "Drew Adams" <drew.adams@oracle.com> writes:
> 
> > If you pop up a menu (e.g. using `x-popup-menu') and one of its items
> > calls `minibuffer-message', the user will never see the message,
> > presumably because the call to `sit-for' in `minibuffer-message' sees
> > the mouse-up event (from choosing the menu item) as user input,
> > canceling the `sit-for' timeout.
> 
> Do you have a test case that demonstrates the problem?

You won't be able to see this on any system but MS-Windows.  On
Windows, the Emacs menus are implemented in a way that has this
unfortunate side effect: we basically preempt the Emacs command loop
for as long as the menu stays open.

If someone wants to redesign how Emacs menus are implemented on
Windows, I think that would be very welcome.  But until that happens,
this is Emacs "functioning as designed", not some easily-fixed bug.

bug#8116: [External] : Re: bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Lars Ingebrigtsen]

Drew Adams <drew.adams@oracle.com> writes:

>> Do you have a test case that demonstrates the problem?
>
> The problem is specified in the bug report.

A test case makes it much more likely that somebody will take the effort
to poke at this problem.  Without a test case, that somebody will first
have to come up with one themselves.

But if you're not interested in being helpful, that's fine, of course.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> You won't be able to see this on any system but MS-Windows.  On
> Windows, the Emacs menus are implemented in a way that has this
> unfortunate side effect: we basically preempt the Emacs command loop
> for as long as the menu stays open.

Ah, I see.  I'll rename the bug report to be more specific, then.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#8116: [External] : Re: bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Drew Adams]

> > > If you pop up a menu (e.g. using `x-popup-menu') and one of its items
> > > calls `minibuffer-message', the user will never see the message,
> > > presumably because the call to `sit-for' in `minibuffer-message' sees
> > > the mouse-up event (from choosing the menu item) as user input,
> > > canceling the `sit-for' timeout.
> >
> > Do you have a test case that demonstrates the problem?
> 
> You won't be able to see this on any system but MS-Windows.  On
> Windows, the Emacs menus are implemented in a way that has this
> unfortunate side effect: we basically preempt the Emacs command loop
> for as long as the menu stays open.
> 
> If someone wants to redesign how Emacs menus are implemented on
> Windows, I think that would be very welcome.  But until that happens,
> this is Emacs "functioning as designed", not some easily-fixed bug.

Thanks for this info.  I guess the bug can then either
be closed or kept open in the wishlist, the fix needing
redesign etc.

I wonder if this effect should be mentioned in the manual?

bug#8116: [External] : Re: bug#8116: 24.0.50; `minibuffer-message': ignore mouse-up event for `sit-for'?

[Drew Adams]

> But if you're not interested in being helpful,
> that's fine, of course.

Yet another snarky disparagement.  Not helpful.

Follows on your "since Drew can't be bothered"
to send a patch - when the bug report provided
the exact replacement for the single function
that needed to be replaced.