RE: poplife-mode

[Drew Adams]

> >> It's visible if it was activated by an earlier command,
> >> of course.  E.g. "mouse-3" followed by "down-mouse-3".
> > Pressing and holding down `mouse-3' does _not_ highlight
> > the region.  That happens only when `mouse-3' is released.
> 
> `mouse-3` is an "up", so it implies releasing the button.

Sure - when we mean the _event_.  `mouse-3' is also the
name of the button - as in "pressing and holding down
`mouse-3'.  If that wasn't clear, please reread what I
wrote, substituting "3rd mouse button" for such occurrences.

> > If you mean first click `mouse-3' (press and release)
> > and then press and hold `mouse-3',
> 
> That's right.
> 
> > then you are talking about _two_ `mouse-3' clicks (it will be
> > released) - similar to what `mouse3.el' does.
> 
> But that's only if you want to use the `mouse-3` event first.
> My suggestion is mostly about providing *another* operation on that same
> button (I expect it will rarely be used together with a mouse-3 event
> (i.e. with mouse-save-then-kill)).

Not sure what you are trying to say.  You seemed to be
saying that the region is highlighted with your feature
also.  My point was that the region is highlighted only
after the button is release, which is as many clicks
with your approach as with the mouse3.el approach (which
shows the menu upon button release, not upon long press).

With your approach, simply pressing and holding the button
down a long time will show the menu, but it won't highlight
the region.  So it is not appropriate for menu items that
act on the region or otherwise take the region into account. 

> > Besides not highlighting the region that is active and
> > can be acted on, your approach has the disadvantage of
> > mistakenly killing the region when an intended long
> > press happens to be a bit too short.
> 
> This risk only exists in the case you use down-mouse-3
> right after mouse-3, which I expect will be rare.

AFAICT, it's the only way, with your approach, to see
the region you will act on get highlighted.  Are you
relegating your approach to menus that don't have items
that act on the region?

> > What's better about it?
> 
> No need to double click, hence (except for a 450ms delay)

As I pointed out, there is just as much need to click twice,
if you want to see the region.  (It's not a double-click
event in either case; it's two click events in both cases.)

> it behaves just like the "standard" context menu of other
> applications.

At least on MS Windows, the apps I use all show the
menu upon button-release, not upon button press.  (And
press-and-hold is surely not standard for showing a menu
anywhere, is it?)

And if there is anywhere where a `mouse-3' menu is used
ubiquitously it is on Windows.  Windows users are used to
`mouse-3' showing a context menu when it is clicked, not
when it is pressed and held.

So I'd say that the mouse3.el approach is at least as
"standard-context-menu"-like as your approach.