RE: Why the `click' modifier for a mouse-wheel event?

[Drew Adams]

> >> In Emacs 22 (and 21, 20...), using the mouse wheel did not 
> >> add any modifiers to the event that was read.  In Emacs 23,
> >> `event-modifiers' for a `mouse-wheel-down-event' or
> >> `mouse-wheel-up-event' returns `(click)'.
> >
> > I think this change was made to make Windows behave like the other
> > ports.  IIRC, this was in response to a bug report by Lennart (?),
> > but I can't find it in the tracker right now.
> 
> It was bug#687, about wheel events triggering mouse-avoidance, since 
> they were not considered as mouse events.

Thanks for tracking that other bug down, Jason. But I don't understand why that
fix to bug #687 is TRT. It seems to me very wrong that a wheel event that does
not in any way involve the user pressing the wheel as a middle mouse button
should be treated as a mouse-button `click' event. I don't understand everything
about bug #687 (and it sounds, from reading it, like others might not either),
but this fix seems off-base.

And I was obviously wrong in my own last post in that bug thread, when I said
that I thought I would not need to change my own code because of your change. As
I mentioned earlier, that change complicates user code - see the examples I
gave. The test for a keyboard-only modifier (shift, control, meta, alt, hyper,
or super) has become downright silly - but perhaps there is a better way to do
it than what I posted?


[OT - BTW, do others find the Emacs bug-maintenance site usable? I found no
field for entering a bug number (e.g. 687), to show a specific bug, so I
searched the page for it. Not finding it, I tried to show the Archived bugs
(there is a field for that, at least). More than 50 minutes later (!), with
still no page displayed, I gave up, just clicked some other bug link, and
changed the bug number in the resulting URL. I can't believe this is the best
bug-maintenance tool we could find, but so be it.]