On 9 November 2016 at 19:43, Eli Zaretskii <eliz@gnu.org> wrote:

>
> I don't think I asked for something so complicated and hard to
> accomplish that it would justify partial fixes on master.  I'm very
> uneasy with having the list of hard-coded events we ignore for this
> purpose continue to grow one event at a time.  Making this controlled
> by a Lisp variable will make its future maintenance much easier,
> including when someone, perhaps you, comes up complaining about yet
> another obscure event.
>

​I already showed I can't see what the correct solution is: I thought you
made a precise suggestion, which I discussed (a list of events for
input-pending-p to ignore), but then you said you did not mean anything so
precise.

So, if you could indicate a precise design that would be acceptable, I will
implement it. That seems to me a good medium between extending a hard fix
that I agree is inelegant, and my spending lots of time trying to
understand yet another area of Emacs.

-- 
http://rrt.sc3d.org