Re: `y-or-n-p' does not wait for input - assumes `n'

[Richard Stallman]

    Wouldn't it be better to fix this in C code so other potential uses of
    y-or-n-p and yes-or-no-p during track-mouse do not need to be fixed
    as well ?

I think that the right solution in this case is for the
specific callers to avoid calling them inside
of track-mouse.

    E.g. what if the undo-overflow question we are discussing in abother
    thread is asked while track-mouse is non-nil ?

I think the right solution for this is to bind track-mouse to nil in
undo-outer-limit-truncate.  That function is a very special case, and
I don't think it should drive everything.  Any other asynchronous
functions that asks questions would have to bind track-mouse too, but
there are very few of those.

Luc wrote:

    The problem is that `track-mouse' not only messes up y-or-n-p and
    yes-or-no-p.  It messes up all functions that are trying to read user
    input without wanting to track mouse motion.

Some kinds of input might ignore the mouse motion events, but
basically you are right.  However, I think the main solution should be
to limit the part of the code that is inside the track-mouse binding,
so that it doesn't affect other activities.