Re: sit-for considered harmful in kbd-macros

[Stefan Monnier]

> > How about short-circuiting sit-for when called from a kbd-macro,
> > as if input was pending ?
> 
> I agree that it should definitely not delay execution (in fact, I
> don't think 20.7 did that).  I don't know if it should perform
> redisplay though.

I think it should behave exactly as if the macro's events were being
entered by the user: i.e. as if there was input pending.

So maybe the change should be done inside detect_input_pending_run_timers
(in keyboard.c).  Looking at requeued_events_pending_p (and the comment
before it), I see that someone has had similar ideas and that
pushing the idea further might not be as straightforward as it seems.

So maybe patching sit-for is good enough for now.


        Stefan