Any way to make sit-for avoid triggering redisplay in the general case?

Any way to make sit-for avoid triggering redisplay in the general case?

[Dmitry Gutov]

And not just "if input is available before it starts". Or is there a 
something I can call that would have a similar effect?

Effect like "wait until either a specified about of time passes, or the 
user types something, while allowing subprocess sentinels and filters to 
run, and hopefully timers as well, all without triggering redisplay".

company-mode has a flickering problem associated with asynchronous 
completion backends, and I have a seemingly easy solution in mind that 
turned out to depend on being able to do this.

Re: Any way to make sit-for avoid triggering redisplay in the general case?

[Noam Postavsky]

On Thu, Feb 8, 2018 at 5:40 PM, Dmitry Gutov <dgutov@yandex.ru> wrote:
> And not just "if input is available before it starts". Or is there a
> something I can call that would have a similar effect?
>
> Effect like "wait until either a specified about of time passes, or the user
> types something, while allowing subprocess sentinels and filters to run, and
> hopefully timers as well, all without triggering redisplay".

Does the NODISP arg not work?

(sit-for SECONDS &optional NODISP)

Redisplay, then wait for SECONDS seconds.  Stop when input is available.
[...]
If optional arg NODISP is t, don't redisplay, just wait for input.

Re: Any way to make sit-for avoid triggering redisplay in the general case?

[Dmitry Gutov]

On 2/9/18 01:49, Noam Postavsky wrote:

> Does the NODISP arg not work?

Nope: (progn (insert "a") (sit-for 0.2 t) (delete-char -1))

I read it the same way, but apparently it's just meant to avoid the 
initial redisplay. Looks pretty useless. Here's from the docstring 
continued:

   Redisplay does not happen if input is available before it starts.

Re: Any way to make sit-for avoid triggering redisplay in the general case?

[Noam Postavsky]

On Thu, Feb 8, 2018 at 5:52 PM, Dmitry Gutov <dgutov@yandex.ru> wrote:

>> Does the NODISP arg not work?
>
> Nope: (progn (insert "a") (sit-for 0.2 t) (delete-char -1))

Ah, how about inhibit-redisplay:

(let ((inhibit-redisplay t))
  (insert "a") (sit-for 0.2 t) (delete-char -1))

Re: Any way to make sit-for avoid triggering redisplay in the general case?

[Dmitry Gutov]

On 2/9/18 01:55, Noam Postavsky wrote:

> Ah, how about inhibit-redisplay:
> 
> (let ((inhibit-redisplay t))
>    (insert "a") (sit-for 0.2 t) (delete-char -1))

Ooh, that works! Thanks a lot.