Re: emacs rendering comparisson between emacs23 and emacs26.3

[Stefan Monnier]

>>> that = skip jit-lock when input is pending.
>> AFAIK jit-lock is never skipped: it's (part of) the redisplay which is
>> skipped (and that in turn may include calls to jit-lock).
> This sounds like splitting hairs; if the part containing it is skipped, then
> jit-lock is skipped as well, obviously.
>
> Does that happen when fast-but-imprecise-scrolling is nil as well?

Skipping redisplay when there's pending input happens regardless of
`fast-but-imprecise-scrolling`.

Now here's where "splitting hairs" becomes important: while the choice
to skip redisplay is not affected by `fast-but-imprecise-scrolling`, the
choice to skip jit-lock *is*.

> Or is the situation that redisplay skips jit-lock when there's pending input
> (a good thing), but scrolling commands go back and force it anyway, to
> compute the new positions to scroll to?

That's right (tho, AFAIK the scrolling commands cause jit-lock *before*
redisplay happens, so "go back" is not quite accurate).


        Stefan