Re: Region cache for bidi paragraphs?

[Stefan Monnier]

> The gains sound marginal to me, especially given that the use case
> where the gains are tangible (Input 1, very long line and no paragraph
> separators anywhere in sight) doesn't look like a frequent use case.
> Even in that case you get a speedup of a few milliseconds per
> redisplay cycle.

I like performance improvements, but I tend to agree that in this case,
I'd first like to see some concrete use-case that benefit from it (and
"benefit" in this case should probably mean "finally keeps up while
scolling" or "makes it useable again").

So, does it help keep up with scrolling in the recent C++ example?
Or dos it make it possible to edit that 1MB file with no newlines?
Or is there some other concrete use case where it makes a tangible
difference (other than benchmarks)?

Of course, maybe in itself this patch doesn't do any such thing, but
combined with some other patches it will.  So, even if it doesn't seem
worth the trouble now, it might be worth keeping it around.


        Stefan