Re: Gap buffer problem?

[Eli Zaretskii]

> Date: Thu, 12 Dec 2024 00:11:54 +0200
> From: Dmitry Gutov <dmitry@gutov.dev>
> 
> > https://coredumped.dev/2023/08/09/text-showdown-gap-buffers-vs-ropes/
> 
> To quote from the bottom of the article:
> 
>    The way I see it, gap buffers are better for searching and memory
>    usage, but ropes are better at non-local editing patterns. Despite
>    their simplicity, gap buffers can hold their own in the modern world.
>    Maybe Emacs was on to something.
> 
> This is also my takeaway from reading a number of other texts on the 
> subject (not benchmarking personally, though, TBF).

Yes.  But one important aspect that blog doesn't touch is the
potential effect of changing the buffer text data structure on the
various Emacs display issues.  Some problems in the current display
code that cause slow redisplay in some situations (mainly, very long
lines) cannot really be solved as long as we stay with buffer text
stored as a long C string, with or without the gap.  This important
aspect of Emacs still awaits serious research of possible
alternatives, IMO.