bug#47711: bug#48841: bug#47711: bug#48841: bug#47711: [PATCH VERSION 2] Add new `completion-filter-completions` API and deferred highlighting

[João Távora]

On Fri, Oct 27, 2023 at 1:11 AM Dmitry Gutov <dmitry@gutov.dev> wrote:

> > Also in the last iteration of the "yoyo" fido-vertical-mode test,
> > results with my latest patch are quite a bit faster.
>
> Hmm, your latest (lazy-hilit-2023-v3.diff) improves the (completing-read
> "" obarray) scenario a lot (over the original two 2023 solutions), but
> not the the 'C-h v' scenario. I don't have an explanation.

The improvement was due to running string-match only once per completion,
if you look at the changes to completion-pcm--all-completions.

It could be this code doesn't kick in in the C-h v scenario.  Notice
that that function already has some optimizations: for example, when
the regexp match is performed by all-completions and its use of
completion-regexp-alist, there's no way to get the regex match data
to compute the score, so it'll have to be postponed to
completion-pcm--hilit-commonality as in the v2.diff case.

Then again, that doesn't explain why in the C-h v scenario, the regression
was fixed precisely by adjust that code that I was conjecturing might
not kick in.

Anyway, I think it's safer to say that my latest patch is at least as fast,
sometimes faster, than the 2023 completion-filter-completions solution.

João