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

[Dmitry Gutov]

On 27/10/2023 02:27, João Távora wrote:
> Dmitry Gutov<dmitry@gutov.dev>  writes:
>
> > My understanding is it's due to the judicious call (copy-sequence
> > orig) that you added before 'put-text-property' is called. While it
> > seems like a good idea to preserve the original value, when almost all
> > of obarray matches the current input (which is the current scenario),
> > a lot of strings will be copied.
> You're right, I reproduced the regression.  I thought I had taken out
> the copy-sequence, but forgot it there.  In an earlier stage I suspected
> that I needed the copy, but I don't think I do.  Please try this new
> patch that removes it.  I've also pushed it to the
> feature/completion-lazy-hilit branch.

Yep, without copy-sequence the regression is gone. Now the input strings 
are routinely mutated, though. ;-(

You could do a copy a little later -- after the match succeeds and the 
score is computed. But for short widely-matching inputs like 'a' that 
would make little difference.

I also experimented with hash-tables for "external" score storage. But 
it still comes out a little slower than either of the proposed solutions.