On Tue, 2025-02-11 at 23:08 +0300, Alexander Borisov wrote:
> What's the result?
>
> I would use Range Binary in Unicode case/normalization. The algorithm
> shows good results. Plus it can be customized (increasing/decreasing)
> the table by allowing empty values.
>
> Also, I got a strong feeling that Encoding could be optimized and
> improved. I am interested in this direction and would try my hand at
> improving Encodings.
>
> But I can't grab it all at once, it takes a systematic approach.
> I suggest starting with Unicode Case, then Unicode Normalization.
> Then experiment and improve Encoding.
Thank you.
That sounds good to me, I'll start looking at the patch details.
Regards,
Jeff Davis
