Update ucaps.h to contain proper case matchings

[Chubercik]

• This PR aims to perform the exorcism implement a fix @reduz mentioned in a Tweet a few days ago.
• It was obvious 5 years ago (😳) that the tables were incomplete: Reorder reverse caps characters table for string lower case conversion #26760 (comment).
• I wrote a script that goes through UnicodeData.txt (UCD 16.0.0) and rebuilds ucaps.h, so that it contains the latest unicode case matchings.

[bruvzg]

Ideally this should be generated from UCD (https://www.unicode.org/reports/tr21/tr21-5.html#UnicodeData, for these functions only singe character mappings are useful, TextServer already do the rest via ICU), not from python (which might use different UCD versions).

[Chubercik]

Why version 3.2.0 specifically?

[bruvzg]

Why version 3.2.0 specifically?

It should be the latest data, not 3.2,0 specifically, but seems like it was incorporated into other parts of the standard. But we need only legacy handling here (without string length changes).

[Chubercik]

Ideally this should be generated from UCD (https://www.unicode.org/reports/tr21/tr21-5.html#UnicodeData, for these functions only singe character mappings are useful, TextServer already do the rest via ICU), not from python (which might use different UCD versions).

Python version	UCD version
3.8	12.1.0 (2019)
3.9 (this PR)	13.0.0 (2020)
3.10	13.0.0 (2020)
3.11	14.0.0 (2021)
3.12	15.0.0 (2022)
3.13	15.1.0 (2023)
dev (3.14)	16.0.0 (2024)

Different UCD versions when using different Python versions is a problem only if we see it as such - if we want to have the latest case matchings, we can always run the script using the latest Python version; however, if this is a dealbreaker, I could write a script that parses UnicodeData.txt (UCD 16.0.0) (also see: UnicodeData.txt's Property Table, UnicodeData File Format).

---

Edit:

[...] I could write a script that parses UnicodeData.txt (UCD 16.0.0) (also see: UnicodeData.txt's Property Table, UnicodeData File Format).

Done, PR is ready for review :)

[Ivorforce]

There are also caps tables in char_range.inc. I don't know where those are from, but it would make sense to have them be up-to-date, and if they were created automatically, do both in the same script.

[Chubercik]

Will look into that later today, thanks for the heads-up.

---

Edit:

• Core: Automate generation of the char_range.inc file #101878

[bruvzg]

Looks good, I would apply the same changes proposed in #101878 (review) (header generation) to this script as well.

[Chubercik]

Sure thing, I'll look into it later today.

---

Edit:

Done, though I couldn't place the // This file was generated using the `ucaps_fetch.py` script. notice above the include guard; it might have something to do with Godot's code style enforced by pre-commit, idk.

[Repiteo]

Thanks!