Capitalizing the German letter ß gives SS instead of ẞ

[septsea]

Describe the bug Capitalizing the German letter ß gives SS instead of ẞ.

To Reproduce Steps to reproduce the behavior:

1. Type any text that contains ß. (For example, Gauß und Weierstraß.)
2. Select the text.
3. Press gU.
4. Gauß und Weierstraß is converted to GAUSS UND WEIERSTRASS.

Expected behavior Since there is already the uppercase version of ß (https://glyphsapp.com/learn/localize-your-font-german-capital-sharp-s), I expect the result to be GAUẞ UND WEIERSTRAẞ.

Screenshots If applicable, add screenshots to help explain your problem. If remapping-related, please attach log output: https://github.com/VSCodeVim/Vim#debugging-remappings.

Environment (please complete the following information):

• Extension (VsCodeVim) version: 1.24.1
• VSCode version: 1.71.2
• OS: Windows_NT x64 10.0.25211

Additional context Add any other context about the problem here.

[J-Fields]

Hmm... looks like in my locale, at least, 'ß'.toUpperCase() and 'ß'.toLocaleUpperCase() both return 'SS'. This could be a locale issue though (this stuff is definitely not my wheelhouse).

@septsea Can you please confirm whether the g~ (toggle case) command works correctly on your machine with 'ß'? I noticed gU and gu use the locale-agnostic methods while g~ uses the locale-sensitive methods.

[septsea]

The g~ command converts ß to SS. @J-Fields

[J-Fields]

Yeah, my testing suggests this is an issue with javascript's toLocaleUppercase (at least in common implementations) even with the German locale.

[septsea]

I did some experiments.

• "Möbelträgerfüße".upper() in Python returns MÖBELTRÄGERFÜSSE.
• "Möbelträgerfüße".toUpperCase() in Java returns MÖBELTRÄGERFÜSSE.
• "Möbelträgerfüße".upcase in Ruby returns MÖBELTRÄGERFÜSSE.
• "Möbelträgerfüße".uppercased() in Swift returns MÖBELTRÄGERFÜSSE.
• "Möbelträgerfüße".to_uppercase() in Rust returns MÖBELTRÄGERFÜSSE.
• "Möbelträgerfüße".ToUpper() in C# returns MÖBELTRÄGERFÜßE (instead of MÖBELTRÄGERFÜẞE).
• strings.ToUpper("Möbelträgerfüße") in Go returns MÖBELTRÄGERFÜßE (instead of MÖBELTRÄGERFÜẞE).

Well. ẞ was nonexistent or nonstandard when toUpperCase (or toLocaleUpperCase, or their counterparts in other programming languages) was created.

[lordofscripts]

Last year I ran into a description somewhere that there was indeed such a problem with THAT German character. To me it was (and is) a problem because I am working on an application where that translation should return ONE character, not two ("ss").

And now in another application with a TabulaRecta things are going amiss thanks to THAT specific character.