Codemirror not highlighting code in locales other than English

[Danita]

Hi, I forked this wonderful project in order to add the Spanish translation, and today I stumbled on this. Please excuse any inaccuracies because I'm not completely familiar with the codebase yet.

As I understand the problem arises from the CM logo lexer cm/logo.js having the English keywords hardcoded. Possible solutions that come to my mind:

• Add a new lexer for each language code. Cons: too much shared code.
• Attempt to monkeypatch the CodeMirror mode indent() and token() after reading the l10n file to support custom regexes.
• Best, I think: programmatically generate the lexer according to the selected language and lazy loading it.

Thoughts?

[inexorabletash]

We really only need the keywords TRUE, FALSE, ELSE, TO and END, correct? (As an aside, the interpreter doesn't treat TRUE, FALSE and TO as keywords; perhaps it should.)

Another option that comes to mind is to have the mode's lexer have hooks that index.js can interact with, much like the interpreter itself. We could assign a keyword map to the global (window) or another object and the mode's lexer could use that rather than hardcoded regular expressions. That might be simpler.

Otherwise, I agree that programmatically generating the lexer might be easiest. I'd definitely be willing to review patches here. Thank you very much for reporting this and investigating possible options!

[Danita]

You're welcome Joshua, I liked your implementation of Logo very much and I'll be collaborating on a time-permitting basis 😄

I agree the hooks approach might be simpler to implement. I'll try this once I finish the Spanish translation.

[Svish]

Shouldn't the keywords (like is usually the case with programming languages) always be in English? Regardless of the language used "around" it?

[inexorabletash]

Since Logo is aimed at introductory programming (and kids) it is often localized, even down to the level of keywords. Particularly when the procedure library is localized, the remaining keywords — and there's only a few, listed above — really stand out.

Standardization and portability of the language was never a big concern.