Feature Request: Text Transformation interface.

[leximayfield]

Pretty simple. It would be really nice if RmlUi had a text transformation interface that allowed users to override the handling of a string transformation.

Right now, text-transform seems to be hardcoded to only handle ASCII properly. It would be nice to be able to override this functionality in order to plug in an i18n function that either uses a locale-aware function, ICU, or some other way of uppercasing/lowercasing/capitalizing a string.

[Paril]

Yeah I brought this up before, but having a word break iterator through here would also be huge. I think it should prob be its own interface API so that we can potentially upstream implementations that use ICU or home-baked versions that handle the majority of cases.

[mikke89]

Yeah, this would be handy for internationalization.

It could make sense to integrate this into the font engine interface. At some point we want to have better API for text shaping, including script detection, normalization and other operations along those lines. Although I would be okay with making a separate interface for now.

[Paril]

Hmm. I don't think you'd want them in the same interface; like you wouldn't want the FreeType font interface to be stuck with UTF8 if you want to use something different. They should be separate things; character sets typically aren't hard-linked to fonts (and vice versa).

[mikke89]

Yeah, I guess it makes some sense to separate them. I'd be okay with it either way I think. Using UTF-8 on the library side and in the interfaces is quite set in stone though, that won't be changing anytime soon.

[TriangulumDesire]

I've had an idea for something similar that I'd like to see. I'm not sure if this is what you'd want in the library, but I'd be fond of a ‘localisation’ interface that would handle localisation/internationalisation tasks such as the following:

• Translating strings.
• Transforming strings (lowercase, uppercase, titlecase, et cetera).
• Normalising strings.
• Text segmentation (script/direction).
• Determining character properties (for example, if a character is whitespace).
• Word breaking.
• Grapheme breaking (for highlighting text).

And the default implementation would just assume that everything is in ASCII.