Advanced locale fallback

[EliteMasterEric]

If these features already exist, I didn't see them in the code and no documentation on it is available.

As seen below, the BCP-47 locale standard (as used by browsers) is multi-layered, as seen below:

Besides the language code, there are also parts for a script code, country code, variant, and extensions.

• zh (language code) — Chinese language
• Hans (script code) — written in simplified characters
• CN (country-code) — as used in China.
• bauddha (variant) — using a Buddhist Hybrid Sanskrit dialect
• u-nu-hanidec (extension) — using Han decimal numbers

The intent of these codes is SPECIFICITY. Each subsequent part of the code, if it is provided, conveys a specific extension on the language before it.

My suggestion is that this module implement the following procedure:

• Add a simple function to split locale codes into up to 5 parts as seen above.
• When localizing strings, first attempt to use the current language.
• If the current language does not contain the specified key, or the current language is not localized, strip the last part and try again.
• If no parts remain, throw an warning in the console and return the default language value.

Here is an example:

• The current language code is zh-hans-cn-bauddha-u-nu-hani.
• The code makes a request to localize-strings with some localization data.
• If the localization data contains a set of data for zh-hans-cn-bauddha-u-nu-hani, respond to that request as appropriate.
• If the localization data does NOT contain data for zh-hans-cn-bauddha-u-nu-hani, or is missing that specific localization string, do not throw an error immediately. Instead, maybe throw a warning, then proceed to the next step.
• Look in the localization data for zh-hans-cn-bauddha. If data isn't there...
• Look in the localization data for zh-hans-cn. If data isn't there...
• Look in the localization data for zh-hans. If data isn't there...
• Look in the localization data for zh. If data isn't there...
• Look in the localization data for the default locale.

The idea is that you could translate for 'zh', creating a baseline translation for all Chinese languages, then adding corrections for more specific locales, without having to copy and paste between languages and without having to make 'zh' your default locale.

[stefalda]

HI @MasterEric, nice analysis... it seems feasible... it just needs to be coded...

[GrantDG]

I definitely vote for this. However, a challenge is working out which part is missing...

I'm in Australia, and our language code is en-AU, which is '[language code]-[country code]' (we don't have alternate scripts). There's definitely a fallback from en-AU to en - as we only differ in a few words from en (also usually considered synonymous with en-US)