Translation: Refactor locale matching to use proper language code

The previous code only parsed the first two characters (potentially reading
out of bounds if input was invalid), but some locales use a 3-letter language
code (e.g. 'nah_MX').

So I refactored the logic a bit to properly parse the locale and extract the
part left of the regional code, if provided (supports both 'en_US' and 'en-US'
style).

I made TranslationServer::get_language_code() public as I'll use it in a
follow up commit.

0 files changed, 0 insertions, 0 deletions