Use 2 letter language codes #975
Conversation
| if country_candidate: | ||
| country_result = _normalize_country_code(country_candidate) | ||
| if country_result is None: | ||
| return None |
There was a problem hiding this comment.
This matches the previous behavior, but I'm now wondering if it's better to keep the language code rather than dropping it on an invalid country code.
There was a problem hiding this comment.
I'm wondering if we have something like name:en="Foo", name:en_XY="Bar" do we want the name:en_XY to normalise to name:en and potentially clash with the other entry?
There was a problem hiding this comment.
Good point. We can set it up so that we'll only set the language code if one doesn't already exist in this scenario, but the added complexity might not be worth it.
There was a problem hiding this comment.
I've changed the behaviour in 4df4863 to keep the most specific language code. At the moment, that just means that any with an omitted country are less specific and won't overwrite a code without a country, e.g: en always beats en_NOEXIST.
There was a problem hiding this comment.
Great idea with priorities! ⚡ My first thoughts were also to just return a more complex value object, but just whether the country code was invalid or not. The priority notion is a better generalization of that.
…iority to the result, so that we can keep the most specific match.
|
👍 |
| 1. `name:de` German | ||
| 1. `name:gr` Greek | ||
| 1. `name:it` Italian | ||
| 1. `name:jp` Japanese |
There was a problem hiding this comment.
This should probably be ja instead of jp.
Use ISO 639-1 two letter language codes for WOF & OSM rather than ISO 639-3 3 letter codes. This partially reverses changes made for #418.
Connects to #972.
@rmarianski could you review, please?