Add language code mapping

[phlax]

No description provided.

[dwaynebailey]

An interesting concept would be to name certain mappings. So e.g. gnu will contain the general mappings for GNU software i.e. ca@venetian, zh_CN, zh_TW and map those to correct codes in Pootle by default.

[phlax]

definitely possible - can you give me some preset lists

[dwaynebailey]

Just flagging that we want to follow BCP47, for which this validator is really useful.

[dwaynebailey]

This Python language code library could be useful.

[dwaynebailey]

Created a langcode spreadsheet to create known mappings, only GNU for now.

[phlax]

@dwaynebailey this now basically works

I havent written tests yet for the presets - but will when i get them. I will probably tidy up the tests a bit at that point. I might also need to add some test for the finder class.

[dwaynebailey]

@phlax nice. Would be good to have a simple way to track the mappings outside of the code otherwise we have to update them in code which is OK I guess but isn't easy to maintain and adjust.

[phlax]

@dwaynebailey usually i would be against having them in code - but in this case they really do want to be constants. Ie if a user creates their mapping based on a preset - we cant change the preset without breaking their store/file associations

[dwaynebailey]

@phlax just fearful of a state where we have to make a release everytime we change the mappings. But yes I agree we pretty much need to keep these static. But lets be realistic, if we get these wring we should be able to migrate them even if a user makes use of them. We just need to make sure we catch those and give the users the tools to migrate things.

[phlax]

k - just trying to explain why i havent already proposed that we put these in the db - maybe we should - but that is the reason.

[phlax]

tho - thinking about it - if we are going to put a set of comprehensive/standard langs/codes in the db - then maybe that is the best place for the data.

[dwaynebailey]

@phlax think you keep missing that mappings I made in the doc pointed to by #7 (comment)

Our previous approach has been to put such shareable data into ttk. But I'm not worried about that now. Just flagging the lets load data into DB thinking as there are other places where we likely want to use these mappings.