Our translations are generated with an automated script. Sometimes, the translations are not good enough and manual corrections must be done.
However, there is a risk of overwriting the corrections if we run the script once again.
Sometimes, such correction would be lost in merging PRs, as these files are regularly conflicted.
Solution
We could implement a system with both translations living in different files
auto/kr.json would be the korean translations auto generated
manual/kr.json would be the manual correction of some keys
And final translation would be taken from the "auto" files with "manual" on top to override some keys.
This will be easier to integrate contributions from community on translations, with much less friction during merge/conflict resolution, etc.
Current problem
Our translations are generated with an automated script. Sometimes, the translations are not good enough and manual corrections must be done.
However, there is a risk of overwriting the corrections if we run the script once again. Sometimes, such correction would be lost in merging PRs, as these files are regularly conflicted.
Solution
We could implement a system with both translations living in different files
And final translation would be taken from the "auto" files with "manual" on top to override some keys.
This will be easier to integrate contributions from community on translations, with much less friction during merge/conflict resolution, etc.