Actually, switching the language (from the GUI Settings) to "Nederlands" will print out those errors:
This is also the case with "Bahasa Indonesia".
Before fixing translations, we should be robust against those errors.
I propose to create a Python script that will check translations files.
The idea is:
- to retrieve possible arguments contained in every sentences from the i18n.json file (it is in english, the reference language);
- then, for each other translations files, check sentences against the reference one.
If the sentence has not the same arguments (count and name), store the error in a list:
At the end of all checks, if the errors list is not empty:
- prints out all errors
- exit with a failure code -1
- A new option, --check-translations, will be added to deploy scripts.
- The script will be called before tests.
When traiting a translation file, it should be a good idea too to store excedant sentences (became obsolete at a certain point).
They will only be printed out, but not making the script to exit with a failure.