Minetest version

Minetest 5.9.0-dev

Summary

The script uses regex to parse translation lines, which is incorrect. Take the following example:

Something@@= Something@@

The script will ignore this line as it thinks the @ is escaping the =, but it is actually adding @. Another issue: the script ignores invalid lines without warning!

The script also assumes that a translation is only ever on one line. This is not the case, given @\n:

A @n newline = Une @
nouvelle ligne

Steps to reproduce

I've written my own Python .tr file parser with unit tests here. Unit tests should be added to mod_translation_updater.py, it's unacceptable to have a parser without tests.

My implemention throws a lot of information away (ie: comments) so couldn't be taken directly, but inspiration can be taken from the approach (which is based on the .cpp code)