Write smarter RTP translation table logic #802
One issue is that different translations map the same string to different assets. This could be solved by matching all RTP-translations against the RTP directory file list and the one with most hits is used as the reference RTP.
What is also missing is that N-to-M mapping. Currently it's Japanese to Other, instead of "Prefered one" to All.
An example case that should ideally be handled by the new code.
What I would also propose is:
Scan the RTP folders and then determine how many % of a RTP overlap with our internal table (hopefully 100% ;)) and then use the detected RTP as the reference for lookup.
But for this we would need a different data structure first to make this more efficient. Currently we have "Any -> Japanese". A table structure like in the Wiki would be better:
1 Line with one RTP per column:
This way you could easier do Any-To-Any mappings.