This was needed because when we add contexts to existing strings, it's considered as a new string and translators will have to translate those again.
To avoid that extra workload on all translators, this script matches contextualized strings with previous ones that had no context at all :)
I also had to replace all occurences of
translate-context="hello/world" (note the leading semicolon and the wrapping
'' were removed. I don't know why I thought it was important in the first place, but it caused some extractions issue, and things seem to work fine without those.