The MUI scraping tools under /tools/locale are quite rough. I tried to get the job done using solely Bash as to not require Python but it ended up being infeasible. The end result now is that the scripts are quite horrific, alongside not really doing the job particularly well.
A few things for me to tackle:
Just don't even bother with Bash, it would be far easier to just use Python entirely for this tool (parsing the Windows resource binaries, and particular the UTF-16 strings etc. is all very craptacular to handle in Bash)
Provide a method of selecting a translation from a particular module from a set of matches, at the moment I sort of bodged it by 'prioritising' certain DLLs so their matches would win out in the case on conflicts... this is naff really, it would be better to manually pick a winner
Further onto this, the source for the translations should be remembered in case they ever need to be re-scraped (I need to do this for handling accelerators for instance)
When re-scaping, somehow we need to avoid bulldozing any custom translations (for new strings that don't exist in Windows itself, or when someone provides a better translation in future)
The MUI scraping tools under
/tools/localeare quite rough. I tried to get the job done using solely Bash as to not require Python but it ended up being infeasible. The end result now is that the scripts are quite horrific, alongside not really doing the job particularly well.A few things for me to tackle: