This is still an issue….
When scanning this set in ClrMame Pro it baulks on these two files containing Ž in their name, as it does when rebuilding (AFAIR). Even if you add them to the archive yourself, it insists the names are wrong (even when they clearly not when you open the archive). Even if it attempts to ‘correct’ them it fails anyway.
This has nothing to do with the OS, the packer (WinRAR/7z/zip), or even CMP itself (all modern versions of the software I’ve just mentioned can easily handle Unicode) – it’s a limitation of the old DAT format.
I raised this issue with Roman to confirm and received the (rather unsurprising) confirmation back:
“You're using a deprecated datfile format. Use XML instead and specify the encoding in the <? xml encoding...> attribute.”
So this leaves us with a few (equally crap) options….
- Remove the set for now (until we are in a position to support such sets in our DATs properly)
- Change the Ž character to something else, potentially breaking the image. As well as the other obvious disadvantage of, you know, it being named wrong…
- Leave as it is (and put up with being asked again and again as to why the TI-83 DAT is broken)