JFS uses UTF8 for encoding, though I guess it gets converted to/from your codepage when read/write.
I think the problem is actually git, which seems to encode in the native codepage. I committed an authors name that used umlauts, seemed to work fine with OS/2, Martin looked at the repository on I assume Windows and the name looked wrong, he changed it and then it was broken on OS/2 but looked correct in web browser using UTF8.
Searching, it seems you can use .gitattributes to control encoding, eg
https://stackoverflow.com/questions/48907049/what-is-the-appropriate-character-encoding-for-a-git-repo or the official documentation,
https://git-scm.com/docs/gitattributes/2.18.0I didn't know about this and need to remember it for similar reasons. I do routinely use .gitattributes to make Rexx scripts have the correct line endings.
.gitattributes from the screensaver,
# CRLF is crucial for OS/2 REXX files (also fine for Windows .cmd)
*.cmd eol=crlf
*.txt eol=crlf
# Be safe by making the Warpin script (*.wis) have crlf
*.wis eol=crlf
Edit: you likely need to experiment, let us know the results