Looks like you're bumping into things using varying text encoding, since your first example looks like UTF-8 interpreted as CP-1252 ('\u2014'.encode('utf-8').decode('windows-1252') to incorrectly interpret a UTF-8 em-dash yields that repeating pattern) and the second looks like UTF-16 also interpreted as CP-1252 ('T'.encode('utf-16').decode('windows-1252') gives you a byte order mark at the beginning then alternating null characters throughout).

A very simple heuristic you could use would be to assume UTF-8 unless there's a UTF-16 byte order mark at the beginning of the file and then use UTF-16. That would correctly handle both of the examples you've given and plain ASCII, but not any less common encodings. More robust would be to automatically guess the correct encoding using a library such as chardet.

I was able to get ticalc running in the browser without NPM or Node. I basically reverse-engineered the way ticalc is initialized, directly from ticalc.link.

also add 8xk support as ticalc.link doesn't support that, I don't know if it is because it is harder to do or they just didn't add it yet