Should TermInfo parsing be a stdlib?

[vchuravy]

#50797 introduced a TermInfo parsing to Base. I think it is only part of 1.11 and not part of 1.10?

Since the code is fairly separated and given #51110 it might be useful to update the terminfo parser,
independently of Julia.

In particular given #50797 (comment) it would be easier to implement more involved behaviours and ship them to the user in a timely fashion if it was a stdlib.

@tecosaur what do you think?

[tecosaur]

I think it is only part of 1.11 and not part of 1.10?

Indeed, and it's not being backported AFAIK.

and given #51110 it might be useful to update the terminfo parser

I feel a need to add some context to this, that issue is with one of the seven terminfo files that have a problem (all screen.X) out of the 2873 terminfo files on my system. The seven problematic files all seem to relate to the same issue, and so I'd be inclined to think once that's solved we should be fine? It's not like the terminfo spec is changing any time soon (it's part of POSIX).

In particular given #50797 (comment) it would be easier to implement more involved behaviours and ship them to the user in a timely fashion if it was a stdlib.

This could make sense, I guess it depends what more involved behaviours we might want to implement? I haven't got any particular ideas at this point, but it seems like a reasonable thought to me.

---

Thinking on this a bit, I think this perhaps makes the most sense if thought of along with the exfiltration of the code involved in actually interacting with the TTY/REPL from Julia (e.g. with #51350). Let me know how this meshes with your own thoughts 🙂.