I think getting it solid is not just the terminfo but mapping Unicode, filtering ANSI escapes, and doing flow control.
Software flow control on Linux basically doesn't work. I'm not sure why but it's just to slow to react to a flow off from model t and then the model T's 64 bytes buffer gets overrun. So you need to use hardware flow control. Hterm addresses these things but it is otherwise very rudimentary. One of these days... But it was purpose built to address issues specifically with Linux that made it almost unusable for me. -- John.
