Right, well, I think we can all argue that it's not particularly useful
having tset in one's .login, and certainly we can instruct users how to
correct their .login file so as to do so in the correct fashion if they feel
they really need to. But that's documentation, and who reads that? The
reality is that a significant percentage of potential users will have tset
commands in their .login file, and rcp.el will not work "out of the box" for
these users. Telling potential users that they have to change their .login
files when they may not have changed their .login files in years may be
off-putting to some users. I think it's worth investigating whether rcp.el
could work around it in some fashion that may minimize or eliminate the
problem. Could rcp.el detect the tset prompt? Or could rcp.el send a CR or
two if no command-line prompt becomes evident within a few seconds of logging
in? Something like that. If the cure is worse than the disease, then by all
means, let's display a well-worded error message and refer the user to the
documentation, but I think it's worth investigating whether rcp.el can work
around the problem to make that unnecessary. Basically, it's about presenting
the best "out of the box" experience for the user.
