Re: telnet-mode with plink doesn't track input fields

[Kevin Rodgers]

> (let ((rlogin-program "c:/Program Files/PuTTY 0.58/PLINK.EXE")
>       (rlogin-explicit-args '("-telnet")))
>   (rlogin "solaris.example.com"))

I'm trying to find out why that works but my plink function doesn't.
(My function is exactly like telnet except that host is specified as an
argument to plink instead of via the telnet "open" command.)

I've compared the coding systems (as reported by `C-h C RET') in both
the *rlogin-HOST* and *plink-HOST* buffers, and they're the same.

I've compared the values of all the variables listed by `C-u
apropos-variable RET telnet\|comint RET' in both buffers, and found only
these differences:

variable                       rlogin (works)               plink (broken)
--------                       --------------               --------------
comint-delimiter-argument-list '(124 38 60 62 40 41 59)     nil
comint-input-autoexpand        'history                     nil
comint-input-filter-functions  '(shell-directory-tracker t) nil
comint-input-ring-file-name    "~/.history"                 nil
comint-input-sender            'comint-simple-send 
'telnet-simple-send
comint-ptyp                    nil                          t
telnet-count                   0                            5

(I made the comparison immediately after logging in to each remote
host.)

It doesn't seem likely that any of those settings would affect whether
comint can identify the remote shell's prompt and my user input.  I
think that's the root of the problem, since comint-use-prompt-regexp is
nil and the plink buffer has no field text properties whereas the rlogin
buffer correctly has field properties set to output and input, resp.

What can I do next to diagnose the failure to identify the prompt output
and user input?

Thanks,
--
Kevin