[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tramp (2.0.36); tramp-rsh-end-of-line needs to be changed in my conf

From: Greek0
Subject: Re: tramp (2.0.36); tramp-rsh-end-of-line needs to be changed in my configuration.
Date: Sun, 6 Jun 2004 12:56:38 +0200

On Sat, 05 Jun 2004 20:55:31 +0200
Michael Albinus <address@hidden> wrote:

> address@hidden writes:
> > I'm using tramp with Emacs on Windows Xp. I want to connect to a
> > linux box via ssh using putty (0.53b). The ssh server is running
> > OpenSSH_3.8p1 Debian 1:3.8p1-2. The ssh/sshx methods using the
> > cygwin OpenSSH binaries didn't work at all out-of-the-box, but I
> > didn't dig into the problem, instead I looked at putty/plink. Using
> > plink the connection is opened, the password is prompted, but after
> > `stty -inlcr -echo kill '^U'` it doesn't recognize the shell prompt
> > any more, and after 30 seconds it gives up. The reason is that
> > tramp-send-command-internal uses tramp-rsh-end-of-line (per default
> > "\n") as line terminator, which seems to expand to CRLF on windows.
> > The server misinterprets this as 2 line breaks, which results in 2
> > command prompts in the new stty mode: "$ $ ". This is the case after
> > every command given, since the wrong line terminator is always used.
> >
> > I fixed this by setting tramp-rsh-end-of-line to \r in my .emacs
> > file, everything works fine thereafter. The documentation to that
> > variable says it normally doesn't need to be changed, and a
> > notification is requested otherwise.
> > I guess it really is a windows specific problem that \n expands to 2
> > characters. Another way to fix it would be to tell stty to recognize
> > CRLF as one line-break instead of two, although I don't know how to
> > do this properly.
> See the discussion in variable `tramp-default-password-end-of-line'
> about this problem. I guess there is no consistent way to set it,
> otherwise Kai wouldn't have introduced it.

But still the tramp-rsh-end-of-line docs say that it shouldn't be
changed normally, perhaps a remark that it might be necessary with plink
would be nice. I don't know how often this is needed with plink, but if
it's more or less common, you should perhaps add paragraph to your

At first it really didn't look that tramp-rsh-end-of-line was the source
of the problem. I tried playing with the tramp-shell-prompt-pattern,
because I thought that's why the prompt isn't recognized. Then I thought
tramp-open-connection-setup-interactive-shell inserted an adiditional \n
somewhere, from there I dug into tramp-send-command-internal, where I
finally found tramp-rsh-end-of-line.
Long story short: It's not obvious to change tramp-rsh-end-of-line, so a
remark somewhere (aside from the documentation of a variable I didn't
even bother to look at in the beginning) would be nice.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]