Re: the telnet/rlogin problem
On 7 May, Philip Hands wrote:
> That's not quite right, [<---] will be generating ASCII BS, which we take to
> mean emacs help, and perhaps (in bash for example) delete-left.
> [Delete] will be generating ASCII DEL which will do delete-left, so they don't
> get to do delete-under-cursor, and [<---] will often insert ^H.
Of course, you're right, I was referring to the reverse situation, when
telnetting to a broken host ([<---] sends ^? (DEL), which will do
delete-under-cursor on the broken system, and [Delete] will send ^[3~,
which will most likely do nothing).
> They just need to set stty in their .profile
> They can make this dependent upon $REMOTEHOST if necessary (i.e. if they
> only sometimes come in from a broken host).
> If that does not work, it is a bug in the application (because it should take
> heed of stty), which needs to be reported.
Right again, it turns out [<---] always works, so it's no problem after
> While such bugs exist, they could always set TERM to xterm-bs (or whatever)
> at the same time as setting stty erase ^h.
I wonder, when is a terminfo entry given a new name? Maybe the modified
xterm entry should be named differently? There seem to be quite a few
already. This is not very important at this point, but it's one of the
little things to be considered at some point. I take it the ultimate
goal is to get it all working in the distribution (after everyting is
tested, cleaned up, made conform the policy, etc.). Is there any plan or
timescale for the operation? What are the problems to be expected?
> possible /etc/profile snippet:
> # deal with connections from terminals who's [<---] key generates ASCII BS
> if grep -q "^$REMOTEHOST\$" /etc/broken-backspace-hosts ; then
> stty erase '^h'
> # handle buggy apps.
> [ "$TERM" = xterm ] && TERM=xterm-bs
I think this can be used to get [Delete] to work (on the broken host).
We'll have to make constructions such as:
for readline/bash (inputrc) and tcsh (cshrc). I take it the REMOTEHOST
variable will have to be exported by the telnet client?
Broken-clients using xterm-bs will have to resort to the rcfile-kludge
method to get the apps that ignore terminfo to work (since these apps
have been made to work with the modified xterminfo).
I have made a wrapper for telnet and rlogin:
if grep -wq "$@" /etc/broken-backspace-hosts ; then
exec kbdfix telnet.real "$@"
exec telnet.real "$@"
Where kbdfix is the expect script. This works fine for me. Are wrappers
like this acceptable?
It seems that everything works on and between "fixed-hosts". What's
To UNSUBSCRIBE, email to email@example.com
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org