Cursor movement screwy on remote sessions.

Cursor movement screwy on remote sessions.

[Ged Haywood]

Hi folks,

First post, joined the list today for just this question.  Read the
faqs, Googled etc, didn't see anything remotely related to this.

I use emacs remotely all over the place.  One particular problem gives
me great grief, only when using emacs remotely.

The cursor doesn't reliably move to the place it's supposed to go to
on the screen, with the result that you're inserting/deleting text at
a place diffeent from the one you see on the screen until you refresh.

The problem occurs mostly when I'm using an instance of Emacs which is
running on a machine in Sheffield, England, but I'm sitting in Ceret,
France running an xterm on my local machine.  The screen is about 190
columns dies and 70 rows.  X is not running on the remote host and the
version of Emacs on there was compiled without X.  Until today it was
version 21.2 but now it's 21.3 and that has made no difference to the
problem.

Ordinary characters malmost always ove the cursor properly AFAICT.

Cursor movement sequences like [ESC-f] will move the cursor in the
most unexpected of ways.  Sometimes it will move back to the beginning
of the line instead of to the end of the next word.  Oddly enough, an
[ESC-b] usually does what's intended.  [ESC-c] will capitalise the
word but not move the cursor to the end of it and it may move it to
somewhere in the middle of it, and it may make the word look garbled.

Sometimes I press the space bar after a word and the cursor doesn't
move at all. [ESC-190-space] will put the cursor in the right place.

Usually [CTRL-L] will put the cursor in the right place.

The cursor generally stay on the right line, just in the wrong column.

I have the feeling that this is some weird interaction with ncurses.

If I run Emacs using the same xterm but over a VPN to a machine at my
office near Sheffield (about 25 miles south of the problem machine) I
have no problems.  I sometimes have simlar problems running Emacs on
the Sheffield machine from my office, but not so bad.

I never have problems like this with Vi.  I hate Vi, but I'll go and
scrub with carbolic now anyway.

Any suggestions?

73,
Ged.

Re: Cursor movement screwy on remote sessions.

[Thomas Dickey]

Ged Haywood <ged@jubileegroup.co.uk> wrote:
> Hi folks,

> First post, joined the list today for just this question.  Read the
> faqs, Googled etc, didn't see anything remotely related to this.

> I use emacs remotely all over the place.  One particular problem gives
> me great grief, only when using emacs remotely.

> The cursor doesn't reliably move to the place it's supposed to go to
> on the screen, with the result that you're inserting/deleting text at
> a place diffeent from the one you see on the screen until you refresh.

it's probably network delays (the escape sequence is read over too long
a time interval to allow the application to interpret it as one chunk).
I don't know what you must tweak in emacs to make this work (it's simple
with ncurses).

> I never have problems like this with Vi.  I hate Vi, but I'll go and
> scrub with carbolic now anyway.

vi usually has a longer timeout (and it's a settable value).
But it's susceptible to the same problem.

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

Re: Cursor movement screwy on remote sessions.

[Ged Haywood]

Hi there,

On Sat, 18 Sep 2004 Thomas Dickey wrote:

> Ged Haywood wrote:
>
> > The cursor doesn't reliably move to the place it's supposed to go to
>
> it's probably network delays (the escape sequence is read over too long
> a time interval to allow the application to interpret it as one chunk).

That makes perfect sense.  Thanks Thomas, I'll look into it.

> I don't know what you must tweak in emacs to make this work (it's
> simple with ncurses).

Now you've said it, a few pennies are starting to drop.  It looks as
though most of the time it's the local end that's getting it wrong -
Emacs usually seems to know where the cursor should be, but the xterm
doesn't - but it's not always that way, which could explain the screwy
nature of the problem and why the cause isn't as clear as it might be.

It probably means I'd be better off running Emacs locally on the
remote file rather than running emacs remotely on the same file.
Unfortunately it's often a remote application that's starting Emacs
for me, but I can easily get around that.

Thanks again.

73,
Ged.

Re: Cursor movement screwy on remote sessions.

[kgold]

Ged Haywood <ged@jubileegroup.co.uk> writes:
> I use emacs remotely all over the place.  One particular problem gives
> me great grief, only when using emacs remotely.
> 
> The cursor doesn't reliably move to the place it's supposed to go to
> on the screen, with the result that you're inserting/deleting text at
> a place diffeent from the one you see on the screen until you refresh.
> 
> The problem occurs mostly when I'm using an instance of Emacs which is
> running on a machine in Sheffield, England, but I'm sitting in Ceret,
> France running an xterm on my local machine.  The screen is about 190
> columns dies and 70 rows.  X is not running on the remote host and the
> version of Emacs on there was compiled without X.  Until today it was
> version 21.2 but now it's 21.3 and that has made no difference to the
> problem.
> [snip]
> Any suggestions?

Another possibility.  Do you have emacs on the local Ceret machine, or
can you install it?

The way I do remote editing is with ftp.  emacs will automatically use
ftp to read and write the file on the remote machine.  Not only should
it eliminate your problem, but performance should be much better,
since editing is done locally.  And you'll be using your emacs with
your .emacs, not what they happen to install in Sheffield.