Strange "1;2600;0c" sequence in startup (message-mode?)
Jesse Sheidlower
I have recently been seeing a strange sequence of characters when I
start Emacs up to send a message, either when sending a new message, or
replying to an existing one. I believe that this started when I began to
use message-mode. Typically it will be "1;2600;0c", inserted at the
start of a message, but I have noticed other numbers close to "2600" in
the middle section.
This only happens when I have started typing before Emacs has fully
launched (usually by hitting return once or twice); if I wait for it to
load, I do not get any such sequence. But it does happen very regularly
when
It's been hard to Google for this. Does this ring a bell for anyone? I
have some hooks for message-mode, but nothing that looks like it could
cause this, although I don't know what the sequence represents.
Thanks.
Jesse Sheidlower
David Kastrup
> I have recently been seeing a strange sequence of characters when I
> start Emacs up to send a message, either when sending a new message,
> or replying to an existing one. I believe that this started when I
> began to use message-mode. Typically it will be "1;2600;0c", inserted
> at the start of a message, but I have noticed other numbers close to
> "2600" in the middle section.
This looks similar to cursor positioning control sequences. Are you
using Emacs on a character terminal rather than in graphical mode?
--
David Kastrup
Pascal J. Bourguignon
It's the end of an escape sequence. You're probably using emacs in a
terminal, and there must be a discrepancy between what the terminal is
really, and what declared in the environment variable TERM. Or perhaps
there's an error in the terminfo (or termcap) database.
--
__Pascal Bourguignon__ http://www.informatimago.com/
A bad day in () is better than a good day in {}.
Jesse Sheidlower
Yes, I am--when this happens I'm ssh'd into a Unix shell, and am
interacting with Emacs there, not in an X display.
Er, so what does this mean in terms of trying to fix it? It's a
relatively new phenomenon.
Jesse Sheidlower
Pascal J. Bourguignon
I told you. Set your TERM environment variable correctly, or correct
the terminfo database on the remote system.
Thomas E. Dickey
wrote:
> jes...@panix.com (Jesse Sheidlower) writes:
> > I have recently been seeing a strange sequence of characters when I
> > start Emacs up to send a message, either when sending a new message, or
> > replying to an existing one. I believe that this started when I began to
> > use message-mode. Typically it will be "1;2600;0c", inserted at the
> > start of a message, but I have noticed other numbers close to "2600" in
> > the middle section.
> It's the end of an escape sequence. You're probably using emacs in a
> terminal, and there must be a discrepancy between what the terminal is
> really, and what declared in the environment variable TERM. Or perhaps
> there's an error in the terminfo (or termcap) database.
It looks more like a response to the "Send Device Attributes" control
sequence.
If you're using gnome-terminal or some other terminal that uses vte,
that's
the likely explanation (can't google on this, since there's no useful
documentation
for vte).
Likewise, it won't be a problem with the terminal description, since
that feature's
only addressed in the (nonstandard) u8 capability, and emacs wouldn't
be using it.
HighCat
Gnome terminal (Ubuntu 10.10).
Very embarrassing sometimes, when you edit some script on server — and
later it crashes because of '1;2600;0c' in the beginning of file.
Should be considered a bug (from the end-user's view), I think..
On Apr 26, 3:29 pm, jes...@panix.com (Jesse Sheidlower) wrote:
> In article <8739l51gce....@fencepost.gnu.org>,
HighCat
Ilya Zakharevich
> I have this issue too, using Emacs 23.3.1, when ssh'ing from standard
> Gnome terminal (Ubuntu 10.10).
> later it crashes because of '1;2600;0c' in the beginning of file.
> Should be considered a bug (from the end-user's view), I think..
??? How can it be NOT a bug?
Here is an excerpt from my decade-old terminal query script:
# "\e[>c" returns "\e[>0;166;0c" on new xterms, and "\e[?1;2c" on Solaris ones
# I.e., Solaris ignores ">" (meaning "Secondary"), and sends the primary.
# Same on HP ones.
# On newer xterm the numbers mean: 0/1=vt100/220 (from decTerminalID),
# patchlevel, 0 [cartridge no]
See "Xterm Control Sequences",
http://invisible-island.net/xterm/ctlseqs/ctlseqs.html
Hope this helps,
Ilya