*shell* fast and loose with what it sends to the shell

*shell* fast and loose with what it sends to the shell

[Dan Jacobson]

Often in the *shell* buffer, the first letter or so of a command gets
stripped when sent to the shell:

08:04 fm$ dlocate -L gdal
08:10 1 fm$ locate: invalid option -- L
Usage: locate [-d path | --database=path] [-e | --existing]

Here I typed "dlocate" but it only sent "locate" to the shell.
This may be because I had been moving the cursor around lots, and have
a fancy shell prompt.

Sure, today was no big deal, but one day it will remove somebody's
almost completed masterpiece, etc.

Re: *shell* fast and loose with what it sends to the shell

[Kevin Rodgers]

Dan Jacobson wrote:
 > Often in the *shell* buffer, the first letter or so of a command gets
 > stripped when sent to the shell:
 >
 > 08:04 fm$ dlocate -L gdal
 > 08:10 1 fm$ locate: invalid option -- L
 > Usage: locate [-d path | --database=path] [-e | --existing]
 >
 > Here I typed "dlocate" but it only sent "locate" to the shell.
 > This may be because I had been moving the cursor around lots,

Well, does it happen if you don't move the cursor around a lot?

 > and have a fancy shell prompt.

Well, does it happen if you don't have a fancy shell prompt?  If not,
what is your shell prompt and what are all the shell-* and comint-*
variables when it happens?

-- 
Kevin Rodgers

Re: *shell* fast and loose with what it sends to the shell

[Dan Jacobson]

emacs-version"21.3.1"
>> Often in the *shell* buffer, the first letter or so of a command gets

K> Well, does it happen if you don't move the cursor around a lot?

No. I recall it is after say some e.g. C-c C-p's, and maybe some
editing of commands not at the bottom of the buffer, and hitting RET
thereupon, or something like that.

>> and have a fancy shell prompt.

K> Well, does it happen if you don't have a fancy shell prompt?  If not,
K> what is your shell prompt 

I have never tried it without my fancy shell prompt, which in .bashrc is
case z-$TERM in
    z-rxvt)
	my_color_on="\[\e[01;35m\]" my_color_off="\[\e[00m\]"
	if test `id -u` = 0
	    then my_color_on="\[\e[01;41;30m\]"
	elif test -n "$GRASS_GUI"
	    then my_color_on="\[\e[01;46;30m\]"
	fi
esac
#Mr. http://cfaj.freeshell.org:
PROMPT_COMMAND='prompt_tricks_z=" $?"; prompt_tricks_z=${prompt_tricks_z# 0}
 case $PWD in $HOME) pwd=\~;; *) pwd=${PWD##*/};; esac'
PS1=$SPECIAL_PS1"$my_color_on\A\$prompt_tricks_z \$pwd\\$""$my_color_off "
PROMPT_COMMAND='prompt_tricks_z=" $?"; prompt_tricks_z=${prompt_tricks_z# 0}
 case $PWD in $HOME) pwd=\~;; *) pwd=${PWD##*/};; esac'
PS1=$SPECIAL_PS1"$my_color_on\A\$prompt_tricks_z \$pwd\\$""$my_color_off "

K> and what are all the shell-* and comint-* variables when it
K> happens?

Hmm, .emacs has
(add-hook 'comint-mode-hook
  (function (lambda ()(setq comint-input-autoexpand nil)))) ;no more "!" surprises
(add-hook 'shell-mode-hook
  (function (lambda ()(setq comint-input-autoexpand nil)))) ;no more "!" surprises
(setq comint-scroll-show-maximum-output t)

The whole thing is in http://jidanni.org/comp/emacs.txt.gz
Maybe I should write back when I get a reproducible example.
Like the scroll bar, I look up and it's gone for the rest of the
session, but I never am aware of just when it goes away.