Re: [Babel] Marker does not point anywhere (when session buffer needs to be created)

From: "Sébastien Vauban" <wxhgmqzgwmuf-geNee64TY+gS+FvcfC7Uqw@public.gmane.org>
To: emacs-orgmode-mXXj517/zsQ@public.gmane.org
Subject: Re: [Babel] Marker does not point anywhere (when session buffer needs to be created)
Date: Thu, 04 Nov 2010 16:51:58 +0100	[thread overview]
Message-ID: <80lj595boh.fsf@mundaneum.com> (raw)
In-Reply-To: 87iq0dhzns.fsf@gmail.com

HI Eric,

"Eric Schulte" wrote:
> Sébastien Vauban <wxhgmqzgwmuf-geNee64TY+gS+FvcfC7Uqw@public.gmane.org> writes:
>> * Marker does not point anywhere
>>
>> Executing the following:
>>
>> #+begin_src sh :session NoSuchSessionYet
>> cd ~
>> ls *.txt
>> #+end_src
>>
>> outputs this in *Messages*:
>>
>> : executing Sh code block...
>> : think it is Cygwin...
>> : ~ 
>> : ansi-color-process-output: Marker does not point anywhere
>>
>> when executed for the *first* time.
>
> This "mark doesn't point anywhere" error is common across a bunch of
> sessions (ruby and python as well as shell). It's simply a matter of Babel
> trying to use the session before it is ready. I haven't found these errors
> to be an issue as all subsequent use of the session works as expected.

OK.

> I suppose that we should try to be more careful on session creation
> explicitly waiting until the buffer is ready, however this would add some
> degree of complication to the code...

Just a stupid idea: adding a sleep for 5 seconds or so, the first time the
session is created?  I wouldn't mind having to wait a bit on the first session
invocation, if I'm sure everything works better that way.

>> All subsequent executions don't exhibit this error (currently, they hang,
>> but that's another story).

The new posting of today refers to this issue.

>> My prompt is colored -- I need it, really!  *None* of the =ls= results are
>> colored, though (in this above case: =ls ~/*.txt=).
>>
>> Can we fix this somehow?

For now, I've given up using a colored shell, just to be sure I try to get sh
sessions working before adding complexity into the game.

So, now, my prompt is without any ANSI color spec, and completely mono-lined
(no extra \n), as you can see in the other post. Don't look at this for now --
except if you find it great to use! ;-)

>> Test the following, and you'll see you need a colored prompt as well, if not
>> yet convinced:
>>
>> #+begin_src sh :tangle .sva-bashrc
>> #*** Controlling the Prompt
>>
>> # define some colors
>> RED='\e[1;31m'
>> GREEN='\e[1;32m'
>> NO_COLOR='\e[0m'
>>
>> # my format of the prompt
>> function my_prompt_command ()
>> {
>>     # colorful prompt, based on whether the previous command succeeded or not
>>     if [[ $? -eq 0 ]]; then
>>         HILIT=${GREEN}
>>     else
>>         HILIT=${RED}
>>     fi
>>
>>     # replace the $HOME prefix by ~ in the current directory
>>     if [[ $HOME == ${PWD:0:${#HOME}} ]]; then
>>         NEWPWD="~${PWD:${#HOME}}"
>>     else
>>         NEWPWD=$PWD
>>     fi
>>
>>     # how many characters of the $PWD should be kept
>>     local pwd_max_length=15
>>
>>     if [[ ${#NEWPWD} -gt $pwd_max_length ]]; then
>>         local pwd_offset=$(( ${#NEWPWD} - $pwd_max_length ))
>>         NEWPWD="...${NEWPWD:$pwd_offset:$pwd_max_length}"
>>     fi
>>
>>     # prompt character
>>     if [[ $(whoami) = "root" ]]; then
>>         local PROMPTCHAR="#"
>>     else
>>         local PROMPTCHAR=">"
>>     fi
>>
>>     case $TERM in
>>         dumb)  # for Tramp?
>>             setenv PS1 "[\u@\h] ${NEWPWD}${PROMPTCHAR}"
>>             ;;
>>         *)
>>             setenv PS1 "\n\[${HILIT}\][\u@\h] ${NEWPWD}${PROMPTCHAR}\[${NO_COLOR}\]"
>>             ;;
>>     esac
>> }
>>
>> # shell prompt
>> setenv PROMPT_COMMAND my_prompt_command
>> echo ""
>> #+end_src

Best regards,
  Seb

-- 
Sébastien Vauban

_______________________________________________
Emacs-orgmode mailing list
Please use `Reply All' to send replies to the list.
Emacs-orgmode-mXXj517/zsQ@public.gmane.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode