prompt problem

prompt problem

[Vincent Lefevre]

After a "zsh -f", source the following:

precmd()
{
  psvar[1]=$?
  if [[ $psvar[1] -gt 128 ]] then
    local sig=$signals[$(($psvar[1]-127))]
    [[ -n $sig ]] && psvar[1]=$sig
  fi
}

TRAPCLD() { [[ -o interactive && -n $TTY ]] && precmd }

PS1="%m:%20<...<%~%<<%(?..[%1v])%(#.#.>) "

then execute

  { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }

and interrupt it with Ctrl-C.

I can see two unexpected behaviors (not always reproducible, one
needs to try several times), as seen here:

ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[INT]> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[0]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%
ypig:~[0]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
^C%

So, sometimes the prompt is output twice. This seems to be a bug.

I sometimes get a "[0]" instead of "[INT]". This one may be due
to the fact that $? is 0 in the trap. I wonder whether this is
correct. At least the behavior doesn't seem to be consistent
with:

ypig:~> sleep 5 &
[1] 9217
ypig:~> blah
zsh: command not found: blah
ypig:~[127]> 
ypig:~[127]> 
[1]  + done       sleep 5
ypig:~[127]> 
ypig:~[127]> echo $?
127

where the value of $? seems to be preserved after "sleep 5" has
terminated.

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / Arénaire project (LIP, ENS-Lyon)

Re: prompt problem

[Vincent Lefevre]

Any news?

On 2010-08-20 14:28:23 +0200, Vincent Lefevre wrote:
> After a "zsh -f", source the following:
> 
> precmd()
> {
>   psvar[1]=$?
>   if [[ $psvar[1] -gt 128 ]] then
>     local sig=$signals[$(($psvar[1]-127))]
>     [[ -n $sig ]] && psvar[1]=$sig
>   fi
> }
> 
> TRAPCLD() { [[ -o interactive && -n $TTY ]] && precmd }
> 
> PS1="%m:%20<...<%~%<<%(?..[%1v])%(#.#.>) "
> 
> then execute
> 
>   { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> 
> and interrupt it with Ctrl-C.
> 
> I can see two unexpected behaviors (not always reproducible, one
> needs to try several times), as seen here:
> 
> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[INT]> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[INT]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[0]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> ypig:~[0]> { sleep 3 } | { sleep 3 } | { sleep 3 } | { sleep 3 }
> ^C%
> 
> So, sometimes the prompt is output twice. This seems to be a bug.
> 
> I sometimes get a "[0]" instead of "[INT]". This one may be due
> to the fact that $? is 0 in the trap. I wonder whether this is
> correct. At least the behavior doesn't seem to be consistent
> with:
> 
> ypig:~> sleep 5 &
> [1] 9217
> ypig:~> blah
> zsh: command not found: blah
> ypig:~[127]> 
> ypig:~[127]> 
> [1]  + done       sleep 5
> ypig:~[127]> 
> ypig:~[127]> echo $?
> 127
> 
> where the value of $? seems to be preserved after "sleep 5" has
> terminated.

-- 
Vincent Lefèvre <vincent@vinc17.net> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / Arénaire project (LIP, ENS-Lyon)

Re: prompt problem

[Bart Schaefer]

On Sep 7,  9:26pm, Vincent Lefevre wrote:
}
} Any news?

My guess would be that this is a race condition of some kind.  In some
cases the TRAPCLD function gets called twice on a single interrupt.
It's somewhat easier to make this happen [and much easier to observe
that it is what happens] if (a) you put a "print" statement in the
precmd function and (b) you vary the number of seconds passed to each
sleep in the pipe.  When called twice, $? differs in each call -- it's
130 (INT) in the first call, and 0 in the second, in every case where
I got two calls to happen at all.

The race probably depends on which part of the pipeline gets signalled
first -- something over which zsh has no control, as it's counting on
the OS to kill the entire process group when managing of a pipeline.

As a further guess, when the process group is killed zsh wakes up and
starts to print the prompt when the SIGCHLD is received for the group
leader.  Another child in the group then exits, catching the internals
of ZLE in a state where the prompt has been printed but it hasn't
recorded that fact yet.

Re: Prompt problem

[gwing]

Alexey Solovyov wrote:
:I just noticed that one-symbol PS1 is not handled properly. If it is a
:feature how to get rid of it?   :)

This was a bug for a short while, quite a while back.  
New versions should not have it.  If so, what version, what machine?
-- 
Geoff Wing [gwing@primenet.com.au]   Technical Manager
  Phone    : +61-3-9818 2977	     PrimeNet - Internet Consultancy
  Facsimile: +61-3-9819 3788	     Web : <URL:http://www.primenet.com.au/>
  Mobile   : 0412 162 441

Re: Prompt problem

[Goran Larsson]

> I just noticed that one-symbol PS1 is not handled properly. If it is a
> feature how to get rid of it?   :)

You are using zsh version 3.0.0 and just found the problem I reported
on the 3:rd of september. The fix from Geoff Wing <mason@primenet.com.au>
is included in version 3.0.1.

-- 
 Goran Larsson                mailto:hoh@approve.se
 I was an atheist,            http://home1.swipnet.se/%7Ew-12153/
 until I found out I was God.

Prompt problem

[Alexey Solovyov]

Hi, there,

I just noticed that one-symbol PS1 is not handled properly. If it is a
feature how to get rid of it?   :)

With respect,
----------------------------------------------------------------------
Alexey A. Solovyov
<alekso@math.ucr.edu>
<alekso@math.uu.se>