job control

job control

[Peter Stephenson]

I was just trying to put together some tests for job control.  I had
problems with this:

while true; do sed -e 's/foo/bar/' nosuchfile >/dev/null; done

I'm not quite sure what's going on, but it takes a long time to kill or
suspend.  You can see lots of ^C's and error messages go sailing by until
eventually it stops.  It works much more reliably when operating on a file
that does exist.  That sort of suggests the problem happens when zsh,
rather than sed, gets the signal.  I had it inside a function to begin
with, but I think the problem was the same one.

Hitting ^Z is worse, since it still takes a long time, but when it
eventually works the whole thing freezes

-- 
Peter Stephenson <pws@ibmth.df.unipi.it>       Tel: +39 050 844536
WWW:  http://www.ifh.de/~pws/
Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy

RE: job control

[Andrej Borsenkow]

> >
> > while true; do sed -e 's/foo/bar/' nosuchfile >/dev/null; done
> >
>
> Damn. ^C works without problems for me (both on this DEC box here and
> on the Linux-Laptop). ^Z *is* a problem for us since we can stop only
> if you hit one of those sed's with the ^Z between the eyes -- only
> then can we find out that someone tried to suspend the loop.

Both work for me. Sometimes I need more than one ^C or ^Z. I believe, ^C should
have the same problem as ^Z.

/andrej

Re: job control

[Sven Wischnowsky]

Peter Stephenson wrote:

> I was just trying to put together some tests for job control.  I had
> problems with this:
> 
> while true; do sed -e 's/foo/bar/' nosuchfile >/dev/null; done
> 
> I'm not quite sure what's going on, but it takes a long time to kill or
> suspend.  You can see lots of ^C's and error messages go sailing by until
> eventually it stops.  It works much more reliably when operating on a file
> that does exist.  That sort of suggests the problem happens when zsh,
> rather than sed, gets the signal.  I had it inside a function to begin
> with, but I think the problem was the same one.
> 
> Hitting ^Z is worse, since it still takes a long time, but when it
> eventually works the whole thing freezes

Damn. ^C works without problems for me (both on this DEC box here and
on the Linux-Laptop). ^Z *is* a problem for us since we can stop only
if you hit one of those sed's with the ^Z between the eyes -- only
then can we find out that someone tried to suspend the loop.
However, this works for me, too: once I've managed to suspend the loop 
I can cleanly fg it, ^C it, and so on.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de