Re: [PATCH 2/2] screen: Implement dead/zombie window polling

On Monday, December 17, 2012 06:51:55 PM Sadrul Habib Chowdhury wrote:
> Hi! From a quick glance, this looks like a good change. Comments inline:
> 
> On Thu, Dec 13, 2012 at 6:28 AM, Thomas Renninger <[email protected]> wrote:
> [snip]
> > --- a/src/doc/screen.1
> > +++ b/src/doc/screen.1
> > @@ -3542,6 +3542,15 @@ Optionally you can put the word \*Qonerror\*U after the keys. This will cause sc
> >  to monitor exit status of the process running in the window. If it exits normally ('0'),
> >  the window disappears. Any other exit value causes the window to become a zombie.
> >
> > +.BR "zombie_timeout" [\fIseconds\fP]
> > +.PP
> > +Per default
> > +.I screen
> > +windows are removed from the window list as soon as
> > +the windows process (e.g. shell) exits. If \fBzombie\fP keys are defined
> > +(compare with above \fBzombie\fP command), it is possible to also set a
> > +timeout when screen tries to automatically reconnect a dead screen window.
> > +
> 
> 'automatically reconnect a dead screen window'  -> 'automatically
> launch the process that was running in the dead window', or
> 'automatically resurrect the dead screen window'.
> 
> The 'zombie' command applies for all windows. But it looks like
> 'zombie_timeout' applies to the currently selected window, and all
> newly created windows? Can you please explicitly state that in the
> man/info pages.

Thanks for the review!
It took me a while to repost, due to Christmas/New Year vacations.

These patches introduce rather separated (no new interface or similar),
but useful features. They are implemented in separated code chunks making use
of existing screen functions/API. Therefore the code changes are
just some lines.
Would be great to see the patches getting applied.

Thanks,

   Thomas