Interrupt causing problems for subsequent completions

Interrupt causing problems for subsequent completions

[Felix Rosencrantz]

>> Also, on a side note, I think there might be a bug in compadd (possibly
>> matching) that causes it to get in a bad state if an interrupt is sent while
it
>> is working.  I typically want to interrupt in situations when completion is
>> slow.  Completion works, but matching seems to have some problems.  (I know,
>> you probably want some those helpful details...)
>
>Indeed, I do.

Thought so...  Sigh, don't have much in the way of details yet.  I did notice
that the signal was being caught by a stack trace that ends like:
      
            strlen         
            modify         [subst.c:2001]
            paramsubst     [subst.c:1538]
            stringsubst    [subst.c:121]
            prefork        [subst.c:70]
            execcmd        [exec.c:1715]

So it looks like compadd is not being interrupted.  But maybe the completion
widget state is somehow messed up.

I have directory /drive2/tools that I attempt to type as "cd /d2/to<TAB>" that
completes properly ( my matching spec has  "r:[^0-9]||[0-9]=**"). But after
interrupting a completion widget, normal completion fails, and it attempts to
do correction, and doesn't expand properly.


-FR.

__________________________________________________
Do You Yahoo!?
Yahoo! Photos -- now, 100 FREE prints!
http://photos.yahoo.com

Re: Interrupt causing problems for subsequent completions

[Sven Wischnowsky]

Felix Rosencrantz wrote:

> >> Also, on a side note, I think there might be a bug in compadd (possibly
> >> matching) that causes it to get in a bad state if an interrupt is sent while
> it
> >> is working.  I typically want to interrupt in situations when completion is
> >> slow.  Completion works, but matching seems to have some problems.  (I know,
> >> you probably want some those helpful details...)
> >
> >Indeed, I do.
> 
> Thought so...  Sigh, don't have much in the way of details yet.  I did notice
> that the signal was being caught by a stack trace that ends like:
>       
>             strlen         
>             modify         [subst.c:2001]
>             paramsubst     [subst.c:1538]
>             stringsubst    [subst.c:121]
>             prefork        [subst.c:70]
>             execcmd        [exec.c:1715]
> 
> So it looks like compadd is not being interrupted.  But maybe the completion
> widget state is somehow messed up.
> 
> I have directory /drive2/tools that I attempt to type as "cd /d2/to<TAB>" that
> completes properly ( my matching spec has  "r:[^0-9]||[0-9]=**"). But after
> interrupting a completion widget, normal completion fails, and it attempts to
> do correction, and doesn't expand properly.

Yes, sounds more like some of the shell-code state not being reset
properly, I'll have a look (it may well be that this is a pure
shell-code problem, we don't catch SIGINT in there either).

Bye
 Sven


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