RE: Command completion

RE: Command completion

[Sven Wischnowsky]

Andrej Borsenkow wrote:

> ...
> 
> Just to make sure we mean the same thing. It *does* insert nslookup.

Of course I tried it. I always try to try things.

> The
> example above was slightly incorrect; the actual screenshot after TAB is
> (with cursor immediately after nslookup)
> 
> bor@itsrm2% nslookup<CURSOR HERE>
> Completing external command
> nslookup
> Completing shell function
> nslookup
> 
> So, in this case I believe, that
> 
> - no list is to be displayed (we have just a single match)
> - exact match should be immediately accepted and space inserted

And that's exactly what I want, too.

> If you like to see list in this case ... another style?

I wondered about this, too, but what I meant was cases where there are 
also other strings so that get a list anyway.

Ok, it would also be cool, somehow, if we could hide matches for one
type when they are overridden by another type. But this would
definitely require C-code support and I have no idea how to do that in 
a generic way. And, as I said, I have the impression that in
completely ambiguous cases I would like to see the list (no, I can't
really explain why).

Bye
 Sven


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

RE: Command completion

[Andrej Borsenkow]

>
> Ok, it would also be cool, somehow, if we could hide matches for one
> type when they are overridden by another type. But this would
> definitely require C-code support and I have no idea how to
> do that in
> a generic way.

I may be wrong, but we have strict preference order, don't we? So, if
both $functions[nslookup] and $commands[nslookup] exist, we need not
offer the latter?

And, as I said, I have the impression that in
> completely ambiguous cases I would like to see the list (no, I can't
> really explain why).
>

Hmm ...

bor@itsrm2% whence -av nslookup
nslookup is a shell function
nslookup is /usr/sbin/nslookup

??

-andrej

RE: Command completion

[Andrej Borsenkow]

> Andrej Borsenkow wrote:
>
> > bor@itsrm2% nslo<TAB>
> > Completing external command
> > nslookup
> > Completing shell function
> > nslookup
> >
> > This obviously makes no sense. We _may_ make completion so
> smart, as to
> > insert full pathname (or command) for external commands -
> but currently
> > nslookup will always execute functions, won't it? So, the
> above has one
> > unambiguous match?
>
> Hmhm. I'd still like to see that information for really unambiguous
> matches, but in cases like this one we should just go ahead
> and accept
> the single possible string.
>
> No patch yet...
>


Just to make sure we mean the same thing. It *does* insert nslookup. The
example above was slightly incorrect; the actual screenshot after TAB is
(with cursor immediately after nslookup)

bor@itsrm2% nslookup<CURSOR HERE>
Completing external command
nslookup
Completing shell function
nslookup

So, in this case I believe, that

- no list is to be displayed (we have just a single match)
- exact match should be immediately accepted and space inserted

If you like to see list in this case ... another style?

-andrej

Re: Command completion

[Sven Wischnowsky]

Andrej Borsenkow wrote:

> bor@itsrm2% nslo<TAB>
> Completing external command
> nslookup
> Completing shell function
> nslookup
> 
> This obviously makes no sense. We _may_ make completion so smart, as to
> insert full pathname (or command) for external commands - but currently
> nslookup will always execute functions, won't it? So, the above has one
> unambiguous match?

Hmhm. I'd still like to see that information for really unambiguous
matches, but in cases like this one we should just go ahead and accept 
the single possible string.

No patch yet...

Bye
 Sven


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

Command completion

[Andrej Borsenkow]

bor@itsrm2% nslo<TAB>
Completing external command
nslookup
Completing shell function
nslookup

This obviously makes no sense. We _may_ make completion so smart, as to
insert full pathname (or command) for external commands - but currently
nslookup will always execute functions, won't it? So, the above has one
unambiguous match?

-andrej

Have a nice DOS!
B >>