[asterisk-dev] [Code Review] Channel Redirect CLI command

[Michiel van Baak]

On 18:18, Sat 13 Dec 08, Johansson Olle E wrote:
> >
> > (I also don't remember the discussion where we were trying to move  
> > away
> > from this type of functionality; do you have a thread you can direct  
> > me to?)
> >
> I can't point you to a specific message, but we did decide to  
> deprecate the "consise"
> CLI's because we felt that we did want to move API-like stuff to  
> manager and keep
> commands for admin usage in the CLI. 

I think we all agree(d) on that.

> I don't see many cases where you  
> have an
> operator or receptionist sitting at the CLI monitoring calls and doing  
> a redirect
> on the cli when they see a bad call. That's just for geeks that test  
> stuff, which might
> be a good reason to have it as an optional module. Normal users would  
> do it
> in a web interface, and the recommendation is that that kind of  
> interfaces use
> the AMI, not the CLI.

And that's why I really like this patch.
If I'm testing stuff on my testbed I dont want another console with the
ami connection open, just to create/redirect/cancel calls.
This one, in combination with the other channel and the console commands
make testing a lot easier.

> 
> That we have a bunch of these still around is no reason to keep  
> accepting new ones.

Amen.

> With that way of reasoning, we could add a lot of stuff that no longer  
> makes sense...
> We have to have an idea of where we want to go, some design to follow.  

I think we are already going that route by deprecating the consise
stuff, and recommend people to use the AMI when the have to parse the
output or use some non-local interface (webbased, application on desktop
etc).

> To copy
> a manager command in the CLI is just adding to the confusion on where  
> to use
> CLI and where to use AMI.

For most stuff I agree.
But this one is already in the manager, and a missing tool in
testing/debugging/designing dialplans.

> 
> To commit within a very short time after initiating a code review is  
> not very
> polite behaviour either... If you really want to have a review and  
> just not do it
> for the sake of the process, please give it time.

There are some reviews on reviewboard for some time now. Most reviews
that get committed quickly after the review is posted are patches that
normally would simply go in without review (or maybe review inside
Digium where we outsiders would have no knowledge of).
I think the requests that need review by more then just one or two
people stay on reviewboard long enough.
Take the calendar api for example. It's there for over a month now and a
lot of different people posted suggestions/corrections/reviews there.

As with all new things, people have to learn how to use it.
So maybe there were some 'too quick oks' on it but at least we now have
a tool to review patches more easily then the comment function on
mantis.

> 
> /O

-- 

Michiel van Baak
michiel at vanbaak.eu
http://michiel.vanbaak.eu
GnuPG key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x71C946BD

"Why is it drug addicts and computer aficionados are both called users?"