old or new style completion?

old or new style completion?

[Atom Smasher]

i've got a verbose copy defined as an alias in my ~/.zshrc:
  alias cpv="rsync -phb --backup-dir=/tmp/rsync -e /dev/null --progress --"

this needs a completion function similar to cp (cpv file to 
file-or-directory). i'm just starting to get my hands dirty with writing 
my own completion functions and i'm curious...

a) if an alias or function defined in ~/.zshrc should the corresponding 
completion widget/function also be in ~/.zshrc?

b) is the old style (compctl) completion system best suited to define a 
completion, if it's defined in ~/.zshrc?

c) being considered an "old" style of completion, is compctl in any danger 
of being deprecated? or does it still compliment the "new" completion 
system?

thanks...


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"A criminal is a person with predatory instincts without
 	 sufficient capital to form a corporation."
 		-- Howard Scott

Re: old or new style completion?

[Stephane Chazelas]

On Sat, Sep 29, 2007 at 11:36:27PM +1200, Atom Smasher wrote:
> i've got a verbose copy defined as an alias in my ~/.zshrc:
>  alias cpv="rsync -phb --backup-dir=/tmp/rsync -e /dev/null --progress --"
[...]

The completion system will complete according to the expanded
text of the alias. Here, it will complete a rsync command using
the _rsync completion function.

Above it doesn't work as you'd expect because of the "--" that
seems to confuse the completion widget.


So it would seem to be a bug in the _rsync completion function.


If you ommit the "--", it seems to work OK.

Cheers,
Stéphane

Re: old or new style completion?

[Atom Smasher]

On Sat, 29 Sep 2007, Stephane Chazelas wrote:

> The completion system will complete according to the expanded text of 
> the alias. Here, it will complete a rsync command using the _rsync 
> completion function.
>
> Above it doesn't work as you'd expect because of the "--" that seems to 
> confuse the completion widget.
>
> So it would seem to be a bug in the _rsync completion function.
>
> If you ommit the "--", it seems to work OK.
===================

1) if the completion widget for rsync farts after the "--" option, that's 
a problem with the _rsync function.

2) whether i declare "cpv" as an alias or a function, the completion 
system "interprets it" as the command is rsync, not cpv. if there's no way 
around that it seems like a problem with the completion system... no?

3) in this particular case, if #1 is addressed things will work fine (with 
some interesting exceptions), but #2 is still a problem.


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"Democracy must be something more than two wolves and a sheep
 	 voting on what to have for dinner."
 		-- James Bovard

Re: old or new style completion?

[Clint Adams]

On Sun, Sep 30, 2007 at 03:30:48AM +1200, Atom Smasher wrote:
> 1) if the completion widget for rsync farts after the "--" option, that's a 
> problem with the _rsync function.

Yes, _rsync is broken.

> 2) whether i declare "cpv" as an alias or a function, the completion system 
> "interprets it" as the command is rsync, not cpv. if there's no way around 
> that it seems like a problem with the completion system... no?

You are incorrect.  Try
cpv() {
rsync "$@"
}

You will not get rsync completion unless you do a "compdef _rsync cpv".
You could demand cp-like completion instead with "compdef _cp cpv",
though that would not be useful.  If you are witnessing different
results, perhaps you still have your alias set.

Re: old or new style completion?

[Atom Smasher]

On Sat, 29 Sep 2007, Clint Adams wrote:

>> 2) whether i declare "cpv" as an alias or a function, the completion 
>> system "interprets it" as the command is rsync, not cpv. if there's no 
>> way around that it seems like a problem with the completion system... 
>> no?
>
> You are incorrect.  Try cpv() { rsync "$@" }
>
> You will not get rsync completion unless you do a "compdef _rsync cpv". 
> You could demand cp-like completion instead with "compdef _cp cpv", 
> though that would not be useful.  If you are witnessing different 
> results, perhaps you still have your alias set.
=====================

yeah, i still had the alias defined; now it works fine as a function. 
since cpv is rsync, but neutered to the point of only copying to/from the 
local system i wouldn't want "compdef _rsync cpv", and i'm not even sure i 
need "compdef _cp cpv", since the default of auto-completing directories 
and files seems to work fine.


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"Facts are stupid things."
 		-- Ronald Reagan (1988)