[slime-devel] Re: [Patch] Improvements for arglist and slime-complete-form

[Matthias Koeppe]

Luke Gorrie <luke at synap.se> writes:

> I know it's late in the game to be asking this, but what's the
> complete-form command really used for?

I type an operator that has a complicated arglist.
Slime-complete-form fills in a template for the arglist.  I can C-M-f
over the formal arguments and replace them with the actual args.  I
can C-M-k unneeded keyword arguments.  When the actual args are
complicated, I often let-bind them to the formal argument name
instead, so I don't even have to modify the inserted template.  

This is more efficient (and feels much less stupid) than looking at
the echo area and typing in what is displayed there.

>> The patch integrates the code for detecting extra keyword arguments
>> for generic functions into the arglist display in the echo area, as
>> suggested by Luke. 
>
> What I would really like is that if I type this:
>
>   CL-USER> (defclass foo () ((x :initarg :x) (y :initarg :y)))
>   #<STANDARD-CLASS FOO>
>   CL-USER> (make-instance 'foo 
>
> then slime-space arglist display shows this:
>
>   (make-instance 'foo &key x y)
>
> basically for slime-space to generate output similarly to
> slime-complete-form. 

Yes, I also want this.  It requires, however, a change in the protocol
of ARGLIST-FOR-ECHO-AREA.  Currently it only gets the list of nested
operator names as strings.  For the new functionality, we would also
need the "nested operands" before point.  And one needs to be more
careful with READing the operands before point, so as not to INTERN
too much.  (Like for reading the operators, where SLIME uses
PARSE-SYMBOL rather than READ.)  Maybe someone would like to help
here. 

> And if we had such good arglist display as this I
> wonder if slime-complete-form would still be neede?

Yes!

Cheers,
-- 
Matthias Koeppe -- http://www.math.uni-magdeburg.de/~mkoeppe