On Mar 30, 2014, at 20:26 , Joeish W <joeish80...@yahoo.com> wrote:
> Here is the definition, but thanks for the HELambdaP idea
> Trust me the opinion of a CFFI expert does matter...could you also show me
> what a HELambdaP definition would look like for this defcfun.
>
> ;; double RNG::uniform(double a, double b) ;; C++
> ;; double cv_RNG_uniform_double(RNG* self, double a, double b) ;; C
> (defcfun ("cv_RNG_uniform_double" uniform-d) :double
> "Returns the next random number sampled from the uniform distribution."
> (self (:pointer rng))
> (a :double)
> (b :double))
What you need is to define (I can actually define it for a next release) a
specialized DOC-BIT for CFFI “cfun”s and a EXTRACT-FORM-DOCUMENTATION method.
I first stab (untested) could be
(in-package “HELAMBDA”)
(defstruct (cffi-cfun-doc-bit (:include function-doc-bit (kind-tag “C
Function”)))
(c-name "" :read-only t :type string)
(cffi-argument-types () :read-only t :type list))
(defmethod extract-form-documentation ((fk (eql 'cffi:defcfun)) (form cons))
(destructuring-bind (defcfun name-and-options return-type &rest spec)
form
(declare (ignore defcfun))
(let ((name (munge-name-and-option-for-name name-and-options))
(c-name (munge-name-and-option-for-c-name name-and-options))
)
(make-cffi-cfun-doc-bit
:name name
:c-name c-name
:lambda-list (mapcar #’first (spec-sans-docstring spec))
:doc-string (if (stringp (first spec)) (first spec))
:values return-type
:cffi-argument-types (munge-cffi-spec spec)
#| other keys here … |#
)))
Note that the missing functions are needed to munge the DEFCFUN encoded
information. At this point, from your example, HELAMBDAP will happily produce
a documentation like the following:
Function uniform-d
Package:
CV_RNG
Syntax:
uniform-d self a b
Arguments and Values:
SELF : a T
A : a T
B : a T
RESULT : a T
Description:
Returns the next random number sampled from the uniform distribution.
If your doc string were the following:
"Returns the next random number sampled from the uniform distribution.
Arguments and Values:
SELF : a pointer to a RNG
A : a DOUBLE
B : a DOUBLE
RESULT : a DOUBLE
"
then the “Arguments and Values:” section would be copied as is (more or less).
Of course, you could define a specialized method to produce the DEFCFUN
documentation.
(defmethod produce-documentation ((format (eql 'html))
(doc-bit cffi-cfun-doc-bit)
(out file-stream)
doc-bits
&key
documentation-title
&allow-other-keys)
#| … |#
)
Of course, YMMV… As I said, I will be very happy to provide a HELambdaP
extension for CFFI. Actually, let me rephrase it: I will be happy to
coordinate a HELambdaP extension for CFFI :)
Cheers
—
MA
> On Sunday, March 30, 2014 9:44 AM, Marco Antoniotti <marc...@cs.nyu.edu>
> wrote:
> What is the CFFI actual definition?
>
> I - but that is a very selfish proposal :) - would just write a proper
> handler for HELambdaP.
>
> Cheers
>
> MA
>
>
>
>
>
> On Mar 30, 2014, at 17:11 , Joeish W <joeish80...@yahoo.com> wrote:
>
> > Which is the better way do write documentation for functions for a CFFI
> > library
> >
> > like this?
> >
> > C++: int RNG::uniform(int a, int b)
> >
> > Common Lisp: (UNIFORM-D (RNG (:POINTER RNG)) (A :DOUBLE) (B :DOUBLE)) =>
> > :DOUBLE
> >
> > or like this without the parenthesis
> >
> > C++: float RNG::uniform(float a, float b)
> >
> > Common Lisp: (UNIFORM-F (RNG :POINTER RNG) A :FLOAT B :FLOAT) => :FLOAT
> >
> > _______________________________________________
> > Cffi-devel mailing list
> > Cffi-devel@common-lisp.net
> > http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel
>
> --
> Marco Antoniotti
>
>
>
>
> _______________________________________________
> Cffi-devel mailing list
> Cffi-devel@common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel
>
>
> _______________________________________________
> Cffi-devel mailing list
> Cffi-devel@common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel
--
Marco Antoniotti
_______________________________________________
Cffi-devel mailing list
Cffi-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel
