Using guile's introspection facilities

Using guile's introspection facilities

[Daniel Ridge]

Guilers,

Here's a snippet of a trivial C module that makes a couple of non- 
visible libguile functions available for your use:

SCM_DEFINE(guile_srcprops_p, "srcprops?", 1, 0, 0,
            (SCM obj),
            "is obj a srcprops?")
#define FUNC_NAME guile_srcprops_p
{
   if(SRCPROPSP(obj)) {
     return SCM_BOOL_T;
   } else {
     return SCM_BOOL_F;
   }
}
#undef FUNC_NAME

SCM_DEFINE(guile_srcprops_to_plist, "srcprops-to-plist", 1, 0, 0,
            (SCM obj),
            "convert a guile srcprops object to a plist")
#define FUNC_NAME guile_srcprops_to_plist
{
   return scm_srcprops_to_plist(obj);
}
#undef FUNC_NAME

I use these in guile 1.8 to allow me to use source-whash as a  
navigation aid. With these bindings, I found
that I could evaluate expressions like:

(let* (
           (propl (hash-fold (lambda (k v p) (cons (srcprops-to-plist  
v) p)) '() source-whash))
           (all-filenames (map cdr (map (lambda (v) (assoc 'filename  
v)) propl)))
           (unique-filenames (unique (sort (delete #f (map (lambda  
(v) (and (string? v) v)) all-filenames)) string>?)))
         )

     (apply with-xml-tag `("<div>"
       ,(make-raw (apply string-append `(
         "There are: " ,#~(length propl) " source terms on file\n"
         "Source files:<br/>"
         ,@(map (lambda (v)
           (string-append v "<br/>")
         ) unique-filenames)
       )))
     ))
   )

that (if you excuse a couple of library functions of mine) generates  
HTML of the list of files associated by the reader with your runtime.  
This function, and functions like it are enormously helpful for me as  
I navigate my own guile environments.

Can anyone think of a way to accomplish this same result without  
exposing srcprops-to-plist? If not, can anyone think  of a problem  
with the above solution?

If there are no problems, I propose adding these bindings to libguile.

Cheers,
	Dan Ridge


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user

Re: Using guile's introspection facilities

[Kevin Ryde]

Daniel Ridge <dskr@mac.com> writes:
>
> SCM_DEFINE(guile_srcprops_p, "srcprops?", 1, 0, 0,

I don't think that's meant to be visible, merely a compact form of
what's normally prsented as an alist.  If you want a test then one
possibility would be that it automagically shows up in goops as a
`<srcprops>' class

	(is-a? my-object <srcprop>)

I've used that on (documented) objects like for instance arbiters
which don't otherwise have a test func.

> (let* (
>           (propl (hash-fold (lambda (k v p) (cons (srcprops-to-plist
> v) p)) '() source-whash))

I'm not sure if source-whash is actually meant to be used.  But if you
do then I think `source-properties' can be called on each key in the
hash to get info in alist form.


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user

Re: Using guile's introspection facilities

[Neil Jerram]

Kevin Ryde <user42@zip.com.au> writes:

> Daniel Ridge <dskr@mac.com> writes:
>>
>> SCM_DEFINE(guile_srcprops_p, "srcprops?", 1, 0, 0,
>
> I don't think that's meant to be visible, merely a compact form of
> what's normally prsented as an alist.

And in any case, your motivating code doesn't use this, does it?

>
>> (let* (
>>           (propl (hash-fold (lambda (k v p) (cons (srcprops-to-plist
>> v) p)) '() source-whash))
>
> I'm not sure if source-whash is actually meant to be used.  But if you
> do then I think `source-properties' can be called on each key in the
> hash to get info in alist form.

Agreed.  So I don't see a remaining argument for exposing these.

Regards,
     Neil



_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user