scm_defined_p(sym, env)

scm_defined_p(sym, env)

[Richard Shann]

I have defined a function with one needed and one optional arg, using

scm_c_define_gsubr (name, 2, 0, 0, callback);

in my function I need to test if the second argument is present, it
looks like I need

scm_defined_p(sym, env)

but, if so, how do I find the value of env for the top-level
environment?

Richard Shann

Re: scm_defined_p(sym, env)

[rm]

On Tue, Aug 23, 2011 at 12:36:44PM +0100, Richard Shann wrote:
> I have defined a function with one needed and one optional arg, using
> 
> scm_c_define_gsubr (name, 2, 0, 0, callback);
> 
> in my function I need to test if the second argument is present, it
> looks like I need
> 
> scm_defined_p(sym, env)
> 
> but, if so, how do I find the value of env for the top-level
> environment?

Not shure I understand you here, Richard. Where would sym come from?
If  your function looks like this:

 (richards-function required-param an-optional-param)

thane your c function needs exactly two parameters:

 ricks_c_func(SCM req, SCM opt) {
 
 ...

  if (!SCM_UNBNDP (opt)) {
   /* opt wasn't provided, so do something here */
  }

  ...

 } 


Does that help? 

 Cheers, Ralf Mattes

> Richard Shann
> 

> 

Re: scm_defined_p(sym, env)

[Richard Shann]

I didn't find the SCM_UNBNDP() that you mention but since mailing the
list I stumbled on SCM_UNDEFINED and tried 
if(opt==SCM_UNDEFINED) ...
and that seems to work. 
So (unless I am doing something bad) I think I am back on course -
thanks!

Richard


On Tue, 2011-08-23 at 13:54 +0200, rm@tuxteam.de wrote:
> On Tue, Aug 23, 2011 at 12:36:44PM +0100, Richard Shann wrote:
> > I have defined a function with one needed and one optional arg, using
> > 
> > scm_c_define_gsubr (name, 2, 0, 0, callback);
> > 
> > in my function I need to test if the second argument is present, it
> > looks like I need
> > 
> > scm_defined_p(sym, env)
> > 
> > but, if so, how do I find the value of env for the top-level
> > environment?
> 
> Not shure I understand you here, Richard. Where would sym come from?
> If  your function looks like this:
> 
>  (richards-function required-param an-optional-param)
> 
> thane your c function needs exactly two parameters:
> 
>  ricks_c_func(SCM req, SCM opt) {
>  
>  ...
> 
>   if (!SCM_UNBNDP (opt)) {
>    /* opt wasn't provided, so do something here */
>   }
> 
>   ...
> 
>  } 
> 
> 
> Does that help? 
> 
>  Cheers, Ralf Mattes
> 
> > Richard Shann
> > 
> 
> > 

Re: scm_defined_p(sym, env)

[Andreas Rottmann]

Richard Shann <richard.shann@virgin.net> writes:

> I didn't find the SCM_UNBNDP() that you mention but since mailing the
> list I stumbled on SCM_UNDEFINED and tried 
> if(opt==SCM_UNDEFINED) ...
> and that seems to work. 
>
IIRC, that this works is not actually guaranteed (it depends on the type
of the SCM values; that type might not have defined/reasonable behavior
wrt. the "==" operator, so the right way to write the above comparison
would be `scm_is_eq(opt, SCM_UNDEFINED)', and incidentially, this is how
SCM_UNBNDP() is defined (see libguile/tags.h, at least on Guile 2.0).

> So (unless I am doing something bad) I think I am back on course -
> thanks!
>
The idiom you used may work for you now, but may be broken at any time,
or even when compiling Guile with a different configuration (search for
SCM_DEBUG_TYPING_STRICTNESS in libguile/tags.h for illustration).

HTH, Rotty

PS: Please consider avoiding top-posts (see for instance
<http://www.usenet-replayer.com/faq/alt.games.generals.html> for a
rationale).
-- 
Andreas Rottmann -- <http://rotty.yi.org/>

Re: scm_defined_p(sym, env)

[Richard Shann]

Thanks for this, I see there is also scm_is_true to take care of the
possibility that SCM_BOOL_T might not be a scalar type, we should be
using that too.

Richard

On Wed, 2011-08-24 at 12:01 -0400, guile-user-request@gnu.org wrote:
> 
> Richard Shann <richard.shann@virgin.net> writes:
> 
> > I didn't find the SCM_UNBNDP() that you mention but since mailing
> the
> > list I stumbled on SCM_UNDEFINED and tried 
> > if(opt==SCM_UNDEFINED) ...
> > and that seems to work. 
> >
> IIRC, that this works is not actually guaranteed (it depends on the
> type
> of the SCM values; that type might not have defined/reasonable
> behavior
> wrt. the "==" operator, so the right way to write the above comparison
> would be `scm_is_eq(opt, SCM_UNDEFINED)', and incidentially, this is
> how
> SCM_UNBNDP() is defined (see libguile/tags.h, at least on Guile 2.0).
> 
> > So (unless I am doing something bad) I think I am back on course -
> > thanks!
> >
> The idiom you used may work for you now, but may be broken at any
> time,
> or even when compiling Guile with a different configuration (search
> for
> SCM_DEBUG_TYPING_STRICTNESS in libguile/tags.h for illustration).
> 
> HTH, Rotty 

Re: scm_defined_p(sym, env)

[rixed]

-[ Tue, Aug 23, 2011 at 12:36:44PM +0100, Richard Shann ]----
> I have defined a function with one needed and one optional arg, using
> 
> scm_c_define_gsubr (name, 2, 0, 0, callback);

You mean :

scm_c_define_gsubr (name, 1, 1, 0, callback);

don't you ?