Re: register_function line

• Date: Thu, 30 May 1996 17:49:07 PDT
• From:
  Diana L Harris <dlh5226@erc.jscc.cc.tn.us>
• Cc: local-moo@terminus.zytor.com, moo-cows@parc.xerox.com
• Content-Type: TEXT/PLAIN; charset=US-ASCII
• In-Reply-To: <4oj1ck$m3@terminus.zytor.com>

On Wed, 29 May 1996, H. Peter Anvin wrote:

> Followup to:  <Pine.LNX.3.91.960529192149.494A-100000@ralin.mistral.co.uk>
> By author:    bartmoss@mistral.co.uk
> In newsgroup: local.moo
> >
> > > float bf_distance(float stop[3], float start[3])
> >
> > The register line looked fine, did you notice that 
> > arglist.v.list[i].v.fnum was a pointer?
> > 
> 
> More importantly, the calling convention for a builtin function is:
> 
> static package
> bf_functionname(Var arglist, Byte next, void *vdata, Objid progr)
> {
> 	...
> }
> 
> In your case you have not only to unroll the arglist (which is a
> variable which is always a list) but also check the constitution of
> the component lists (since that is not done automatically).
> 
> Completely untested top-of-my-head code:
> 
> static package
> bf_distance(Var arglist, Byte next, void *vdata, Objid progr)
> {
>   double stop[3], start[3], *dp, dist;
>   Var r;
>   int i, j;
> 
>   for ( i = 1 ; i <= 2 ; i++ )
>     {
>       dp = (i==1) ? stop : start;
> 
>       r = arglist.v.list[i];
>       if (r.v.list[0].num != 3)
> 	{
> 	  free_var(arglist);
> 	  return make_error_pack(E_INVARG);
> 	}
>       
>       for ( j = 1 ; i <= 3 ; i++ )
> 	{
> 	  if ( r.v.list[j].type != TYPE_FLOAT )
> 	    {
> 	      free_var(arglist);
> 	      return make_error_pack(E_INVARG);
> 	    }
> 	  dp[j-1] = *r.v.list[j].v.fnum;
> 	}
>     }
> 
>   /* Do your thing, for example... */
> 
>   dist = 0.0;
>   for ( i = 0 ; i < 3 ; i++ )
>
****** Wouldn't I need a { here?
>     dist += (start[i]-stop[i])*(start[i]-stop[i]);
And } here?
>   dist = sqrt(dist);
> 
>   return make_var_pack(new_float(dist));
> }

extensions.c: In function `bf_distance':
extensions.c:190: structure has no member named `num' 
extensions.c:210: warning: implicit declaration of function `newfloat' 
extensions.c:211: incompatible type for argument 1 of `make_var_pack' 
*** Error code 1
Then it stopped...What's wrong with it? 

ALSO: When I compiled another MOO, the version of GCC didn't support the '-g'
option. I got hundreds of warnings - I mean lots of 'em. It still 
compiled and is running fine so far, but will this have any effect on the 
moo's operation or anything else? Is there a way around it (besides getting
a new version of GCC, of which I have no control over)?

Thanks again.

• Re: register_function line
• From: Thomas Wouters <zonny@dds.nl>
• Re: register_function line
• From: cadams@weather.brockport.edu (Chuck Adams)

• Re: register_function line
• From: hpa@trantor.zytor.com (H. Peter Anvin)

• Prev: Re: register_function line
• Next: Re: register_function line
• Index: MOO-cows
• Thread: MOO-cows