Re: A86: Non-floating point parameters


[Prev][Next][Index][Thread]

Re: A86: Non-floating point parameters




Yeah, those error messages could be helpful...  I invoked a Domain error in asaptrig when the parameter
to TrigMode is incorrect by loading OP1 with 0, then taking the LN of it...  That works nicely, but an
error message would be better, and would maybe give a more accurate message.  The string wasn't
considered the 4th parameter, the variable was $0C (string), $03 (3 letters in name) "$",0,0 (name)...

Dux Gregis wrote:

> Stephen Hicks wrote:
>
> > Yeah - i just checked out the VAT.  # and ! always exist.  One holds whatever is currently on the
> > homescreen prompt (not sure which, but it'd be easy to figure out).  However, when you run a
> > token that takes a parameter (at least if the eostable uses $03, not sure about the others), it
> > creates a number of hidden vars equal to the number of parameters passed.  The variable "$",00,00
> > holds the 1st parameter; "$",01,00 holds the 2nd, and so on.
> >
> > This works for everything i've tried except reals...?
>
> Ah.  Very good to know.  Not sure why it would treat a string as a fourth parameter, but if it always
> does, you can go ahead and use rst 10h.  You will probably have to also load the type ($0c for
> strings) and invoke an error message if it's not the type you want...   If you need to know the call
> for a particular error message, just ask; I have them all written down.



--
Stephen Hicks
mailto:shicks@mindspring.com
ICQ:5453914
IRC:Kupopo



References: