Test regression in rtest_trig with cmucl



On Fri, Jul 20, 2012 at 8:46 AM, Stavros Macrakis <macrakis at alum.mit.edu>wrote:

> Well, if we consistently treated (N1+N2*%i) as a 'number', then all sorts
> of things would automatically simplify, regardless of the value of numer
> and float, e.g. 2*(1+%i) => 2+2*%i, 2*(1.0+%i) => 2.0+1.0*%i (?), etc.


I was kind of avoiding having 2*(1+%i) become 2+2*%i.  Sometimes the former
is what you really want.  But when I say float(<hairy expression>), I
almost always want x+%i*y.


>
> By the way, won't calling rectform within this case be an infinite
> recursion?  The problem came up within rectform/polarform.
>

Oh, I wasn't suggesting changing the code, but changing the test suite.
Then there's not infinite recursion.

Ray