On Fri, Jun 20, 2008 at 8:46 AM, Raymond Toy (RT/EUS)
<raymond.toy at ericsson.com> wrote:
> This doesn't quite seem to fit plan B, since you can't compare complex
> numbers. What are you going to do for %i .. 2+2*%i?
Agreed. If a or b are complex, either leave a nounform or give an error.
>> (%i12) 40.0 .. 42.0b0;
>> (%o12) [40.0,41.0,42.0]
Well, while we're talking about fun edge cases, what about 2.0^53 ..
(2.0b0^53+1)? Should this be [2.0^53, 2.0^53+1] == [2.0^53, 2.0^53] or
[2.0b0^53, 2.0b0^53+1]? And how about 2.0^53 .. (2.0^53+2)? Or with
bfloats of different precision? (The clever thing to do is to
represent the results with enough fpprec to be accurate....) But in
reality, who cares?
-s