Re: [j-devel] patch for complex acos

[Erik H. <eh...@gm...>]

Committed. Thanks again!

Bye,

Erik.

On Thu, Mar 19, 2009 at 4:43 PM, Erik Huelsmann <eh...@gm...> wrote:
> I should have replied to all.
>
> On Thu, Mar 19, 2009 at 4:42 PM, Erik Huelsmann <eh...@gm...> wrote:
>> On Thu, Mar 19, 2009 at 7:13 AM, Robert Dodier <rob...@ya...> wrote:
>>>
>>> Hello, something turned up by the Maxima test suite
>>> (problem 60 in rtest_trig). The ACOS function for
>>> complex double argument seems to be calculated incorrectly.
>>> The real part appears to be single precision
>>> (while the imaginary part is double precision). I think
>>> the problem is fixed by the patch shown below.
>>> What do you think?
>>
>> I think it's great you found the source of the issue! Thanks!
>>
>> The solution however, is a bit hairier than the one you suggest,
>> because !DoubleFloat does not only mean SingleFloat, but also Ratio,
>> Fixnum or Bignum. So, what I did is make this out of it:
>>
>>        if (!(arg instanceof DoubleFloat)) {
>>            if (arg instanceof Complex &&
>>                    ((Complex)arg).getRealPart() instanceof DoubleFloat) {
>>                    // do nothing; we want to keep the double float value
>>            }
>>            else
>>                result = new SingleFloat((float)((DoubleFloat)result).value);
>>        }
>>
>> I'm running the test suite now and if it passes both the ansi and the
>> maxima tests, I'll go on to commit later tonight.
>>
>> Thanks again for finding the issue!
>>
>> Bye,
>>
>> Erik.
>>
>>>
>>> Hope this helps
>>>
>>> Robert Dodier
>>>
>>>
>>> Index: src/org/armedbear/lisp/MathFunctions.java
>>> ===================================================================
>>> --- src/org/armedbear/lisp/MathFunctions.java   (revision 11712)
>>> +++ src/org/armedbear/lisp/MathFunctions.java   (working copy)
>>> @@ -168,7 +168,7 @@
>>>                 return new SingleFloat((float)Math.acos(f));
>>>         }
>>>         LispObject result = new DoubleFloat(Math.PI/2);
>>> -        if (!(arg instanceof DoubleFloat))
>>> +        if (arg instanceof SingleFloat || (arg instanceof Complex && ((Complex) arg).getRealPart () instanceof SingleFloat))
>>>             result = new SingleFloat((float)((DoubleFloat)result).value);
>>>         result = result.subtract(asin(arg));
>>>         if (result instanceof Complex) {
>>>
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>>> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
>>> software that enables intelligent coding and step-through debugging.
>>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>>> _______________________________________________
>>> armedbear-j-devel mailing list
>>> arm...@li...
>>> https://lists.sourceforge.net/lists/listinfo/armedbear-j-devel
>>>
>>
>