Matt Juntunen commented on MATH-1449:

Yes, I suppose it does look kind of strange with the tolerance. I didn't use 
Double.isInfinite() because I wanted to distinguish between positive and 
negative infinity. The tolerance value is just there to keep Assert happy. I 
could rewrite it as 

Assert.assertTrue(Double.POSITIVE_INFINITY == poly.getSize());
to make it clearer if needed.

Also, I disagree that the boolean statements read better with the 
assertTrue/False methods in this case. When asserting a method return value, I 
prefer to have the expected return value listed explicitly rather than encoded 
in the name of the assertion method. I think it stands out better that way. 
But, this is just a preference thing and really makes no difference otherwise.

Let me know if you'd like me to make either of these changes. Note that I'm 
pretty sure I used both of the syntaxes you mentioned above all over my last 
batch of unit test changes, so, if I do change these, I should probably update 
the other tests as well to match.

> PolygonsSet.getBarycenter() returns incorrect value for empty region
> --------------------------------------------------------------------
>                 Key: MATH-1449
>                 URL: https://issues.apache.org/jira/browse/MATH-1449
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 4.0, 3.6
>            Reporter: Matt Juntunen
>            Priority: Major
>             Fix For: 4.0
> The o.a.c.m.geometry.euclidean.twod.PolygonsSet.getBarycenter() method 
> returns the point (0, 0) when the region is empty. It should return (NaN, 
> NaN) since the barycenter does not exist. This is the behavior of 
> IntervalsSet and PolyhedronsSet.
> Pull request: https://github.com/apache/commons-math/pull/80

This message was sent by Atlassian JIRA

Reply via email to