You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Gilles (JIRA)" <ji...@apache.org> on 2012/07/13 13:24:34 UTC

[jira] [Commented] (MATH-821) SparseRealVectorTest.testMap and testMapToSelf fail because zero entries lose their sign

    [ https://issues.apache.org/jira/browse/MATH-821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13413659#comment-13413659 ] 

Gilles commented on MATH-821:
-----------------------------

In revision 1361164, I disabled the two unit tests that demonstrate this bug.
When the bug is fixed, the tests should be made active again.

                
> SparseRealVectorTest.testMap and testMapToSelf fail because zero entries lose their sign
> ----------------------------------------------------------------------------------------
>
>                 Key: MATH-821
>                 URL: https://issues.apache.org/jira/browse/MATH-821
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 3.0
>            Reporter: Sébastien Brisard
>              Labels: linear, sparse
>
> Mapping {{Inverse}} to an {{OpenMapRealVector}} can lead to wrong answers, because {{1.0 / 0.0}} should return {{+/-Infinity}} depending on the sign of the zero entry. Since the sign is lost in {{OpenMapRealVector}}, the answer must be wrong if the entry is truly {{-0.0}}.
> This is a difficult bug, because it potentially affects any function passed to {{OpenMapRealVector.map()}} or {{mapToSelf()}}. I would recommend we relax the requirements in the unit tests of this class, and make people aware of the issue in the class documentation.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira