[
https://issues.apache.org/jira/browse/MATH-261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12698927#action_12698927
]
Sebb commented on MATH-261:
---------------------------
Re removing @SuppressWarnings:
The (T) casts cannot be removed from the obj parameter - the code won't compile.
Not sure why the compiler is objecting to the "instanceof " check and the
(Integer) cast - seems like a bug to me.
Regarding type "mixing": yes, I was trying to preserve the existing
(documented) behaviour as far as possible. That was the whole point.
It would be possible to treat Long and Integer as Objects, and that would
simplify the class.
But what about primitive types - should long and int and short be comparable?
Or should they all be treated separately?
That is a separate discussion, which is probably better dealt with on the
mailing list.
AFAICT, the only change is to insist that the Objects added to the class
implement Comparable.
That was true previously, except for the trivial case of adding a single object
- but then the getXXX methods could fail to return sensible values, see
MATH-259.
It was never possible to add two items unless they were both Comparable and
also Comparable to each other.
I don't think that has changed - if you have a test case that worked previously
and now fails, please add it.
> Add generics to Frequency
> -------------------------
>
> Key: MATH-261
> URL: https://issues.apache.org/jira/browse/MATH-261
> Project: Commons Math
> Issue Type: Improvement
> Reporter: Sebb
> Fix For: 2.0
>
> Attachments: Frequency.patch
>
>
> I've attempted to add generics to Frequency.
> This requires a minor change to the API, in that Objects to be used with the
> class need to be Comparable.
> This will break some code, e.g. one change was needed to the test cases.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
