[ 
https://issues.apache.org/jira/browse/JENA-1514?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16437352#comment-16437352
 ] 

Andy Seaborne commented on JENA-1514:
-------------------------------------

I can't see a way of doing anything here that is not a change that will impact 
applications.

The contracts around XSDBaseNumericType.cannonicalise which impact memory graph 
indexing, also look to be such that a change will affect behaviour there as 
well.

Can we close this until a concrete proposal is available when we can explore 
the impact on a concrete case?

> Class cast exception with BigInteger literals.
> ----------------------------------------------
>
>                 Key: JENA-1514
>                 URL: https://issues.apache.org/jira/browse/JENA-1514
>             Project: Apache Jena
>          Issue Type: Bug
>          Components: Jena
>    Affects Versions: Jena 3.6.0, Jena 3.7.0
>            Reporter: Claude Warren
>            Priority: Minor
>         Attachments: BigIntIssue.java
>
>
> When saving a BigInteger as a literal it is converted into an 
> XSDBaseNumericType.
> When that  type is read back it is converted into the smallest numeric data 
> type that can hold the value; a narrowing if you will.
> While  this works fine for primitive and primitive wrapping types (e.g. int 
> and Integer) there is no automatic boxing available for BigInteger so 
> attempting to retrieve the value of a BigInteger that is smaller than the 
> Long.MAX_VALUE will result in a class cast exception when the primitive type 
> is cast to the BigInteger.
> I suspect the similar issues arise when BigDecimal is also used but I have 
> not tested.
> Should we address this issue by creating a BigInteger and BigDecimal type 
> that perhaps extends XSDBaseNumericType?  Should we address this issue at all 
> and just post a warning?
> Test code is included to show the issue.
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to