[
https://issues.apache.org/jira/browse/TRINIDAD-1489?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12712312#action_12712312
]
Gabrielle Crawford commented on TRINIDAD-1489:
----------------------------------------------
Currently there's this code in UIXEditableValue
protected boolean compareValues(Object previous, Object value)
{
if (isEmpty(previous))
return !isEmpty(value);
return !previous.equals(value);
}
The fix being proposed is that if the .equals fails we'll check if the two
values are
1. the same type
2. implement comparable
and if 1 and 2 are true we'll use compareTo
> get a valueChangeEvent for bigDecimal even though user didn't make a change.
> ----------------------------------------------------------------------------
>
> Key: TRINIDAD-1489
> URL: https://issues.apache.org/jira/browse/TRINIDAD-1489
> Project: MyFaces Trinidad
> Issue Type: Bug
> Reporter: Gabrielle Crawford
>
> When attribute data type is BigDecimal and <af:convertNumber> is used Trin
> treats the attribute as if the attribute value is changed even though the
> attribute value has not been changed.
> Under the covers the numberConverter is using the java decimalFormat class,
> and things can get a little funny when you use bigdecimal, because bigdecimal
> remembers formatting information like scale. So 2.0 is not equal to 2.00 in
> bigdecimal.
> We can add logic to UIXEditableValueHolder that if .equals fails and if the
> values are the same type and implement comparable we should then check
> compareTo.
> There are 2 workarounds for now.
> 1] apply the pattern to the bigdecimal in the getter, so in the salary
> example code above return newSal instead of sal.
> 2] check compareTo in the setter, and don't set if you get 0.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
