[ 
https://issues.apache.org/jira/browse/TAMAYA-327?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16353523#comment-16353523
 ] 

ASF subversion and git services commented on TAMAYA-327:
--------------------------------------------------------

Commit ca4de3dc49fefa020e89073b8032be7de20ab27d in incubator-tamaya's branch 
refs/heads/master from [~wlieurance]
[ https://git-wip-us.apache.org/repos/asf?p=incubator-tamaya.git;h=ca4de3d ]

TAMAYA-327: Consistent signature creating ConversionContexts

Comments in ConversionContext lead me to believe that supportedFormats
should have their order maintained after import.  The HashSet that the
Builder object uses as a backer will not accomplish that goal.  This
change converts the Builder to use a LinkedHashSet supportedFormats, as
well as the object proper.

Some additional tests have been added here in pursuit of TAMAYA-288's
additional code coverage.


> Inconsistent signature for creating ConversionContext
> -----------------------------------------------------
>
>                 Key: TAMAYA-327
>                 URL: https://issues.apache.org/jira/browse/TAMAYA-327
>             Project: Tamaya
>          Issue Type: Bug
>            Reporter: William Lieurance
>            Priority: Major
>
> During some additional testing in the API added as part of TAMAYA-288, I came 
> across an inconsistency between the ConversionContext and the 
> ConversionContext.Builder.  Namely, the comments say that supportedFormats 
> have their order maintained, which is consistent with the use of an ArrayList 
> here 
> [https://github.com/apache/incubator-tamaya/blob/master/code/api/src/main/java/org/apache/tamaya/spi/ConversionContext.java#L42]
>  .  The builder, however implements the same supportedFormats with a HashSet, 
> which does not ensure order.  See 
> [https://github.com/apache/incubator-tamaya/blob/master/code/api/src/main/java/org/apache/tamaya/spi/ConversionContext.java#L148]
>  .
> I'm planning on creating a pull request with the additional tests for that 
> object as well as a patch to more consistently maintain order (all 
> ArrayLists).  Please let me know if the better answer is to more consistently 
> leave order unspecified (all HashSets).



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

Reply via email to