[ 
https://issues.apache.org/jira/browse/DERBY-6634?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Knut Anders Hatlen updated DERBY-6634:
--------------------------------------

    Attachment: d6634-1a-dead-code.diff

The patch [^d6634-1a-dead-code.diff] removes some dead, untested code that 
originally was there in order to compile XPath queries in trigger actions the 
first time the trigger fired. The code was made redundant by the fix for 
DERBY-3870, when SqlXmlUtil stopped implementing Formatable and was no longer 
stored as part of the trigger action's stored prepared statement.

All regression tests ran cleanly with the patch.

> Improve test coverage of SqlXmlUtil.java
> ----------------------------------------
>
>                 Key: DERBY-6634
>                 URL: https://issues.apache.org/jira/browse/DERBY-6634
>             Project: Derby
>          Issue Type: Improvement
>          Components: Test
>    Affects Versions: 10.11.0.0
>            Reporter: Knut Anders Hatlen
>            Assignee: Knut Anders Hatlen
>            Priority: Minor
>         Attachments: d6634-1a-dead-code.diff
>
>
> The test coverage report for SqlXmlUtil indicates that we don't have tests 
> for:
> - recompilation of XML queries used in triggers (the code looks dead, and can 
> probably be removed)
> - serialization of sequences of atomic values returned by an XPath query (as 
> far as I can see, we only support single atomic values and sequences of 
> documents, so probably that's also code that can be removed, possibly 
> replaced by an assert)
> - XMLEXISTS operations where the XPath query returns a scalar (after 
> DERBY-6624, it will take the same code path as a query that returns a 
> non-empty sequence, so it won't show up in the report anymore, but it's still 
> useful to have a test case for it)
> - XMLQUERY operations where the XPath query returns a boolean (after 
> DERBY-6624, the same code path will be taken for all data types, so it won't 
> show up in the report anymore, but it's still useful to have a test case for 
> it)



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to