[
https://issues.apache.org/jira/browse/PHOENIX-2612?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15106734#comment-15106734
]
Karan Singhal commented on PHOENIX-2612:
----------------------------------------
@Cody Marcel this failure never happened before because generally
test-scenario.xml and user_defined_scenario.xml are picked up first by
pattern-matcher and luckily both the xmls had same CREATED_DATE column which is
being tested in RuleGeneartorTest so that's why it never failed.
> RuleGeneartor test in phoenix-pherf passes and fails in different
> environments.
> ---------------------------------------------------------------------------------
>
> Key: PHOENIX-2612
> URL: https://issues.apache.org/jira/browse/PHOENIX-2612
> Project: Phoenix
> Issue Type: Bug
> Reporter: Karan Singhal
> Assignee: Karan Singhal
>
> This is happening because regex used in test for searching path of
> test-scenario.xml is not specific. The pattern being used is
> *scenario.*scenario.*.xml due to which pattern matcher finds all the
> different xmls in the package. After that all the xml's are parsed if
> pattern matcher finds test-scenario.xml in path first then test-scenario.xml
> is parsed first and test passes because test is always reading datamodel at
> index 0 (hard coded in test) from list of datamodels and if any other xml is
> parsed first it fails because it did not get right datamodel.
> Solution is simple just make matcher scenario in test to be more specific so
> that every time test-scenario.xml is picked up. After looking at some other
> tests I found .*test_scenario.xml is used in lot of other tests which is the
> most appropriate for searching test_sceanario.xml and can be used for this
> test too.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
