[WFCORE-839] : Don't store state in InMemoryAuditLogHandlerResourceDefinition.

[ehsavoie]

Updating ManagedAuditLogger to be able to access the last log entries with only a real implemntation in the InMemoryAuditLogHander.

Jira: https://issues.jboss.org/browse/WFCORE-839

[wildfly-ci]

Core - Full Integration Build 1418 is now running using a merge of cf6913c

[wildfly-ci]

Windows Build 1801 is now running using a merge of cf6913c

[wildfly-ci]

Linux Build 2281 is now running using a merge of cf6913c

[wildfly-ci]

Windows Build 1801 outcome was SUCCESS using a merge of cf6913c
Summary: Tests passed: 3421, ignored: 64 Build time: 0:35:09

[wildfly-ci]

Linux Build 2281 outcome was SUCCESS using a merge of cf6913c
Summary: Tests passed: 3421, ignored: 64 Build time: 0:36:28

[wildfly-ci]

Core - Full Integration Build 1418 outcome was SUCCESS using a merge of cf6913c
Summary: Tests passed: 2983, ignored: 358 Build time: 0:48:15

[jamezp]

The rest of this class isn't thread safe. I don't think volatile is needed here.

[wildfly-ci]

Core - Full Integration Build 1450 is now running using a merge of a892dd1

[wildfly-ci]

Windows Build 1831 is now running using a merge of a892dd1

[wildfly-ci]

Linux Build 2314 is now running using a merge of a892dd1

[wildfly-ci]

Windows Build 1831 outcome was SUCCESS using a merge of a892dd1
Summary: Tests passed: 3423, ignored: 64 Build time: 0:35:11

[wildfly-ci]

Linux Build 2314 outcome was SUCCESS using a merge of a892dd1
Summary: Tests passed: 3423, ignored: 64 Build time: 0:36:14

[wildfly-ci]

Core - Full Integration Build 1450 outcome was SUCCESS using a merge of a892dd1
Summary: Tests passed: 3098, ignored: 368 Build time: 0:50:15

[bstansberry]

We're all these changes necessary for WFCORE-839, or was there some other problem that had to be fixed? This is a lot bigger than I expected; I thought it would be trivial.

The handler doesn't seem thread safe now.

[ehsavoie]

The handler isn't thread safe as all handlers. Thread safety is managed by ManagedAuditLoggerImpl. Since the map was already existing i switched back to using ManagedAuditLogger for all operations on the InMemoryHandler as was 'proper'.

[bstansberry]

Thanks; I'll have another look now that I get the idea. :)

[bstansberry]

Looking at ManagedAuditLogger, it already has handler-specific modification methods, e.g. updateSyslogHandlerAppName, updateSyslogHandlerFacility, updateSyslogHandlerReconnectTimeout. That's a bit yuck, but since they are already there I don't see why another one can't be added for setting the max history.

ManagedAuditLogger isn't usable outside the kernel so we can be a bit casual about its API.

[wildfly-ci]

Linux Build 2398 is now running using a merge of e730e78

[wildfly-ci]

Linux Build 2398 outcome was FAILURE using a merge of e730e78
Summary: Tests failed: 1 (1 new), passed: 3400, ignored: 64 Build time: 00:37:13

Failed tests

org.jboss.as.test.integration.mgmt.access.InMemoryAuditReportTestCase.testChangeMaxHistory: java.lang.AssertionError: 
Expected: is <2>
     but: was <3>
    at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:20)
    at org.junit.Assert.assertThat(Assert.java:865)
    at org.junit.Assert.assertThat(Assert.java:832)
    at org.jboss.as.test.integration.mgmt.access.InMemoryAuditReportTestCase.testChangeMaxHistory(InMemoryAuditReportTestCase.java:176)

[wildfly-ci]

WildFly pull requests Build 1827 outcome was FAILURE using a merge of e730e78
Summary: Snapshot dependencies failed: 2 (new) Build time: 00:00:01

[wildfly-ci]

Linux Build 2399 is now running using a merge of fec40cb

[wildfly-ci]

Linux Build 2399 outcome was SUCCESS using a merge of fec40cb
Summary: Tests passed: 3401, ignored: 64 Build time: 00:37:35

[wildfly-ci]

WildFly pull requests Build 1828 outcome was FAILURE using a merge of fec40cb
Summary: Snapshot dependency failed: ... WildFly Core Full - Integration (new) Build time: 00:00:01