WFLY-3947: fixing calendar based timeout; also replacing problematic bui...

[emmartins]

...ld test case with one matching original issue WFLY-1468

[wildfly-ci]

Linux with security manager Build 172 is now running using a merge of bea57f3

[wildfly-ci]

Windows Build 287 is now running using a merge of bea57f3

[wildfly-ci]

Linux Build 5141 is now running using a merge of bea57f3

[wildfly-ci]

Linux with security manager Build 172 outcome was SUCCESS using a merge of bea57f3
Summary: Tests passed: 793, ignored: 242 Build time: 0:06:40

[wildfly-ci]

Linux Build 5141 outcome was SUCCESS using a merge of bea57f3
Summary: Tests passed: 3088, ignored: 260 Build time: 0:56:02

[wildfly-ci]

Windows Build 287 outcome was SUCCESS using a merge of bea57f3
Summary: Tests passed: 3088, ignored: 260 Build time: 0:56:07

[emmartins]

Some notes about the issues fixed by the PR:
Major:
* truncate of calendar does not keeps DST offset
* testsuite fix
Minor:
* truncate not needed if setting of time fields is done properly
* excessive calendar cloning
* excessive checks for no more timeouts
* first and next timeout computation logic duplicated

[stuartwdouglas]

Has this been tested against the timers TCK?

[emmartins]

yes, it passes the 141 tests in com/sun/ts/tests/ejb30/timer/schedule

[wildfly-ci]

Linux with security manager Build 221 is now running using a merge of db5e9e0

[wildfly-ci]

Windows Build 336 is now running using a merge of db5e9e0

[wildfly-ci]

Linux Build 5212 is now running using a merge of db5e9e0

[wildfly-ci]

Linux with security manager Build 221 outcome was SUCCESS using a merge of db5e9e0
Summary: Tests passed: 792, ignored: 244 Build time: 0:06:19

[wildfly-ci]

Windows Build 336 outcome was SUCCESS using a merge of db5e9e0
Summary: Tests passed: 3021, ignored: 238, muted: 1 Build time: 0:54:41

[wildfly-ci]

Linux Build 5212 outcome was SUCCESS using a merge of db5e9e0
Summary: Tests passed: 3021, ignored: 238, muted: 1 Build time: 0:56:58