LOGBACK-116: RollingCalendarTest fails when user.timezone is not CEST

simplification of SocketApender connection logic

LOGBACK-909 Fix race condition in SMTPAppenderBase

better fix for broken AbstractSocketAppenderTest

Update javadoc, cannot find CustomDBNameResolver

Fixed typo in configuration.html

AppenderBase started variable made volatile

AppenderBase started variable made volatile as it is mutable, access to it is not synchronized and it is shared between threads.
Fix for http://jira.qos.ch/browse/LOGBACK-1037

Correct RollingCalendarTest. Modify time-zone offset handling.

Fix for Prudent FileAppender and thread interruptions (LOGBACK-875)

Prevent logging on an interrupted thread from shutting down prudent file
appenders due to a FileLockInterruptedException

This reverts commit 147ed05b.

added missing unit test for re-establishing the socket connection when stream creation fails with IOException

fixed broken test setup for addsInfoMessageWhenEventIsBeingDroppedBecauseOfConnectionProblemAndDequeCapacityLimitReached

AsyncAppenderBaseTest: The appender was not allowed enough time to
flush the appender on #stop(), causing the empty-check of the appender's
queue to fail. The test appender's max-flush time is now extended from
30ms to >1000ms.

AbstractSocketAppenderIntegrationTest: A unit test expected
deque.takeFirst() to be called only once after a log-event is appended,
but the call actually occurs twice in a while-loop, where the second
call is intended to block, waiting for the next log-event.

AbstractSocketAppenderTest: A unit test that verified appender
behavior when its deque was at capacity neglected to fill the deque
before testing for errors.

Added note for LOGBACK-128

Add notes for #193 (LOBGACK-977), provided by Sebastian Gröbler

LOGBACK-977 - AbstractSocketAppender loses event for every socket connection break

extended documentation to reflect new SocketAppender behaviour towards synchronous/asynchronous queueing