Skip to content
Commits on Feb 19, 2010
  1. @afeinberg

    Switched version to 0.80

    afeinberg committed
  2. @afeinberg

    Updated release notes.

    afeinberg committed
Commits on Feb 18, 2010
  1. @afeinberg
Commits on Feb 17, 2010
  1. @afeinberg
Commits on Feb 15, 2010
  1. @afeinberg
  2. @afeinberg
  3. @afeinberg
  4. @kirktrue
  5. @afeinberg
  6. @kirktrue

    Fixed some oversights in use of System.currentTimeMillis() vs. System…

    kirktrue committed
    ….nanoTime() for FailureDetector as found by Ismael.
  7. @afeinberg
  8. @afeinberg
Commits on Feb 14, 2010
  1. @afeinberg
Commits on Feb 13, 2010
  1. @kirktrue
Commits on Feb 12, 2010
  1. @kirktrue

    Fix for issue 207 (FailureDetector.recordSuccess could take startTime…

    kirktrue committed
    … instead or in addition to requestTime). Changing the API required a lot of changes. Instead I just changed the two users of the class to use nanoseconds instead and refactored into separate private methods in each class which makes it cleaner in general.
  2. @kirktrue
  3. @kirktrue

    Separated out the "normal" NIO selector count from the "admin service…

    kirktrue committed
    …" NIO selector count as the latter may tend to need to be bigger on account of the long-lived requests.
  4. @kirktrue
Commits on Feb 10, 2010
  1. @afeinberg

    Javadoc updates.

    afeinberg committed
  2. @afeinberg
  3. @afeinberg

    Minor javadoc fixes.

    afeinberg committed
Commits on Feb 9, 2010
  1. @kirktrue

    Merge remote branch 'upstream/master' into nio-server2-cleanup

    kirktrue committed
    Conflicts:
    	src/java/voldemort/server/protocol/admin/AdminServiceRequestHandler.java
    	test/integration/voldemort/socketpool/SimpleSocketPoolTest.java
    	test/unit/voldemort/client/AdminServiceBasicTest.java
    	test/unit/voldemort/client/AdminServiceFailureTest.java
  2. @kirktrue
  3. @kirktrue
  4. @kirktrue
  5. @kirktrue

    Copyright updates.

    kirktrue committed
  6. @kirktrue

    Got rid of that terrible lastCompleteStreamPosition stuff by compacti…

    kirktrue committed
    …ng the buffer back to a zero-based offset before the next read on partial streaming reads. Makes the code cleaner and avoids unnecessary buffer allocation. This at the cost of an (internal) buffer copy from position() to 0.
  7. @afeinberg
  8. @afeinberg

    Merge branch 'li-r1008'

    afeinberg committed
    Conflicts:
    	.classpath
Commits on Feb 8, 2010
  1. @afeinberg
  2. @afeinberg

    Fixed Issue 209.

    afeinberg committed
Commits on Feb 7, 2010
  1. @ijuma
  2. @ijuma
  3. @ijuma
Commits on Feb 5, 2010
  1. @afeinberg
Something went wrong with that request. Please try again.