• Apr 27, 2011

    0.8.3

    Release of 0.8.3
    This release includes bugfixes for formatting of size_t output and
    stops shadowing a global from time.h.  It also removes an assertion
    check causing a warning with some toolchains.
  • Jan 5, 2011

    0.8.2

    Release of 0.8.2
    This release includes some bugfixes for reporting
    errors from the population phase, and some other fixes
    regarding building with c99 on Linux and performance.
  • Dec 3, 2010

    0.8.1

    Version 0.8.1
    The changes here are in being more verbose about which operations have
    failed and why they have failed.  This allows for more testing where
    SERVER_ERROR may be expected.
  • Nov 27, 2010

    0.8.0

    Version 0.8.0
    The new feature in 0.8 is support for a key
    prefix.  This allows one to run multiple instances
    of memcachetest with different sizes.
  • Oct 12, 2010

    0.7.1

    Refactored backoff, fix for missing data.
    Trond wrote some things that refactored backoff a bit better, so it
    is cleaner in the code and you get a better idea of what has gone
    wrong.  Plus, now that some values may be 0, the missing data is
    allowed to go 0.
  • Sep 26, 2010

    0.7.0

    Release of memcachetest 0.7.0
    This release adds support for a backoff per membase, and output
    of 99th percentile latency stats.
  • Aug 29, 2010

    0.6.6

    Fixes for ascii populate and handling 0 length messages. Portability.
    
  • Aug 25, 2010

    0.6.5

    Stats gathered per thread. Improvements in useability.
    
  • Aug 18, 2010

    0.6.0

    Better error handling, rudimentary vbucket support.
    This release has better error handling in a number of ways,
    stats code ripped out (to be replaced soon) and has
    very, very, very rudimentary vbucket support.
    
    It is also using pandorabuild but is also disabling
    the stop on error because there are thousands of minor
    bugs to be cleaned up.
    
    Okay, thousands may be a bit many.
  • Jul 19, 2010

    0.5

    Version 0.5.
    Why 0.5?  Because it's more mature than 0.1, but it's hardly ready
    for much of anything.