Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on Oct 11, 2012
  1. @tadmarshall
  2. @svenstaro @andy10gen

    SERVER-4314 Migrate to Boost Filesystem v3, away from deprecated v2.

    svenstaro authored andy10gen committed
    Original author: Sven-Hendrik Haase <sh@lutzhaase.com>
    Merged by: Andy Schwerin <schwerin@10gen.com>
    Signed-off-by: Andy Schwerin <schwerin@10gen.com>
  3. @kchodorow
  4. @kchodorow
  5. @dwight
  6. @dwight

    SERVER-7329 also now checks at the beginning, after the first write, …

    dwight authored
    …in addition to ad the end
  7. @dwight
  8. @tadmarshall

    SERVER-6207 protect printWindowsStackTrace() with critical section

    tadmarshall authored
    Use SimpleMutex::scoped_lock to make printWindowsStackTrace() run in
    only one thread at a time.
  9. @tadmarshall

    SERVER-7289 add inexpensive tests to validate command

    tadmarshall authored
    Checks added to the validate command:
    1) The 'xprev' pointer in each extent in the extent list should point
       to the previous extent, or be null for the first extent
    2) The 'lastExtent' pointer in NamespaceDetails should point to the last extent
    3) If an exception is thrown while counting extents, identify the failing extent
    4) If an exception is thrown while counting extents, report the count
    5) If an exception is thrown while counting extents, the first extent should be
       validated and the check for 'xprev' in the first extent being null should be done
    6) If an invalid signature is found in an extent, report the value found and
       identify the extent
    7) The "self-pointer" 'myLoc' in an extent should be validated
    8) The extent size should be compared with Extent::minSize(), not zero
  10. @tadmarshall
  11. @tadmarshall
  12. @astaple
  13. @milkie
  14. @milkie
Commits on Oct 10, 2012
  1. @tadmarshall

    fix Windows compile

    tadmarshall authored
  2. @matulef

    SERVER-2001 change shard key validation to allow hashed shard keys

    matulef authored
    This changes the top-level shard key validation to allow shard keys
    such as {a : "hashed"}. It also adds some helper functions for
    determining when a unique index is compatible with a given shard
    key, and a variety of unit tests and a js test.
  3. @matulef

    SERVER-2001 make modifiedRangeBound accept index patterns without 1/-…

    matulef authored
    …1 vals
    
    To allow more complex index types, modifiedRangeBound needs to make sense
    of index patterns like {a : "hashed"} that don't have 1 or -1 for the
    field values.  With this change, it will treat bounds over a hashed
    keyspace as ascending over the range of hashed values.
  4. @stbrody
  5. @milkie
  6. @milkie
  7. @andy10gen

    SERVER-5211 Separate of CmdLine::store into two functions.

    andy10gen authored
    The new functions are CmdLine::store, with the same signature as before, and
    mongo::initializeServerGlobalState(), which should take no parameters, but does
    take an optional bool to deal with a knot in mongod --shutdown that I haven't
    yet untangled.
    
    This separation makes it easier to test functions in the command line parsing
    part of the library, and provides a little extra clarity on the role of the two
    functions.
  8. @renctan

    SERVER-7230 timing error in repl_monitor_stress.js

    renctan authored
    Change the test to keep on trying until we confirmed that reconfig actually succeeded
  9. @stbrody
  10. @matulef

    SERVER-2001 shardChunkManager no longer modifies shard key with ascen…

    matulef authored
    …ding fields
    
    Validation of proper shard keys is now done at the top-level, so there
    is no need to replace field values with "1.0". Moreover, hashed keys
    like { a : "hashed" } will not have numeric field vals.
  11. @matulef

    Fix for js test.

    matulef authored
    Previously the index was not actually created because it was declared
    as unique, but there were duplicate x values.  Thus, this test was
    passing for the wrong reasons.  Removing the uniqueness condition
    fixes that.
  12. @IanWhalen
  13. @astaple
  14. @astaple
  15. @astaple

    SERVER-3104 Allow intersection of sibling multikey index field ranges…

    astaple authored
    … having a shared $elemMatch parent.
  16. @andy10gen
  17. @erh

    use new simple Random class

    erh authored
  18. @erh
  19. @andy10gen
  20. @andy10gen
Something went wrong with that request. Please try again.