Commits on Dec 31, 2012
  1. Upgrade plucky to latest.

    New version of plucky includes a few performance tweaks.
    jnunemaker committed Dec 31, 2012
Commits on Dec 28, 2012
  1. Update plucky version.

    jnunemaker committed Dec 28, 2012
  2. Merge branch 'query_methods_match_correct_docs' of…

    …/balexand/mongomapper into balexand-query_methods_match_correct_docs
    jnunemaker committed Dec 28, 2012
  3. Merge pull request #485 from nviennot/reverse_merge

    reverse_merge! -> reverse_merge
    jnunemaker committed Dec 28, 2012
Commits on Dec 25, 2012
  1. reverse_merge! -> reverse_merge

    Tampering with the caller's argument might not be a good idea.
    Have our own copy.
    nviennot committed Dec 25, 2012
  2. Some source files were executable

    cleaned up with:
        find lib test -type f | xargs chmod 644
    nviennot committed Dec 25, 2012
Commits on Dec 7, 2012
  1. Fix options in the template

    cheald committed Dec 7, 2012
  2. Instead of forcing safety on by default, let the connection settings …

    …specify it. Add w, pool_size, and ssl option defaults to the config.
    cheald committed Dec 7, 2012
  3. Upgrade straggler connection references.

    Upgrade safe semantics to be consistent with the new MongoClient safe semantics - legacy usage (:safe => true) is transparently upgraded to the new system.
    BIG CHANGE: In keeping with the driver defaults, safety is now ON by default (:w => 1) rather than off by default. Use `safe false` to turn off safety in a model.
    Fixed tests.
    cheald committed Dec 7, 2012
  4. Fix legacy mongo class names, that are in deprecation as of 1.8.0.

    Mongo::Connection -> Mongo::MongoClient
    Mongo::ReplSetConnection -> Mongo::MongoReplicaSetClient
    Blaž Hrastnik committed with cheald Dec 1, 2012
Commits on Nov 25, 2012
Commits on Nov 3, 2012
Commits on Oct 26, 2012
Commits on Oct 22, 2012
  1. Correctly define _id as bson ordered hash.

    The reason the modifier tests were failing is because in #attributes, _id was being converted to a hash with indifferent access instead of staying a bson ordered hash.
    jnunemaker committed Oct 22, 2012
  2. Save typecast/uncast values and use those where possible, rather than…

    … recomputing them every time they're accessed.
    Use #key? rather than #keys.include? for better performance
    Significantly reduce calls to #read_keys by avoiding #attributes calls from ActiveModel::Dirty. Improve the speed of #read_keys by avoiding string interns and ivar lookups where possible. Eliminate extraneous array creations in Keys#attributes and Associations#embedded_associations. Improve the speed of Dirty#write_key by using ivars directly rather than going through instance methods.
    Revert to the slower-but-more-semantic key reads for #attributes now that the Dirty attributes thrashing is fixed.
    attr_accessor -> attr_reader for semantic correctness
    cheald committed Oct 21, 2012
Commits on Oct 19, 2012
  1. Don't iterate the whole cursor twice when using IdentityMap with #all…

    …. Improve performance by avoiding explicit block bindings, extraneous method calls, and extraneous array creation.
    cheald committed Oct 19, 2012
Commits on Oct 3, 2012
  1. Merge pull request #462 from jnunemaker/association-each-not-yielding

    Provide a fix for many associations not yielding to each in callbacks.
    jnunemaker committed Oct 3, 2012
Commits on Oct 2, 2012
  1. Provide a fix for many associations not yielding to each in callbacks.

    Ran into an issue when iterating over a many association in a
    before_save callback on a document. For an association named foos,
    `foos.inspect` would show the items but `foos.each` would never yield to
    the given block since it appears to be empty.
    The original issue I was tracking down was that instead of not yielding
    it was yielding instances of Mongo::Cursor instead of the association
    object. I haven't been able to recreate that case.
    gaffneyc committed Oct 2, 2012
  2. Merge pull request #459 from jnunemaker/support-string-ids-in-modifiers

    Support non-ObjectID ids being given to modifiers.
    jnunemaker committed Oct 2, 2012
Commits on Oct 1, 2012
  1. Support non-ObjectID ids being given to modifiers.

    The partially broken by cf82ec6 where it allowed a single ID to be
    passed and it would work as expected. This also adds support for
    multiple IDs to be provided.
    gaffneyc committed Oct 1, 2012
Commits on Sep 28, 2012
Commits on Sep 21, 2012
  1. A few modifier test tweaks.

    jnunemaker committed Sep 21, 2012