Commits on Apr 1, 2013
  1. New snapshot

    amalloy committed Apr 1, 2013
  2. Merge branch 'develop'

    amalloy committed Apr 1, 2013
  3. New snapshot

    amalloy committed Mar 29, 2013
Commits on Feb 15, 2013
  1. 0.11.0

    ninjudd committed Feb 15, 2013
  2. make flatland.jiraph.core more flexible for non-existent layers

    ninjudd committed Feb 15, 2013
    do not throw an error when calling read methods with layer-specs that
    don't exist. return nil instead. if you want the more strict behavior,
    you can use flatland.jiraph.graph directly.
  3. s/writeable/writable/

    ninjudd committed Feb 15, 2013
Commits on Feb 14, 2013
  1. support layer specs instead of just names in flatland.jiraph.core

    ninjudd committed Feb 14, 2013
    If layer-spec is a vector, get-layer descends through the child layers
    after looking up the initial layer in *graph*.
    Also add get-edge-ids that returns edge ids where :exists is true. This
    lays the groundwork for getting rid of get-incoming as a special method
    in jiraph.
    Instead of:
        (get-incoming :foo "id-1")
    You will use:
        (get-edge-ids [:foo :incoming] "id-1")
Commits on Feb 13, 2013
  1. Merge branch 'exists' into develop

    ninjudd committed Feb 8, 2013
  2. bump to 0.11.0-SNAPSHOT

    ninjudd committed Feb 13, 2013
Commits on Feb 11, 2013
  1. Fix for returned revision numbers.

    amalloy committed Feb 11, 2013
    Since we're now reading these in reverse, we have to "un-reverse" them
    before giving them out to the caller.
  2. 0.10.0 snapshot

    ninjudd committed with amalloy Feb 8, 2013
Commits on Feb 8, 2013
  1. rename layer namespaces

    ninjudd committed Feb 8, 2013
    - backend layer implementations are under flatland.jiraph.layer.*
    - layer wrappers are under flatland.jiraph.*
Commits on Jan 7, 2013
  1. Document some confusing functions.

    amalloy committed Jan 7, 2013
    They're still confusing, but such is life.
Commits on Jan 3, 2013
  1. Fix protobuf-format test.

    amalloy committed Jan 3, 2013
    The protobuf format no longer works when not writing with a revision,
    so the tests that lazily use no revision must be updated to instead
    use the "basic" protobuf format.
  2. Tidy up implementation of header-hopping.

    amalloy committed Jan 3, 2013
    Added a bunch of docstrings, and pulled several pieces out of the
    one big, ugly function into their own separate functions.
Commits on Jan 2, 2013
  1. Use trailers, not headers.

    amalloy committed Jan 2, 2013
Commits on Dec 31, 2012
  1. Fix code formatting

    amalloy committed Dec 31, 2012
  2. Work properly for empty arrays

    amalloy committed Dec 31, 2012
  3. Update test .proto files.

    amalloy committed Dec 21, 2012
    We no longer need each of our protobufs to manually specify how to
    encode revision and proto_length.
  4. New binary format for protobufs.

    amalloy committed Dec 19, 2012
    Instead of letting protobuf encode the revision numbers and
    byte-lengths, we encode them by hand in a format that is simpler to
    parse and makes reading historical data faster. We smuggle it into the
    existing protbuf bytestream such that a protobuf parser will ignore it.