Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
branch: master
Commits on May 24, 2011
Commits on May 23, 2011
  1. @davisp

    Upgraded to LevelDB 1.2

    davisp authored
Commits on May 17, 2011
  1. @davisp
Commits on May 16, 2011
  1. @davisp

    Implement destroy_db(Db).

    davisp authored
    Destroys a database when it closes.
Commits on May 15, 2011
  1. @davisp
Commits on May 14, 2011
  1. @davisp

    Forgot to add Fabian to the THANKS file.

    davisp authored
    Thanks for the build fix.
  2. @davisp

    Temporary fix for building on Linux.

    davisp authored
    This appears to work on a semi recent Ubuntu. I'm not entirely
    certain how widespread the -std=c++0x support is. If it turns
    out that enough distros don't have a new enough GCC I can
    go back and port the necessary parts from Chromium.
  3. @davisp

    Remove nonexistant app entry point.

    davisp authored
    Thanks to Kevin Smith for spotting this.
  4. @davisp
  5. @davisp
Commits on May 13, 2011
  1. @davisp
  2. @davisp

    Slight type fix.

    davisp authored
  3. @davisp

    Removed extra whitespace.

    davisp authored
  4. @davisp

    Slightly modified dialyzer patch from Kevin Smith.

    davisp authored
    Made it part of make check because of the PLT file stuff. Also tries
    to warn when that file doesn't exist.
  5. @davisp
Commits on May 12, 2011
  1. @davisp

    Synchronize access to write batches and iterators.

    davisp authored
    Just your average mutexes to ensure that we only have one thread
    manipulating the objects at any given time. In general I don't
    expect these to stray out of a single Erlang process so there
    shouldn't be much lock contention.
  2. @davisp
  3. @davisp

    Refactored README.

    davisp authored
  4. @davisp

    Add a README.

    davisp authored
Commits on May 11, 2011
  1. @davisp
  2. @davisp

    Fixed whitespace.

    Robert Dionne authored davisp committed
  3. @davisp

    Specify the plan count.

    davisp authored
  4. @davisp

    Support for database snapshots.

    davisp authored
    Snapshots allow users to read specific versions of the database
    at a given point in time. They can be read from a random database
    state or they can be returned from calls that modify the database
    to refer to a snapshot just after that update occurs but before
    any other update.
  5. @davisp
  6. @davisp

    Minor whitespace cleanup.

    davisp authored
  7. @davisp

    Minor touch up for readability.

    davisp authored
  8. @davisp

    Add tests for iterator options.

    davisp authored
  9. @davisp

    Finished option work for all methods.

    davisp authored
    put, del, wb_write all accept write options:
    
      * sync             - Call fsync after writing data to disk.
      * snapshot         - Return {ok, Snapshot} for use in read options.
    
    get, iter accept read options
    
      * verify_checksums - Make sure data checksums match for data read.
      * skip_cache       - Do not cache the data for this read which is useful
                           when doing bulk reads.
      * snapshot         - Issue a read using the provided snapshot.
Commits on May 10, 2011
  1. @davisp

    Fix plan count.

    davisp authored
  2. @davisp

    Fix test for badarg failure.

    davisp authored
  3. @davisp

    Support for setting options for key lookups.

    davisp authored
    Also includes most of the infrastructure I'll need for doing
    snapshots. But first I'm going to add write option support
    cause its borring and wanting to do snapshots will motivate
    me to finish them.
  4. @davisp

    Support database creation options.

    davisp authored
      * create_if_missing      - Create a new database if one doesn't exist.
      * error_if_exists        - Error if a db already exists with this name.
      * paranoid_checks        - Check extensively for file corruption.
      * write_buffer_size      - Amount of memory to buffer before sorting and
                                 writing to disk.
      * max_open_files         - Limit the number of active files open at once.
      * cache_size             - Size of an LRU cache for commonly read blocks.
      * block_size             - Approximate size of user data packed per block.
      * block_restart_interval - Number of keys to delta encode prefixes.
  5. @davisp

    Write batch support for bulk operations.

    davisp authored
    This allows users to create a batch and fill it with put/delete
    requests before committing them atomically. Larger batches will
    make for better performance.
    
    Caveat: I have not tested the concurrency of this or the itertor
    construct. I'm guessing they need locks.
Something went wrong with that request. Please try again.