Permalink
Commits on Apr 28, 2011
  1. fixed a couple of memory leaks that were causing seg faults and added…

    … a threaded signal handler to allow for in place load of a snapshot file
    dwayn committed Apr 28, 2011
Commits on Sep 10, 2010
  1. rewritten the core of the memory model to greatly improve performance…

    … and capacity
    
     - the score pools are now a binary search tree
     - the linked lists of items are now double-linked and head and tail are kept track of to speed in adding and removal
     - the score pools hold the head and tail of a list
     - this has yeilded a queue that is from 10x - 5000x+ faster...the largest gain coming from very large populated queue
    Dwayn Matthies committed with Jul 4, 2010
  2. rewritten the core of the memory model to greatly improve performance…

    … and capacity
    
     - the score pools are now a binary search tree
     - the linked lists of items are now double-linked and head and tail are kept track of to speed in adding and removal
     - the score pools hold the head and tail of a list
     - this has yeilded a queue that is from 10x - 5000x+ faster...the largest gain coming from very large populated queue
    Dwayn Matthies committed with Jul 4, 2010
Commits on Jun 18, 2010
  1. Removing the sending of the QUIT command to the server: The server do…

    …es not have an implementation of this.
    
    Causes an issue of the server throwing Connection reset by peer
    
    Explanation:
    Client establishes socket
    Client sends command (next, update, peek, etc)
    Client waits for response
    Server sends response (+OK, -ERROR, etc)
    Client receives response
    Client sends QUIT
    Client closes socket
    Server receives QUIT 
    Server responds -ERROR (valid error handling because QUIT is not a valid command)
      - Socket is closed so server throws ECONNECTRESET (and a SIGPIPE I think)
    
    Server handles a few of these alright, but if the server is being sent commands a lot then the errors get stacked quickly and the eventual effect is a Seg Fault
    dwayn committed with Jun 18, 2010
Commits on Mar 16, 2010
Commits on Mar 9, 2010
Commits on Mar 8, 2010
  1. Refactoring command parsing into a separate source file. Adding a pro…

    …per .gitignore file. Misc code cleanup and organization.
    committed Mar 8, 2010
  2. Fixed a bug in stats returned by the 'INFO' command. Misc code cleanu…

    …p and organization changes.
    committed Mar 8, 2010
Commits on Mar 2, 2010
Commits on Feb 25, 2010
  1. fixing bug in sync timeout.

    committed Feb 25, 2010
  2. Bumping version to 0.2.1

    committed Feb 25, 2010
  3. Updating benchmark application with new protocol syntax and adding su…

    …pport for the info command to the client application.
    committed Feb 25, 2010
  4. Adding the client library to the build and added update, next and pea…

    …k functionality to it.
    committed Feb 25, 2010
Commits on Feb 23, 2010
  1. Updating the API to use a Redis-like line-wire protocol. Added a PHP …

    …extension client library.
    Nick committed Feb 23, 2010
Commits on Feb 18, 2010
  1. Updating readme.

    Nick committed Feb 18, 2010
  2. Updating docs.

    Nick committed Feb 18, 2010
  3. Barbershop will now periodically save state to disk. On startup it wi…

    …ll also attempt to load state from disk.
    Nick committed Feb 18, 2010
Commits on Feb 17, 2010
  1. Adding support for scattered/random priority inserts. Bug fixing and …

    …refactoring.
    Nick committed Feb 17, 2010
  2. Updating the todo list.

    committed Feb 17, 2010
Commits on Feb 16, 2010
  1. Fixed the bugs.

    Nick committed Feb 16, 2010
Commits on Feb 3, 2010
Commits on Feb 2, 2010
  1. Working on bugs

    Nick committed Feb 2, 2010
Commits on Feb 1, 2010
  1. Added next and stats functionality. Added garbage collection queues a…

    …lthough actual garbage collection in items and scores is unimplemented.
    committed Feb 1, 2010
Commits on Jan 31, 2010
Commits on Jan 30, 2010
  1. Removing unused files.

    committed Jan 30, 2010