Permalink
Commits on Oct 14, 2009
  1. Dropping an eval() call for send().

    committed Oct 14, 2009
  2. Speed improvements for each().

    committed Oct 14, 2009
Commits on Oct 13, 2009
  1. todo : reenabled #copy for Tyrant (Edo and Tokyo) - closes gh-19

    (would be good to have one at hand for tyrant tables)
    jmettraux committed Oct 13, 2009
  2. officer on the deck

    jmettraux committed Oct 13, 2009
  3. removed trailing white spaces

    jmettraux committed Oct 13, 2009
Commits on Oct 12, 2009
  1. Added the ability to select the built-in :lexical or :decimal comparison

    functions for B-tree databases, or even set a custom Proc to order the keys.
    This feature is currently only available through the Edo interface.
    
    I did try to add it to the Tokyo side as well, but it's not possible without a
    lot of work.  The reason is that comparison function must be set before the
    B-tree database is opened.  Because the Tokyo interface uses the abstract
    interface though, the underlying B-tree database isn't created until the open()
    call.  Thus, the abstract interface would need to be abandoned for B-tree
    databases to support this feature.
    committed Oct 12, 2009
Commits on Oct 11, 2009
  1. Fixing multiple bugs in fixed-width databases. First, Edo didn't supp…

    …ort any
    
    tuning parameters due to an extra argument in the list of supported options.
    Also, keys() and iteration didn't work on these databases, due to fwmkeys() not
    being supported for them.  I switched to using range() on fixed-width databases
    only to get keys() and iteration working.
    committed Oct 11, 2009
  2. Fixing multiple bugs in keys() when called on B-tree databases. First…

    …, keys()
    
    without :prefix would just crash when using the Edo interface.  Also, keys()
    without :prefix would return duplicate keys due to the use of iteration whereas
    keys() with :prefix (and the underlying C library) would not.
    committed Oct 11, 2009
  3. Provide a class level open() method for all database classes. Previou…

    …sly, only
    
    the Cabinet type classes had this.  It has now been added to Table type classes
    and Dystopia.
    
    This change also makes the class level open() function like Ruby's IO.open() by
    returning the last value in the block.  This introduces a small incompatibility
    since older versions returned nil, but I doubt any code was counting on that.
    committed Oct 11, 2009
Commits on Sep 25, 2009
Commits on Sep 24, 2009
Commits on Sep 23, 2009
  1. tightened hmethods' spec

    jmettraux committed Sep 23, 2009
Commits on Sep 22, 2009
  1. db#mget(k0, k1, ...) now OK.

    jmettraux committed Sep 22, 2009
Commits on Sep 17, 2009
  1. release 1.0.1

    jmettraux committed Sep 17, 2009
Commits on Sep 3, 2009
  1. version numbers update

    jmettraux committed Sep 3, 2009
  2. rdoc for metasearch

    jmettraux committed Sep 3, 2009
Commits on Aug 28, 2009