Permalink
Commits on Jan 6, 2010
  1. 2010

    jmettraux committed Jan 6, 2010
  2. 1.0.5

    jmettraux committed Jan 6, 2010
  3. tcfree for Tyrant as well

    jmettraux committed Jan 6, 2010
Commits on Jan 5, 2010
  1. integrating Aaron's test

    jmettraux committed Jan 5, 2010
Commits on Dec 25, 2009
  1. release 1.0.4

    jmettraux committed Dec 25, 2009
Commits on Dec 24, 2009
  1. removed trailing whitespaces

    jmettraux committed Dec 24, 2009
Commits on Dec 22, 2009
  1. removed .tdb examples. Closes gh-25

    jmettraux committed Dec 22, 2009
  2. updated links (1978th.net)

    jmettraux committed Dec 22, 2009
Commits on Nov 18, 2009
Commits on Nov 16, 2009
  1. release 1.0.3

    jmettraux committed Nov 16, 2009
Commits on Nov 2, 2009
Commits on Nov 1, 2009
  1. provided a gem.description

    jmettraux committed Nov 1, 2009
Commits on Oct 30, 2009
  1. preparing 1.0.2

    jmettraux committed Oct 30, 2009
Commits on Oct 14, 2009
  1. Dropping an eval() call for send().

    JEG2 committed Oct 14, 2009
  2. Speed improvements for each().

    JEG2 committed Oct 14, 2009
Commits on Oct 13, 2009
  1. todo : reenabled #copy for Tyrant (Edo and Tokyo) - closes gh-19

    jmettraux committed Oct 13, 2009
    (would be good to have one at hand for tyrant tables)
  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

    JEG2 committed Oct 12, 2009
    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.
Commits on Oct 11, 2009
  1. Fixing multiple bugs in fixed-width databases. First, Edo didn't supp…

    JEG2 committed Oct 11, 2009
    …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.
  2. Fixing multiple bugs in keys() when called on B-tree databases. First…

    JEG2 committed Oct 11, 2009
    …, 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.
  3. Provide a class level open() method for all database classes. Previou…

    JEG2 committed Oct 11, 2009
    …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.