v0.7.7 (1st October 2009)
- optimise sed usage in ONIX::Normaliser. *huge* speed improvement on
  large files.

v0.7.6 (21st September 2009)
- provide access to the PackQuantity element

v0.7.5 (8th September 2009)
- Don't raise an exception on malformed dates when reading files

v0.7.4 (2nd September 2009)
- Expand ONIX::Normaliser
  - strip control chars
  - add encoding declaration to valid utf-8 files that aren't declared
    as such

v0.7.3 (19th August 2009)
- Switch from java to xsltproc to convert short tag ONIX files
  to reference tags

v0.7.2 (19th August 2009)
- Added ONIX::Normaliser class
  - for normalising various ONIX files into a form that makes them easy
    to process. Shouldn't be necesary to pre-process files like this, but
    I'm sick of trying to wrestle the libxml ruby bindings

v0.7.1 (24th June 2009)
- Small tweak to ordering of elements in the Product group

v0.7.0 (17th June 2009)
- try using LibXML for reader again
  - retrieving the ONIX version of the input file is currently disabled, as 
    that seems to be the source of our instability
- Various Ruby 1.9 compatability tweaks
  - add source file coding declarations. All source files are UTF-8
  - ONIX::Reader ensures all input data is converted to UTF-8
  - the ROXML based objects seem to forget the encoding when they're marshalled,
    so force string based attributes *back* to UTF-8

v0.6.7 (Unreleased)
- add some accessors to the Title composite

v0.6.6 (Unreleased)
- Forget the S on an element name

v0.6.5 (Unreleased)
- Ruby 1.9 compat

v0.6.4 (Unreleased)
- Add APAProduct#price

v0.6.3(Unreleased)
- Bump ROXML dependency to 2.5.3 to get libxml-ruby 1.1.3 compatibility

v0.6.2 (Unreleased)
- Fix a small typo in APAProduct

v0.6.1 (Unreleased)
- Stopped using LibXMLs Reader class as the basis for our reader.
  - We were getting too many segfaults (even 1 is too many!)
  - until we resolve it, reverted to manual string parsing
  - This is a fairly major regression of functionality. For 99% of files
    it won't matter, but for some corner cases it will. ie UTF-16 encoded
    files
  - Will also be noticeably slower
  - Hopefully only a short term fix, until I work out what is going on with
    libxml

v0.6.0 (18th March 2009)
- remove use of threads in ONIX::Reader
  - a producer/consumer pattern was useful in the REXML stream parsing days, but
    now LibXML's Reader binding provides a better alternative
  - API left unchanged, this was all under the hood
- bump required ROXML version to 2.5.2

v0.5.1 (4th March 2009)
- Fix a single letter typo

v0.5 (2nd March 2009)
- Switch ROXML dependency from a patched version to vanilla
  - Vanilla ROXML now has all the features we need
  - This change should be transparent to ONIX gem users

v0.4.7 (9th December 2008)
- Contributor sub elements should match the order specified in the DTD

v0.4.6 (2nd December 2008)
- 2 new accessors on the contributor class - bio and corporate name

v0.4.5 (21st November 2008)
- APAProduct wrapper should generate valid MediaFile composites

v0.4.4 (19th November 2008)
- Added support for more elements from MarketRepresentation

v0.4.3 (11th November 2008)
- Added support for AgentName and MarketPublishingStatus

v0.4.2 (1st November 2008)
- Remove final remnants of REXML code
- Minor reordering of elements to match DTD

v0.4.1 (UNRELEASED)
- Added accessors to various product measurements. Height, weight, etc.
- Reduced time for an ONIX::Reader class to initialise

v0.4.0 (28th October 2008)
- Major rework: now based on ROXML instead of xml-mapping
  - Mostly API Compatible
  - StreamReader and StreamWriter renamed to Reader and Writer
  - ROXML is based on libxml, so things should be significantly faster

v0.2.7 (Unreleased)
- Add line breaks after each product

v0.2.5 (Unreleased)
- Make PublishingStatus a two_digit_node

v0.2.4 (Unreleased)
- Initialise the media files array of a new product correctly

v0.2.3 (Unreleased)
- Switch a few more fields over to TwoDigitNodes
- Make the product availability field accessible from APAProduct

v0.2.2 (Unreleased)
- Add a new nodetype (DateNode) for YYYYMMDD fields

v0.2.1 (Unreleased)
- Add a new nodetype (TwoDigitNode) for two digit codes

v0.2.0 (16th July 2008)
- Add support for reading and storing subject codes (BIC, BISAC, etc)

v0.1.0 (12th June 2008)
- Initial Release