= RELEASE HISTORY == 2.4.0 / 2010-09-02 All engines go! QED has not been tested against 1.8.6, 1.8.7 and 1.9.2. Underthehood steps are not organized in doubely-linked lists, which makes them much more robust and flexible. This release also improves scoping, test counts, and inline documentation parsing. Changes: * Use new doubly-linked list step design. * Fix -r option on command line. * Provide #instance_exec core extension for Ruby 1.8.6. * Scope is extended by and includes applique. == 2.3.0 / 2010-07-14 Bug to the scurry! QED has broken through the code/document ceiling and is cracking ectoskeletons all the way to the bank. A proverbal can of Roach-Be-Gone this is! What's that you say? I will exlpain. QED can now run directly against code comments. Simply slip the qed command the -c option and feed it some ruby scripts, and presto watch you commnets fail ;) I think you can figure out what to do next. In addition to this coolness QED has been improved under the floor boards as well. The parser, which is much faster, now blocks commentary paragraphs and code examples togeher in one-to-one pairings. Not only does this clean-up the code, but it opens up the potential for Around advice in a future version. Changes: * NEW! Ruby script comment run mode. * Better parsing system uses commentary-example pairs. * Colon can also be used to specify plain text (along with ellipsis). * Now distributed under the more permissive Apache 2.0 license. == 2.2.2 / 2010-06-21 An issue was reported in which the a code block at the very top of a demo was being ignored. This release fixes this issue by rewriting the parser (much better now thanks!). At the same time the Data and Table methods have been polished, both of which can now pick up sample data relative to the current demo. Changes: * Rewrite parser and fix top code issue. * Data method cannot write data, instead executes block. * Data and Table methods look for file relative to demo first. * Added -R option to run demos relative to project root. == 2.2.1 / 2010-06-20 Remove dependencies to Tilt and Nokogiri. Should have done this in last release but alas --there is so much to do. Changes: * Removed HTML parsing dependencies. * Reduce Advice to a single class. == 2.2.0 / 2010-06-19 This release returns to a text-based evaluator, rather then use HTML. Processing HTML proved to have too many edge cases to be effective --both in implementation and in end-usage. So to remedy the situation QED has return to supportting simple markup formats such as RDoc and Markup. This release also adds multi-pattern advice. Instead of a single pattern, multiple patterns can be matched sequentially. This make it a easier to match large text descriptions without restoring to regular expressions. In addition QED now supports raw text blocks. By ending a description section in ellipsis (...), the subsequent code setion becomes a plain text section and is passed into the argument list of any matching When advice. This makes it easy to scaffold fixture files, for example. Finally, this release also refines the evaluation scopes. Where before, a new binding was being created, each was attached to the TOPLEVEL, and therefore not truly isolated on a per-dcoument basis. To correct, QED now mocks the TOPLEVEL providing a new instance of this mock object for each document. Changes: * No longer uses HTML for document processing. * Support for plain text blocks using ellipsis. * New sequential multi-pattern matches. * Mock TOPLEVEL at both the demo and applique levels. * Adjust color support for latest ANSI release. == 2.1.1 / 2010-04-08 Fixed bug introduced in the last version that executed all scripts in a single binding. There needed to be a binding for each script. Changes: * Fixed cross-script bug by moving binding instantiation into Script class. == 2.1.0 / 2010-04-07 QED documents are now run in the TOPLEVEL context, rather than in a subclass of Scope. This ensures code runs as one would expect it too in the wild. Changes: * Scope.new redirect to TOPLEVEL. * DomainLanguage module is added to include into TOPLEVEL. == 2.0.0 / 2010-03-04 This is a major new release of QED. All demonstration documents are now converted to HTML via Tilt (http://github.com/tilt) before being run through the test runner. So QED now supports any markup format supported by Tilt, as well as ordinary HTML. Simply stated, QED process pre tags as code and everything else as comments. Nokogiri is used to parse the HTML. Changes: * HTML serves as a common format. * New dependencies: Tilt and Nokogiri. * New system of version numbers. == 1.2 / 2009-12-07 This release adds a significant new feature, Comment Matchers. These work like Cucumber allowing for background code to be run when matching comments occur --a much better solution for setup and teardown. Changes: * 2 Major Enhancements * Added command matchers via #When method. * All QED methods are now capitalized. * 2 Minor Enhancements * Use OptionParser for qed exectuable. * Verbatim reporter is literally verbatim. == 1.1 / 2009-09-05 This release Changes: * 2 Major Enhancements * Helpers are provided by bottom code. * Added Markdown header support. * 2 Minor Enhancements * Use Ansi project for color output. * Use latest RDoc version. == 1.0 / 2009-06-30 QED has found itself. It took some time to really figure out what this project "was" and how it should best be utilized. This release is the initial release that puts QED in proper perpective. Changes: * 2 Major Enhancement * Partial rewrite of a project that was once called "Quarry". * Now uese AE for assertions.