merging in current codebase

History.txt

@@ -1,40 +1,100 @@
-== Trunk
+== Trunk (Git)
 
+* _IMPORTANT_ Breaking change:
+
+  * Removed init.rb auto-require of webrat/rails
+  * Removed auto-require of webrat/rails when requiring webrat when RAILS_ENV is
+    defined
+
+  In your env.rb file, ensure you have:
+
+    require "webrat/rails"
+
 * Major enhancements
 
-  * Added selects_time, selects_date, and selects_datetime to API. (Ben Mabey, Ticket #36)
+  * Added select_time, select_date, and select_datetime to API. [#36] (Ben Mabey)
+  * Use Hpricot and REXML when not parsing with Nokogiri (on JRuby, for example)
 
 * Minor enhancements
-  * Now sets the mode with configuration.mode in the block (gaffo, Ticket #85)
-  * Allow clicking links by id and id regexp (gaffo)
-  * Raise Webrat::PageLoadError when a failure occurs so that application exceptions can be more accurately tested (Ryan Briones)
-  * Helpful error message for missing option in select box. (Ben Mabey, Ticket #40)
+  * Now sets the mode with configuration.mode in the block (Mike Gaffney, Ticket #85)
+  * Detect if the document is XML or HTML using the Content-Type when in Rails mode
+  * Expose #selenium method for direct access to Selenium client
+  * Check nokogiri gem version before requiring nokogiri
+  * Include the Selenium server jar file in the gem (Bryan Helmkamp, Ben Schwarz)
+  * Added key_down, key_up and fire_event to Selenium session (Fernando Garcia)
+  * Fix outputing README during Rails plugin install (Fernando Garcia)
+  * Strip newlines when matching label text (Miha Filej)
+  * Add simple support for accessing Webrat's matchers from RSpec by requiring
+    "webrat/rspec-rails" (David Chelimsky)
+  * Support save_and_open_page in Windows and Cygwin (Mike Gaffney)
+  * Added RadioField#checked? to indicated whether or not a radio button is checked
+    (Luke Melia)
+  * Add spec:jruby rake task for running Webrat's spec suite with JRuby
+  * Added field_by_xpath to locate a Webrat::Field using arbitrary XPath expressions
+  * Helpful error message for missing option values [#40] (Ben Mabey)
+  * Add set_hidden_field method (Noah Davis, Bryan Helmkamp)
+  * Add submit_form method for submitting a form by ID (Noah Davis, Bryan Helmkamp)
+  * Switch to using Nokogiri.parse for simple XML/XHTML autodetection [#66]
+  * Removed Webrat.root method, which seemed to be unused
+  * Added Webrat.configure method for global Webrat configuration [#33]
+    (Mike Gaffney)
+  * Added automatic starting and stopping of the Selenium server and a Mongrel Rails
+    app server when using webrat/selenium
+  * Switch to using selenium-client gem and vendor selenium-server.jar (Luke Melia)
+  * Added gemspec so the gem builds on GitHub now
+  * Deprecate old style methods (fills_in is deprecated in favor of fill_in, etc.)
+  * Improcements to the README and RDoc (Bryan Helmkamp, Mike Gaffney)
+  * Allow clicking links by id and id regexp (Mike Gaffney)
+  * Raise Webrat::DisabledFieldError when attempting to manipulate a disabled field
+  * Raise Webrat::NotFoundErrors when an element is not found
+  * Raise Webrat::PageLoadError when a failure occurs so that application exceptions
+    can be more accurately tested (Ryan Briones)
+  * Helpful error message for missing option in select box. [#40] (Ben Mabey)
 
+* Bug fixes
+
+  * Match against link _text_ which decodes character references.
+    Useful for multibyte languages like Japanese (moronatural@gmail.com)
+  * Fix params hash generation for Mechanize when Merb is not defined [#62]
+  * Expose some Webrat methods that were missing from the Webrat::Methods module
+    (Low Chin Chau)
+  * Allow mechanize session to pass through basic auth (Ryan Briones)
+  * Improvements to the Mechanize support (Jeremy Burks)
+  * Fix following fully qualified local links (Lawrence Pit)
+  * Fixed bug where Webrat would lose complex parameters (like foo[bar[baz]][])
+    in Merb due to not correctly merging Mashes. (Drew Colthorp)
+  * Extend Rails' ActionController::IntegrationTest instead of
+    ActionController::Integration::Session (Fixes using Webrat's #select method and 
+    avoids usage of method_missing)
+
 == 0.3.2 / 2008-11-08
 
-* Minor enhancements
+* 1 Minor enhancement
 
-  * Fixes behavior or have_tag when a block is passed. It passes the matched node(s) to the block for further specs again. (Carl Lerche)
+  * Fixes behavior or have_tag when a block is passed. It passes the matched node(s)
+    to the block for further specs again. (Carl Lerche)
 
 == 0.3.1 / 2008-11-07
 
-  * Minor enhancements
+* 1 Minor enhancement
 
-    * Use @_webrat_session instance variable instead of @session for Merb integration to avoid collisions
+  * Use @_webrat_session instance variable instead of @session for Merb integration
+    to avoid collisions
 
 == 0.3.0 / 2008-11-07
 
-* Major enhancements
+* 4 Major enhancements
 
   * Added Merb support (Gwyn Morfey, Jeremy Burks, Rob Kaufman, Yehuda Katz)
   * Added experimental Selenium support (Luke Melia)
   * Add have_selector, have_xpath, have_tag and contain matchers from Merb
   * Switch from Hpricot to Nokogiri for XML parsing (thanks, Aaron Patterson)
 
-* Minor enhancements
+* 37 Minor enhancements
 
   * Added #within for manipulating the current page within a selector scope
-  * Add support for file fields via #attaches_file method (Rails only at the moment) (Kyle Hargraves)
+  * Add support for file fields via #attaches_file method (Rails only at the moment)
+    (Kyle Hargraves)
   * Add support for simulating SSL requests (Luke Melia)
   * Added #basic_auth(user, pass) to support HTTP Basic Auth (Aslak Hellesøy)
   * Added support for Sinatra and Rack (Aslak Hellesøy)
@@ -47,32 +107,37 @@
   * Allow clicking links by a regular expression
   * Add #http_accept for including MIME type HTTP "Accept" headers (Ryan Briones)
   * Add #header to support inclusion of custom HTTP headers (Ryan Briones)
-  * Consider response codes 200-499 as successful enough to not raise a Webrat error (David Leal)
+  * Consider response codes 200-499 as successful enough to not raise a Webrat error
+    (David Leal)
   * Add support for clicking areas of an image map (Alex Lang)
   * Support relative links, including href="?foo=bar" (Kyle Hargraves)
-  * Separated Rails-specific code from the Webrat core to make it easier to use Webrat with other environments
+  * Separated Rails-specific code from the Webrat core to make it easier to use
+    Webrat with other environments
   * Alias visits as visit, clicks_link as click_link, etc. for better readability
   * Raise error when trying to interact with a disabled form element (Luke Melia)
   * Don't send disabled form elements to the server (Nicholas A. Evans)
   * Display response body when the page load is not successful (David Leal)
   * CGI escape form field values (Miha Filej)
-  * Add support for redirect_to :back by sending HTTP_REFERER headers (Hendrik Volkmer)
+  * Add support for redirect_to :back by sending HTTP_REFERER headers
+    (Hendrik Volkmer)
   * Expose current DOM (as an Hpricot object) as #current_dom
-  * Add support for disabling JavaScript when clicking a link to enable testing of both JS
-    and non-JS implementations (Luke Melia and Bryan Helmkamp)
+  * Add support for disabling JavaScript when clicking a link to enable testing of
+    both JS and non-JS implementations (Luke Melia and Bryan Helmkamp)
   * Support &nbsp's as spaces in matching link text (Luke Melia)
   * Support button elements (Nick Sieger)
   * Support matching select options by regexp (Kyle Hargraves)
-  * save_and_open_page rewrites css and image references to provide a friendlier debugging experience (Luke Melia)
-  * Added support for matching alt attributes in fields (primarly for clicks_button) (Aaron Quint)
+  * save_and_open_page rewrites css and image references to provide a friendlier
+    debugging experience (Luke Melia)
+  * Added support for matching alt attributes in fields (primarly for clicks_button)
+    (Aaron Quint)
   * Support '&' in submitted values (Kyle Hargraves)
   * Support clicking links by title (Dan Barry)
   * Added missing spec for clicking image buttons (Tim Harper)
   * Switched tests to specs, and from Mocha to RSpec's mocking library
   * Add support to click_button for IDs (Gwyn Morfey)
   * Miscellaneous core refactorings (Jan Suchal)
 
-* Bug fixes
+* 8 Bug fixes
 
   * Fix initialization of WWW::Mechanize (Derek Kastner)
   * Don't open blank pages in the browser (Kyle Hargraves)
@@ -81,7 +146,8 @@
   * Fix bug with empty select list option (Kyle Hargraves)
   * Fix regression of not sending default values in password fields
   * Don't explode if encountering inputs with no type attribute (assume text)
-  * Fix bug where choosing a radio button in a series with a default submitted the incorrect field value (Luke Melia)
+  * Fix bug where choosing a radio button in a series with a default submitted the
+    incorrect field value (Luke Melia)
 
 == 0.2.0 / 2008-04-04
 
@@ -98,25 +164,29 @@
   * Added reloads method to reload the page (Kamal Fariz Mahyuddi)
   * Prevent making a request if clicking on local anchor link (Kamal Fariz Mahyuddi)
   * Added clicks_link_within(selector, link_text), allowing restricting link search 
-  to within a given css selector (Luke Melia)
+    to within a given css selector (Luke Melia)
   * Allow specifying the input name/label when doing a select (David Chelimsky)
-  * Raise a specific exception if the developer tries to manipulate form elements before loading a page (James Deville)
+  * Raise a specific exception if the developer tries to manipulate form elements
+    before loading a page (James Deville)
   * Add support for alternate POST, PUT and DELETE link clicking (Kyle Hargraves)
   * Change clicks_link to find the shortest matching link (Luke Melia)
   * Improve matching for labels in potentially ambiguous cases
 
 * 7 Bug fixes
 
-  * Fix incorrect serializing of collection inputs, i.e. name contains [] (Kamal Fariz Mahyuddi)
+  * Fix incorrect serializing of collection inputs, i.e. name contains []
+    (Kamal Fariz Mahyuddi)
   * Serialize empty text field values just like browsers (Kamal Fariz Mahyuddi)
   * Quick fix to avoid @dom not initialized warnings (Kamal Fariz Mahyuddi)
   * Docfix: bad reference to #select method in README (Luke Melia)
-  * Ensure Rails-style checkboxes work properly (checkboxes followed by a hidden input with the same name)
+  * Ensure Rails-style checkboxes work properly (checkboxes followed by a hidden
+    input with the same name)
   * Fix Edge Rails (a.k.a. 2.0 RC) compatibility (David Chelimsky)
   * Support param hashes nested more than one level (David Chelimsky)
 
 == 0.1.0 / 2007-11-28
 
 * 1 major enhancement
+
   * Birthday!
 

README.rdoc

@@ -4,6 +4,7 @@
 - http://groups.google.com/group/webrat
 - http://webrat.lighthouseapp.com/
 - http://github.com/brynary/webrat
+- #webrat on Freenode
 
 == Description
 
@@ -15,7 +16,7 @@ web application.
 * Browser Simulator for expressive, high level acceptance testing without the
   performance hit and browser dependency of Selenium or Watir (See Webrat::Session)
 * Use the same API for Browser Simulator and real Selenium tests using
-  Webrat::SeleniumSession when necessary (eg. for testing AJAX interactions)
+  Webrat::Selenium when necessary (eg. for testing AJAX interactions)
 * Supports multiple Ruby web frameworks: Rails, Merb and Sinatra
 * Supports popular test frameworks: RSpec, Cucumber, Test::Unit and Shoulda
 * Webrat::Matchers API for verifying rendered HTML using CSS, XPath, etc.
@@ -40,6 +41,14 @@ Behind the scenes, Webrat will ensure:
 * If a URL is invalid, the test will fail.
 * If a page load or form submission is unsuccessful, the test will fail.
 
+== Installing Nokogiri
+
+Users of Debian Linux (e.g. Ubuntu) need to run:
+
+  sudo apt-get install libxslt1-dev libxml2-dev.
+
+Otherwise the Nokogiri gem, which Webrat depends on, won't install properly.
+
 == Install for Rails
 
 To install the latest release as a gem:

Rakefile

@@ -21,14 +21,14 @@ spec = Gem::Specification.new do |s|
   s.bindir       = "bin"
   s.description  = s.summary
   s.require_path = "lib"
-  s.files        = %w(History.txt init.rb install.rb MIT-LICENSE.txt README.rdoc Rakefile) + Dir["lib/**/*"]
+  s.files        = %w(History.txt install.rb MIT-LICENSE.txt README.rdoc Rakefile) + Dir["lib/**/*"] + Dir["vendor/**/*"]
 
   # rdoc
   s.has_rdoc         = true
   s.extra_rdoc_files = %w(README.rdoc MIT-LICENSE.txt)
 
   # Dependencies
-  s.add_dependency "nokogiri", ">= 1.0.3"
+  s.add_dependency "nokogiri", ">= 1.0.6"
 
   s.rubyforge_project = "webrat"
 end
@@ -75,8 +75,13 @@ task :install_gem => [:clean, :package] do
   sh "sudo gem install --local #{gem}"
 end
 
+desc "Delete generated RDoc"
+task :clobber_docs do
+  FileUtils.rm_rf("doc")
+end
+
 desc "Generate RDoc"
-task :docs do
+task :docs => :clobber_docs do
   system "hanna --title 'Webrat #{Webrat::VERSION} API Documentation'"
 end
 
@@ -85,4 +90,15 @@ task "spec:jruby" do
   system "jruby -S rake spec"
 end
 
-task :default => :spec
+desc "Run each spec in isolation to test for dependency issues"
+task :spec_deps do
+  Dir["spec/**/*_spec.rb"].each do |test|
+    if !system("spec #{test} &> /dev/null")
+      puts "Dependency Issues: #{test}"
+    end
+  end
+end
+
+task :default => :spec
+
+task :precommit => ["spec", "spec:jruby"]

lib/webrat.rb

@@ -10,14 +10,16 @@ class WebratError < StandardError
   VERSION = '0.3.2.1'
 
   def self.require_xml
+    gem "nokogiri", ">= 1.0.6"
+
     if on_java?
       # We need Nokogiri's CSS to XPath support, even if using REXML and Hpricot for parsing and searching
       require "nokogiri/css"
       require "hpricot"
       require "rexml/document"
     else
       require "nokogiri"
-      require "webrat/core/nokogiri"
+      require "webrat/core/xml/nokogiri"
     end
   end
 

lib/webrat/core.rb

@@ -1,13 +1,13 @@
 require "webrat/core/configuration"
 require "webrat/core/xml"
-require "webrat/core/nokogiri"
+require "webrat/core/xml/nokogiri"
 require "webrat/core/logging"
-require "webrat/core/form"
+require "webrat/core/elements/form"
 require "webrat/core/scope"
-require "webrat/core/link"
-require "webrat/core/area"
-require "webrat/core/label"
-require "webrat/core/select_option"
+require "webrat/core/elements/link"
+require "webrat/core/elements/area"
+require "webrat/core/elements/label"
+require "webrat/core/elements/select_option"
 require "webrat/core/session"
 require "webrat/core/methods"
 require "webrat/core/matchers"

lib/webrat/core/configuration.rb

@@ -7,10 +7,15 @@ def self.configure(configuration = Webrat::Configuration.new)
     @@configuration = configuration
   end
 
-  def self.configuration
+  def self.configuration # :nodoc:
     @@configuration ||= Webrat::Configuration.new
   end
-
+
+  # Webrat can be configured using the Webrat.configure method. For example:
+  # 
+  #   Webrat.configure do |config|
+  #     config.parse_with_nokogiri = false
+  #   end
   class Configuration
 
     RAILS_MODE = :rails
@@ -24,7 +29,7 @@ class Configuration
     attr_writer :parse_with_nokogiri
 
     # Webrat's mode, set automatically when requiring webrat/rails, webrat/merb, etc.
-    attr_accessor :mode
+    attr_accessor :mode # :nodoc:
 
     # Save and open pages with error status codes (500-599) in a browser? Defualts to true.
     attr_writer :open_error_files

lib/webrat/core/area.rb → lib/webrat/core/elements/area.rb

@@ -1,34 +1,21 @@
+require "webrat/core/elements/element"
+
 module Webrat
-  class Area #:nodoc:
+  class Area < Element #:nodoc:
 
-    def initialize(session, element)
-      @session  = session
-      @element  = element
+    def self.xpath_search
+      ".//area"
     end
 
     def click(method = nil, options = {})
       @session.request_page(absolute_href, :get, {})
     end
 
-    def matches_text?(id_or_title)
-      matcher = /#{Regexp.escape(id_or_title.to_s)}/i
-      title =~ matcher || id =~ matcher
-    end
-
-    protected
+  protected
 
     def href
-      @element["href"]
-    end
-
-    def title
-      @element["title"]
+      Webrat::XML.attribute(@element, "href")
     end
-
-    def id
-      @element["id"]
-    end
-
 
     def absolute_href
       if href =~ /^\?/