glue 0.17.0 → 0.18.0

data/lib/html/tokenizer.rb

@@ -0,0 +1,103 @@
+require 'strscan'
+
+module HTML#:nodoc:
+  
+  # A simple HTML tokenizer. It simply breaks a stream of text into tokens, where each
+  # token is a string. Each string represents either "text", or an HTML element.
+  #
+  # This currently assumes valid XHTML, which means no free < or > characters.
+  #
+  # Usage:
+  #
+  #   tokenizer = HTML::Tokenizer.new(text)
+  #   while token = tokenizer.next
+  #     p token
+  #   end
+  class Tokenizer#:nodoc:
+    
+    # The current (byte) position in the text
+    attr_reader :position
+    
+    # The current line number
+    attr_reader :line
+    
+    # Create a new Tokenizer for the given text.
+    def initialize(text)
+      @scanner = StringScanner.new(text)
+      @position = 0
+      @line = 0
+      @current_line = 1
+    end
+
+    # Return the next token in the sequence, or +nil+ if there are no more tokens in
+    # the stream.
+    def next
+      return nil if @scanner.eos?
+      @position = @scanner.pos
+      @line = @current_line
+      if @scanner.check(/<\S/)
+        update_current_line(scan_tag)
+      else
+        update_current_line(scan_text)
+      end
+    end
+  
+    private
+
+      # Treat the text at the current position as a tag, and scan it. Supports
+      # comments, doctype tags, and regular tags, and ignores less-than and
+      # greater-than characters within quoted strings.
+      def scan_tag
+        tag = @scanner.getch
+        if @scanner.scan(/!--/) # comment
+          tag << @scanner.matched
+          tag << @scanner.scan_until(/--\s*>/)
+        elsif @scanner.scan(/!/) # doctype
+          tag << @scanner.matched
+          tag << consume_quoted_regions
+        else
+          tag << consume_quoted_regions
+        end
+        tag
+      end
+
+      # Scan all text up to the next < character and return it.
+      def scan_text
+        @scanner.getch + (@scanner.scan(/[^<]*/) || "")
+      end
+      
+      # Counts the number of newlines in the text and updates the current line
+      # accordingly.
+      def update_current_line(text)
+        @current_line += text.scan(/\r\n|\r|\n/).length
+        text
+      end
+      
+      # Skips over quoted strings, so that less-than and greater-than characters
+      # within the strings are ignored.
+      def consume_quoted_regions
+        text = ""
+        loop do
+          match = @scanner.scan_until(/['"<>]/) or break
+
+          delim = @scanner.matched
+          if delim == "<"
+            match = match.chop
+            @scanner.pos -= 1
+          end
+
+          text << match
+          break if delim == "<" || delim == ">"
+
+          # consume the conqued region
+          while match = @scanner.scan_until(/[\\#{delim}]/)
+            text << match
+            break if @scanner.matched == delim
+            text << @scanner.getch # skip the escaped character
+          end
+        end
+        text
+      end
+  end
+  
+end

data/lib/html/version.rb

@@ -0,0 +1,11 @@
+module HTML#:nodoc:
+  module Version#:nodoc:
+
+    MAJOR = 0
+    MINOR = 5
+    TINY  = 0
+
+    STRING = [ MAJOR, MINOR, TINY ].join(".")
+
+  end
+end

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.8.10
 specification_version: 1
 name: glue
 version: !ruby/object:Gem::Version 
-  version: 0.17.0
-date: 2005-05-16
+  version: 0.18.0
+date: 2005-06-01
 summary: Glue utilities
 require_paths: 
   - lib
@@ -36,6 +36,7 @@ files:
   - doc/LICENSE
   - doc/AUTHORS
   - lib/glue
+  - lib/html
   - lib/glue.rb
   - lib/glue/logger.rb
   - lib/glue/array.rb
@@ -43,18 +44,20 @@ files:
   - lib/glue/flexob.rb
   - lib/glue/hash.rb
   - lib/glue/property.rb
-  - lib/glue/inflector.rb
+  - lib/glue/sanitize.rb
   - lib/glue/number.rb
-  - lib/glue/dynamic_include.rb
   - lib/glue/aspects.rb
   - lib/glue/misc.rb
   - lib/glue/time.rb
   - lib/glue/attribute.rb
-  - lib/glue/cache.rb
   - lib/glue/string.rb
   - lib/glue/object.rb
   - lib/glue/mixins.rb
   - lib/glue/pool.rb
+  - lib/html/document.rb
+  - lib/html/node.rb
+  - lib/html/version.rb
+  - lib/html/tokenizer.rb
   - test/glue
   - test/glue/tc_strings.rb
   - test/glue/tc_validation.rb
@@ -63,33 +66,9 @@ files:
   - test/glue/tc_logger.rb
   - test/glue/tc_aspects.rb
   - test/glue/tc_property_type_checking.rb
-  - test/glue/tc_cache.rb
   - test/glue/tc_hash.rb
   - test/glue/tc_attribute.rb
   - test/glue/tc_property.rb
-  - vendor/extensions
-  - vendor/binding_of_caller.rb
-  - vendor/breakpoint.rb
-  - vendor/blankslate.rb
-  - vendor/README
-  - vendor/breakpoint_client.rb
-  - vendor/extensions/kernel.rb
-  - vendor/extensions/array.rb
-  - vendor/extensions/enumerable.rb
-  - vendor/extensions/hash.rb
-  - vendor/extensions/module.rb
-  - vendor/extensions/numeric.rb
-  - vendor/extensions/ostruct.rb
-  - vendor/extensions/class.rb
-  - vendor/extensions/symbol.rb
-  - vendor/extensions/string.rb
-  - vendor/extensions/object.rb
-  - vendor/extensions/io.rb
-  - vendor/extensions/all.rb
-  - vendor/extensions/_base.rb
-  - vendor/extensions/binding.rb
-  - vendor/extensions/continuation.rb
-  - vendor/extensions/_template.rb
 test_files: []
 rdoc_options: 
   - "--main"
@@ -115,5 +94,5 @@ dependencies:
         - 
           - ">="
           - !ruby/object:Gem::Version 
-            version: 0.7.1
+            version: 0.7.2
       version: 

data/lib/glue/cache.rb

@@ -1,136 +0,0 @@
-# * George Moschovitis  <gm@navel.gr>
-# * Anastasios Koutoumanos  <ak@navel.gr>
-
-module Glue
-
-# A cache utilizing a simple LRU (Least Recently Used) policy.
-# The items managed by this cache must respond to the #key method.
-# Attempts to optimize reads rather than inserts!
-#
-# LRU semantics are enforced by inserting the items in a queue.
-# The lru item is always at the tail. Two special sentinels
-# (head, tail) are used to simplify (?) the code.
-
-class LRUCache < Hash
-
-	# Mix this in your class to make LRU-managable.
-	
-	module Item
-		attr_accessor :lru_key, :lru_prev, :lru_next
-	end
-
-	# head-tail sentinels
-	
-	class Sentinel; include Item; end
-
-	# the maximum number of items in the cache.
-
-	attr_accessor :max_items
-
-	# the head sentinel
-	
-	attr :head
-
-	# the tail sentinel, tail.prev points to the lru item.
-
-	attr :tail
-
-	def initialize(max_items)
-		@max_items = max_items
-		lru_clear()
-	end
-
-	# Lookup an item in the cache.
-	
-	def [](key)
-		if item = super
-			return lru_touch(item)
-		end
-	end
-
-	# The inserted item is considered mru!
-	
-	def []=(key, item)
-		item = super
-		item.lru_key = key
-		lru_insert(item)
-	end
-
-	# Delete an item from the cache.
-	
-	def delete(key)
-		if item = super
-			lru_delete(item)
-		end
-	end
-
-	# Clear the cache.
-	
-	def clear
-		super
-		lru_clear()
-	end
-
-	# The first (mru) element in the cache.
-	
-	def first
-		@head.lru_next
-	end
-
-	# The last (lru) element in the cache.
-	
-	def last
-		@tail.lru_prev
-	end
-	alias_method :lru, :last
-
-	private
-
-	# Delete an item from the lru list.
-	
-	def lru_delete(item)
-		lru_join(item.lru_prev, item.lru_next)
-		return item
-	end
-
-	# Join two items in the lru list.
-	# Return y to allow for chaining.
-	
-	def lru_join(x, y)
-		x.lru_next = y
-		y.lru_prev = x
-		return y
-	end
-
-	# Append a child item to a parent item in the lru list
-	# (Re)inserts the child in the list.
-	
-	def lru_append(parent, child)
-		lru_join(child, parent.lru_next)
-		lru_join(parent, child)
-	end
-
-	# Insert an item
-
-	def lru_insert(item)
-		delete(last.lru_key) if size() > @max_items
-		lru_append(@head, item)
-	end
-	
-	# Touch an item, make mru!
-	# Returns the item.
-	
-	def lru_touch(item)
-		lru_append(@head, lru_delete(item))
-	end
-	
-	# Clear the lru.
-	
-	def lru_clear
-		@head = Sentinel.new
-		@tail = Sentinel.new
-		lru_join(@head, @tail)
-	end	
-end
-
-end 

data/lib/glue/dynamic_include.rb

@@ -1,45 +0,0 @@
-# = Dynamic include
-#
-# Allows you to include dyanamic mixins.
-#
-# === Example
-#
-# module Mixin
-#		def self.append_dynamic_features(base, options)
-#			base.class_eval %{
-#				def hello
-#					puts 'Hello from #{options[:name]}'
-#				end
-#			}
-#		end
-#	end
-#
-# class MyClass
-#		include Mixin, :name => 'tml'
-#	end
-#
-# m = MyClass.new
-# m.hello -> 'Hello from tml'
-#
-# * George Moschovitis  <gm@navel.gr>
-# (c) 2004-2005 Navel, all rights reserved.
-# $Id: dynamic_include.rb 1 2005-04-11 11:04:30Z gmosx $
-
-class Module 
-
-	alias_method :__include_without_options__, :include
-
-	def include(*args)
-		options = args.last.is_a?(Hash) ? args.pop : {}
-
-		for mod in args
-			if mod.respond_to?(:append_dynamic_features)
-				mod.append_dynamic_features(self, options)
-			end
-		end
-
-		__include_without_options__(*args)
-	end
-	alias_method :dynamic_include, :include
-	
-end

data/lib/glue/inflector.rb

@@ -1,91 +0,0 @@
-# Code from RubyOnRails (http://www.rubyonrails.com)
-# Copyright (c) 2004-2005 David Heinemeier Hansson.
-
-module Glue
-
-# The Inflector transforms words from singular to plural, 
-# class names to table names, modulized class names to ones without,
-# and class names to foreign keys.
-
-module Inflector 
-  extend self
-
-  def pluralize(word)
-    result = word.dup
-    plural_rules.each do |(rule, replacement)|
-      break if result.gsub!(rule, replacement)
-    end
-    return result
-  end
-
-  def singularize(word)
-    result = word.dup
-    singular_rules.each do |(rule, replacement)|
-      break if result.gsub!(rule, replacement)
-    end
-    return result
-  end
-
-  def camelize(lower_case_and_underscored_word)
-    lower_case_and_underscored_word.gsub(/(^|_)(.)/){$2.upcase}
-  end
-  
-  def underscore(camel_cased_word)
-    camel_cased_word.gsub(/([A-Z]+)([A-Z])/,'\1_\2').gsub(/([a-z])([A-Z])/,'\1_\2').downcase
-  end
-
-  def demodulize(class_name_in_module)
-    class_name_in_module.gsub(/^.*::/, '')
-  end
-
-  def foreign_key(class_name, separate_class_name_and_id_with_underscore = true)
-    Inflector.underscore(Inflector.demodulize(class_name)) + 
-      (separate_class_name_and_id_with_underscore ? "_id" : "id")
-  end
-  
-	# Convert a class to a name.
-	
-	def name(klass)
-		Inflector.underscore(Inflector.demodulize(klass.to_s))
-	end
-	
-	# Convert a class to a name in plural
-	
-	def plural_name(klass)
-		Inflector.pluralize(Inflector.underscore(Inflector.demodulize(klass.to_s)))
-	end
-	
-  private
-    def plural_rules #:doc:
-      [
-        [/(x|ch|ss)$/, '\1es'],               # search, switch, fix, box, process, address
-        [/([^aeiouy]|qu)y$/, '\1ies'],        # query, ability, agency
-        [/(?:([^f])fe|([lr])f)$/, '\1\2ves'], # half, safe, wife
-        [/sis$/, 'ses'],                      # basis, diagnosis
-        [/([ti])um$/, '\1a'],                 # datum, medium
-        [/person$/, 'people'],                # person, salesperson
-        [/man$/, 'men'],                      # man, woman, spokesman
-        [/child$/, 'children'],               # child
-        [/s$/, 's'],                          # no change (compatibility)
-        [/$/, 's']
-      ]
-    end
-
-    def singular_rules #:doc:
-      [
-          [/(x|ch|ss)es$/, '\1'],
-          [/([^aeiouy]|qu)ies$/, '\1y'],
-          [/([lr])ves$/, '\1f'],
-          [/([^f])ves$/, '\1fe'],
-          [/(analy|ba|diagno|parenthe|progno|synop|the)ses$/, '\1sis'],
-          [/([ti])a$/, '\1um'],
-          [/people$/, 'person'],
-          [/men$/, 'man'],
-          [/status$/, 'status'],
-          [/children$/, 'child'],
-          [/s$/, '']
-        ]
-    end
-end
-
-end