wax 0.9.5 → 0.9.7

data/lib/address.rb

@@ -6,11 +6,25 @@ class Address
   end
 
   def to_xml(wax)
-    wax.start('address').
-      child('street', @street).
-      child('city', @city).
-      child('state', @state).
-      child('zip', @zip).
+    # Chaining approach
+    #wax.start('address').
+    #  child('street', @street).
+    #  child('city', @city).
+    #  child('state', @state).
+    #  child('zip', @zip).
+    #  end!
+
+    # Non-chaining approach
+    # Put the current object in a local variable
+    # so it can be accessed in the block passed to the write method.
+    address = self
+    wax.write do
+      start 'address'
+      child 'street', address.street
+      child 'city', address.city
+      child 'state', address.state
+      child 'zip', address.zip
       end!
+    end
   end
 end

data/lib/car.rb

@@ -6,10 +6,23 @@ class Car
   end
 
   def to_xml(wax)
-    wax.start('car').attr('year', @year).
-      child('make', @make).
-      child('model', @model).
+    # Chaining approach
+    #wax.start('car').attr('year', @year).
+    #  child('make', @make).
+    #  child('model', @model).
+    #  end!
+
+    # Non-chaining approach
+    # Put the current object in a local variable
+    # so it can be accessed in the block passed to the write method.
+    car = self
+    wax.write do
+      start 'car'
+      attr 'year', car.year
+      child 'make', car.make
+      child 'model', car.model
       end!
+    end
   end
 end
 

data/lib/cd_demo.rb

@@ -2,15 +2,15 @@ require 'wax'
 
 url = "http://www.ociweb.com"
 WAX.write($stdout, "1.0") do
+  xslt "artist.xslt"
+  dtd "#{url}/xml/music.dtd"
+  start "artist"
+  attr "name", "Gardot, Melody"
+  namespace nil, "#{url}/music", "#{url}/xml/music.xsd"
+  namespace "date", "#{url}/date", "#{url}/xml/date.xsd"
   comment "This is one of my favorite CDs!"
-  dtd "#{url}/xml/cd.dtd"
-  xslt "cd.xslt"
   start "cd"
   attr "year", 2008
-  namespace nil, "#{url}/music", "#{url}/xml/cd.xsd"
-  namespace "date", "#{url}/date", "#{url}/xml/date.xsd"
-  start "artist"
-  attr "name", "Gardot, Melody"
   child "title", "Worrisome Heart"
   child "date", "purchaseDate", "4/3/2008"
 end

data/lib/person.rb

@@ -6,13 +6,27 @@ class Person
   end
 
   def to_xml(wax)
-    wax.start('person').attr('birthdate', @birthdate).child('name', @name)
-    @address.to_xml(wax)
-    wax.end!
+    # Chaining approach
+    #wax.start('person').attr('birthdate', @birthdate).child('name', @name)
+    #@address.to_xml(wax)
+    #wax.end!
+
+    # Non-chaining approach
+    # Put the current object in a local variable
+    # so it can be accessed in the block passed to the write method.
+    person = self
+    wax.write do
+      start 'person'
+      attr 'birthdate', person.birthdate
+      child 'name', person.name
+      person.address.to_xml(self) # self is the WAX object here
+      end!
+    end
   end
 end
 
-if $PROGRAM_NAME == __FILE__ || $PROGRAM_NAME =~ /\/rake/
+if $PROGRAM_NAME == __FILE__
+# || $PROGRAM_NAME =~ /\/rake/
   require 'address'
   require 'wax'
 

data/lib/tutorial.rb

@@ -53,7 +53,7 @@ end
 
 out 'Without indentation, on a single line:'
 WAX.write do
-  set_indent nil
+  no_indents_or_crs
   start 'car'
   child 'model', 'Prius'
 end
@@ -155,7 +155,7 @@ out 'Associate an XML Schema:'
 WAX.write do
   start 'car'
   attr 'year', 2008
-  namespace nil, 'http://www.ociweb.com/cars', 'car.xsd'
+  namespace '', 'http://www.ociweb.com/cars', 'car.xsd'
   child 'model', 'Prius'
 end
 # <car year='2008'
@@ -169,7 +169,7 @@ out 'Associate multiple XML Schemas:'
 WAX.write do
   start 'car'
   attr 'year', 2008
-  namespace nil, 'http://www.ociweb.com/cars', 'car.xsd'
+  namespace '', 'http://www.ociweb.com/cars', 'car.xsd'
   namespace 'm', 'http://www.ociweb.com/model', 'model.xsd'
   child 'm', 'model', 'Prius'
 end

data/lib/wax.rb

@@ -32,6 +32,11 @@ class WAX
   # If they are, a RuntimeError is raised.
   STATES = [:in_prolog, :in_start_tag, :in_element, :after_root]
 
+  NONWINDOWS_CR = "\n"
+  WINDOWS_CR = "\r\n"
+
+  attr_reader :state
+
   # Creates a WAX object, invokes the specified block on it
   # and calls close on the WAX object.
   # The writer can be a String file path, an IO object such as a File,
@@ -42,6 +47,11 @@ class WAX
     wax = WAX.new(writer, version)
     wax.instance_exec(&proc)
     wax.close
+    wax # return instance so caller can get configuration info. like @cr
+  end
+
+  def write(writer=$stdout, version=nil, &proc)
+    instance_eval(&proc)
   end
 
   # Initializes new instances of this class.
@@ -49,6 +59,8 @@ class WAX
     @attr_on_new_line = false
     @check_me = true
     @close_stream = writer != $stdout
+    @dtd_specified = false
+    @in_commented_start = false
     @dtd_file_path = nil
     @entity_defs = []
     @has_content = @has_indented_content = false
@@ -59,6 +71,8 @@ class WAX
     @prefixes_stack = []
     @state = :in_prolog
     @writer = writer 
+    @xslt_specified = false
+    use_non_windows_cr
     write_xml_declaration(version)
   end
 
@@ -69,9 +83,9 @@ class WAX
   # to indicate whether the attribute should be written on a new line.
   def attr(p1, p2, p3=nil, p4=nil)
     if (p3 == nil)
-      prefix, name, value, new_line = nil, p1, p2, false
+      prefix, name, value, new_line = nil, p1, p2, @attr_on_new_line
     elsif (p4 == nil)
-      prefix, name, value, new_line = p1, p2, p3, false
+      prefix, name, value, new_line = p1, p2, p3, @attr_on_new_line
     else
       prefix, name, value, new_line = p1, p2, p3, p4
     end
@@ -80,11 +94,11 @@ class WAX
       bad_state("attr") unless @state == :in_start_tag
 
       unless prefix == nil
-        XMLUtil.verify_nmtoken(prefix)
+        XMLUtil.verify_name(prefix)
         @pending_prefixes << prefix
       end
 
-       XMLUtil.verify_nmtoken(name)
+       XMLUtil.verify_name(name)
        value = XMLUtil.escape(value)
     end
 
@@ -168,19 +182,40 @@ class WAX
     write_indent if @parent_stack.size > 0
 
     out "<!-- #{text} -->"
-    out "\n" if will_indent and @parent_stack.size == 0
+    out "#{@cr}" if will_indent and @parent_stack.size == 0
+
+    self
+  end
+
+  # Writes a commented start tag for a given element name,
+  # but doesn't terminate it.
+  def commented_start(p1, p2=nil)
+    if (p2 == nil)
+      # only specified element name
+      prefix, name = nil, p1
+    else
+      # specified element namespace prefix, name and text
+      prefix, name = p1, p2
+    end
+
+    @in_commented_start = true
+    start(prefix, name)
+    @in_commented_start = false
 
     self
   end
 
   # Writes a DOCTYPE that associates a DTD with the XML document.
   def dtd(file_path)
+    raise "can't specify more than one DTD" if @dtd_specified
+
     if @check_me
       bad_state("dtd") unless @state == :in_prolog
       XMLUtil.verify_uri(file_path)
     end
 
     @dtd_file_path = file_path
+    @dtd_specified = true
     self
   end
 
@@ -202,11 +237,17 @@ class WAX
     # are no longer in scope.
     @prefixes_stack.pop
 
+    # Check for hypen at beginning of element name
+    # which indicates that the commentedStart method was used.
+    was_commented_start = name[0] == '-'[0]
+
     if @has_content
       write_indent if @has_indented_content
-      out "</#{name}>"
+      out '</'
+      out was_commented_start ? name[1..-1] : name
+      out was_commented_start ? '-->' : '>'
     else
-      out "/>"
+      out was_commented_start ? '/-->' : '/>'
     end
 
     @has_content = @has_indented_content = true # new setting for parent
@@ -228,6 +269,11 @@ class WAX
     entity_def(name + " SYSTEM", file_path)
   end
 
+  # Gets the carriage return characters currently being used.
+  def get_cr
+    @cr
+  end
+
   # Gets the indentation characters being used.
   def get_indent
     @indent
@@ -277,7 +323,7 @@ class WAX
     if @check_me
       bad_state("namespace") unless @state == :in_start_tag
 
-      XMLUtil.verify_nmtoken(prefix) if has_prefix
+      XMLUtil.verify_name(prefix) if has_prefix
       XMLUtil.verify_uri(uri)
       XMLUtil.verify_uri(schema_path) unless schema_path == nil
     end
@@ -321,11 +367,26 @@ class WAX
     text text, true, @check_me
   end
 
+  # Don't output indent output or write carriage returns.
+  # Write out the XML on a single line.
+  def no_indents_or_crs
+    set_indent nil
+  end
+
+  # Writes the to_s value of an Object to the writer.
+  def out(data)
+    raise "attempting to out XML after close has been called" unless @writer
+    @writer.write(data.to_s)
+  end
+
   # Writes a processing instruction.
   def processing_instruction(target, data)
     if @check_me
       bad_state("pi") if @state == :after_root
-      XMLUtil.verify_nmtoken(target)
+
+      # Special handling for this processing instruction
+      # since starting with "xml" is reserved.
+      XMLUtil.verify_name(target) unless target == "xml-stylesheet"
     end
         
     @has_content = @has_indented_content = true
@@ -333,17 +394,17 @@ class WAX
     write_indent if @parent_stack.size > 0
 
     out "<?#{target} #{data}?>"
-    out("\n") if will_indent and @parent_stack.size == 0
+    out("#{@cr}") if will_indent and @parent_stack.size == 0
 
     self
   end
 
   # Sets the indentation characters to use.
   # The only valid values are
-  # a single tab, one or more spaces, an empty string, or null.
+  # a single tab, one or more spaces, an empty string, or nil.
   # Passing "" causes elements to be output on separate lines,
   # but not indented.
-  # Passing null causes all output to be on a single line.
+  # Passing nil causes all output to be on a single line.
   def set_indent(indent)
     if indent == nil
       @indent = indent
@@ -355,7 +416,7 @@ class WAX
         raise ArgumentError, "can't indent a negative number of spaces"
       end
 
-      if count > 4
+      if @check_me && count > 4
         raise ArgumentError, "#{count} is an unreasonable indentation"
       end
 
@@ -364,23 +425,27 @@ class WAX
       return
 
     elsif indent.kind_of?(String)
-      # Note that the parens on the next line are necessary
-      # because the assignment operator has higher precedence than "or".
-      valid = (indent == nil or indent.length == 0 or indent == "\t")
-      
-      unless valid
-        # It can only be valid now if every character is a space.
-        valid = true
-        for i in 0...indent.length
-          unless indent[i] == 32 # space
-            valid = false
-            break
+      if @check_me
+        # Note that the parens on the next line are necessary
+        # because the assignment operator has higher precedence than "or".
+        valid = (indent == nil or indent.length == 0 or indent == "\t")
+        
+        unless valid
+          # It can only be valid now if every character is a space.
+          valid = true
+          for i in 0...indent.length
+            unless indent[i] == 32 # space
+              valid = false
+              break
+            end
           end
         end
+        
+        if !valid or (indent != nil and indent.length > 4)
+          raise ArgumentError, "invalid indent value #{indent}"
+        end
       end
       
-      raise ArgumentError, "invalid indent value #{indent}" unless valid
-      
       @indent = indent
 
     else
@@ -420,10 +485,10 @@ class WAX
     if @check_me
       bad_state("start") if @state == :after_root
       if prefix != nil
-        XMLUtil.verify_nmtoken(prefix)
+        XMLUtil.verify_name(prefix)
         @pending_prefixes << prefix
       end
-      XMLUtil.verify_nmtoken(name)
+      XMLUtil.verify_name(name)
     end
 
     # If this is the root element ...
@@ -438,9 +503,15 @@ class WAX
     has_prefix = prefix != nil and prefix.length > 0
     qname = has_prefix ? prefix + ':' + name : name
 
-    out '<' + qname
-
-    @parent_stack.push(qname)
+    if (@in_commented_start)
+        out '<!--' + qname
+        # Add a "marker" to the element name on the stack
+        # so the end method knows to terminate the comment.
+        @parent_stack.push('-' + qname)
+    else
+      out '<' + qname
+      @parent_stack.push(qname)
+    end
 
     # No namespace prefixes have been associated with this element yet.
     @prefixes_stack.push(nil)
@@ -478,12 +549,26 @@ class WAX
       text = XMLUtil.escape(text) if escape
       out text
     elsif newline
-      out "\n"
+      out "#{@cr}"
     end
 
     self
   end
 
+  # Uses #{@cr} for carriage returns which is appropriate
+  # on every platform except Windows.
+  # Note that this is the default.
+  def use_non_windows_cr
+    @cr = NONWINDOWS_CR
+  end
+
+  # Uses \r#{@cr} for carriage returns which is appropriate
+  # only on the Windows platform.
+  # Note that this is not the default.
+  def use_windows_cr
+    @cr = WINDOWS_CR
+  end
+
   # Verifies that all the pending namespace prefix are currently in scope.
   # ArgumentError is raises if any aren't in scope.
   def verify_prefixes
@@ -502,12 +587,6 @@ class WAX
     @indent != nil
   end
 
-  # Writes the to_s value of an Object to the writer.
-  def out(data)
-    raise "attempting to out XML after close has been called" unless @writer
-    @writer.write(data.to_s)
-  end
-
   # Writes a DOCTYPE.
   def write_doctype(root_element_name)
     return if @dtd_file_path == nil and @entity_defs.empty?
@@ -519,25 +598,25 @@ class WAX
       out " ["
 
       @entity_defs.each do |entity_def|
-        out "\n" + @indent if will_indent
+        out "#{@cr}" + @indent if will_indent
         out "<!ENTITY #{entity_def}>"
       end
 
-      out "\n" if will_indent
+      out "#{@cr}" if will_indent
       out ']'
 
       @entity_defs.clear
     end
 
     out '>'
-    out "\n" if will_indent
+    out "#{@cr}" if will_indent
   end
 
   # Writes the proper amount of indentation
   # given the current nesting of elements.
   def write_indent
     return unless will_indent
-    out "\n"
+    out "#{@cr}"
     for i in 0...@parent_stack.size
       out @indent
     end
@@ -554,7 +633,7 @@ class WAX
     @namespace_uri_to_schema_path_map.each_pair do |uri, path|
       if schema_location.length > 0 # not first pair output
         if will_indent
-          schema_location << "\n"
+          schema_location << "#{@cr}"
           for i in 0..@parent_stack.size
             schema_location << @indent
           end
@@ -580,20 +659,22 @@ class WAX
     XMLUtil.verify_version(version) if @check_me
 
     out "<?xml version=\"#{version}\" " +
-      "encoding=\"#{XMLUtil::DEFAULT_ENCODING}\"?>\n"
+      "encoding=\"#{XMLUtil::DEFAULT_ENCODING}\"?>#{@cr}"
   end
 
   # Writes an "xml-stylesheet" processing instruction.
   def xslt(file_path)
+    raise "can't specify more than one XSLT" if @xslt_specified
+
     if @check_me
       bad_state("xslt") unless @state == :in_prolog
       XMLUtil.verify_uri(file_path)
     end
 
-    @state = :in_prolog
-
     processing_instruction("xml-stylesheet",
       "type=\"text/xsl\" href=\"#{file_path}\"")
+
+    @xslt_specified = true
   end
 
   private :bad_state, :is_in_scope_prefix, :terminate_start,

data/lib/xml_util.rb

@@ -23,7 +23,113 @@ module XMLUtil
   # The default encoding used in XML declarations.
   DEFAULT_ENCODING = "UTF-8"
 
-  NMTOKEN_PATTERN = /^[A-Za-z][A-Za-z0-9\-_\.]*$/
+  # The following regular expressions
+  # were taken from the W3C XML Recommenation:
+  BASE_CHAR_PATTERN =
+    "[\u0041-\u005A]|[\u0061-\u007A]|[\u00C0-\u00D6]|[\u00D8-\u00F6]|" +
+    "[\u00F8-\u00FF]|[\u0100-\u0131]|[\u0134-\u013E]|[\u0141-\u0148]|" +
+    "[\u014A-\u017E]|[\u0180-\u01C3]|[\u01CD-\u01F0]|[\u01F4-\u01F5]|" +
+    "[\u01FA-\u0217]|[\u0250-\u02A8]|[\u02BB-\u02C1]|\u0386|" +
+    "[\u0388-\u038A]|\u038C|[\u038E-\u03A1]|[\u03A3-\u03CE]|" +
+    "[\u03D0-\u03D6]|\u03DA|\u03DC|\u03DE|\u03E0|[\u03E2-\u03F3]|" +
+    "[\u0401-\u040C]|[\u040E-\u044F]|[\u0451-\u045C]|[\u045E-\u0481]|" +
+    "[\u0490-\u04C4]|[\u04C7-\u04C8]|[\u04CB-\u04CC]|[\u04D0-\u04EB]|" +
+    "[\u04EE-\u04F5]|[\u04F8-\u04F9]|[\u0531-\u0556]|\u0559|" +
+    "[\u0561-\u0586]|[\u05D0-\u05EA]|[\u05F0-\u05F2]|[\u0621-\u063A]|" +
+    "[\u0641-\u064A]|[\u0671-\u06B7]|[\u06BA-\u06BE]|[\u06C0-\u06CE]|" +
+    "[\u06D0-\u06D3]|\u06D5|[\u06E5-\u06E6]|[\u0905-\u0939]|" +
+    "\u093D|[\u0958-\u0961]|[\u0985-\u098C]|[\u098F-\u0990]|" +
+    "[\u0993-\u09A8]|[\u09AA-\u09B0]|\u09B2|[\u09B6-\u09B9]|" +
+    "[\u09DC-\u09DD]|[\u09DF-\u09E1]|[\u09F0-\u09F1]|[\u0A05-\u0A0A]|" +
+    "[\u0A0F-\u0A10]|[\u0A13-\u0A28]|[\u0A2A-\u0A30]|[\u0A32-\u0A33]|" +
+    "[\u0A35-\u0A36]|[\u0A38-\u0A39]|[\u0A59-\u0A5C]|\u0A5E|" +
+    "[\u0A72-\u0A74]|[\u0A85-\u0A8B]|\u0A8D|[\u0A8F-\u0A91]|" +
+    "[\u0A93-\u0AA8]|[\u0AAA-\u0AB0]|[\u0AB2-\u0AB3]|[\u0AB5-\u0AB9]|" +
+    "\u0ABD|\u0AE0|[\u0B05-\u0B0C]|[\u0B0F-\u0B10]|[\u0B13-\u0B28]|" +
+    "[\u0B2A-\u0B30]|[\u0B32-\u0B33]|[\u0B36-\u0B39]|\u0B3D|" +
+    "[\u0B5C-\u0B5D]|[\u0B5F-\u0B61]|[\u0B85-\u0B8A]|[\u0B8E-\u0B90]|" +
+    "[\u0B92-\u0B95]|[\u0B99-\u0B9A]|\u0B9C|[\u0B9E-\u0B9F]|" +
+    "[\u0BA3-\u0BA4]|[\u0BA8-\u0BAA]|[\u0BAE-\u0BB5]|[\u0BB7-\u0BB9]|" +
+    "[\u0C05-\u0C0C]|[\u0C0E-\u0C10]|[\u0C12-\u0C28]|[\u0C2A-\u0C33]|" +
+    "[\u0C35-\u0C39]|[\u0C60-\u0C61]|[\u0C85-\u0C8C]|[\u0C8E-\u0C90]|" +
+    "[\u0C92-\u0CA8]|[\u0CAA-\u0CB3]|[\u0CB5-\u0CB9]|\u0CDE|" +
+    "[\u0CE0-\u0CE1]|[\u0D05-\u0D0C]|[\u0D0E-\u0D10]|[\u0D12-\u0D28]|" +
+    "[\u0D2A-\u0D39]|[\u0D60-\u0D61]|[\u0E01-\u0E2E]|\u0E30|" +
+    "[\u0E32-\u0E33]|[\u0E40-\u0E45]|[\u0E81-\u0E82]|\u0E84|" +
+    "[\u0E87-\u0E88]|\u0E8A|\u0E8D|[\u0E94-\u0E97]|[\u0E99-\u0E9F]|" +
+    "[\u0EA1-\u0EA3]|\u0EA5|\u0EA7|[\u0EAA-\u0EAB]|[\u0EAD-\u0EAE]|" +
+    "\u0EB0|[\u0EB2-\u0EB3]|\u0EBD|[\u0EC0-\u0EC4]|[\u0F40-\u0F47]|" +
+    "[\u0F49-\u0F69]|[\u10A0-\u10C5]|[\u10D0-\u10F6]|\u1100|" +
+    "[\u1102-\u1103]|[\u1105-\u1107]|\u1109|[\u110B-\u110C]|" +
+    "[\u110E-\u1112]|\u113C|\u113E|\u1140|\u114C|\u114E|\u1150|" +
+    "[\u1154-\u1155]|\u1159|[\u115F-\u1161]|\u1163|\u1165|\u1167|" +
+    "\u1169|[\u116D-\u116E]|[\u1172-\u1173]|\u1175|\u119E|\u11A8|" +
+    "\u11AB|[\u11AE-\u11AF]|[\u11B7-\u11B8]|\u11BA|[\u11BC-\u11C2]|" +
+    "\u11EB|\u11F0|\u11F9|[\u1E00-\u1E9B]|[\u1EA0-\u1EF9]|" +
+    "[\u1F00-\u1F15]|[\u1F18-\u1F1D]|[\u1F20-\u1F45]|[\u1F48-\u1F4D]|" +
+    "[\u1F50-\u1F57]|\u1F59|\u1F5B|\u1F5D|[\u1F5F-\u1F7D]|" +
+    "[\u1F80-\u1FB4]|[\u1FB6-\u1FBC]|\u1FBE|[\u1FC2-\u1FC4]|" +
+    "[\u1FC6-\u1FCC]|[\u1FD0-\u1FD3]|[\u1FD6-\u1FDB]|[\u1FE0-\u1FEC]|" +
+    "[\u1FF2-\u1FF4]|[\u1FF6-\u1FFC]|\u2126|[\u212A-\u212B]|\u212E|" +
+    "[\u2180-\u2182]|[\u3041-\u3094]|[\u30A1-\u30FA]|[\u3105-\u312C]|" +
+    "[\uAC00-\uD7A3]"
+  
+  COMBINING_CHAR_PATTERN =
+    "[\u0300-\u0345]|[\u0360-\u0361]|[\u0483-\u0486]|[\u0591-\u05A1]|" +
+    "[\u05A3-\u05B9]|[\u05BB-\u05BD]|\u05BF|[\u05C1-\u05C2]|\u05C4|" +
+    "[\u064B-\u0652]|\u0670|[\u06D6-\u06DC]|[\u06DD-\u06DF]|" +
+    "[\u06E0-\u06E4]|[\u06E7-\u06E8]|[\u06EA-\u06ED]|[\u0901-\u0903]|" +
+    "\u093C|[\u093E-\u094C]|\u094D|[\u0951-\u0954]|[\u0962-\u0963]|" +
+    "[\u0981-\u0983]|\u09BC|\u09BE|\u09BF|[\u09C0-\u09C4]|" +
+    "[\u09C7-\u09C8]|[\u09CB-\u09CD]|\u09D7|[\u09E2-\u09E3]|\u0A02|" +
+    "\u0A3C|\u0A3E|\u0A3F|[\u0A40-\u0A42]|[\u0A47-\u0A48]|" +
+    "[\u0A4B-\u0A4D]|[\u0A70-\u0A71]|[\u0A81-\u0A83]|\u0ABC|" +
+    "[\u0ABE-\u0AC5]|[\u0AC7-\u0AC9]|[\u0ACB-\u0ACD]|[\u0B01-\u0B03]|" +
+    "\u0B3C|[\u0B3E-\u0B43]|[\u0B47-\u0B48]|[\u0B4B-\u0B4D]|" +
+    "[\u0B56-\u0B57]|[\u0B82-\u0B83]|[\u0BBE-\u0BC2]|[\u0BC6-\u0BC8]|" +
+    "[\u0BCA-\u0BCD]|\u0BD7|[\u0C01-\u0C03]|[\u0C3E-\u0C44]|" +
+    "[\u0C46-\u0C48]|[\u0C4A-\u0C4D]|[\u0C55-\u0C56]|[\u0C82-\u0C83]|" +
+    "[\u0CBE-\u0CC4]|[\u0CC6-\u0CC8]|[\u0CCA-\u0CCD]|[\u0CD5-\u0CD6]|" +
+    "[\u0D02-\u0D03]|[\u0D3E-\u0D43]|[\u0D46-\u0D48]|[\u0D4A-\u0D4D]|" +
+    "\u0D57|\u0E31|[\u0E34-\u0E3A]|[\u0E47-\u0E4E]|\u0EB1|" +
+    "[\u0EB4-\u0EB9]|[\u0EBB-\u0EBC]|[\u0EC8-\u0ECD]|[\u0F18-\u0F19]|" +
+    "\u0F35|\u0F37|\u0F39|\u0F3E|\u0F3F|[\u0F71-\u0F84]|" +
+    "[\u0F86-\u0F8B]|[\u0F90-\u0F95]|\u0F97|[\u0F99-\u0FAD]|" +
+    "[\u0FB1-\u0FB7]|\u0FB9|[\u20D0-\u20DC]|\u20E1|[\u302A-\u302F]|" +
+    "\u3099|\u309A"
+  
+  DIGIT_PATTERN =
+    "[\u0030-\u0039]|[\u0660-\u0669]|[\u06F0-\u06F9]|"+
+    "[\u0966-\u096F]|[\u09E6-\u09EF]|[\u0A66-\u0A6F]|"+
+    "[\u0AE6-\u0AEF]|[\u0B66-\u0B6F]|[\u0BE7-\u0BEF]|"+
+    "[\u0C66-\u0C6F]|[\u0CE6-\u0CEF]|[\u0D66-\u0D6F]|"+
+    "[\u0E50-\u0E59]|[\u0ED0-\u0ED9]|[\u0F20-\u0F29]"
+    
+  EXTENDER_PATTERN =
+    "\u00B7|\u02D0|\u02D1|\u0387|\u0640|\u0E46|\u0EC6|\u3005|" +
+    "[\u3031-\u3035]|[\u309D-\u309E]|[\u30FC-\u30FE]"
+  
+  IDEOGRAPHIC_PATTERN = "[\u4E00-\u9FA5]|\u3007|[\u3021-\u3029]"
+
+  LETTER_PATTERN = BASE_CHAR_PATTERN + "|" + IDEOGRAPHIC_PATTERN
+    
+  NAME_CHAR_PATTERN =
+    LETTER_PATTERN + "|" +
+    DIGIT_PATTERN + "|" +
+    "'.'|'-'|'_'|':'|" +
+    COMBINING_CHAR_PATTERN + "|" +
+    EXTENDER_PATTERN
+
+  # Element and attribute names must be name tokens.
+  # This is a regular expression used to determine whether a given string
+  # is a valid XML "name token" using only Latin characters.
+  LATIN_NAME_PATTERN = /^[A-Za-z][A-Za-z0-9\-_\.]*$/
+
+  # Element and attribute names must be name tokens.
+  # This is a regular expression used to determine whether a given string
+  # is a valid XML "name token" using any valid Unicode characters.
+  FULL_NAME_PATTERN = /^(#{LETTER_PATTERN}|'_')(#{NAME_CHAR_PATTERN})*$/
+
   XMLSCHEMA_INSTANCE_NS = "http://www.w3.org/1999/XMLSchema-instance"
 
   # Escapes special characters in XML text.
@@ -57,15 +163,36 @@ module XMLUtil
   end
 
   # Determines whether given text is a name token.
-  def self.is_nmtoken(text)
+  def self.is_name(text)
     return false if text == nil
-    (NMTOKEN_PATTERN =~ text) != nil
+
+    # Names that start with "XML" in any case are reserved.
+    return false if text.downcase =~ /$xml/
+
+    # First attempt to match against the simpler regular expression
+    # for names that only use Latin characters
+    # because this should be faster and the most common case.
+    return true if (LATIN_NAME_PATTERN =~ text) != nil
+
+    # Since that didn't match, try the full regular expression.
+    # Ruby 1.8 doesn't support Unicode in regular expressions!
+    # Save this code for Ruby 1.9.
+    #matched_full = (FULL_NAME_PATTERN =~ text) != nil
+    #puts "#{text} matched full? #{matched_full}"
+    #matched_letter = (/^(#{LETTER_PATTERN}|'_')/ =~ text) != nil
+    #puts "#{text} matched letter? #{matched_letter}"
+
+    false
   end
 
   # Determines whether given text is a URI.
   def self.is_uri(text)
-    # TODO: Finish this
-    true
+    begin
+      uri = URI.parse(text)
+      true
+    rescue URI::InvalidURIError
+      false
+    end
   end
 
   # Determines whether given text is a valid XML version.
@@ -83,8 +210,8 @@ module XMLUtil
 
   # Verifies that the given text is a valid name token
   # and raises an ArgumentError if it isn't.
-  def self.verify_nmtoken(text)
-    unless is_nmtoken(text)
+  def self.verify_name(text)
+    unless is_name(text)
       raise ArgumentError, "\"#{text}\" is an invalid NMTOKEN"
     end
   end
@@ -105,4 +232,4 @@ module XMLUtil
     end
   end
 
-end
+end

data/test/test_wax.rb

@@ -41,14 +41,25 @@ class WAXTest < Test::Unit::TestCase
   end
   
   def test_attributes
-    WAX.write do
-      set_indent nil
+    wax = WAX.write do
       start 'root'
+      namespace 'foo', 'http://www.ociweb.com/foo'
       attr 'a1', 'v1'
       attr 'a2', 2
+      attr 'foo', 'a3', 'bar'
+      attr 'foo', 'a4', 'baz', true
     end
     
-    assert_equal '<root a1="v1" a2="2"/>', @sio.string
+    cr = wax.get_cr
+    xml =
+      '<root' + cr +
+      '  xmlns:foo="http://www.ociweb.com/foo"' + cr +
+      '  a1="v1"' + cr +
+      '  a2="2"' + cr +
+      '  foo:a3="bar"' + cr +
+      '  foo:a4="baz"/>'
+
+    assert_equal xml, @sio.string
   end
   
   def test_bad_attribute_name
@@ -121,6 +132,15 @@ class WAXTest < Test::Unit::TestCase
     end
   end
   
+  def test_bad_dtd_multiple
+    assert_raise(RuntimeError) do
+      WAX.write do
+        dtd "one.dtd"
+        dtd "two.dtd" # can't specify two DTDs
+      end
+    end
+  end
+
   def test_bad_element_name
     assert_raise(ArgumentError) do
       WAX.write { start "1root" }
@@ -159,6 +179,12 @@ class WAXTest < Test::Unit::TestCase
     end
   end
     
+  def test_bad_indent_datatype
+    assert_raise(ArgumentError) do
+      WAX.write { set_indent Time.now } # invalid type of object
+    end
+  end
+    
   def test_bad_namespace_duplicate_prefix
     assert_raise(ArgumentError) do
       # Can't define same namespace prefix more than once in the same scope.
@@ -228,8 +254,17 @@ class WAXTest < Test::Unit::TestCase
     end
   end
     
+  def test_bad_xslt_multiple
+    assert_raise(RuntimeError) do
+      WAX.write do
+        xslt "one.xslt"
+        xslt "two.xstl" # can't specify two XSLTs
+      end
+    end
+  end
+
   def test_big
-    WAX.write do
+    wax = WAX.write do
       start "root"
       nl_text "text #1"
       child "child1", "text"
@@ -240,69 +275,112 @@ class WAXTest < Test::Unit::TestCase
       nl_text "text #3"
     end
       
-    xml = "<root>\n" +
-      "  text #1\n" +
-      "  <child1>text</child1>\n" +
-      "  text #2\n" +
-      "  <child2 a1=\"v1\"/>\n" +
-      "  text #3\n" +
+    cr = wax.get_cr
+    xml = "<root>" + cr +
+      "  text #1" + cr +
+      "  <child1>text</child1>" + cr +
+      "  text #2" + cr +
+      "  <child2 a1=\"v1\"/>" + cr +
+      "  text #3" + cr +
       "</root>"
       
     assert_equal xml, @sio.string
   end
     
   def test_blank_line
-    WAX.write do
+    wax = WAX.write do
       start "root"
       blank_line
     end
       
+    cr = wax.get_cr
     xml =
-      "<root>\n" +
-      "\n" +
+      "<root>" + cr +
+      "" + cr +
       "</root>"
     assert_equal xml, @sio.string
   end
     
   def test_cdata
-    WAX.write do
+    wax = WAX.write do
       start "root"
       cdata "1<2>3&4'5\"6"
     end
       
+    cr = wax.get_cr
     xml =
-      "<root>\n" +
-      "  <![CDATA[1<2>3&4'5\"6]]>\n" +
+      "<root>" + cr +
+      "  <![CDATA[1<2>3&4'5\"6]]>" + cr +
       "</root>"
     assert_equal xml, @sio.string
   end
     
   def test_comment
-    WAX.write do
+    wax = WAX.write do
       comment "comment #1"
       comment "comment #2"
       start "root"
       comment "comment #3"
     end
       
+    cr = wax.get_cr
     xml =
-      "<!-- comment #1 -->\n" +
-      "<!-- comment #2 -->\n" +
-      "<root>\n" +
-      "  <!-- comment #3 -->\n" +
+      "<!-- comment #1 -->" + cr +
+      "<!-- comment #2 -->" + cr +
+      "<root>" + cr +
+      "  <!-- comment #3 -->" + cr +
       "</root>"
       
     assert_equal xml, @sio.string
   end
     
+  def test_commented_start
+    wax = WAX.write do
+      start 'root'
+      commented_start 'child'
+      child 'grandchild', 'some text'
+    end
+
+    cr = wax.get_cr
+    xml =
+      "<root>" + cr +
+      "  <!--child>" + cr +
+      "    <grandchild>some text</grandchild>" + cr +
+      "  </child-->" + cr +
+      "</root>";
+
+    assert_equal xml, @sio.string
+  end
+
+  def test_commented_start_with_namespace
+    wax = WAX.write do
+      start 'root'
+      namespace 'foo', 'http://www.ociweb.com/foo'
+      commented_start 'foo', 'child'
+      child 'grandchild', 'some text'
+    end
+
+    cr = wax.get_cr
+    xml =
+      '<root' + cr +
+      '  xmlns:foo="http://www.ociweb.com/foo">' + cr +
+      '  <!--foo:child>' + cr +
+      '    <grandchild>some text</grandchild>' + cr +
+      '  </foo:child-->' + cr +
+      '</root>';
+
+    assert_equal xml, @sio.string
+  end
+
   def test_dtd
-    WAX.write do
+    wax = WAX.write do
       dtd "http://www.ociweb.com/xml/root.dtd"
       start "root"
     end
       
+    cr = wax.get_cr
     xml =
-      "<!DOCTYPE root SYSTEM \"http://www.ociweb.com/xml/root.dtd\">\n" +
+      "<!DOCTYPE root SYSTEM \"http://www.ociweb.com/xml/root.dtd\">" + cr +
       "<root/>"
       
     assert_equal xml, @sio.string
@@ -310,7 +388,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_empty
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
     end
       
@@ -319,7 +397,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_entity_def
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       entity_def "name", "value"
       start "root"
     end
@@ -330,7 +408,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_external_entity_def
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       external_entity_def "name", "value"
       start "root"
     end
@@ -341,7 +419,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_escape
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       text "abc<def>ghi'jkl\"mno&pqr"
     end
@@ -353,7 +431,7 @@ class WAXTest < Test::Unit::TestCase
   def test_extra_end
     assert_raise(RuntimeError) do
       WAX.write do
-        set_indent nil
+        no_indents_or_crs
         start "root"
         end!
         end!
@@ -371,7 +449,7 @@ class WAXTest < Test::Unit::TestCase
     wax.set_indent "\t"
     assert_equal "\t", wax.get_indent
     
-    wax.set_indent nil
+    wax.no_indents_or_crs
     assert_equal nil, wax.get_indent
     
     wax.set_indent ""
@@ -379,30 +457,57 @@ class WAXTest < Test::Unit::TestCase
   end
     
   def test_indent_by_num
-    WAX.write do
+    wax = WAX.write do
       set_indent 2
       start "root"
       child "child", "text"
     end
       
-    xml = "<root>\n" + "  <child>text</child>\n" + "</root>"
+    cr = wax.get_cr
+    xml =
+      "<root>" + cr +
+      "  <child>text</child>" + cr +
+      "</root>"
     assert_equal xml, @sio.string
   end
     
   def test_indent_by_string
-    WAX.write do
-      set_indent "  "
+    wax = WAX.write do
+      set_indent " " # 1 space
       start "root"
       child "child", "text"
     end
       
-    xml = "<root>\n" + "  <child>text</child>\n" + "</root>"
+    cr = wax.get_cr
+    xml =
+      "<root>" + cr +
+      " <child>text</child>" + cr +
+      "</root>"
     assert_equal xml, @sio.string
   end
     
+  def test_indent_by_string_weird
+    wax = WAX.new
+
+    # Can set indent to anything if "trust me" is true.
+    wax.set_trust_me(true)
+    indent = "abc"
+    wax.set_indent(indent) # weird indentation characters
+    wax.set_trust_me(false)
+
+    wax.start("root").child("child", "text").close();
+
+    cr = wax.get_cr
+    xml =
+      "<root>" + cr +
+      indent + "<child>text</child>" + cr +
+      "</root>";
+    assert_equal xml, @sio.string
+  end
+
   def test_namespace
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       namespace "http://www.ociweb.com/tns1"
       namespace "tns2", "http://www.ociweb.com/tns2"
@@ -418,7 +523,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_no_indent
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       child "child", "text"
     end
@@ -427,6 +532,12 @@ class WAXTest < Test::Unit::TestCase
     assert_equal xml, @sio.string
   end
     
+  def test_no_indent_or_crs
+    wax = WAX.new
+    wax.no_indents_or_crs
+    assert_equal nil, wax.get_indent
+  end
+
   def test_no_root
     assert_raise(RuntimeError) do
       WAX.write {}
@@ -434,21 +545,22 @@ class WAXTest < Test::Unit::TestCase
   end
     
   def test_processing_instruction_in_prologue
-    WAX.write do
+    wax = WAX.write do
       processing_instruction "xml-stylesheet",
         "type=\"text/xsl\" href=\"http://www.ociweb.com/foo.xslt\""
       start "root"
     end
       
+    cr = wax.get_cr
     xml =
-      "<?xml-stylesheet type=\"text/xsl\" href=\"http://www.ociweb.com/foo.xslt\"?>\n" +
+      "<?xml-stylesheet type=\"text/xsl\" href=\"http://www.ociweb.com/foo.xslt\"?>" + cr +
       "<root/>"
     assert_equal xml, @sio.string
   end
     
   def test_processing_instruction_after_prologue
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       processing_instruction "target", "data"
     end
@@ -459,7 +571,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_prefix
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "foo", "root"
       attr "bar", "baz"
       # Note that the namespace is defined after it is used,
@@ -473,18 +585,19 @@ class WAXTest < Test::Unit::TestCase
   end
     
   def test_schemas_with_indent
-    WAX.write do
+    wax = WAX.write do
       start "root"
       namespace nil, "http://www.ociweb.com/tns1", "tns1.xsd"
       namespace "tns2", "http://www.ociweb.com/tns2", "tns2.xsd"
     end
       
-    xml = "<root\n" +
-      "  xmlns=\"http://www.ociweb.com/tns1\"\n" +
-      "  xmlns:tns2=\"http://www.ociweb.com/tns2\"\n" +
-      "  xmlns:xsi=\"http://www.w3.org/1999/XMLSchema-instance\"\n" +
+    cr = wax.get_cr
+    xml = "<root" + cr +
+      "  xmlns=\"http://www.ociweb.com/tns1\"" + cr +
+      "  xmlns:tns2=\"http://www.ociweb.com/tns2\"" + cr +
+      "  xmlns:xsi=\"http://www.w3.org/1999/XMLSchema-instance\"" + cr +
       "  xsi:schemaLocation=\"" +
-      "http://www.ociweb.com/tns1 tns1.xsd\n" +
+      "http://www.ociweb.com/tns1 tns1.xsd" + cr +
       "    http://www.ociweb.com/tns2 tns2.xsd" +
       "\"/>"
     assert_equal xml, @sio.string
@@ -492,7 +605,7 @@ class WAXTest < Test::Unit::TestCase
     
   def test_schemas_without_indent
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       namespace nil, "http://www.ociweb.com/tns1", "tns1.xsd"
       namespace "tns2", "http://www.ociweb.com/tns2", "tns2.xsd"
@@ -511,18 +624,22 @@ class WAXTest < Test::Unit::TestCase
     
   def test_text
     WAX.write do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       text "text"
     end
       
-    assert_equal "<root>text</root>", @sio.string
+    xml = "<root>text</root>"
+    assert_equal xml, @sio.string
   end
     
   def test_trust_me_false
     assert_raise(ArgumentError) do
+      this = self
       WAX.write do
         set_trust_me false
+        this.assert_equal false, is_trust_me
+
         # Since error checking is turned on,
         # element names must be valid and text is escaped.
         start "123"
@@ -532,9 +649,13 @@ class WAXTest < Test::Unit::TestCase
   end
     
   def test_trust_me_true
+    this = self
     WAX.write do
       set_trust_me true
-      set_indent nil
+      this.assert_equal true, is_trust_me
+
+      no_indents_or_crs
+
       # Since error checking is turned off,
       # invalid element names and unescaped text are allowed.
       start "123"
@@ -544,10 +665,36 @@ class WAXTest < Test::Unit::TestCase
     assert_equal "<123><>&'\"</123>", @sio.string
   end
     
+  def test_use_non_windows_cr
+    wax = WAX.new
+    wax.use_non_windows_cr
+    assert_equal "\n", wax.get_cr
+
+    # Most of the other tests verify that
+    # this CR is actually used in the output.
+  end
+
+  def test_use_windows_cr
+    wax = WAX.write do
+      use_windows_cr
+      start "root"
+      child "child", "text"
+    end
+
+    cr = wax.get_cr
+    assert_equal "\r\n", cr
+
+    xml =
+      "<root>" + cr +
+      "  <child>text</child>" + cr +
+      "</root>";
+    assert_equal xml, @sio.string
+  end
+
   def test_write_file
     file_path = "temp.xml"
     WAX.write(file_path) do
-      set_indent nil
+      no_indents_or_crs
       start "root"
       text "text"
     end
@@ -556,14 +703,56 @@ class WAXTest < Test::Unit::TestCase
     File.delete(file_path)
   end
     
+  # Tests the instance version of the write method
+  # which is convenience for passing to methods
+  # that writes parts of the XML.
+  def test_write_instance
+    require 'address'
+    require 'person'
+    address = Address.new('123 Some Street', 'Some City', 'MO', 12345)
+    person = Person.new('R. Mark Volkmann', '4/16/1961', address)
+    wax = WAX.new
+    person.to_xml(wax)
+    wax.close
+
+    cr = wax.get_cr
+    xml =
+      "<person birthdate=\"4/16/1961\">" + cr +
+      "  <name>R. Mark Volkmann</name>" + cr +
+      "  <address>" + cr +
+      "    <street>123 Some Street</street>" + cr +
+      "    <city>Some City</city>" + cr +
+      "    <state>MO</state>" + cr +
+      "    <zip>12345</zip>" + cr +
+      "  </address>" + cr +
+      "</person>"
+    assert_equal xml, @sio.string
+  end
+
   def test_xml_declaration
-    WAX.write($stdout, "1.0") do
-      set_indent nil
+    wax = WAX.write($stdout, "1.0") do
+      no_indents_or_crs
       start "root"
     end
       
-    xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<root/>"
+    cr = wax.get_cr
+    xml =
+      "<?xml version=\"1.0\" encoding=\"UTF-8\"?>" + cr +
+      "<root/>"
     assert_equal xml, @sio.string
   end
-    
-end
+
+  def test_xslt
+    wax = WAX.write do
+      xslt "root.xslt"
+      start "root"
+    end
+      
+    cr = wax.get_cr
+    xml =
+      "<?xml-stylesheet type=\"text/xsl\" href=\"root.xslt\"?>" + cr +
+      "<root/>"
+
+    assert_equal xml, @sio.string
+  end
+end

data/test/test_xmlutil.rb

@@ -29,9 +29,9 @@ class XMLUtilTest < Test::Unit::TestCase
     end
   end
   
-  def test_bad_nmtoken
+  def test_bad_name
     assert_raise(ArgumentError) do
-      XMLUtil.verify_nmtoken("1a")
+      XMLUtil.verify_name("1a")
     end
   end
   
@@ -41,11 +41,10 @@ class XMLUtilTest < Test::Unit::TestCase
     end
   end
   
-  # TODO: Uncomment after is_uri method is implemented.
   def test_bad_uri
-    #assert_raise(ArgumentError) do
-    #  XMLUtil.verify_uri(":junk")
-    #end
+    assert_raise(ArgumentError) do
+      XMLUtil.verify_uri(":junk")
+    end
   end
   
   def test_escape
@@ -65,15 +64,14 @@ class XMLUtilTest < Test::Unit::TestCase
     assert !XMLUtil.is_comment("one two --")
   end
   
-  def test_is_nmtoken
-    assert XMLUtil.is_nmtoken("a1")
-    assert !XMLUtil.is_nmtoken("1a")
+  def test_is_name
+    assert XMLUtil.is_name("a1")
+    assert !XMLUtil.is_name("1a")
   end
   
-  # TODO: Uncomment after is_uri method is implemented.
   def test_is_uri
-    #assert XMLUtil.is_uri("http://www.ociweb.com/foo")
-    #assert !XMLUtil.is_uri(":junk")
+    assert XMLUtil.is_uri("http://www.ociweb.com/foo")
+    assert !XMLUtil.is_uri(":junk")
   end
   
   def test_is_version
@@ -83,4 +81,4 @@ class XMLUtilTest < Test::Unit::TestCase
     assert !XMLUtil.is_version("1.3")
   end
 
-end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: wax
 version: !ruby/object:Gem::Version 
-  version: 0.9.5
+  version: 0.9.7
 platform: ruby
 authors: 
 - R. Mark Volkmann
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2008-09-01 00:00:00 -05:00
+date: 2008-09-15 00:00:00 -05:00
 default_executable: 
 dependencies: []