Class: HTree::Text

Inherits:
Object
  • Object
show all
Includes:
Leaf, Trav
Defined in:
lib/htree/output.rb,
lib/htree/loc.rb,
lib/htree/text.rb,
lib/htree/rexml.rb,
lib/htree/modules.rb,
lib/htree/modules.rb,
lib/htree/modules.rb,
lib/htree/equality.rb,
lib/htree/raw_string.rb,
lib/htree/raw_string.rb,
lib/htree/extract_text.rb

Overview

:stopdoc:

Defined Under Namespace

Modules: Trav Classes: Loc

Constant Summary collapse

ChRef = 
{
  '>' => '>',
  '<' => '&lt;',
  '"' => '&quot;',
}

Constants included from HTree

DefaultContext, ElementContent, ElementExclusions, ElementInclusions, EmptyBindingObject, HTMLContext, NamedCharacters, NamedCharactersPattern, OmittedAttrName

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Trav

#traverse_text_internal

Methods included from Leaf::Trav

#traverse_all_element, #traverse_some_element, #traverse_text_internal

Methods included from Traverse

#bogusetag?, #comment?, #doc?, #doctype?, #elem?, #get_subnode, #procins?, #text?, #traverse_text, #xmldecl?

Methods included from Leaf

#init_raw_string, #pretty_print, #raw_string_internal

Methods included from Node

#display_html, #display_xml, #generate_xml_output_code, #make_loc, #raw_string, #subst, #subst_internal, #to_node, #to_rexml

Methods included from HTree

#==, build_node, #check_equality, compile_template, #exact_equal?, #exact_equal_object, expand_template, fix_element, fix_structure_list, frozen_string, #hash, parse, parse_as, parse_pairs, parse_xml, scan, #usual_equal_object, with_frozen_string_hash

Constructor Details

#initialize(rcdata, normalized_rcdata = internal_normalize(rcdata)) ⇒ Text

:notnew:



29
30
31
32
33
 
# File 'lib/htree/text.rb', line 29

def initialize(rcdata, normalized_rcdata=internal_normalize(rcdata)) # :notnew:
  init_raw_string
  @rcdata = rcdata && HTree.frozen_string(rcdata)
  @normalized_rcdata = @rcdata == normalized_rcdata ? @rcdata : normalized_rcdata
end

Instance Attribute Details

#normalized_rcdata ⇒ Object (readonly)

Returns the value of attribute normalized_rcdata.



34
35
36
 
# File 'lib/htree/text.rb', line 34

def normalized_rcdata
  @normalized_rcdata
end

#rcdata ⇒ Object (readonly)

Returns the value of attribute rcdata.



34
35
36
 
# File 'lib/htree/text.rb', line 34

def rcdata
  @rcdata
end

Class Method Details

.concat(*args) ⇒ Object

HTree::Text.concat returns a text which is concatenation of arguments.

An argument should be one of follows.

  • String

  • HTree::Text

  • HTree::Location which points HTree::Text



102
103
104
105
106
107
108
109
110
111
112
113
 
# File 'lib/htree/text.rb', line 102

def Text.concat(*args)
  rcdata = ''
  args.each {|arg|
    arg = arg.to_node if HTree::Location === arg
    if Text === arg
      rcdata << arg.rcdata
    else
      rcdata << arg.gsub(/&/, '&amp;')
    end
  }
  new_internal rcdata
end

.new(arg) ⇒ Object

:startdoc:



16
17
18
19
20
21
22
23
24
25
26
27
 
# File 'lib/htree/text.rb', line 16

def Text.new(arg)
  arg = arg.to_node if HTree::Location === arg
  if Text === arg
    new_internal arg.rcdata, arg.normalized_rcdata
  elsif String === arg
    arg2 = arg.gsub(/&/, '&amp;')
    arg = arg2.freeze if arg != arg2
    new_internal arg
  else
    raise TypeError, "cannot initialize Text with #{arg.inspect}"
  end
end

.new_internal ⇒ Object 



12
 
# File 'lib/htree/text.rb', line 12

alias new_internal new

.parse_cdata_content(raw_string) ⇒ Object 



329
330
331
332
333
 
# File 'lib/htree/parse.rb', line 329

def Text.parse_cdata_content(raw_string)
  result = Text.new(raw_string)
  result.raw_string = raw_string
  result
end

.parse_cdata_section(raw_string) ⇒ Object 



335
336
337
338
339
340
341
342
343
344
345
 
# File 'lib/htree/parse.rb', line 335

def Text.parse_cdata_section(raw_string)
  unless /\A#{Pat::CDATA_C}\z/o =~ raw_string
    raise HTree::Error, "cannot recognize as CDATA section: #{raw_string.inspect}"
  end

  content = $1

  result = Text.new(content)
  result.raw_string = raw_string
  result
end

.parse_pcdata(raw_string) ⇒ Object 



305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
 
# File 'lib/htree/parse.rb', line 305

def Text.parse_pcdata(raw_string)
  fixed = raw_string.gsub(/&(?:(?:#[0-9]+|#x[0-9a-fA-F]+|([A-Za-z][A-Za-z0-9]*));?)?/o) {|s|
    name = $1
    case s
    when /;\z/
      s
    when /\A&#/
      "#{s};"
    when '&'
      '&amp;'
    else 
      if NamedCharactersPattern =~ name
        "&#{name};"
      else
        "&amp;#{name}"
      end
    end
  }
  fixed = raw_string if fixed == raw_string
  result = Text.new_internal(fixed)
  result.raw_string = raw_string
  result
end

Instance Method Details

#eliminate_raw_string ⇒ Object 



86
87
88
 
# File 'lib/htree/raw_string.rb', line 86

def eliminate_raw_string
  Text.new_internal(@rcdata)
end

#empty? ⇒ Boolean

Returns:

  • (Boolean) 


80
81
82
 
# File 'lib/htree/text.rb', line 80

def empty?
  @normalized_rcdata.empty?
end

#extract_text ⇒ Object 



32
33
34
 
# File 'lib/htree/extract_text.rb', line 32

def extract_text
  self
end

#make_exact_equal_object ⇒ Object 



155
156
157
 
# File 'lib/htree/equality.rb', line 155

def make_exact_equal_object
  [@raw_string, @rcdata]
end

#make_usual_equal_object ⇒ Object 



159
160
161
 
# File 'lib/htree/equality.rb', line 159

def make_usual_equal_object
  @normalized_rcdata
end

#node_test_string ⇒ Object 



92
 
# File 'lib/htree/loc.rb', line 92

def node_test_string() 'text()' end

#output(out, context = nil) ⇒ Object 



17
18
19
 
# File 'lib/htree/output.rb', line 17

def output(out, context=nil)
  out.output_text @rcdata.gsub(/[<>]/) {|s| ChRef[s] }
end

#output_attvalue(out, context) ⇒ Object 



25
26
27
28
29
 
# File 'lib/htree/output.rb', line 25

def output_attvalue(out, context)
  out.output_string '"'
  out.output_text to_attvalue_content
  out.output_string '"'
end

#output_cdata(out) ⇒ Object 



31
32
33
34
35
36
37
 
# File 'lib/htree/output.rb', line 31

def output_cdata(out)
  str = self.to_s
  if %r{</} =~ str
    raise ArgumentError, "CDATA cannot contain '</': #{str.inspect}"
  end
  out.output_string(str)
end

#raw_string=(arg) ⇒ Object 



52
53
54
55
56
57
58
 
# File 'lib/htree/raw_string.rb', line 52

def raw_string=(arg)
  if arg == @rcdata then
    @raw_string = @rcdata
  else
    super
  end
end

#strip ⇒ Object 



84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/htree/text.rb', line 84

def strip
  rcdata = @normalized_rcdata.dup
  rcdata.sub!(/\A(?:\s|&nbsp;)+/, '')
  rcdata.sub!(/(?:\s|&nbsp;)+\z/, '')
  if rcdata == @normalized_rcdata
    self
  else
    rcdata.freeze
    Text.new_internal(rcdata, rcdata)
  end
end

#to_attvalue_content ⇒ Object 



21
22
23
 
# File 'lib/htree/output.rb', line 21

def to_attvalue_content
  @rcdata.gsub(/[<>"]/) {|s| ChRef[s] }
end

#to_rexml_internal(parent, context) ⇒ Object 



90
91
92
93
 
# File 'lib/htree/rexml.rb', line 90

def to_rexml_internal(parent, context)
  rcdata = self.rcdata.gsub(/[<>]/) { Encoder::ChRef[$&] }
  REXML::Text.new(rcdata, true, parent, true)
end

#to_s ⇒ Object

HTree::Text#to_s converts the text to a string.

  • character references are decoded as much as possible.

  • undecodable character reference are converted to ‘?’ character.



69
70
71
72
73
74
75
76
77
78
 
# File 'lib/htree/text.rb', line 69

def to_s
  @normalized_rcdata.gsub(/&(?:#([0-9]+));/o) {|s|
    u = $1.to_i
    if 0 <= u && u <= 0x7f
      [u].pack("C")
    else
      '?'
    end
  }
end