Class: MaRuKu::In::Markdown::SpanLevelParser::HTMLHelper

Inherits:
Object
  • Object
show all
Includes:
Strings
Defined in:
lib/maruku/input/html_helper.rb

Overview

I tried to do this with REXML, but wasn’t able to. (suggestions?)

Constant Summary collapse

Tag = 
%r{^<(/)?(\w+)\s*([^>]*)>}m
PartialTag = 
%r{^<.*}m
EverythingElse = 
%r{^[^<]+}m
CommentStart = 
%r{^<!--}x
CommentEnd = 
%r{^.*-->}
TO_SANITIZE = 
['img','hr','br']

Constants included from Strings

Strings::Abbreviation, Strings::AttributeDefinitionList, Strings::Definition, Strings::EMailAddress, Strings::FootnoteText, Strings::HeaderWithAttributes, Strings::HeaderWithId, Strings::IncompleteLink, Strings::InlineAttributeList, Strings::LinkRegex, Strings::MightBeTableHeader, Strings::Sep, Strings::TabSize, Strings::TableSeparator

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Strings

#add_tabs, #dbg_describe_ary, #force_linebreak?, #line_md_type, #normalize_key_and_value, #num_leading_hashes, #number_of_leading_spaces, #parse_email_headers, #sanitize_ref_id, #spaces_before_first_char, #split_lines, #strip_hashes, #strip_indent, #unquote

Constructor Details

#initialize ⇒ HTMLHelper

Returns a new instance of HTMLHelper.



45
46
47
48
49
50
51
 
# File 'lib/maruku/input/html_helper.rb', line 45

def initialize 
  @rest = ""
  @tag_stack = []
  @m = nil
  @already = ""
  self.state = :inside_element
end

Instance Attribute Details

#rest ⇒ Object (readonly)

Returns the value of attribute rest.



39
40
41
 
# File 'lib/maruku/input/html_helper.rb', line 39

def rest
  @rest
end

#state ⇒ Object

:inside_element, :inside_tag, :inside_comment,



53
54
55
 
# File 'lib/maruku/input/html_helper.rb', line 53

def state
  @state
end

Instance Method Details

#eat_this(line) ⇒ Object 



55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
 
# File 'lib/maruku/input/html_helper.rb', line 55

def eat_this(line)
  @rest = line  + @rest
  things_read = 0
  until @rest.empty?
    case self.state
      when :inside_comment
        if @m = CommentEnd.match(@rest)
          @already += @m.pre_match + @m.to_s
          @rest = @m.post_match
          self.state = :inside_element
        else 
          @already += @rest 
          @rest = ""
          self.state = :inside_comment
        end
      when :inside_element 
        if @m = CommentStart.match(@rest)
          things_read += 1
          @already += @m.pre_match + @m.to_s
          @rest = @m.post_match
          self.state = :inside_comment
        elsif @m = Tag.match(@rest) then
          my_debug "#{@state}: Tag: #{@m.to_s.inspect}"
          things_read += 1
          handle_tag
          self.state = :inside_element
        elsif @m = PartialTag.match(@rest) then
          my_debug "#{@state}: PartialTag: #{@m.to_s.inspect}"
          @already += @m.pre_match 
          @rest = @m.post_match
          @partial_tag = @m.to_s
          self.state = :inside_tag
        elsif @m = EverythingElse.match(@rest)
          my_debug "#{@state}: Everything: #{@m.to_s.inspect}"
          @already += @m.pre_match + @m.to_s
          @rest = @m.post_match
          self.state = :inside_element
        else
          error "Malformed HTML: not complete: #{@rest.inspect}"
        end
      when :inside_tag
        if @m = /^[^>]*>/.match(@rest) then
          my_debug "#{@state}: inside_tag: matched #{@m.to_s.inspect}"
          @partial_tag += @m.to_s
          my_debug "#{@state}: inside_tag: matched TOTAL: #{@partial_tag.to_s.inspect}"
          @rest = @partial_tag + @m.post_match
          @partial_tag = nil
          self.state = :inside_element
        else
          @partial_tag += @rest
          @rest = ""
          self.state = :inside_tag
        end
      else
        raise "Bug bug: state = #{self.state.inspect}"
    end # not inside comment
    
#       puts inspect
#       puts "Read: #{@tag_stack.inspect}"
    break if is_finished? and things_read>0 
  end
end

#error(s) ⇒ Object

Raises:



163
164
165
 
# File 'lib/maruku/input/html_helper.rb', line 163

def error(s)
  raise Exception, "Error: #{s} \n"+ inspect, caller
end

#handle_tag ⇒ Object 



118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
 
# File 'lib/maruku/input/html_helper.rb', line 118

def handle_tag()
  @already += @m.pre_match
  @rest = @m.post_match

  is_closing = !!@m[1]
  tag = @m[2]
  attributes = @m[3].to_s

  is_single = false
  if attributes[-1] == ?/ # =~ /\A(.*)\/\Z/
    attributes = attributes[0, attributes.size-1]
    is_single = true
  end

  my_debug "Attributes: #{attributes.inspect}"
  my_debug "READ TAG #{@m.to_s.inspect} tag = #{tag} closing? #{is_closing} single = #{is_single}"
  
  if TO_SANITIZE.include? tag 
    attributes.strip!
#   puts "Attributes: #{attributes.inspect}"
    if attributes.size > 0
      @already +=  '<%s %s />' % [tag, attributes]
    else
      @already +=  '<%s />' % [tag]
    end
  elsif is_closing
    @already += @m.to_s
    if @tag_stack.empty?
      error "Malformed: closing tag #{tag.inspect} "+
            "in empty list"
    end 
    if @tag_stack.last != tag
      error "Malformed: tag <#{tag}> "+
            "closes <#{@tag_stack.last}>"
    end
    @tag_stack.pop
  else 
    @already += @m.to_s
    
    if not is_single
      @tag_stack.push(tag) 
      my_debug "Pushing #{tag.inspect} when read #{@m.to_s.inspect}"
    end
  end
end

#inspect ⇒ Object 



167
168
169
170
171
172
173
174
175
 
# File 'lib/maruku/input/html_helper.rb', line 167

def inspect; "HTML READER\n state=#{self.state} "+
  "match=#{@m.to_s.inspect}\n"+
  "Tag stack = #{@tag_stack.inspect} \n"+
  "Before:\n"+
  add_tabs(@already,1,'|')+"\n"+
  "After:\n"+
  add_tabs(@rest,1,'|')+"\n"
  
end

#is_finished? ⇒ Boolean

Returns:

  • (Boolean) 


184
185
186
 
# File 'lib/maruku/input/html_helper.rb', line 184

def is_finished?
  (self.state == :inside_element)  and @tag_stack.empty?
end

#my_debug(s) ⇒ Object 



41
42
43
 
# File 'lib/maruku/input/html_helper.rb', line 41

def my_debug(s)
#     puts "---"*10+"\n"+inspect+"\t>>>\t"s
end

#stuff_you_read ⇒ Object 



178
179
180
 
# File 'lib/maruku/input/html_helper.rb', line 178

def stuff_you_read
  @already
end