Class: HTML::SGMLParser

Inherits:
Object
  • Object
show all
Defined in:
lib/html/sgml-parser.rb

Direct Known Subclasses

StackingParser

Constant Summary collapse

Interesting =

Regular expressions used for parsing:

/[&<]/
Incomplete = 
Regexp.compile('&([a-zA-Z][a-zA-Z0-9]*|#[0-9]*)?|' +
'<([a-zA-Z][^<>]*|/([a-zA-Z][^<>]*)?|' +
'![^<>]*)?')
Entityref = 
/&([a-zA-Z][-.a-zA-Z0-9]*)[^-.a-zA-Z0-9]/
Charref = 
/&#([0-9]+)[^0-9]/
Starttagopen = 
/<[>a-zA-Z]/
Endtagopen = 
/<\/[<>a-zA-Z]/
Endbracket = 
/[<>]/
Special = 
/<![^<>]*>/
Commentopen = 
/<!--/
Commentclose = 
/--[ \t\n]*>/
Tagfind = 
/[a-zA-Z][a-zA-Z0-9.-]*/
Attrfind = 
Regexp.compile('[\s,]*([a-zA-Z_][a-zA-Z_0-9.-]*)' +
'(\s*=\s*' +
"('[^']*'" +
'|"[^"]*"' +
'|[-~a-zA-Z0-9,./:+*%?!()_#=]*))?')
Endtagfind = 
/\s*\/\s*>/
Entitydefs = 
{'lt'=>'<', 'gt'=>'>', 'amp'=>'&', 'quot'=>'"', 'apos'=>'\''}

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(verbose = false) ⇒ SGMLParser

Returns a new instance of SGMLParser.



45
46
47
48
 
# File 'lib/html/sgml-parser.rb', line 45

def initialize(verbose=false)
  @verbose = verbose
  reset
end

Instance Attribute Details

#src_range ⇒ Object (readonly)

Returns the value of attribute src_range.



18
19
20
 
# File 'lib/html/sgml-parser.rb', line 18

def src_range
  @src_range
end

Instance Method Details

#close ⇒ Object 



91
92
93
 
# File 'lib/html/sgml-parser.rb', line 91

def close
  goahead(true)
end

#feed(data) ⇒ Object 



86
87
88
89
 
# File 'lib/html/sgml-parser.rb', line 86

def feed(data)
  @rawdata << data
  goahead(false)
end

#finish_endtag(tag) ⇒ Object 



279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
 
# File 'lib/html/sgml-parser.rb', line 279

def finish_endtag(tag)
  if tag == ''
    found = @stack.length - 1
    if found < 0
      unknown_endtag(tag)
      return
    end
  else
    unless @stack.include? tag
      method = 'end_' + tag
      unless self.respond_to?(method)
        unknown_endtag(tag)
      end
      return
    end
    found = @stack.index(tag) #or @stack.length
  end
  while @stack.length > found
    tag = @stack[-1]
    method = 'end_' + tag
    if respond_to?(method)
      handle_endtag(tag, method)
    else
      unknown_endtag(tag)
    end
    @stack.pop
  end
end

#finish_starttag(tag, attrs) ⇒ Object 



261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
 
# File 'lib/html/sgml-parser.rb', line 261

def finish_starttag(tag, attrs)
  method = 'start_' + tag
  if self.respond_to?(method)
    @stack << tag
    handle_starttag(tag, method, attrs)
    return 1
  else
    method = 'do_' + tag
    if self.respond_to?(method)
      handle_starttag(tag, method, attrs)
      return 0
    else
      unknown_starttag(tag, attrs)
      return -1
    end
  end
end

#get_source(range) ⇒ Object 



60
61
62
63
64
65
66
 
# File 'lib/html/sgml-parser.rb', line 60

def get_source(range)
  start = range.first
  end_index = range.end
  exclusive = range.exclude_end?
  offset_range = Range.new(start-@offset, end_index-@offset, exclusive)
  return @rawdata[offset_range]
end

#goahead(_end) ⇒ Object 



103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
 
# File 'lib/html/sgml-parser.rb', line 103

def goahead(_end)
  rawdata = @rawdata
  i = 0
  n = rawdata.length
  while i < n
    if @nomoretags
      i = handle_data_range(rawdata, i, n)
      break
    end
    j = rawdata.index(Interesting, i)
    j = n unless j
    i = handle_data_range(rawdata, i, j)
    break if (i == n)
    if rawdata[i] == ?< #
      if rawdata.index(Starttagopen, i) == i
        if @literal
          i = handle_data_range(rawdata, i, i+1)
          next
        end
        k = parse_starttag(i)
        break unless k
        i = k
        next
      end
      if rawdata.index(Endtagopen, i) == i
        k = parse_endtag(i)
        break unless k
        i = k
        @literal = false
        next
      end
      if rawdata.index(Commentopen, i) == i
        if @literal
          i = handle_data_range(rawdata, i, i+1)
          next
        end
        k = parse_comment(i)
        break unless k
        i += k
        next
      end
      if rawdata.index(Special, i) == i
        if @literal
          i = handle_data_range(rawdata, i, i+1)
          next
        end
        k = parse_special(i)
        break unless k
        i += k
        next
      end
    elsif rawdata[i] == ?& #
      if rawdata.index(Charref, i) == i
        end_index = i + $&.length
        end_index -= 1 unless rawdata[end_index-1] == ?;
        set_range(i, end_index)
        handle_charref($1)
        i = end_index
        next
      end
      if rawdata.index(Entityref, i) == i
        end_index = i + $&.length
        end_index -= 1 unless rawdata[end_index-1] == ?;
        set_range(i, end_index)
        handle_entityref($1)
        i = end_index
        next
      end
    else
      raise RuntimeError, 'neither < nor & ??'
    end
    # We get here only if incomplete matches but
    # nothing else
    match = rawdata.index(Incomplete, i)
    unless match == i
      i = handle_data_range(rawdata, i, i+1)
      next
    end
    j = match + $&.length
    break if j == n # Really incomplete
    i = handle_data_range(rawdata, i, j)
  end
  # end while
  if _end and i < n
    i = handle_data_range(rawdata, i, n)
  end
  @rawdata = rawdata[i..-1]
  @offset += i
end

#handle_charref(name) ⇒ Object 



334
335
336
337
338
339
340
341
 
# File 'lib/html/sgml-parser.rb', line 334

def handle_charref(name)
  n = Integer(name)
  if !(0 <= n && n <= 255)
    unknown_charref(name)
    return
  end
  handle_data(n.chr)
end

#handle_comment(data) ⇒ Object 



356
357
 
# File 'lib/html/sgml-parser.rb', line 356

def handle_comment(data)
end

#handle_data(data) ⇒ Object 



353
354
 
# File 'lib/html/sgml-parser.rb', line 353

def handle_data(data)
end

#handle_data_range(rawdata, start, end_index) ⇒ Object 



95
96
97
98
99
100
101
 
# File 'lib/html/sgml-parser.rb', line 95

def handle_data_range(rawdata, start, end_index)
  if end_index > start
    set_range(start, end_index)
    handle_data(rawdata[start...end_index])
  end
  return end_index
end

#handle_endtag(tag, method) ⇒ Object 



323
324
325
 
# File 'lib/html/sgml-parser.rb', line 323

def handle_endtag(tag, method)
  self.send(method)
end

#handle_entityref(name) ⇒ Object 



343
344
345
346
347
348
349
350
351
 
# File 'lib/html/sgml-parser.rb', line 343

def handle_entityref(name)
  table = Entitydefs
  if table.include?(name)
    handle_data(table[name])
  else
    unknown_entityref(name)
    return
  end
end

#handle_special(data) ⇒ Object 



359
360
 
# File 'lib/html/sgml-parser.rb', line 359

def handle_special(data)
end

#handle_starttag(tag, method, attrs) ⇒ Object 



319
320
321
 
# File 'lib/html/sgml-parser.rb', line 319

def handle_starttag(tag, method, attrs)
  self.send(method, attrs)
end

#has_context(gi) ⇒ Object 



73
74
75
 
# File 'lib/html/sgml-parser.rb', line 73

def has_context(gi)
  @stack.include? gi
end

#parse_comment(i) ⇒ Object 



193
194
195
196
197
198
199
200
201
202
203
204
205
206
 
# File 'lib/html/sgml-parser.rb', line 193

def parse_comment(i)
  rawdata = @rawdata
  if rawdata[i, 4] != '<!--'
    raise RuntimeError, 'unexpected call to handle_comment'
  end
  match = rawdata.index(Commentclose, i)
  return nil unless match
  matched_length = $&.length
  j = match
  src_length = match + matched_length - i
  set_range(i, i + src_length)
  handle_comment(rawdata[i+4..(j-1)])
  return src_length
end

#parse_endtag(i) ⇒ Object 



248
249
250
251
252
253
254
255
256
257
258
259
 
# File 'lib/html/sgml-parser.rb', line 248

def parse_endtag(i)
  rawdata = @rawdata
  j = rawdata.index(Endbracket, i + 1)
  return nil unless j
  tag = (rawdata[i+2..j-1].strip).downcase
  if rawdata[j] == ?> #
    j += 1
  end
  set_range(i, j)
  finish_endtag(tag)
  return j
end

#parse_special(i) ⇒ Object 



308
309
310
311
312
313
314
315
316
317
 
# File 'lib/html/sgml-parser.rb', line 308

def parse_special(i)
  rawdata = @rawdata
  match = rawdata.index(Endbracket, i+1)
  return nil unless match
  matched_length = $&.length
  src_length = match - i + matched_length
  set_range(i, i + src_length)
  handle_special(rawdata[i+1..(match-1)])
  return src_length
end

#parse_starttag(i) ⇒ Object 



208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
 
# File 'lib/html/sgml-parser.rb', line 208

def parse_starttag(i)
  rawdata = @rawdata
  j = rawdata.index(Endbracket, i + 1)
  return nil unless j
  attrs = []
  if rawdata[i+1] == ?> #
    # SGML shorthand: <> == <last open tag seen>
    k = j
    tag = @lasttag
  else
    match = rawdata.index(Tagfind, i + 1)
    unless match
      raise RuntimeError, 'unexpected call to parse_starttag'
    end
    k = i + 1 + ($&.length)
    tag = $&.downcase
    @lasttag = tag
  end
  while k < j
    break if rawdata.index(Endtagfind, k)
    break unless rawdata.index(Attrfind, k)
    matched_length = $&.length
    attrname, rest, attrvalue = $1, $2, $3
    if not rest
      attrvalue = '' # was: = attrname
    elsif (attrvalue[0] == ?' && attrvalue[-1] == ?') or
        (attrvalue[0] == ?" && attrvalue[-1] == ?")
      attrvalue = attrvalue[1..-2]
    end
    attrs << [attrname.downcase, attrvalue]
    k += matched_length
  end
  if rawdata[j] == ?> #
    j += 1
  end
  set_range(i, j)
  finish_starttag(tag, attrs)
  return j
end

#report_unbalanced(tag) ⇒ Object 



327
328
329
330
331
332
 
# File 'lib/html/sgml-parser.rb', line 327

def report_unbalanced(tag)
  if @verbose
    print '*** Unbalanced </' + tag + '>', "\n"
    print '*** Stack:', self.stack, "\n"
  end
end

#reset ⇒ Object 



50
51
52
53
54
55
56
57
58
 
# File 'lib/html/sgml-parser.rb', line 50

def reset
  @rawdata = ''
  @stack = []
  @lasttag = '???'
  @nomoretags = false
  @literal = false
  @offset = 0
  @ranges = []
end

#set_range(start, end_index) ⇒ Object 



68
69
70
71
 
# File 'lib/html/sgml-parser.rb', line 68

def set_range(start, end_index)
  @src_range = Range.new(start+@offset, end_index+@offset, exclusive = true)
  #puts "setting range #{@src_range}, text = \"#{get_source(src_range)}\""
end

#setliteral(*args) ⇒ Object 



82
83
84
 
# File 'lib/html/sgml-parser.rb', line 82

def setliteral(*args)
  @literal = true
end

#setnomoretags ⇒ Object 



77
78
79
80
 
# File 'lib/html/sgml-parser.rb', line 77

def setnomoretags
  @nomoretags = true
  @literal = true
end

#unknown_charref(ref) ⇒ Object 



366
367
 
# File 'lib/html/sgml-parser.rb', line 366

def unknown_charref(ref)
end

#unknown_endtag(tag) ⇒ Object 



364
365
 
# File 'lib/html/sgml-parser.rb', line 364

def unknown_endtag(tag)
end

#unknown_entityref(ref) ⇒ Object 



368
369
 
# File 'lib/html/sgml-parser.rb', line 368

def unknown_entityref(ref)
end

#unknown_starttag(tag, attrs) ⇒ Object 



362
363
 
# File 'lib/html/sgml-parser.rb', line 362

def unknown_starttag(tag, attrs)
end