Class: HTML::SGMLParser

Inherits:
Object
  • Object
show all
Defined in:
lib/html/htmlparser.rb

Overview

A parser for SGML, using the derived class as static DTD.

Direct Known Subclasses

HTMLParser

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 =

Assaf: fixed to allow tag to close itself (XHTML)

/<|>|\/>/
Special = 
/<![^<>]*>/
Commentopen = 
/<!--/
Commentclose = 
/--[ \t\n]*>/
Tagfind = 
/[a-zA-Z][a-zA-Z0-9.-]*/
Attrfind =

Assaf: / is no longer part of allowed attribute value

Regexp.compile('[\s,]*([a-zA-Z_][a-zA-Z_0-9.-]*)' +
'(\s*=\s*' +
"('[^']*'" +
'|"[^"]*"' +
'|[-~a-zA-Z0-9,.:+*%?!()_#=]*))?')
Entitydefs = 
{'lt'=>'<', 'gt'=>'>', 'amp'=>'&', 'quot'=>'"', 'apos'=>'\''}

Instance Method Summary collapse

Constructor Details

#initialize(verbose = false) ⇒ SGMLParser

Returns a new instance of SGMLParser.



34
35
36
37
 
# File 'lib/html/htmlparser.rb', line 34

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

Instance Method Details

#close ⇒ Object 



65
66
67
 
# File 'lib/html/htmlparser.rb', line 65

def close
    goahead(true)
end

#feed(data) ⇒ Object 



60
61
62
63
 
# File 'lib/html/htmlparser.rb', line 60

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

#finish_endtag(tag) ⇒ Object 



255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
 
# File 'lib/html/htmlparser.rb', line 255

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 



237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
 
# File 'lib/html/htmlparser.rb', line 237

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

#goahead(_end) ⇒ Object 



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
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
 
# File 'lib/html/htmlparser.rb', line 69

def goahead(_end)
    rawdata = @rawdata
    i = 0
    n = rawdata.length
    while i < n
    if @nomoretags
        handle_data(rawdata[i..(n-1)])
        i = n
        break
    end
    j = rawdata.index(Interesting, i)
    j = n unless j
    if i < j
        handle_data(rawdata[i..(j-1)])
    end
    i = j
    break if (i == n)
    if rawdata[i] == ?< #
        if rawdata.index(Starttagopen, i) == i
        if @literal
            handle_data(rawdata[i, 1])
            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
            handle_data(rawdata[i,1])
            i += 1
            next
        end
        k = parse_comment(i)
        break unless k
        i += k
        next
        end
        if rawdata.index(Special, i) == i
        if @literal
            handle_data(rawdata[i, 1])
            i += 1
            next
        end
        k = parse_special(i)
        break unless k
        i += k
        next
        end
    elsif rawdata[i] == ?& #
        if rawdata.index(Charref, i) == i
        i += $&.length
        handle_charref($1)
        i -= 1 unless rawdata[i-1] == ?;
        next
        end
        if rawdata.index(Entityref, i) == i
        i += $&.length
        handle_entityref($1)
        i -= 1 unless rawdata[i-1] == ?;
        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
        handle_data(rawdata[i, 1])
        i += 1
        next
    end
    j = match + $&.length
    break if j == n # Really incomplete
    handle_data(rawdata[i..(j-1)])
    i = j
    end
    # end while
    if _end and i < n
    handle_data(@rawdata[i..(n-1)])
    i = n
    end
    @rawdata = rawdata[i..-1]
end

#handle_charref(name) ⇒ Object 



308
309
310
311
312
313
314
315
 
# File 'lib/html/htmlparser.rb', line 308

def handle_charref(name)
    n = Integer(name) rescue -1
    if !(0 <= n && n <= 255)
    unknown_charref(name)
    return
    end
    handle_data(n.chr)
end

#handle_comment(data) ⇒ Object 



330
331
 
# File 'lib/html/htmlparser.rb', line 330

def handle_comment(data)
end

#handle_data(data) ⇒ Object 



327
328
 
# File 'lib/html/htmlparser.rb', line 327

def handle_data(data)
end

#handle_endtag(tag, method) ⇒ Object 



297
298
299
 
# File 'lib/html/htmlparser.rb', line 297

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

#handle_entityref(name) ⇒ Object 



317
318
319
320
321
322
323
324
325
 
# File 'lib/html/htmlparser.rb', line 317

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 



333
334
 
# File 'lib/html/htmlparser.rb', line 333

def handle_special(data)
end

#handle_starttag(tag, method, attrs) ⇒ Object 



293
294
295
 
# File 'lib/html/htmlparser.rb', line 293

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

#has_context(gi) ⇒ Object 



47
48
49
 
# File 'lib/html/htmlparser.rb', line 47

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

#parse_comment(i) ⇒ Object 



164
165
166
167
168
169
170
171
172
173
174
175
176
 
# File 'lib/html/htmlparser.rb', line 164

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
    handle_comment(rawdata[i+4..(j-1)])
    j = match + matched_length
    return j-i
end

#parse_endtag(i) ⇒ Object 



225
226
227
228
229
230
231
232
233
234
235
 
# File 'lib/html/htmlparser.rb', line 225

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
    finish_endtag(tag)
    return j
end

#parse_special(i) ⇒ Object 



284
285
286
287
288
289
290
291
 
# File 'lib/html/htmlparser.rb', line 284

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

#parse_starttag(i) ⇒ Object 



178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
 
# File 'lib/html/htmlparser.rb', line 178

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
    # Assaf: fixed to allow tag to close itself (XHTML)
    break unless idx = rawdata.index(Attrfind, k) and idx < j
    matched_length = $&.length
    attrname, rest, attrvalue = $1, $2, $3
    if not rest
        attrvalue = '' # was: = attrname
    # Assaf: fixed to handle double quoted attribute values properly
    elsif (attrvalue[0] == ?' && attrvalue[-1] == ?') or
        (attrvalue[0] == ?" && attrvalue[-1] == ?")
        attrvalue = attrvalue[1..-2]
    end
    attrs << [attrname.downcase, attrvalue]
    k += matched_length
    end
    # Assaf: fixed to allow tag to close itself (XHTML)
    if rawdata[j,2] == '/>'
    j += 2
    finish_starttag(tag, attrs)
    finish_endtag(tag)
    else
    if rawdata[j] == ?> #
        j += 1
    end
    finish_starttag(tag, attrs)
    end
    return j
end

#report_unbalanced(tag) ⇒ Object 



301
302
303
304
305
306
 
# File 'lib/html/htmlparser.rb', line 301

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

#reset ⇒ Object 



39
40
41
42
43
44
45
 
# File 'lib/html/htmlparser.rb', line 39

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

#setliteral(*args) ⇒ Object 



56
57
58
 
# File 'lib/html/htmlparser.rb', line 56

def setliteral(*args)
    @literal = true
end

#setnomoretags ⇒ Object 



51
52
53
54
 
# File 'lib/html/htmlparser.rb', line 51

def setnomoretags
    @nomoretags = true
    @literal = true
end

#unknown_charref(ref) ⇒ Object 



340
341
 
# File 'lib/html/htmlparser.rb', line 340

def unknown_charref(ref)
end

#unknown_endtag(tag) ⇒ Object 



338
339
 
# File 'lib/html/htmlparser.rb', line 338

def unknown_endtag(tag)
end

#unknown_entityref(ref) ⇒ Object 



342
343
 
# File 'lib/html/htmlparser.rb', line 342

def unknown_entityref(ref)
end

#unknown_starttag(tag, attrs) ⇒ Object 



336
337
 
# File 'lib/html/htmlparser.rb', line 336

def unknown_starttag(tag, attrs)
end