nokogumbo 1.3.0 → 1.4.1

Sign up to get free protection for your applications and to get access to all the features.
Files changed (20) hide show
  1. data/README.md +1 -1
  2. data/ext/nokogumboc/nokogumbo.c +1 -0
  3. data/gumbo-parser/src/error.c +6 -3
  4. data/gumbo-parser/src/gumbo.h +36 -170
  5. data/gumbo-parser/src/parser.c +1030 -779
  6. data/gumbo-parser/src/string_buffer.c +8 -1
  7. data/gumbo-parser/src/string_buffer.h +5 -0
  8. data/gumbo-parser/src/tag.c +35 -162
  9. data/gumbo-parser/src/tag.in +150 -0
  10. data/gumbo-parser/src/tag_enum.h +150 -0
  11. data/gumbo-parser/src/tag_gperf.h +343 -0
  12. data/gumbo-parser/src/tag_sizes.h +1 -0
  13. data/gumbo-parser/src/tag_strings.h +150 -0
  14. data/gumbo-parser/src/token_type.h +1 -0
  15. data/gumbo-parser/src/tokenizer.c +29 -21
  16. data/gumbo-parser/src/utf8.c +9 -8
  17. data/gumbo-parser/src/vector.c +1 -1
  18. data/gumbo-parser/visualc/include/strings.h +2 -1
  19. data/test-nokogumbo.rb +140 -0
  20. metadata +16 -10
data/gumbo-parser/src/tag_gperf.h ADDED
@@ -0,0 +1,343 @@
1
+ static unsigned int tag_hash(register const char *str, register unsigned int len)
2
+ {
3
+ static unsigned short asso_values[] =
4
+ {
5
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
6
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
7
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
8
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
9
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 6,
10
+ 4, 3, 1, 1, 0, 1, 0, 0, 296, 296,
11
+ 296, 296, 296, 296, 296, 22, 73, 151, 4, 13,
12
+ 59, 65, 2, 69, 0, 134, 9, 16, 52, 55,
13
+ 28, 101, 0, 1, 6, 63, 126, 104, 93, 124,
14
+ 296, 296, 296, 296, 296, 296, 296, 22, 73, 151,
15
+ 4, 13, 59, 65, 2, 69, 0, 134, 9, 16,
16
+ 52, 55, 28, 101, 0, 1, 6, 63, 126, 104,
17
+ 93, 124, 296, 296, 296, 296, 296, 296, 296, 296,
18
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
19
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
20
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
21
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
22
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
23
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
24
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
25
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
26
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
27
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
28
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
29
+ 296, 296, 296, 296, 296, 296, 296, 296, 296, 296,
30
+ 296, 296, 296, 296, 296, 296, 296, 296, 296
31
+ };
32
+ register unsigned int hval = len;
33
+
34
+ switch (hval)
35
+ {
36
+ default:
37
+ hval += asso_values[(unsigned char)str[1]+3];
38
+ /*FALLTHROUGH*/
39
+ case 1:
40
+ hval += asso_values[(unsigned char)str[0]];
41
+ break;
42
+ }
43
+ return hval + asso_values[(unsigned char)str[len - 1]];
44
+ }
45
+
46
+ static const unsigned char kGumboTagMap[] = {
47
+ GUMBO_TAG_LAST,
48
+ GUMBO_TAG_LAST,
49
+ GUMBO_TAG_LAST,
50
+ GUMBO_TAG_S,
51
+ GUMBO_TAG_H6,
52
+ GUMBO_TAG_H5,
53
+ GUMBO_TAG_H4,
54
+ GUMBO_TAG_H3,
55
+ GUMBO_TAG_SPACER,
56
+ GUMBO_TAG_H2,
57
+ GUMBO_TAG_HEADER,
58
+ GUMBO_TAG_H1,
59
+ GUMBO_TAG_HEAD,
60
+ GUMBO_TAG_LAST,
61
+ GUMBO_TAG_DETAILS,
62
+ GUMBO_TAG_SELECT,
63
+ GUMBO_TAG_DIR,
64
+ GUMBO_TAG_LAST,
65
+ GUMBO_TAG_DEL,
66
+ GUMBO_TAG_LAST,
67
+ GUMBO_TAG_SOURCE,
68
+ GUMBO_TAG_LEGEND,
69
+ GUMBO_TAG_DATALIST,
70
+ GUMBO_TAG_METER,
71
+ GUMBO_TAG_MGLYPH,
72
+ GUMBO_TAG_LAST,
73
+ GUMBO_TAG_MATH,
74
+ GUMBO_TAG_LABEL,
75
+ GUMBO_TAG_TABLE,
76
+ GUMBO_TAG_TEMPLATE,
77
+ GUMBO_TAG_LAST,
78
+ GUMBO_TAG_RP,
79
+ GUMBO_TAG_TIME,
80
+ GUMBO_TAG_TITLE,
81
+ GUMBO_TAG_DATA,
82
+ GUMBO_TAG_APPLET,
83
+ GUMBO_TAG_HGROUP,
84
+ GUMBO_TAG_SAMP,
85
+ GUMBO_TAG_TEXTAREA,
86
+ GUMBO_TAG_ABBR,
87
+ GUMBO_TAG_MARQUEE,
88
+ GUMBO_TAG_LAST,
89
+ GUMBO_TAG_MENUITEM,
90
+ GUMBO_TAG_SMALL,
91
+ GUMBO_TAG_META,
92
+ GUMBO_TAG_A,
93
+ GUMBO_TAG_LAST,
94
+ GUMBO_TAG_LAST,
95
+ GUMBO_TAG_LAST,
96
+ GUMBO_TAG_LAST,
97
+ GUMBO_TAG_EMBED,
98
+ GUMBO_TAG_MAP,
99
+ GUMBO_TAG_LAST,
100
+ GUMBO_TAG_PARAM,
101
+ GUMBO_TAG_LAST,
102
+ GUMBO_TAG_LAST,
103
+ GUMBO_TAG_NOBR,
104
+ GUMBO_TAG_P,
105
+ GUMBO_TAG_SPAN,
106
+ GUMBO_TAG_EM,
107
+ GUMBO_TAG_LAST,
108
+ GUMBO_TAG_NOFRAMES,
109
+ GUMBO_TAG_SECTION,
110
+ GUMBO_TAG_NOEMBED,
111
+ GUMBO_TAG_NEXTID,
112
+ GUMBO_TAG_FOOTER,
113
+ GUMBO_TAG_NOSCRIPT,
114
+ GUMBO_TAG_HR,
115
+ GUMBO_TAG_LAST,
116
+ GUMBO_TAG_FONT,
117
+ GUMBO_TAG_DL,
118
+ GUMBO_TAG_TR,
119
+ GUMBO_TAG_SCRIPT,
120
+ GUMBO_TAG_MO,
121
+ GUMBO_TAG_LAST,
122
+ GUMBO_TAG_DD,
123
+ GUMBO_TAG_MAIN,
124
+ GUMBO_TAG_TD,
125
+ GUMBO_TAG_FOREIGNOBJECT,
126
+ GUMBO_TAG_FORM,
127
+ GUMBO_TAG_OBJECT,
128
+ GUMBO_TAG_LAST,
129
+ GUMBO_TAG_FIELDSET,
130
+ GUMBO_TAG_LAST,
131
+ GUMBO_TAG_BGSOUND,
132
+ GUMBO_TAG_MENU,
133
+ GUMBO_TAG_TFOOT,
134
+ GUMBO_TAG_FIGURE,
135
+ GUMBO_TAG_RB,
136
+ GUMBO_TAG_LI,
137
+ GUMBO_TAG_LISTING,
138
+ GUMBO_TAG_BASEFONT,
139
+ GUMBO_TAG_OPTGROUP,
140
+ GUMBO_TAG_LAST,
141
+ GUMBO_TAG_BASE,
142
+ GUMBO_TAG_ADDRESS,
143
+ GUMBO_TAG_MI,
144
+ GUMBO_TAG_LAST,
145
+ GUMBO_TAG_PLAINTEXT,
146
+ GUMBO_TAG_LAST,
147
+ GUMBO_TAG_PROGRESS,
148
+ GUMBO_TAG_LAST,
149
+ GUMBO_TAG_LAST,
150
+ GUMBO_TAG_LAST,
151
+ GUMBO_TAG_ACRONYM,
152
+ GUMBO_TAG_ARTICLE,
153
+ GUMBO_TAG_LAST,
154
+ GUMBO_TAG_PRE,
155
+ GUMBO_TAG_LAST,
156
+ GUMBO_TAG_LAST,
157
+ GUMBO_TAG_LAST,
158
+ GUMBO_TAG_AREA,
159
+ GUMBO_TAG_RT,
160
+ GUMBO_TAG_LAST,
161
+ GUMBO_TAG_OPTION,
162
+ GUMBO_TAG_IMAGE,
163
+ GUMBO_TAG_DT,
164
+ GUMBO_TAG_LAST,
165
+ GUMBO_TAG_TT,
166
+ GUMBO_TAG_HTML,
167
+ GUMBO_TAG_WBR,
168
+ GUMBO_TAG_OL,
169
+ GUMBO_TAG_LAST,
170
+ GUMBO_TAG_STYLE,
171
+ GUMBO_TAG_STRIKE,
172
+ GUMBO_TAG_SUP,
173
+ GUMBO_TAG_MULTICOL,
174
+ GUMBO_TAG_U,
175
+ GUMBO_TAG_DFN,
176
+ GUMBO_TAG_UL,
177
+ GUMBO_TAG_FIGCAPTION,
178
+ GUMBO_TAG_MTEXT,
179
+ GUMBO_TAG_LAST,
180
+ GUMBO_TAG_VAR,
181
+ GUMBO_TAG_LAST,
182
+ GUMBO_TAG_LAST,
183
+ GUMBO_TAG_FRAMESET,
184
+ GUMBO_TAG_LAST,
185
+ GUMBO_TAG_BR,
186
+ GUMBO_TAG_I,
187
+ GUMBO_TAG_FRAME,
188
+ GUMBO_TAG_LAST,
189
+ GUMBO_TAG_DIV,
190
+ GUMBO_TAG_LAST,
191
+ GUMBO_TAG_TH,
192
+ GUMBO_TAG_MS,
193
+ GUMBO_TAG_ANNOTATION_XML,
194
+ GUMBO_TAG_B,
195
+ GUMBO_TAG_TBODY,
196
+ GUMBO_TAG_THEAD,
197
+ GUMBO_TAG_BIG,
198
+ GUMBO_TAG_BLOCKQUOTE,
199
+ GUMBO_TAG_XMP,
200
+ GUMBO_TAG_LAST,
201
+ GUMBO_TAG_KBD,
202
+ GUMBO_TAG_LAST,
203
+ GUMBO_TAG_LINK,
204
+ GUMBO_TAG_IFRAME,
205
+ GUMBO_TAG_MARK,
206
+ GUMBO_TAG_CENTER,
207
+ GUMBO_TAG_OUTPUT,
208
+ GUMBO_TAG_DESC,
209
+ GUMBO_TAG_CANVAS,
210
+ GUMBO_TAG_COL,
211
+ GUMBO_TAG_MALIGNMARK,
212
+ GUMBO_TAG_IMG,
213
+ GUMBO_TAG_ASIDE,
214
+ GUMBO_TAG_LAST,
215
+ GUMBO_TAG_CODE,
216
+ GUMBO_TAG_LAST,
217
+ GUMBO_TAG_SUB,
218
+ GUMBO_TAG_MN,
219
+ GUMBO_TAG_LAST,
220
+ GUMBO_TAG_LAST,
221
+ GUMBO_TAG_INS,
222
+ GUMBO_TAG_AUDIO,
223
+ GUMBO_TAG_STRONG,
224
+ GUMBO_TAG_CITE,
225
+ GUMBO_TAG_LAST,
226
+ GUMBO_TAG_LAST,
227
+ GUMBO_TAG_LAST,
228
+ GUMBO_TAG_INPUT,
229
+ GUMBO_TAG_LAST,
230
+ GUMBO_TAG_LAST,
231
+ GUMBO_TAG_LAST,
232
+ GUMBO_TAG_NAV,
233
+ GUMBO_TAG_LAST,
234
+ GUMBO_TAG_COLGROUP,
235
+ GUMBO_TAG_LAST,
236
+ GUMBO_TAG_LAST,
237
+ GUMBO_TAG_LAST,
238
+ GUMBO_TAG_LAST,
239
+ GUMBO_TAG_LAST,
240
+ GUMBO_TAG_SVG,
241
+ GUMBO_TAG_KEYGEN,
242
+ GUMBO_TAG_VIDEO,
243
+ GUMBO_TAG_BDO,
244
+ GUMBO_TAG_LAST,
245
+ GUMBO_TAG_LAST,
246
+ GUMBO_TAG_LAST,
247
+ GUMBO_TAG_LAST,
248
+ GUMBO_TAG_BODY,
249
+ GUMBO_TAG_LAST,
250
+ GUMBO_TAG_Q,
251
+ GUMBO_TAG_LAST,
252
+ GUMBO_TAG_LAST,
253
+ GUMBO_TAG_LAST,
254
+ GUMBO_TAG_LAST,
255
+ GUMBO_TAG_TRACK,
256
+ GUMBO_TAG_LAST,
257
+ GUMBO_TAG_BDI,
258
+ GUMBO_TAG_LAST,
259
+ GUMBO_TAG_LAST,
260
+ GUMBO_TAG_LAST,
261
+ GUMBO_TAG_CAPTION,
262
+ GUMBO_TAG_LAST,
263
+ GUMBO_TAG_LAST,
264
+ GUMBO_TAG_LAST,
265
+ GUMBO_TAG_LAST,
266
+ GUMBO_TAG_LAST,
267
+ GUMBO_TAG_LAST,
268
+ GUMBO_TAG_RUBY,
269
+ GUMBO_TAG_LAST,
270
+ GUMBO_TAG_LAST,
271
+ GUMBO_TAG_BUTTON,
272
+ GUMBO_TAG_SUMMARY,
273
+ GUMBO_TAG_LAST,
274
+ GUMBO_TAG_LAST,
275
+ GUMBO_TAG_LAST,
276
+ GUMBO_TAG_LAST,
277
+ GUMBO_TAG_LAST,
278
+ GUMBO_TAG_LAST,
279
+ GUMBO_TAG_LAST,
280
+ GUMBO_TAG_LAST,
281
+ GUMBO_TAG_LAST,
282
+ GUMBO_TAG_LAST,
283
+ GUMBO_TAG_LAST,
284
+ GUMBO_TAG_LAST,
285
+ GUMBO_TAG_LAST,
286
+ GUMBO_TAG_LAST,
287
+ GUMBO_TAG_LAST,
288
+ GUMBO_TAG_LAST,
289
+ GUMBO_TAG_LAST,
290
+ GUMBO_TAG_LAST,
291
+ GUMBO_TAG_LAST,
292
+ GUMBO_TAG_LAST,
293
+ GUMBO_TAG_LAST,
294
+ GUMBO_TAG_LAST,
295
+ GUMBO_TAG_LAST,
296
+ GUMBO_TAG_LAST,
297
+ GUMBO_TAG_LAST,
298
+ GUMBO_TAG_LAST,
299
+ GUMBO_TAG_LAST,
300
+ GUMBO_TAG_LAST,
301
+ GUMBO_TAG_LAST,
302
+ GUMBO_TAG_LAST,
303
+ GUMBO_TAG_LAST,
304
+ GUMBO_TAG_LAST,
305
+ GUMBO_TAG_RTC,
306
+ GUMBO_TAG_LAST,
307
+ GUMBO_TAG_LAST,
308
+ GUMBO_TAG_LAST,
309
+ GUMBO_TAG_LAST,
310
+ GUMBO_TAG_LAST,
311
+ GUMBO_TAG_LAST,
312
+ GUMBO_TAG_LAST,
313
+ GUMBO_TAG_LAST,
314
+ GUMBO_TAG_BLINK,
315
+ GUMBO_TAG_LAST,
316
+ GUMBO_TAG_LAST,
317
+ GUMBO_TAG_LAST,
318
+ GUMBO_TAG_LAST,
319
+ GUMBO_TAG_LAST,
320
+ GUMBO_TAG_LAST,
321
+ GUMBO_TAG_LAST,
322
+ GUMBO_TAG_LAST,
323
+ GUMBO_TAG_LAST,
324
+ GUMBO_TAG_LAST,
325
+ GUMBO_TAG_LAST,
326
+ GUMBO_TAG_LAST,
327
+ GUMBO_TAG_LAST,
328
+ GUMBO_TAG_LAST,
329
+ GUMBO_TAG_LAST,
330
+ GUMBO_TAG_LAST,
331
+ GUMBO_TAG_LAST,
332
+ GUMBO_TAG_LAST,
333
+ GUMBO_TAG_LAST,
334
+ GUMBO_TAG_LAST,
335
+ GUMBO_TAG_LAST,
336
+ GUMBO_TAG_LAST,
337
+ GUMBO_TAG_LAST,
338
+ GUMBO_TAG_LAST,
339
+ GUMBO_TAG_LAST,
340
+ GUMBO_TAG_LAST,
341
+ GUMBO_TAG_LAST,
342
+ GUMBO_TAG_ISINDEX
343
+ };
data/gumbo-parser/src/tag_sizes.h ADDED
@@ -0,0 +1 @@
1
+ 4, 4, 5, 4, 4, 4, 5, 6, 8, 8, 4, 7, 7, 3, 5, 2, 2, 2, 2, 2, 2, 6, 6, 6, 7, 1, 2, 3, 10, 2, 2, 2, 2, 2, 2, 6, 10, 4, 3, 1, 2, 6, 5, 1, 4, 1, 3, 4, 4, 4, 4, 3, 4, 3, 3, 3, 1, 1, 1, 4, 4, 2, 2, 3, 3, 4, 2, 3, 3, 3, 5, 3, 6, 5, 6, 5, 5, 5, 6, 5, 6, 3, 4, 4, 2, 2, 2, 2, 5, 6, 10, 14, 3, 13, 4, 5, 7, 8, 3, 5, 5, 5, 2, 2, 2, 4, 8, 6, 5, 5, 6, 6, 8, 8, 6, 8, 6, 6, 8, 5, 7, 7, 4, 8, 6, 7, 7, 3, 5, 8, 8, 7, 7, 3, 6, 7, 9, 2, 6, 8, 3, 5, 6, 4, 7, 8, 4, 6, 2, 3,
data/gumbo-parser/src/tag_strings.h ADDED
@@ -0,0 +1,150 @@
1
+ "html",
2
+ "head",
3
+ "title",
4
+ "base",
5
+ "link",
6
+ "meta",
7
+ "style",
8
+ "script",
9
+ "noscript",
10
+ "template",
11
+ "body",
12
+ "article",
13
+ "section",
14
+ "nav",
15
+ "aside",
16
+ "h1",
17
+ "h2",
18
+ "h3",
19
+ "h4",
20
+ "h5",
21
+ "h6",
22
+ "hgroup",
23
+ "header",
24
+ "footer",
25
+ "address",
26
+ "p",
27
+ "hr",
28
+ "pre",
29
+ "blockquote",
30
+ "ol",
31
+ "ul",
32
+ "li",
33
+ "dl",
34
+ "dt",
35
+ "dd",
36
+ "figure",
37
+ "figcaption",
38
+ "main",
39
+ "div",
40
+ "a",
41
+ "em",
42
+ "strong",
43
+ "small",
44
+ "s",
45
+ "cite",
46
+ "q",
47
+ "dfn",
48
+ "abbr",
49
+ "data",
50
+ "time",
51
+ "code",
52
+ "var",
53
+ "samp",
54
+ "kbd",
55
+ "sub",
56
+ "sup",
57
+ "i",
58
+ "b",
59
+ "u",
60
+ "mark",
61
+ "ruby",
62
+ "rt",
63
+ "rp",
64
+ "bdi",
65
+ "bdo",
66
+ "span",
67
+ "br",
68
+ "wbr",
69
+ "ins",
70
+ "del",
71
+ "image",
72
+ "img",
73
+ "iframe",
74
+ "embed",
75
+ "object",
76
+ "param",
77
+ "video",
78
+ "audio",
79
+ "source",
80
+ "track",
81
+ "canvas",
82
+ "map",
83
+ "area",
84
+ "math",
85
+ "mi",
86
+ "mo",
87
+ "mn",
88
+ "ms",
89
+ "mtext",
90
+ "mglyph",
91
+ "malignmark",
92
+ "annotation-xml",
93
+ "svg",
94
+ "foreignobject",
95
+ "desc",
96
+ "table",
97
+ "caption",
98
+ "colgroup",
99
+ "col",
100
+ "tbody",
101
+ "thead",
102
+ "tfoot",
103
+ "tr",
104
+ "td",
105
+ "th",
106
+ "form",
107
+ "fieldset",
108
+ "legend",
109
+ "label",
110
+ "input",
111
+ "button",
112
+ "select",
113
+ "datalist",
114
+ "optgroup",
115
+ "option",
116
+ "textarea",
117
+ "keygen",
118
+ "output",
119
+ "progress",
120
+ "meter",
121
+ "details",
122
+ "summary",
123
+ "menu",
124
+ "menuitem",
125
+ "applet",
126
+ "acronym",
127
+ "bgsound",
128
+ "dir",
129
+ "frame",
130
+ "frameset",
131
+ "noframes",
132
+ "isindex",
133
+ "listing",
134
+ "xmp",
135
+ "nextid",
136
+ "noembed",
137
+ "plaintext",
138
+ "rb",
139
+ "strike",
140
+ "basefont",
141
+ "big",
142
+ "blink",
143
+ "center",
144
+ "font",
145
+ "marquee",
146
+ "multicol",
147
+ "nobr",
148
+ "spacer",
149
+ "tt",
150
+ "rtc",