PageRenderTime 52ms CodeModel.GetById 19ms RepoModel.GetById 0ms app.codeStats 0ms

/peek-build/src/netdepends/hubbub-0.0.2/src/tokeniser/tokeniser.c

https://bitbucket.org/C0deMaver1ck/peeklinux
C | 3370 lines | 2902 code | 319 blank | 149 comment | 615 complexity | 6530fbcb7409a5b2eafc2432fd3ae059 MD5 | raw file
Possible License(s): GPL-2.0, LGPL-2.1, LGPL-2.0 
    

  1. /*
    2. 

  2.  * This file is part of Hubbub.
    3. 

  3.  * Licensed under the MIT License,
    4. 

  4.  *                http://www.opensource.org/licenses/mit-license.php
    5. 

  5.  * Copyright 2007 John-Mark Bell <jmb@netsurf-browser.org>
    6. 

  6.  * Copyright 2008 Andrew Sidwell <takkaria@netsurf-browser.org>
    7. 

  7.  */
    8. 

  8. #include <assert.h>
    9. 

  9. #include <stdbool.h>
    10. 

  10. #include <string.h>
    11. 

  11. 

  12. #include <stdio.h>
    13. 

  13. 

  14. #include <parserutils/charset/utf8.h>
    15. 

  15. 

  16. #include "utils/parserutilserror.h"
    17. 

  17. #include "utils/utils.h"
    18. 

  18. 

  19. #include "tokeniser/entities.h"
    20. 

  20. #include "tokeniser/tokeniser.h"
    21. 

  21. 

  22. /**
    23. 

  23.  * Table of mappings between Windows-1252 codepoints 128-159 and UCS4
    24. 

  24.  */
    25. 

  25. static const uint32_t cp1252Table[32] = {
    26. 

  26. 	0x20AC, 0xFFFD, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021,
    27. 

  27. 	0x02C6, 0x2030, 0x0160, 0x2039, 0x0152, 0xFFFD, 0x017D, 0xFFFD,
    28. 

  28. 	0xFFFD, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014,
    29. 

  29. 	0x02DC, 0x2122, 0x0161, 0x203A, 0x0153, 0xFFFD, 0x017E, 0x0178
    30. 

  30. };
    31. 

  31. 

  32. /**
    33. 

  33.  * UTF-8 encoding of U+FFFD REPLACEMENT CHARACTER
    34. 

  34.  */
    35. 

  35. static const uint8_t u_fffd[3] = { '\xEF', '\xBF', '\xBD' };
    36. 

  36. static const hubbub_string u_fffd_str = { u_fffd, sizeof(u_fffd) };
    37. 

  37. 

  38. 

  39. /**
    40. 

  40.  * String for when we want to emit newlines
    41. 

  41.  */
    42. 

  42. static const uint8_t lf = '\n';
    43. 

  43. static const hubbub_string lf_str = { &lf, 1 };
    44. 

  44. 

  45. 

  46. /**
    47. 

  47.  * Tokeniser states
    48. 

  48.  */
    49. 

  49. typedef enum hubbub_tokeniser_state {
    50. 

  50. 	STATE_DATA,
    51. 

  51. 	STATE_CHARACTER_REFERENCE_DATA,
    52. 

  52. 	STATE_TAG_OPEN,
    53. 

  53. 	STATE_CLOSE_TAG_OPEN,
    54. 

  54. 	STATE_TAG_NAME,
    55. 

  55. 	STATE_BEFORE_ATTRIBUTE_NAME,
    56. 

  56. 	STATE_ATTRIBUTE_NAME,
    57. 

  57. 	STATE_AFTER_ATTRIBUTE_NAME,
    58. 

  58. 	STATE_BEFORE_ATTRIBUTE_VALUE,
    59. 

  59. 	STATE_ATTRIBUTE_VALUE_DQ,
    60. 

  60. 	STATE_ATTRIBUTE_VALUE_SQ,
    61. 

  61. 	STATE_ATTRIBUTE_VALUE_UQ,
    62. 

  62. 	STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE,
    63. 

  63. 	STATE_AFTER_ATTRIBUTE_VALUE_Q,
    64. 

  64. 	STATE_SELF_CLOSING_START_TAG,
    65. 

  65. 	STATE_BOGUS_COMMENT,
    66. 

  66. 	STATE_MARKUP_DECLARATION_OPEN,
    67. 

  67. 	STATE_MATCH_COMMENT,
    68. 

  68. 	STATE_COMMENT_START,
    69. 

  69. 	STATE_COMMENT_START_DASH,
    70. 

  70. 	STATE_COMMENT,
    71. 

  71. 	STATE_COMMENT_END_DASH,
    72. 

  72. 	STATE_COMMENT_END,
    73. 

  73. 	STATE_MATCH_DOCTYPE,
    74. 

  74. 	STATE_DOCTYPE,
    75. 

  75. 	STATE_BEFORE_DOCTYPE_NAME,
    76. 

  76. 	STATE_DOCTYPE_NAME,
    77. 

  77. 	STATE_AFTER_DOCTYPE_NAME,
    78. 

  78. 	STATE_MATCH_PUBLIC,
    79. 

  79. 	STATE_BEFORE_DOCTYPE_PUBLIC,
    80. 

  80. 	STATE_DOCTYPE_PUBLIC_DQ,
    81. 

  81. 	STATE_DOCTYPE_PUBLIC_SQ,
    82. 

  82. 	STATE_AFTER_DOCTYPE_PUBLIC,
    83. 

  83. 	STATE_MATCH_SYSTEM,
    84. 

  84. 	STATE_BEFORE_DOCTYPE_SYSTEM,
    85. 

  85. 	STATE_DOCTYPE_SYSTEM_DQ,
    86. 

  86. 	STATE_DOCTYPE_SYSTEM_SQ,
    87. 

  87. 	STATE_AFTER_DOCTYPE_SYSTEM,
    88. 

  88. 	STATE_BOGUS_DOCTYPE,
    89. 

  89. 	STATE_MATCH_CDATA,
    90. 

  90. 	STATE_CDATA_BLOCK,
    91. 

  91. 	STATE_NUMBERED_ENTITY,
    92. 

  92. 	STATE_NAMED_ENTITY
    93. 

  93. } hubbub_tokeniser_state;
    94. 

  94. 

  95. /**
    96. 

  96.  * Context for tokeniser
    97. 

  97.  */
    98. 

  98. typedef struct hubbub_tokeniser_context {
    99. 

  99. 	size_t pending;				/**< Count of pending chars */
    100. 

  100. 

  101. 	hubbub_string current_comment;		/**< Current comment text */
    102. 

  102. 

  103. 	hubbub_token_type current_tag_type;	/**< Type of current_tag */
    104. 

  104. 	hubbub_tag current_tag;			/**< Current tag */
    105. 

  105. 	hubbub_doctype current_doctype;		/**< Current doctype */
    106. 

  106. 	hubbub_tokeniser_state prev_state;	/**< Previous state */
    107. 

  107. 

  108. 	uint8_t last_start_tag_name[10];	/**< Name of the last start tag
    109. 

  109. 						 * emitted */
    110. 

  110. 	size_t last_start_tag_len;		/**< Length of last start tag */
    111. 

  111. 

  112. 	struct {
    113. 

  113. 		uint32_t count;
    114. 

  114. 		bool match;
    115. 

  115. 	} close_tag_match;			/**< State for matching close 
    116. 

  116. 						 * tags */
    117. 

  117. 

  118. 	struct {
    119. 

  119. 		uint32_t count;			/**< Index into "DOCTYPE" */
    120. 

  120. 	} match_doctype;			/**< State for matching doctype */
    121. 

  121. 

  122. 	struct {
    123. 

  123. 		uint32_t count;			/**< Index into "[CDATA[" */
    124. 

  124. 		uint32_t end;			/**< Index into "]]>" */
    125. 

  125. 	} match_cdata;				/**< State for matching cdata */
    126. 

  126. 

  127. 	struct {
    128. 

  128. 		size_t offset;			/**< Offset in buffer */
    129. 

  129. 		uint32_t length;		/**< Length of entity */
    130. 

  130. 		uint32_t codepoint;		/**< UCS4 codepoint */
    131. 

  131. 		bool complete;			/**< True if match complete */
    132. 

  132. 

  133. 		uint32_t poss_length;		/**< Optimistic length
    134. 

  134. 						 * when matching named
    135. 

  135. 						 * character references */
    136. 

  136. 		uint8_t base;			/**< Base for numeric
    137. 

  137. 						 * entities */
    138. 

  138. 		void *context;			/**< Context for named
    139. 

  139. 						 * entity search */
    140. 

  140. 		size_t prev_len;		/**< Previous byte length
    141. 

  141. 						 * of str */
    142. 

  142. 		bool had_data;			/**< Whether we read
    143. 

  143. 						 * anything after &#(x)? */
    144. 

  144. 		bool overflow;			/**< Whether this entity has
    145. 

  145. 						 * has overflowed the maximum
    146. 

  146. 						 * numeric entity value */
    147. 

  147. 		hubbub_tokeniser_state return_state;	/**< State we were
    148. 

  148. 							 * called from */
    149. 

  149. 	} match_entity;				/**< Entity matching state */
    150. 

  150. 

  151. 	struct {
    152. 

  152. 		uint32_t line;			/**< Current line of input */
    153. 

  153. 		uint32_t col;			/**< Current character in
    154. 

  154. 						 * line */
    155. 

  155. 	} position;				/**< Position in source data */
    156. 

  156. 

  157. 	uint32_t allowed_char;			/**< Used for quote matching */
    158. 

  158. } hubbub_tokeniser_context;
    159. 

  159. 

  160. /**
    161. 

  161.  * Tokeniser data structure
    162. 

  162.  */
    163. 

  163. struct hubbub_tokeniser {
    164. 

  164. 	hubbub_tokeniser_state state;	/**< Current tokeniser state */
    165. 

  165. 	hubbub_content_model content_model;	/**< Current content
    166. 

  166. 						 * model flag */
    167. 

  167. 	bool escape_flag;		/**< Escape flag **/
    168. 

  168. 	bool process_cdata_section;	/**< Whether to process CDATA sections*/
    169. 

  169. 

  170. 	parserutils_inputstream *input;	/**< Input stream */
    171. 

  171. 	parserutils_buffer *buffer;	/**< Input buffer */
    172. 

  172. 

  173. 	hubbub_tokeniser_context context;	/**< Tokeniser context */
    174. 

  174. 

  175. 	hubbub_token_handler token_handler;	/**< Token handling callback */
    176. 

  176. 	void *token_pw;				/**< Token handler data */
    177. 

  177. 

  178. 	hubbub_error_handler error_handler;	/**< Error handling callback */
    179. 

  179. 	void *error_pw;				/**< Error handler data */
    180. 

  180. 

  181. 	hubbub_allocator_fn alloc;	/**< Memory (de)allocation function */
    182. 

  182. 	void *alloc_pw;			/**< Client private data */
    183. 

  183. };
    184. 

  184. 

  185. static hubbub_error hubbub_tokeniser_handle_data(hubbub_tokeniser *tokeniser);
    186. 

  186. static hubbub_error hubbub_tokeniser_handle_character_reference_data(
    187. 

  187. 		hubbub_tokeniser *tokeniser);
    188. 

  188. static hubbub_error hubbub_tokeniser_handle_tag_open(
    189. 

  189. 		hubbub_tokeniser *tokeniser);
    190. 

  190. static hubbub_error hubbub_tokeniser_handle_close_tag_open(
    191. 

  191. 		hubbub_tokeniser *tokeniser);
    192. 

  192. static hubbub_error hubbub_tokeniser_handle_tag_name(
    193. 

  193. 		hubbub_tokeniser *tokeniser);
    194. 

  194. static hubbub_error hubbub_tokeniser_handle_before_attribute_name(
    195. 

  195. 		hubbub_tokeniser *tokeniser);
    196. 

  196. static hubbub_error hubbub_tokeniser_handle_attribute_name(
    197. 

  197. 		hubbub_tokeniser *tokeniser);
    198. 

  198. static hubbub_error hubbub_tokeniser_handle_after_attribute_name(
    199. 

  199. 		hubbub_tokeniser *tokeniser);
    200. 

  200. static hubbub_error hubbub_tokeniser_handle_before_attribute_value(
    201. 

  201. 		hubbub_tokeniser *tokeniser);
    202. 

  202. static hubbub_error hubbub_tokeniser_handle_attribute_value_dq(
    203. 

  203. 		hubbub_tokeniser *tokeniser);
    204. 

  204. static hubbub_error hubbub_tokeniser_handle_attribute_value_sq(
    205. 

  205. 		hubbub_tokeniser *tokeniser);
    206. 

  206. static hubbub_error hubbub_tokeniser_handle_attribute_value_uq(
    207. 

  207. 		hubbub_tokeniser *tokeniser);
    208. 

  208. static hubbub_error hubbub_tokeniser_handle_character_reference_in_attribute_value(
    209. 

  209. 		hubbub_tokeniser *tokeniser);
    210. 

  210. static hubbub_error hubbub_tokeniser_handle_after_attribute_value_q(
    211. 

  211. 		hubbub_tokeniser *tokeniser);
    212. 

  212. static hubbub_error hubbub_tokeniser_handle_self_closing_start_tag(
    213. 

  213. 		hubbub_tokeniser *tokeniser);
    214. 

  214. static hubbub_error hubbub_tokeniser_handle_bogus_comment(
    215. 

  215. 		hubbub_tokeniser *tokeniser);
    216. 

  216. static hubbub_error hubbub_tokeniser_handle_markup_declaration_open(
    217. 

  217. 		hubbub_tokeniser *tokeniser);
    218. 

  218. static hubbub_error hubbub_tokeniser_handle_match_comment(
    219. 

  219. 		hubbub_tokeniser *tokeniser);
    220. 

  220. static hubbub_error hubbub_tokeniser_handle_comment(
    221. 

  221. 		hubbub_tokeniser *tokeniser);
    222. 

  222. static hubbub_error hubbub_tokeniser_handle_match_doctype(
    223. 

  223. 		hubbub_tokeniser *tokeniser);
    224. 

  224. static hubbub_error hubbub_tokeniser_handle_doctype(
    225. 

  225. 		hubbub_tokeniser *tokeniser);
    226. 

  226. static hubbub_error hubbub_tokeniser_handle_before_doctype_name(
    227. 

  227. 		hubbub_tokeniser *tokeniser);
    228. 

  228. static hubbub_error hubbub_tokeniser_handle_doctype_name(
    229. 

  229. 		hubbub_tokeniser *tokeniser);
    230. 

  230. static hubbub_error hubbub_tokeniser_handle_after_doctype_name(
    231. 

  231. 		hubbub_tokeniser *tokeniser);
    232. 

  232. static hubbub_error hubbub_tokeniser_handle_match_public(
    233. 

  233. 		hubbub_tokeniser *tokeniser);
    234. 

  234. static hubbub_error hubbub_tokeniser_handle_before_doctype_public(
    235. 

  235. 		hubbub_tokeniser *tokeniser);
    236. 

  236. static hubbub_error hubbub_tokeniser_handle_doctype_public_dq(
    237. 

  237. 		hubbub_tokeniser *tokeniser);
    238. 

  238. static hubbub_error hubbub_tokeniser_handle_doctype_public_sq(
    239. 

  239. 		hubbub_tokeniser *tokeniser);
    240. 

  240. static hubbub_error hubbub_tokeniser_handle_after_doctype_public(
    241. 

  241. 		hubbub_tokeniser *tokeniser);
    242. 

  242. static hubbub_error hubbub_tokeniser_handle_match_system(
    243. 

  243. 		hubbub_tokeniser *tokeniser);
    244. 

  244. static hubbub_error hubbub_tokeniser_handle_before_doctype_system(
    245. 

  245. 		hubbub_tokeniser *tokeniser);
    246. 

  246. static hubbub_error hubbub_tokeniser_handle_doctype_system_dq(
    247. 

  247. 		hubbub_tokeniser *tokeniser);
    248. 

  248. static hubbub_error hubbub_tokeniser_handle_doctype_system_sq(
    249. 

  249. 		hubbub_tokeniser *tokeniser);
    250. 

  250. static hubbub_error hubbub_tokeniser_handle_after_doctype_system(
    251. 

  251. 		hubbub_tokeniser *tokeniser);
    252. 

  252. static hubbub_error hubbub_tokeniser_handle_bogus_doctype(
    253. 

  253. 		hubbub_tokeniser *tokeniser);
    254. 

  254. static hubbub_error hubbub_tokeniser_handle_match_cdata(
    255. 

  255. 		hubbub_tokeniser *tokeniser);
    256. 

  256. static hubbub_error hubbub_tokeniser_handle_cdata_block(
    257. 

  257. 		hubbub_tokeniser *tokeniser);
    258. 

  258. static hubbub_error hubbub_tokeniser_consume_character_reference(
    259. 

  259. 		hubbub_tokeniser *tokeniser, size_t off);
    260. 

  260. static hubbub_error hubbub_tokeniser_handle_numbered_entity(
    261. 

  261. 		hubbub_tokeniser *tokeniser);
    262. 

  262. static hubbub_error hubbub_tokeniser_handle_named_entity(
    263. 

  263. 		hubbub_tokeniser *tokeniser);
    264. 

  264. 

  265. static inline hubbub_error emit_character_token(hubbub_tokeniser *tokeniser,
    266. 

  266. 		const hubbub_string *chars);
    267. 

  267. static inline hubbub_error emit_current_chars(hubbub_tokeniser *tokeniser);
    268. 

  268. static inline hubbub_error emit_current_tag(hubbub_tokeniser *tokeniser);
    269. 

  269. static inline hubbub_error emit_current_comment(hubbub_tokeniser *tokeniser);
    270. 

  270. static inline hubbub_error emit_current_doctype(hubbub_tokeniser *tokeniser,
    271. 

  271. 		bool force_quirks);
    272. 

  272. static hubbub_error hubbub_tokeniser_emit_token(hubbub_tokeniser *tokeniser,
    273. 

  273. 		hubbub_token *token);
    274. 

  274. 

  275. /**
    276. 

  276.  * Create a hubbub tokeniser
    277. 

  277.  *
    278. 

  278.  * \param input      Input stream instance
    279. 

  279.  * \param alloc      Memory (de)allocation function
    280. 

  280.  * \param pw         Pointer to client-specific private data (may be NULL)
    281. 

  281.  * \param tokeniser  Pointer to location to receive tokeniser instance
    282. 

  282.  * \return HUBBUB_OK on success,
    283. 

  283.  *         HUBBUB_BADPARM on bad parameters,
    284. 

  284.  *         HUBBUB_NOMEM on memory exhaustion
    285. 

  285.  */
    286. 

  286. hubbub_error hubbub_tokeniser_create(parserutils_inputstream *input,
    287. 

  287. 		hubbub_allocator_fn alloc, void *pw, 
    288. 

  288. 		hubbub_tokeniser **tokeniser)
    289. 

  289. {
    290. 

  290. 	parserutils_error perror;
    291. 

  291. 	hubbub_tokeniser *tok;
    292. 

  292. 

  293. 	if (input == NULL || alloc == NULL || tokeniser == NULL)
    294. 

  294. 		return HUBBUB_BADPARM;
    295. 

  295. 

  296. 	tok = alloc(NULL, sizeof(hubbub_tokeniser), pw);
    297. 

  297. 	if (tok == NULL)
    298. 

  298. 		return HUBBUB_NOMEM;
    299. 

  299. 

  300. 	perror = parserutils_buffer_create(alloc, pw, &tok->buffer);
    301. 

  301. 	if (perror != PARSERUTILS_OK) {
    302. 

  302. 		alloc(tok, 0, pw);
    303. 

  303. 		return hubbub_error_from_parserutils_error(perror);
    304. 

  304. 	}
    305. 

  305. 

  306. 	tok->state = STATE_DATA;
    307. 

  307. 	tok->content_model = HUBBUB_CONTENT_MODEL_PCDATA;
    308. 

  308. 

  309. 	tok->escape_flag = false;
    310. 

  310. 	tok->process_cdata_section = false;
    311. 

  311. 

  312. 	tok->input = input;
    313. 

  313. 

  314. 	tok->token_handler = NULL;
    315. 

  315. 	tok->token_pw = NULL;
    316. 

  316. 

  317. 	tok->error_handler = NULL;
    318. 

  318. 	tok->error_pw = NULL;
    319. 

  319. 

  320. 	tok->alloc = alloc;
    321. 

  321. 	tok->alloc_pw = pw;
    322. 

  322. 

  323. 	memset(&tok->context, 0, sizeof(hubbub_tokeniser_context));
    324. 

  324. 

  325. 	*tokeniser = tok;
    326. 

  326. 

  327. 	return HUBBUB_OK;
    328. 

  328. }
    329. 

  329. 

  330. /**
    331. 

  331.  * Destroy a hubbub tokeniser
    332. 

  332.  *
    333. 

  333.  * \param tokeniser  The tokeniser instance to destroy
    334. 

  334.  * \return HUBBUB_OK on success, appropriate error otherwise
    335. 

  335.  */
    336. 

  336. hubbub_error hubbub_tokeniser_destroy(hubbub_tokeniser *tokeniser)
    337. 

  337. {
    338. 

  338. 	if (tokeniser == NULL)
    339. 

  339. 		return HUBBUB_BADPARM;
    340. 

  340. 

  341. 	if (tokeniser->context.current_tag.attributes != NULL) {
    342. 

  342. 		tokeniser->alloc(tokeniser->context.current_tag.attributes,
    343. 

  343. 				0, tokeniser->alloc_pw);
    344. 

  344. 	}
    345. 

  345. 

  346. 	parserutils_buffer_destroy(tokeniser->buffer);
    347. 

  347. 

  348. 	tokeniser->alloc(tokeniser, 0, tokeniser->alloc_pw);
    349. 

  349. 

  350. 	return HUBBUB_OK;
    351. 

  351. }
    352. 

  352. 

  353. /**
    354. 

  354.  * Configure a hubbub tokeniser
    355. 

  355.  *
    356. 

  356.  * \param tokeniser  The tokeniser instance to configure
    357. 

  357.  * \param type       The option type to set
    358. 

  358.  * \param params     Option-specific parameters
    359. 

  359.  * \return HUBBUB_OK on success, appropriate error otherwise
    360. 

  360.  */
    361. 

  361. hubbub_error hubbub_tokeniser_setopt(hubbub_tokeniser *tokeniser,
    362. 

  362. 		hubbub_tokeniser_opttype type,
    363. 

  363. 		hubbub_tokeniser_optparams *params)
    364. 

  364. {
    365. 

  365. 	if (tokeniser == NULL || params == NULL)
    366. 

  366. 		return HUBBUB_BADPARM;
    367. 

  367. 

  368. 	switch (type) {
    369. 

  369. 	case HUBBUB_TOKENISER_TOKEN_HANDLER:
    370. 

  370. 		tokeniser->token_handler = params->token_handler.handler;
    371. 

  371. 		tokeniser->token_pw = params->token_handler.pw;
    372. 

  372. 		break;
    373. 

  373. 	case HUBBUB_TOKENISER_ERROR_HANDLER:
    374. 

  374. 		tokeniser->error_handler = params->error_handler.handler;
    375. 

  375. 		tokeniser->error_pw = params->error_handler.pw;
    376. 

  376. 		break;
    377. 

  377. 	case HUBBUB_TOKENISER_CONTENT_MODEL:
    378. 

  378. 		tokeniser->content_model = params->content_model.model;
    379. 

  379. 		break;
    380. 

  380. 	case HUBBUB_TOKENISER_PROCESS_CDATA:
    381. 

  381. 		tokeniser->process_cdata_section = params->process_cdata;
    382. 

  382. 		break;
    383. 

  383. 	}
    384. 

  384. 

  385. 	return HUBBUB_OK;
    386. 

  386. }
    387. 

  387. 

  388. /**
    389. 

  389.  * Process remaining data in the input stream
    390. 

  390.  *
    391. 

  391.  * \param tokeniser  The tokeniser instance to invoke
    392. 

  392.  * \return HUBBUB_OK on success, appropriate error otherwise
    393. 

  393.  */
    394. 

  394. hubbub_error hubbub_tokeniser_run(hubbub_tokeniser *tokeniser)
    395. 

  395. {
    396. 

  396. 	hubbub_error cont = HUBBUB_OK;
    397. 

  397. 

  398. 	if (tokeniser == NULL)
    399. 

  399. 		return HUBBUB_BADPARM;
    400. 

  400. 

  401. #if 0
    402. 

  402. #define state(x) \
    403. 

  403. 		case x: \
    404. 

  404. 			printf( #x "\n");
    405. 

  405. #else
    406. 

  406. #define state(x) \
    407. 

  407. 		case x:
    408. 

  408. #endif
    409. 

  409. 

  410. 	while (cont == HUBBUB_OK) {
    411. 

  411. 		switch (tokeniser->state) {
    412. 

  412. 		state(STATE_DATA)
    413. 

  413. 			cont = hubbub_tokeniser_handle_data(tokeniser);
    414. 

  414. 			break;
    415. 

  415. 		state(STATE_CHARACTER_REFERENCE_DATA)
    416. 

  416. 			cont = hubbub_tokeniser_handle_character_reference_data(
    417. 

  417. 					tokeniser);
    418. 

  418. 			break;
    419. 

  419. 		state(STATE_TAG_OPEN)
    420. 

  420. 			cont = hubbub_tokeniser_handle_tag_open(tokeniser);
    421. 

  421. 			break;
    422. 

  422. 		state(STATE_CLOSE_TAG_OPEN)
    423. 

  423. 			cont = hubbub_tokeniser_handle_close_tag_open(
    424. 

  424. 					tokeniser);
    425. 

  425. 			break;
    426. 

  426. 		state(STATE_TAG_NAME)
    427. 

  427. 			cont = hubbub_tokeniser_handle_tag_name(tokeniser);
    428. 

  428. 			break;
    429. 

  429. 		state(STATE_BEFORE_ATTRIBUTE_NAME)
    430. 

  430. 			cont = hubbub_tokeniser_handle_before_attribute_name(
    431. 

  431. 					tokeniser);
    432. 

  432. 			break;
    433. 

  433. 		state(STATE_ATTRIBUTE_NAME)
    434. 

  434. 			cont = hubbub_tokeniser_handle_attribute_name(
    435. 

  435. 					tokeniser);
    436. 

  436. 			break;
    437. 

  437. 		state(STATE_AFTER_ATTRIBUTE_NAME)
    438. 

  438. 			cont = hubbub_tokeniser_handle_after_attribute_name(
    439. 

  439. 					tokeniser);
    440. 

  440. 			break;
    441. 

  441. 		state(STATE_BEFORE_ATTRIBUTE_VALUE)
    442. 

  442. 			cont = hubbub_tokeniser_handle_before_attribute_value(
    443. 

  443. 					tokeniser);
    444. 

  444. 			break;
    445. 

  445. 		state(STATE_ATTRIBUTE_VALUE_DQ)
    446. 

  446. 			cont = hubbub_tokeniser_handle_attribute_value_dq(
    447. 

  447. 					tokeniser);
    448. 

  448. 			break;
    449. 

  449. 		state(STATE_ATTRIBUTE_VALUE_SQ)
    450. 

  450. 			cont = hubbub_tokeniser_handle_attribute_value_sq(
    451. 

  451. 					tokeniser);
    452. 

  452. 			break;
    453. 

  453. 		state(STATE_ATTRIBUTE_VALUE_UQ)
    454. 

  454. 			cont = hubbub_tokeniser_handle_attribute_value_uq(
    455. 

  455. 					tokeniser);
    456. 

  456. 			break;
    457. 

  457. 		state(STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE)
    458. 

  458. 			cont = hubbub_tokeniser_handle_character_reference_in_attribute_value(
    459. 

  459. 					tokeniser);
    460. 

  460. 			break;
    461. 

  461. 		state(STATE_AFTER_ATTRIBUTE_VALUE_Q)
    462. 

  462. 			cont = hubbub_tokeniser_handle_after_attribute_value_q(
    463. 

  463. 					tokeniser);
    464. 

  464. 			break;
    465. 

  465. 		state(STATE_SELF_CLOSING_START_TAG)
    466. 

  466. 			cont = hubbub_tokeniser_handle_self_closing_start_tag(
    467. 

  467. 					tokeniser);
    468. 

  468. 			break;
    469. 

  469. 		state(STATE_BOGUS_COMMENT)
    470. 

  470. 			cont = hubbub_tokeniser_handle_bogus_comment(
    471. 

  471. 					tokeniser);
    472. 

  472. 			break;
    473. 

  473. 		state(STATE_MARKUP_DECLARATION_OPEN)
    474. 

  474. 			cont = hubbub_tokeniser_handle_markup_declaration_open(
    475. 

  475. 					tokeniser);
    476. 

  476. 			break;
    477. 

  477. 		state(STATE_MATCH_COMMENT)
    478. 

  478. 			cont = hubbub_tokeniser_handle_match_comment(
    479. 

  479. 					tokeniser);
    480. 

  480. 			break;
    481. 

  481. 		case STATE_COMMENT_START:
    482. 

  482. 		case STATE_COMMENT_START_DASH:
    483. 

  483. 		case STATE_COMMENT:
    484. 

  484. 		case STATE_COMMENT_END_DASH:
    485. 

  485. 		case STATE_COMMENT_END:
    486. 

  486. 			cont = hubbub_tokeniser_handle_comment(tokeniser);
    487. 

  487. 			break;
    488. 

  488. 		state(STATE_MATCH_DOCTYPE)
    489. 

  489. 			cont = hubbub_tokeniser_handle_match_doctype(
    490. 

  490. 					tokeniser);
    491. 

  491. 			break;
    492. 

  492. 		state(STATE_DOCTYPE)
    493. 

  493. 			cont = hubbub_tokeniser_handle_doctype(tokeniser);
    494. 

  494. 			break;
    495. 

  495. 		state(STATE_BEFORE_DOCTYPE_NAME)
    496. 

  496. 			cont = hubbub_tokeniser_handle_before_doctype_name(
    497. 

  497. 					tokeniser);
    498. 

  498. 			break;
    499. 

  499. 		state(STATE_DOCTYPE_NAME)
    500. 

  500. 			cont = hubbub_tokeniser_handle_doctype_name(
    501. 

  501. 					tokeniser);
    502. 

  502. 			break;
    503. 

  503. 		state(STATE_AFTER_DOCTYPE_NAME)
    504. 

  504. 			cont = hubbub_tokeniser_handle_after_doctype_name(
    505. 

  505. 					tokeniser);
    506. 

  506. 			break;
    507. 

  507. 

  508. 		state(STATE_MATCH_PUBLIC)
    509. 

  509. 			cont = hubbub_tokeniser_handle_match_public(
    510. 

  510. 					tokeniser);
    511. 

  511. 			break;
    512. 

  512. 		state(STATE_BEFORE_DOCTYPE_PUBLIC)
    513. 

  513. 			cont = hubbub_tokeniser_handle_before_doctype_public(
    514. 

  514. 					tokeniser);
    515. 

  515. 			break;
    516. 

  516. 		state(STATE_DOCTYPE_PUBLIC_DQ)
    517. 

  517. 			cont = hubbub_tokeniser_handle_doctype_public_dq(
    518. 

  518. 					tokeniser);
    519. 

  519. 			break;
    520. 

  520. 		state(STATE_DOCTYPE_PUBLIC_SQ)
    521. 

  521. 			cont = hubbub_tokeniser_handle_doctype_public_sq(
    522. 

  522. 					tokeniser);
    523. 

  523. 			break;
    524. 

  524. 		state(STATE_AFTER_DOCTYPE_PUBLIC)
    525. 

  525. 			cont = hubbub_tokeniser_handle_after_doctype_public(
    526. 

  526. 					tokeniser);
    527. 

  527. 			break;
    528. 

  528. 		state(STATE_MATCH_SYSTEM)
    529. 

  529. 			cont = hubbub_tokeniser_handle_match_system(
    530. 

  530. 					tokeniser);
    531. 

  531. 			break;
    532. 

  532. 		state(STATE_BEFORE_DOCTYPE_SYSTEM)
    533. 

  533. 			cont = hubbub_tokeniser_handle_before_doctype_system(
    534. 

  534. 					tokeniser);
    535. 

  535. 			break;
    536. 

  536. 		state(STATE_DOCTYPE_SYSTEM_DQ)
    537. 

  537. 			cont = hubbub_tokeniser_handle_doctype_system_dq(
    538. 

  538. 					tokeniser);
    539. 

  539. 			break;
    540. 

  540. 		state(STATE_DOCTYPE_SYSTEM_SQ)
    541. 

  541. 			cont = hubbub_tokeniser_handle_doctype_system_sq(
    542. 

  542. 					tokeniser);
    543. 

  543. 			break;
    544. 

  544. 		state(STATE_AFTER_DOCTYPE_SYSTEM)
    545. 

  545. 			cont = hubbub_tokeniser_handle_after_doctype_system(
    546. 

  546. 					tokeniser);
    547. 

  547. 			break;
    548. 

  548. 		state(STATE_BOGUS_DOCTYPE)
    549. 

  549. 			cont = hubbub_tokeniser_handle_bogus_doctype(
    550. 

  550. 					tokeniser);
    551. 

  551. 			break;
    552. 

  552. 		state(STATE_MATCH_CDATA)
    553. 

  553. 			cont = hubbub_tokeniser_handle_match_cdata(
    554. 

  554. 					tokeniser);
    555. 

  555. 			break;
    556. 

  556. 		state(STATE_CDATA_BLOCK)
    557. 

  557. 			cont = hubbub_tokeniser_handle_cdata_block(
    558. 

  558. 					tokeniser);
    559. 

  559. 			break;
    560. 

  560. 		state(STATE_NUMBERED_ENTITY)
    561. 

  561. 			cont = hubbub_tokeniser_handle_numbered_entity(
    562. 

  562. 					tokeniser);
    563. 

  563. 			break;
    564. 

  564. 		state(STATE_NAMED_ENTITY)
    565. 

  565. 			cont = hubbub_tokeniser_handle_named_entity(
    566. 

  566. 					tokeniser);
    567. 

  567. 			break;
    568. 

  568. 		}
    569. 

  569. 	}
    570. 

  570. 

  571. 	return (cont == HUBBUB_NEEDDATA) ? HUBBUB_OK : cont;
    572. 

  572. }
    573. 

  573. 

  574. 

  575. /**
    576. 

  576.  * Various macros for manipulating buffers.
    577. 

  577.  *
    578. 

  578.  * \todo make some of these inline functions (type-safety)
    579. 

  579.  * \todo document them properly here
    580. 

  580.  */
    581. 

  581. 

  582. #define START_BUF(str, cptr, length) \
    583. 

  583. 	do { \
    584. 

  584. 		parserutils_error perror; \
    585. 

  585. 		perror = parserutils_buffer_append(tokeniser->buffer, \
    586. 

  586. 				(uint8_t *) (cptr), (length)); \
    587. 

  587. 		if (perror != PARSERUTILS_OK) \
    588. 

  588. 			return hubbub_error_from_parserutils_error(perror); \
    589. 

  589. 		(str).len = (length); \
    590. 

  590. 	} while (0)
    591. 

  591. 

  592. #define COLLECT(str, cptr, length) \
    593. 

  593. 	do { \
    594. 

  594. 		parserutils_error perror; \
    595. 

  595. 		assert(str.len != 0); \
    596. 

  596. 		perror = parserutils_buffer_append(tokeniser->buffer, \
    597. 

  597. 				(uint8_t *) (cptr), (length)); \
    598. 

  598. 		if (perror != PARSERUTILS_OK) \
    599. 

  599. 			return hubbub_error_from_parserutils_error(perror); \
    600. 

  600. 		(str).len += (length); \
    601. 

  601. 	} while (0)
    602. 

  602. 

  603. #define COLLECT_MS(str, cptr, length) \
    604. 

  604. 	do { \
    605. 

  605. 		parserutils_error perror; \
    606. 

  606. 		perror = parserutils_buffer_append(tokeniser->buffer, \
    607. 

  607. 				(uint8_t *) (cptr), (length)); \
    608. 

  608. 		if (perror != PARSERUTILS_OK) \
    609. 

  609. 			return hubbub_error_from_parserutils_error(perror); \
    610. 

  610. 		(str).len += (length); \
    611. 

  611. 	} while (0)
    612. 

  612. 

  613. 

  614. /* this should always be called with an empty "chars" buffer */
    615. 

  615. hubbub_error hubbub_tokeniser_handle_data(hubbub_tokeniser *tokeniser)
    616. 

  616. {
    617. 

  617. 	parserutils_error error;
    618. 

  618. 	hubbub_token token;
    619. 

  619. 	const uint8_t *cptr;
    620. 

  620. 	size_t len;
    621. 

  621. 

  622. 	while ((error = parserutils_inputstream_peek(tokeniser->input,
    623. 

  623. 			tokeniser->context.pending, &cptr, &len)) ==
    624. 

  624. 					PARSERUTILS_OK) {
    625. 

  625. 		const uint8_t c = *cptr;
    626. 

  626. 

  627. 		if (c == '&' &&
    628. 

  628. 				(tokeniser->content_model == HUBBUB_CONTENT_MODEL_PCDATA ||
    629. 

  629. 				tokeniser->content_model == HUBBUB_CONTENT_MODEL_RCDATA) &&
    630. 

  630. 				tokeniser->escape_flag == false) {
    631. 

  631. 			tokeniser->state =
    632. 

  632. 					STATE_CHARACTER_REFERENCE_DATA;
    633. 

  633. 			/* Don't eat the '&'; it'll be handled by entity
    634. 

  634. 			 * consumption */
    635. 

  635. 			break;
    636. 

  636. 		} else if (c == '-' &&
    637. 

  637. 				tokeniser->escape_flag == false &&
    638. 

  638. 				(tokeniser->content_model ==
    639. 

  639. 						HUBBUB_CONTENT_MODEL_RCDATA ||
    640. 

  640. 				tokeniser->content_model ==
    641. 

  641. 						HUBBUB_CONTENT_MODEL_CDATA) &&
    642. 

  642. 				tokeniser->context.pending >= 3) {
    643. 

  643. 			size_t ignore;
    644. 

  644. 			error = parserutils_inputstream_peek(
    645. 

  645. 					tokeniser->input,
    646. 

  646. 					tokeniser->context.pending - 3,
    647. 

  647. 					&cptr,
    648. 

  648. 					&ignore);
    649. 

  649. 

  650. 			assert(error == PARSERUTILS_OK);
    651. 

  651. 

  652. 			if (strncmp((char *)cptr,
    653. 

  653. 					"<!--", SLEN("<!--")) == 0) {
    654. 

  654. 				tokeniser->escape_flag = true;
    655. 

  655. 			}
    656. 

  656. 

  657. 			tokeniser->context.pending += len;
    658. 

  658. 		} else if (c == '<' && (tokeniser->content_model ==
    659. 

  659. 						HUBBUB_CONTENT_MODEL_PCDATA ||
    660. 

  660. 					((tokeniser->content_model ==
    661. 

  661. 						HUBBUB_CONTENT_MODEL_RCDATA ||
    662. 

  662. 					tokeniser->content_model ==
    663. 

  663. 						HUBBUB_CONTENT_MODEL_CDATA) &&
    664. 

  664. 				tokeniser->escape_flag == false))) {
    665. 

  665. 			if (tokeniser->context.pending > 0) {
    666. 

  666. 				/* Emit any pending characters */
    667. 

  667. 				emit_current_chars(tokeniser);
    668. 

  668. 			}
    669. 

  669. 

  670. 			/* Buffer '<' */
    671. 

  671. 			tokeniser->context.pending = len;
    672. 

  672. 			tokeniser->state = STATE_TAG_OPEN;
    673. 

  673. 			break;
    674. 

  674. 		} else if (c == '>' && tokeniser->escape_flag == true &&
    675. 

  675. 				(tokeniser->content_model ==
    676. 

  676. 						HUBBUB_CONTENT_MODEL_RCDATA ||
    677. 

  677. 				tokeniser->content_model ==
    678. 

  678. 						HUBBUB_CONTENT_MODEL_CDATA)) {
    679. 

  679. 			/* no need to check that there are enough characters,
    680. 

  680. 			 * since you can only run into this if the flag is
    681. 

  681. 			 * true in the first place, which requires four
    682. 

  682. 			 * characters. */
    683. 

  683. 			error = parserutils_inputstream_peek(
    684. 

  684. 					tokeniser->input,
    685. 

  685. 					tokeniser->context.pending - 2,
    686. 

  686. 					&cptr,
    687. 

  687. 					&len);
    688. 

  688. 

  689. 			assert(error == PARSERUTILS_OK);
    690. 

  690. 

  691. 			if (strncmp((char *) cptr, "-->", SLEN("-->")) == 0) {
    692. 

  692. 				tokeniser->escape_flag = false;
    693. 

  693. 			}
    694. 

  694. 

  695. 			tokeniser->context.pending += len;
    696. 

  696. 		} else if (c == '\0') {
    697. 

  697. 			if (tokeniser->context.pending > 0) {
    698. 

  698. 				/* Emit any pending characters */
    699. 

  699. 				emit_current_chars(tokeniser);
    700. 

  700. 			}
    701. 

  701. 

  702. 			/* Emit a replacement character */
    703. 

  703. 			emit_character_token(tokeniser, &u_fffd_str);
    704. 

  704. 

  705. 			/* Advance past NUL */
    706. 

  706. 			parserutils_inputstream_advance(tokeniser->input, 1);
    707. 

  707. 		} else if (c == '\r') {
    708. 

  708. 			error = parserutils_inputstream_peek(
    709. 

  709. 					tokeniser->input,
    710. 

  710. 					tokeniser->context.pending + len,
    711. 

  711. 					&cptr,
    712. 

  712. 					&len);
    713. 

  713. 

  714. 			if (error != PARSERUTILS_OK && 
    715. 

  715. 					error != PARSERUTILS_EOF) {
    716. 

  716. 				break;
    717. 

  717. 			}
    718. 

  718. 

  719. 			if (tokeniser->context.pending > 0) {
    720. 

  720. 				/* Emit any pending characters */
    721. 

  721. 				emit_current_chars(tokeniser);
    722. 

  722. 			}
    723. 

  723. 

  724. 			if (error == PARSERUTILS_EOF ||	*cptr != '\n') {
    725. 

  725. 				/* Emit newline */
    726. 

  726. 				emit_character_token(tokeniser, &lf_str);
    727. 

  727. 			}
    728. 

  728. 

  729. 			/* Advance over */
    730. 

  730. 			parserutils_inputstream_advance(tokeniser->input, 1);
    731. 

  731. 		} else {
    732. 

  732. 			/* Just collect into buffer */
    733. 

  733. 			tokeniser->context.pending += len;
    734. 

  734. 		}
    735. 

  735. 	}
    736. 

  736. 

  737. 	if (tokeniser->state != STATE_TAG_OPEN &&
    738. 

  738. 		(tokeniser->state != STATE_DATA || error == PARSERUTILS_EOF) &&
    739. 

  739. 			tokeniser->context.pending > 0) {
    740. 

  740. 		/* Emit any pending characters */
    741. 

  741. 		emit_current_chars(tokeniser);
    742. 

  742. 	}
    743. 

  743. 

  744. 	if (error == PARSERUTILS_EOF) {
    745. 

  745. 		token.type = HUBBUB_TOKEN_EOF;
    746. 

  746. 		hubbub_tokeniser_emit_token(tokeniser, &token);
    747. 

  747. 	}
    748. 

  748. 

  749. 	if (error == PARSERUTILS_EOF) {
    750. 

  750. 		return HUBBUB_NEEDDATA;
    751. 

  751. 	} else {
    752. 

  752. 		return hubbub_error_from_parserutils_error(error);
    753. 

  753. 	}
    754. 

  754. }
    755. 

  755. 

  756. /* emit any pending tokens before calling */
    757. 

  757. hubbub_error hubbub_tokeniser_handle_character_reference_data(
    758. 

  758. 		hubbub_tokeniser *tokeniser)
    759. 

  759. {
    760. 

  760. 	assert(tokeniser->context.pending == 0);
    761. 

  761. 

  762. 	if (tokeniser->context.match_entity.complete == false) {
    763. 

  763. 		return hubbub_tokeniser_consume_character_reference(tokeniser,
    764. 

  764. 				tokeniser->context.pending);
    765. 

  765. 	} else {
    766. 

  766. 		hubbub_token token;
    767. 

  767. 

  768. 		uint8_t utf8[6];
    769. 

  769. 		uint8_t *utf8ptr = utf8;
    770. 

  770. 		size_t len = sizeof(utf8);
    771. 

  771. 

  772. 		token.type = HUBBUB_TOKEN_CHARACTER;
    773. 

  773. 

  774. 		if (tokeniser->context.match_entity.codepoint) {
    775. 

  775. 			parserutils_charset_utf8_from_ucs4(
    776. 

  776. 				tokeniser->context.match_entity.codepoint,
    777. 

  777. 				&utf8ptr, &len);
    778. 

  778. 

  779. 			token.data.character.ptr = utf8;
    780. 

  780. 			token.data.character.len = sizeof(utf8) - len;
    781. 

  781. 

  782. 			hubbub_tokeniser_emit_token(tokeniser, &token);
    783. 

  783. 

  784. 			/* +1 for ampersand */
    785. 

  785. 			parserutils_inputstream_advance(tokeniser->input,
    786. 

  786. 					tokeniser->context.match_entity.length
    787. 

  787. 							+ 1);
    788. 

  788. 		} else {
    789. 

  789. 			parserutils_error error;
    790. 

  790. 			const uint8_t *cptr = NULL;
    791. 

  791. 			error = parserutils_inputstream_peek(
    792. 

  792. 					tokeniser->input,
    793. 

  793. 					tokeniser->context.pending,
    794. 

  794. 					&cptr,
    795. 

  795. 					&len);
    796. 

  796. 

  797. 			assert(error == PARSERUTILS_OK);
    798. 

  798. 

  799. 			token.data.character.ptr = cptr;
    800. 

  800. 			token.data.character.len = len;
    801. 

  801. 

  802. 			hubbub_tokeniser_emit_token(tokeniser, &token);
    803. 

  803. 			parserutils_inputstream_advance(tokeniser->input, len);
    804. 

  804. 		}
    805. 

  805. 

  806. 		/* Reset for next time */
    807. 

  807. 		tokeniser->context.match_entity.complete = false;
    808. 

  808. 

  809. 		tokeniser->state = STATE_DATA;
    810. 

  810. 	}
    811. 

  811. 

  812. 	return HUBBUB_OK;
    813. 

  813. }
    814. 

  814. 

  815. /* this state always switches to another state straight away */
    816. 

  816. /* this state expects the current character to be '<' */
    817. 

  817. hubbub_error hubbub_tokeniser_handle_tag_open(hubbub_tokeniser *tokeniser)
    818. 

  818. {
    819. 

  819. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    820. 

  820. 

  821. 	size_t len;
    822. 

  822. 	const uint8_t *cptr;
    823. 

  823. 	parserutils_error error;
    824. 

  824. 	uint8_t c;
    825. 

  825. 

  826. 	assert(tokeniser->context.pending == 1);
    827. 

  827. /*	assert(tokeniser->context.chars.ptr[0] == '<'); */
    828. 

  828. 

  829. 	error = parserutils_inputstream_peek(tokeniser->input, 
    830. 

  830. 			tokeniser->context.pending, &cptr, &len);
    831. 

  831. 

  832. 	if (error != PARSERUTILS_OK) {
    833. 

  833. 		if (error == PARSERUTILS_EOF) {
    834. 

  834. 			/* Return to data state with '<' still in "chars" */
    835. 

  835. 			tokeniser->state = STATE_DATA;
    836. 

  836. 			return HUBBUB_OK;
    837. 

  837. 		} else {
    838. 

  838. 			return hubbub_error_from_parserutils_error(error);
    839. 

  839. 		}
    840. 

  840. 	}
    841. 

  841. 

  842. 	c = *cptr;
    843. 

  843. 

  844. 	if (c == '/') {
    845. 

  845. 		tokeniser->context.pending += len;
    846. 

  846. 

  847. 		tokeniser->context.close_tag_match.match = false;
    848. 

  848. 		tokeniser->context.close_tag_match.count = 0;
    849. 

  849. 

  850. 		tokeniser->state = STATE_CLOSE_TAG_OPEN;
    851. 

  851. 	} else if (tokeniser->content_model == HUBBUB_CONTENT_MODEL_RCDATA ||
    852. 

  852. 			tokeniser->content_model ==
    853. 

  853. 					HUBBUB_CONTENT_MODEL_CDATA) {
    854. 

  854. 		/* Return to data state with '<' still in "chars" */
    855. 

  855. 		tokeniser->state = STATE_DATA;
    856. 

  856. 	} else if (tokeniser->content_model == HUBBUB_CONTENT_MODEL_PCDATA) {
    857. 

  857. 		if (c == '!') {
    858. 

  858. 			parserutils_inputstream_advance(tokeniser->input,
    859. 

  859. 					SLEN("<!"));
    860. 

  860. 

  861. 			tokeniser->context.pending = 0;
    862. 

  862. 			tokeniser->state = STATE_MARKUP_DECLARATION_OPEN;
    863. 

  863. 		} else if ('A' <= c && c <= 'Z') {
    864. 

  864. 			uint8_t lc = (c + 0x20);
    865. 

  865. 

  866. 			START_BUF(ctag->name, &lc, len);
    867. 

  867. 			ctag->n_attributes = 0;
    868. 

  868. 			tokeniser->context.current_tag_type =
    869. 

  869. 					HUBBUB_TOKEN_START_TAG;
    870. 

  870. 

  871. 			tokeniser->context.pending += len;
    872. 

  872. 

  873. 			tokeniser->state = STATE_TAG_NAME;
    874. 

  874. 		} else if ('a' <= c && c <= 'z') {
    875. 

  875. 			START_BUF(ctag->name, cptr, len);
    876. 

  876. 			ctag->n_attributes = 0;
    877. 

  877. 			tokeniser->context.current_tag_type =
    878. 

  878. 					HUBBUB_TOKEN_START_TAG;
    879. 

  879. 

  880. 			tokeniser->context.pending += len;
    881. 

  881. 

  882. 			tokeniser->state = STATE_TAG_NAME;
    883. 

  883. 		} else if (c == '>') {
    884. 

  884. 			/** \todo parse error */
    885. 

  885. 

  886. 			tokeniser->context.pending += len;
    887. 

  887. 			tokeniser->state = STATE_DATA;
    888. 

  888. 		} else if (c == '?') {
    889. 

  889. 			/** \todo parse error */
    890. 

  890. 

  891. 			/* Cursor still at "<", need to advance past it */
    892. 

  892. 			parserutils_inputstream_advance(
    893. 

  893. 					tokeniser->input, SLEN("<"));
    894. 

  894. 			tokeniser->context.pending = 0;
    895. 

  895. 

  896. 			tokeniser->state = STATE_BOGUS_COMMENT;
    897. 

  897. 		} else {
    898. 

  898. 			/* Return to data state with '<' still in "chars" */
    899. 

  899. 			tokeniser->state = STATE_DATA;
    900. 

  900. 		}
    901. 

  901. 	}
    902. 

  902. 

  903. 	return HUBBUB_OK;
    904. 

  904. }
    905. 

  905. 

  906. /* this state expects tokeniser->context.chars to be "</" */
    907. 

  907. /* this state never stays in this state for more than one character */
    908. 

  908. hubbub_error hubbub_tokeniser_handle_close_tag_open(hubbub_tokeniser *tokeniser)
    909. 

  909. {
    910. 

  910. 	hubbub_tokeniser_context *ctx = &tokeniser->context;
    911. 

  911. 

  912. 	size_t len;
    913. 

  913. 	const uint8_t *cptr;
    914. 

  914. 	parserutils_error error;
    915. 

  915. 	uint8_t c;
    916. 

  916. 

  917. 	assert(tokeniser->context.pending == 2);
    918. 

  918. /*	assert(tokeniser->context.chars.ptr[0] == '<'); */
    919. 

  919. /*	assert(tokeniser->context.chars.ptr[1] == '/'); */
    920. 

  920. 

  921. 	/**\todo fragment case */
    922. 

  922. 

  923. 	if (tokeniser->content_model == HUBBUB_CONTENT_MODEL_RCDATA ||
    924. 

  924. 			tokeniser->content_model ==
    925. 

  925. 					HUBBUB_CONTENT_MODEL_CDATA) {
    926. 

  926. 		uint8_t *start_tag_name =
    927. 

  927. 			tokeniser->context.last_start_tag_name;
    928. 

  928. 		size_t start_tag_len =
    929. 

  929. 			tokeniser->context.last_start_tag_len;
    930. 

  930. 

  931. 		while ((error = parserutils_inputstream_peek(tokeniser->input,
    932. 

  932. 					ctx->pending +
    933. 

  933. 						ctx->close_tag_match.count,
    934. 

  934. 					&cptr,
    935. 

  935. 					&len)) == PARSERUTILS_OK) {
    936. 

  936. 			c = *cptr;
    937. 

  937. 

  938. 			if ((start_tag_name[ctx->close_tag_match.count] & ~0x20)
    939. 

  939. 					!= (c & ~0x20)) {
    940. 

  940. 				break;
    941. 

  941. 			}
    942. 

  942. 

  943. 			ctx->close_tag_match.count += len;
    944. 

  944. 

  945. 			if (ctx->close_tag_match.count == start_tag_len) {
    946. 

  946. 				ctx->close_tag_match.match = true;
    947. 

  947. 				break;
    948. 

  948. 			}
    949. 

  949. 		}
    950. 

  950. 

  951. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    952. 

  952. 			return hubbub_error_from_parserutils_error(error);
    953. 

  953. 		}
    954. 

  954. 

  955. 		if (ctx->close_tag_match.match == true) {
    956. 

  956. 			error = parserutils_inputstream_peek(
    957. 

  957. 			 		tokeniser->input,
    958. 

  958. 			 		ctx->pending +
    959. 

  959. 				 		ctx->close_tag_match.count,
    960. 

  960. 					&cptr,
    961. 

  961. 			 		&len);
    962. 

  962. 

  963. 			if (error != PARSERUTILS_OK && 
    964. 

  964. 					error != PARSERUTILS_EOF) {
    965. 

  965. 				return hubbub_error_from_parserutils_error(
    966. 

  966. 						error);
    967. 

  967. 			} else if (error != PARSERUTILS_EOF) {
    968. 

  968. 				c = *cptr;
    969. 

  969. 

  970. 				if (c != '\t' && c != '\n' && c != '\f' &&
    971. 

  971. 						c != ' ' && c != '>' &&
    972. 

  972. 						c != '/') {
    973. 

  973. 					ctx->close_tag_match.match = false;
    974. 

  974. 				}
    975. 

  975. 			}
    976. 

  976. 		}
    977. 

  977. 	}
    978. 

  978. 

  979. 	if (ctx->close_tag_match.match == false &&
    980. 

  980. 			tokeniser->content_model !=
    981. 

  981. 					HUBBUB_CONTENT_MODEL_PCDATA) {
    982. 

  982. 		/* We should emit "</" here, but instead we leave it in the
    983. 

  983. 		 * buffer so the data state emits it with any characters
    984. 

  984. 		 * following it */
    985. 

  985. 		tokeniser->state = STATE_DATA;
    986. 

  986. 	} else {
    987. 

  987. 		error = parserutils_inputstream_peek(tokeniser->input,
    988. 

  988. 				tokeniser->context.pending, &cptr, &len);
    989. 

  989. 

  990. 		if (error == PARSERUTILS_EOF) {
    991. 

  991. 			/** \todo parse error */
    992. 

  992. 

  993. 			/* Return to data state with "</" pending */
    994. 

  994. 			tokeniser->state = STATE_DATA;
    995. 

  995. 			return HUBBUB_OK;
    996. 

  996. 		} else if (error != PARSERUTILS_OK) {
    997. 

  997. 			return hubbub_error_from_parserutils_error(error);
    998. 

  998. 		}
    999. 

  999. 

  1000. 		c = *cptr;
    1001. 

  1001. 

  1002. 		if ('A' <= c && c <= 'Z') {
    1003. 

  1003. 			uint8_t lc = (c + 0x20);
    1004. 

  1004. 			START_BUF(tokeniser->context.current_tag.name,
    1005. 

  1005. 					&lc, len);
    1006. 

  1006. 			tokeniser->context.current_tag.n_attributes = 0;
    1007. 

  1007. 

  1008. 			tokeniser->context.current_tag_type =
    1009. 

  1009. 					HUBBUB_TOKEN_END_TAG;
    1010. 

  1010. 

  1011. 			tokeniser->context.pending += len;
    1012. 

  1012. 

  1013. 			tokeniser->state = STATE_TAG_NAME;
    1014. 

  1014. 		} else if ('a' <= c && c <= 'z') {
    1015. 

  1015. 			START_BUF(tokeniser->context.current_tag.name,
    1016. 

  1016. 					cptr, len);
    1017. 

  1017. 			tokeniser->context.current_tag.n_attributes = 0;
    1018. 

  1018. 

  1019. 			tokeniser->context.current_tag_type =
    1020. 

  1020. 					HUBBUB_TOKEN_END_TAG;
    1021. 

  1021. 

  1022. 			tokeniser->context.pending += len;
    1023. 

  1023. 

  1024. 			tokeniser->state = STATE_TAG_NAME;
    1025. 

  1025. 		} else if (c == '>') {
    1026. 

  1026. 			/* Cursor still at "</", need to collect ">" */
    1027. 

  1027. 			tokeniser->context.pending += len;
    1028. 

  1028. 

  1029. 			/* Now need to advance past "</>" */
    1030. 

  1030. 			parserutils_inputstream_advance(tokeniser->input,
    1031. 

  1031. 					tokeniser->context.pending);
    1032. 

  1032. 			tokeniser->context.pending = 0;
    1033. 

  1033. 

  1034. 			/** \todo parse error */
    1035. 

  1035. 			tokeniser->state = STATE_DATA;
    1036. 

  1036. 		} else {
    1037. 

  1037. 			/** \todo parse error */
    1038. 

  1038. 

  1039. 			/* Cursor still at "</", need to advance past it */
    1040. 

  1040. 			parserutils_inputstream_advance(tokeniser->input,
    1041. 

  1041. 					tokeniser->context.pending);
    1042. 

  1042. 			tokeniser->context.pending = 0;
    1043. 

  1043. 

  1044. 			tokeniser->state = STATE_BOGUS_COMMENT;
    1045. 

  1045. 		}
    1046. 

  1046. 	}
    1047. 

  1047. 

  1048. 	return HUBBUB_OK;
    1049. 

  1049. }
    1050. 

  1050. 

  1051. /* this state expects tokeniser->context.current_tag to already have its
    1052. 

  1052.    first character set */
    1053. 

  1053. hubbub_error hubbub_tokeniser_handle_tag_name(hubbub_tokeniser *tokeniser)
    1054. 

  1054. {
    1055. 

  1055. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1056. 

  1056. 

  1057. 	size_t len;
    1058. 

  1058. 	const uint8_t *cptr;
    1059. 

  1059. 	parserutils_error error;
    1060. 

  1060. 	uint8_t c;
    1061. 

  1061. 

  1062. 	assert(tokeniser->context.pending > 0);
    1063. 

  1063. /*	assert(tokeniser->context.chars.ptr[0] == '<'); */
    1064. 

  1064. 	assert(ctag->name.len > 0);
    1065. 

  1065. /*	assert(ctag->name.ptr); */
    1066. 

  1066. 

  1067. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1068. 

  1068. 			tokeniser->context.pending, &cptr, &len);
    1069. 

  1069. 

  1070. 	if (error != PARSERUTILS_OK) {
    1071. 

  1071. 		if (error == PARSERUTILS_EOF) {
    1072. 

  1072. 			tokeniser->state = STATE_DATA;
    1073. 

  1073. 			return emit_current_tag(tokeniser);
    1074. 

  1074. 		} else {
    1075. 

  1075. 			return hubbub_error_from_parserutils_error(error);
    1076. 

  1076. 		}
    1077. 

  1077. 	}
    1078. 

  1078. 

  1079. 	c = *cptr;
    1080. 

  1080. 

  1081. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1082. 

  1082. 		tokeniser->context.pending += len;
    1083. 

  1083. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_NAME;
    1084. 

  1084. 	} else if (c == '>') {
    1085. 

  1085. 		tokeniser->context.pending += len;
    1086. 

  1086. 		tokeniser->state = STATE_DATA;
    1087. 

  1087. 		return emit_current_tag(tokeniser);
    1088. 

  1088. 	} else if (c == '\0') {
    1089. 

  1089. 		COLLECT(ctag->name, u_fffd, sizeof(u_fffd));
    1090. 

  1090. 		tokeniser->context.pending += len;
    1091. 

  1091. 	} else if (c == '/') {
    1092. 

  1092. 		tokeniser->context.pending += len;
    1093. 

  1093. 		tokeniser->state = STATE_SELF_CLOSING_START_TAG;
    1094. 

  1094. 	} else if ('A' <= c && c <= 'Z') {
    1095. 

  1095. 		uint8_t lc = (c + 0x20);
    1096. 

  1096. 		COLLECT(ctag->name, &lc, len);
    1097. 

  1097. 		tokeniser->context.pending += len;
    1098. 

  1098. 	} else {
    1099. 

  1099. 		COLLECT(ctag->name, cptr, len);
    1100. 

  1100. 		tokeniser->context.pending += len;
    1101. 

  1101. 	}
    1102. 

  1102. 

  1103. 	return HUBBUB_OK;
    1104. 

  1104. }
    1105. 

  1105. 

  1106. hubbub_error hubbub_tokeniser_handle_before_attribute_name(
    1107. 

  1107. 		hubbub_tokeniser *tokeniser)
    1108. 

  1108. {
    1109. 

  1109. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1110. 

  1110. 

  1111. 	size_t len;
    1112. 

  1112. 	const uint8_t *cptr;
    1113. 

  1113. 	parserutils_error error;
    1114. 

  1114. 	uint8_t c;
    1115. 

  1115. 

  1116. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1117. 

  1117. 			tokeniser->context.pending, &cptr, &len);
    1118. 

  1118. 

  1119. 	if (error != PARSERUTILS_OK) {
    1120. 

  1120. 		if (error == PARSERUTILS_EOF) {
    1121. 

  1121. 			tokeniser->state = STATE_DATA;
    1122. 

  1122. 			return emit_current_tag(tokeniser);
    1123. 

  1123. 		} else {
    1124. 

  1124. 			return hubbub_error_from_parserutils_error(error);
    1125. 

  1125. 		}
    1126. 

  1126. 	}
    1127. 

  1127. 

  1128. 	c = *cptr;
    1129. 

  1129. 

  1130. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1131. 

  1131. 		/* pass over in silence */
    1132. 

  1132. 		tokeniser->context.pending += len;
    1133. 

  1133. 	} else if (c == '>') {
    1134. 

  1134. 		tokeniser->context.pending += len;
    1135. 

  1135. 		tokeniser->state = STATE_DATA;
    1136. 

  1136. 		return emit_current_tag(tokeniser);
    1137. 

  1137. 	} else if (c == '/') {
    1138. 

  1138. 		tokeniser->context.pending += len;
    1139. 

  1139. 		tokeniser->state = STATE_SELF_CLOSING_START_TAG;
    1140. 

  1140. 	} else {
    1141. 

  1141. 		hubbub_attribute *attr;
    1142. 

  1142. 

  1143. 		if (c == '"' || c == '\'' || c == '=') {
    1144. 

  1144. 			/** \todo parse error */
    1145. 

  1145. 		}
    1146. 

  1146. 

  1147. 		attr = tokeniser->alloc(ctag->attributes,
    1148. 

  1148. 				(ctag->n_attributes + 1) *
    1149. 

  1149. 					sizeof(hubbub_attribute),
    1150. 

  1150. 				tokeniser->alloc_pw);
    1151. 

  1151. 		if (attr == NULL)
    1152. 

  1152. 			return HUBBUB_NOMEM;
    1153. 

  1153. 

  1154. 		ctag->attributes = attr;
    1155. 

  1155. 

  1156. 		if ('A' <= c && c <= 'Z') {
    1157. 

  1157. 			uint8_t lc = (c + 0x20);
    1158. 

  1158. 			START_BUF(attr[ctag->n_attributes].name, &lc, len);
    1159. 

  1159. 		} else if (c == '\0') {
    1160. 

  1160. 			START_BUF(attr[ctag->n_attributes].name,
    1161. 

  1161. 					u_fffd, sizeof(u_fffd));
    1162. 

  1162. 		} else {
    1163. 

  1163. 			START_BUF(attr[ctag->n_attributes].name, cptr, len);
    1164. 

  1164. 		}
    1165. 

  1165. 

  1166. 		attr[ctag->n_attributes].ns = HUBBUB_NS_NULL;
    1167. 

  1167. 		attr[ctag->n_attributes].value.ptr = NULL;
    1168. 

  1168. 		attr[ctag->n_attributes].value.len = 0;
    1169. 

  1169. 

  1170. 		ctag->n_attributes++;
    1171. 

  1171. 

  1172. 		tokeniser->context.pending += len;
    1173. 

  1173. 		tokeniser->state = STATE_ATTRIBUTE_NAME;
    1174. 

  1174. 	}
    1175. 

  1175. 

  1176. 	return HUBBUB_OK;
    1177. 

  1177. }
    1178. 

  1178. 

  1179. hubbub_error hubbub_tokeniser_handle_attribute_name(hubbub_tokeniser *tokeniser)
    1180. 

  1180. {
    1181. 

  1181. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1182. 

  1182. 

  1183. 	size_t len;
    1184. 

  1184. 	const uint8_t *cptr;
    1185. 

  1185. 	parserutils_error error;
    1186. 

  1186. 	uint8_t c;
    1187. 

  1187. 

  1188. 	assert(ctag->attributes[ctag->n_attributes - 1].name.len > 0);
    1189. 

  1189. 

  1190. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1191. 

  1191. 			tokeniser->context.pending, &cptr, &len);
    1192. 

  1192. 

  1193. 	if (error != PARSERUTILS_OK) {
    1194. 

  1194. 		if (error == PARSERUTILS_EOF) {
    1195. 

  1195. 			tokeniser->state = STATE_DATA;
    1196. 

  1196. 			return emit_current_tag(tokeniser);
    1197. 

  1197. 		} else {
    1198. 

  1198. 			return hubbub_error_from_parserutils_error(error);
    1199. 

  1199. 		}
    1200. 

  1200. 	}
    1201. 

  1201. 

  1202. 	c = *cptr;
    1203. 

  1203. 

  1204. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1205. 

  1205. 		tokeniser->context.pending += len;
    1206. 

  1206. 		tokeniser->state = STATE_AFTER_ATTRIBUTE_NAME;
    1207. 

  1207. 	} else if (c == '=') {
    1208. 

  1208. 		tokeniser->context.pending += len;
    1209. 

  1209. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_VALUE;
    1210. 

  1210. 	} else if (c == '>') {
    1211. 

  1211. 		tokeniser->context.pending += len;
    1212. 

  1212. 		tokeniser->state = STATE_DATA;
    1213. 

  1213. 		return emit_current_tag(tokeniser);
    1214. 

  1214. 	} else if (c == '/') {
    1215. 

  1215. 		tokeniser->context.pending += len;
    1216. 

  1216. 		tokeniser->state = STATE_SELF_CLOSING_START_TAG;
    1217. 

  1217. 	} else if (c == '\0') {
    1218. 

  1218. 		COLLECT(ctag->attributes[ctag->n_attributes - 1].name,
    1219. 

  1219. 				u_fffd, sizeof(u_fffd));
    1220. 

  1220. 		tokeniser->context.pending += len;
    1221. 

  1221. 	} else if ('A' <= c && c <= 'Z') {
    1222. 

  1222. 		uint8_t lc = (c + 0x20);
    1223. 

  1223. 		COLLECT(ctag->attributes[ctag->n_attributes - 1].name,
    1224. 

  1224. 				&lc, len);
    1225. 

  1225. 		tokeniser->context.pending += len;
    1226. 

  1226. 	} else {
    1227. 

  1227. 		COLLECT(ctag->attributes[ctag->n_attributes - 1].name,
    1228. 

  1228. 				cptr, len);
    1229. 

  1229. 		tokeniser->context.pending += len;
    1230. 

  1230. 	}
    1231. 

  1231. 

  1232. 	return HUBBUB_OK;
    1233. 

  1233. }
    1234. 

  1234. 

  1235. hubbub_error hubbub_tokeniser_handle_after_attribute_name(
    1236. 

  1236. 		hubbub_tokeniser *tokeniser)
    1237. 

  1237. {
    1238. 

  1238. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1239. 

  1239. 

  1240. 	size_t len;
    1241. 

  1241. 	const uint8_t *cptr;
    1242. 

  1242. 	parserutils_error error;
    1243. 

  1243. 	uint8_t c;
    1244. 

  1244. 

  1245. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1246. 

  1246. 			tokeniser->context.pending, &cptr, &len);
    1247. 

  1247. 

  1248. 	if (error != PARSERUTILS_OK) {
    1249. 

  1249. 		if (error == PARSERUTILS_EOF) {
    1250. 

  1250. 			tokeniser->state = STATE_DATA;
    1251. 

  1251. 			return emit_current_tag(tokeniser);
    1252. 

  1252. 		} else {
    1253. 

  1253. 			return hubbub_error_from_parserutils_error(error);
    1254. 

  1254. 		}
    1255. 

  1255. 	}
    1256. 

  1256. 

  1257. 	c = *cptr;
    1258. 

  1258. 

  1259. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1260. 

  1260. 		tokeniser->context.pending += len;
    1261. 

  1261. 	} else if (c == '=') {
    1262. 

  1262. 		tokeniser->context.pending += len;
    1263. 

  1263. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_VALUE;
    1264. 

  1264. 	} else if (c == '>') {
    1265. 

  1265. 		tokeniser->context.pending += len;
    1266. 

  1266. 

  1267. 		tokeniser->state = STATE_DATA;
    1268. 

  1268. 		return emit_current_tag(tokeniser);
    1269. 

  1269. 	} else if (c == '/') {
    1270. 

  1270. 		tokeniser->context.pending += len;
    1271. 

  1271. 		tokeniser->state = STATE_SELF_CLOSING_START_TAG;
    1272. 

  1272. 	} else {
    1273. 

  1273. 		hubbub_attribute *attr;
    1274. 

  1274. 

  1275. 		if (c == '"' || c == '\'') {
    1276. 

  1276. 			/** \todo parse error */
    1277. 

  1277. 		}
    1278. 

  1278. 

  1279. 		attr = tokeniser->alloc(ctag->attributes,
    1280. 

  1280. 				(ctag->n_attributes + 1) *
    1281. 

  1281. 					sizeof(hubbub_attribute),
    1282. 

  1282. 				tokeniser->alloc_pw);
    1283. 

  1283. 		if (attr == NULL)
    1284. 

  1284. 			return HUBBUB_NOMEM;
    1285. 

  1285. 

  1286. 		ctag->attributes = attr;
    1287. 

  1287. 

  1288. 		if ('A' <= c && c <= 'Z') {
    1289. 

  1289. 			uint8_t lc = (c + 0x20);
    1290. 

  1290. 			START_BUF(attr[ctag->n_attributes].name, &lc, len);
    1291. 

  1291. 		} else if (c == '\0') {
    1292. 

  1292. 			START_BUF(attr[ctag->n_attributes].name,
    1293. 

  1293. 					u_fffd, sizeof(u_fffd));
    1294. 

  1294. 		} else {
    1295. 

  1295. 			START_BUF(attr[ctag->n_attributes].name, cptr, len);
    1296. 

  1296. 		}
    1297. 

  1297. 

  1298. 		attr[ctag->n_attributes].ns = HUBBUB_NS_NULL;
    1299. 

  1299. 		attr[ctag->n_attributes].value.ptr = NULL;
    1300. 

  1300. 		attr[ctag->n_attributes].value.len = 0;
    1301. 

  1301. 

  1302. 		ctag->n_attributes++;
    1303. 

  1303. 

  1304. 		tokeniser->context.pending += len;
    1305. 

  1305. 		tokeniser->state = STATE_ATTRIBUTE_NAME;
    1306. 

  1306. 	}
    1307. 

  1307. 

  1308. 	return HUBBUB_OK;
    1309. 

  1309. }
    1310. 

  1310. 

  1311. /* this state is only ever triggered by an '=' */
    1312. 

  1312. hubbub_error hubbub_tokeniser_handle_before_attribute_value(
    1313. 

  1313. 		hubbub_tokeniser *tokeniser)
    1314. 

  1314. {
    1315. 

  1315. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1316. 

  1316. 

  1317. 	size_t len;
    1318. 

  1318. 	const uint8_t *cptr;
    1319. 

  1319. 	parserutils_error error;
    1320. 

  1320. 	uint8_t c;
    1321. 

  1321. 

  1322. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1323. 

  1323. 			tokeniser->context.pending, &cptr, &len);
    1324. 

  1324. 

  1325. 	if (error != PARSERUTILS_OK) {
    1326. 

  1326. 		if (error == PARSERUTILS_EOF) {
    1327. 

  1327. 			/** \todo parse error */
    1328. 

  1328. 			tokeniser->state = STATE_DATA;
    1329. 

  1329. 			return emit_current_tag(tokeniser);
    1330. 

  1330. 		} else {
    1331. 

  1331. 			return hubbub_error_from_parserutils_error(error);
    1332. 

  1332. 		}
    1333. 

  1333. 	}
    1334. 

  1334. 

  1335. 	c = *cptr;
    1336. 

  1336. 

  1337. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1338. 

  1338. 		tokeniser->context.pending += len;
    1339. 

  1339. 	} else if (c == '"') {
    1340. 

  1340. 		tokeniser->context.pending += len;
    1341. 

  1341. 		tokeniser->state = STATE_ATTRIBUTE_VALUE_DQ;
    1342. 

  1342. 	} else if (c == '&') {
    1343. 

  1343. 		/* Don't consume the '&' -- reprocess in UQ state */
    1344. 

  1344. 		tokeniser->state = STATE_ATTRIBUTE_VALUE_UQ;
    1345. 

  1345. 	} else if (c == '\'') {
    1346. 

  1346. 		tokeniser->context.pending += len;
    1347. 

  1347. 		tokeniser->state = STATE_ATTRIBUTE_VALUE_SQ;
    1348. 

  1348. 	} else if (c == '>') {
    1349. 

  1349. 		/** \todo parse error */
    1350. 

  1350. 		tokeniser->context.pending += len;
    1351. 

  1351. 

  1352. 		tokeniser->state = STATE_DATA;
    1353. 

  1353. 		return emit_current_tag(tokeniser);
    1354. 

  1354. 	} else if (c == '\0') {
    1355. 

  1355. 		START_BUF(ctag->attributes[ctag->n_attributes - 1].value,
    1356. 

  1356. 				u_fffd, sizeof(u_fffd));
    1357. 

  1357. 		tokeniser->context.pending += len;
    1358. 

  1358. 		tokeniser->state = STATE_ATTRIBUTE_VALUE_UQ;
    1359. 

  1359. 	} else {
    1360. 

  1360. 		if (c == '=') {
    1361. 

  1361. 			/** \todo parse error */
    1362. 

  1362. 		}
    1363. 

  1363. 

  1364. 		START_BUF(ctag->attributes[ctag->n_attributes - 1].value,
    1365. 

  1365. 				cptr, len);
    1366. 

  1366. 

  1367. 		tokeniser->context.pending += len;
    1368. 

  1368. 		tokeniser->state = STATE_ATTRIBUTE_VALUE_UQ;
    1369. 

  1369. 	}
    1370. 

  1370. 

  1371. 	return HUBBUB_OK;
    1372. 

  1372. }
    1373. 

  1373. 

  1374. hubbub_error hubbub_tokeniser_handle_attribute_value_dq(
    1375. 

  1375. 		hubbub_tokeniser *tokeniser)
    1376. 

  1376. {
    1377. 

  1377. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1378. 

  1378. 

  1379. 	size_t len;
    1380. 

  1380. 	const uint8_t *cptr;
    1381. 

  1381. 	parserutils_error error;
    1382. 

  1382. 	uint8_t c;
    1383. 

  1383. 

  1384. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1385. 

  1385. 			tokeniser->context.pending, &cptr, &len);
    1386. 

  1386. 

  1387. 	if (error != PARSERUTILS_OK) {
    1388. 

  1388. 		if (error == PARSERUTILS_EOF) {
    1389. 

  1389. 			tokeniser->state = STATE_DATA;
    1390. 

  1390. 			return emit_current_tag(tokeniser);
    1391. 

  1391. 		} else {
    1392. 

  1392. 			return hubbub_error_from_parserutils_error(error);
    1393. 

  1393. 		}
    1394. 

  1394. 	}
    1395. 

  1395. 

  1396. 	c = *cptr;
    1397. 

  1397. 

  1398. 	if (c == '"') {
    1399. 

  1399. 		tokeniser->context.pending += len;
    1400. 

  1400. 		tokeniser->state = STATE_AFTER_ATTRIBUTE_VALUE_Q;
    1401. 

  1401. 	} else if (c == '&') {
    1402. 

  1402. 		tokeniser->context.prev_state = tokeniser->state;
    1403. 

  1403. 		tokeniser->state = STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE;
    1404. 

  1404. 		tokeniser->context.allowed_char = '"';
    1405. 

  1405. 		/* Don't eat the '&'; it'll be handled by entity consumption */
    1406. 

  1406. 	} else if (c == '\0') {
    1407. 

  1407. 		COLLECT_MS(ctag->attributes[ctag->n_attributes - 1].value,
    1408. 

  1408. 				u_fffd, sizeof(u_fffd));
    1409. 

  1409. 		tokeniser->context.pending += len;
    1410. 

  1410. 	} else if (c == '\r') {
    1411. 

  1411. 		error = parserutils_inputstream_peek(
    1412. 

  1412. 				tokeniser->input,
    1413. 

  1413. 				tokeniser->context.pending + len,
    1414. 

  1414. 				&cptr,
    1415. 

  1415. 				&len);
    1416. 

  1416. 

  1417. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    1418. 

  1418. 			return hubbub_error_from_parserutils_error(error);
    1419. 

  1419. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    1420. 

  1420. 			COLLECT_MS(ctag->attributes[
    1421. 

  1421. 					ctag->n_attributes - 1].value,
    1422. 

  1422. 					&lf, sizeof(lf));
    1423. 

  1423. 		}
    1424. 

  1424. 

  1425. 		/* Consume '\r' */
    1426. 

  1426. 		tokeniser->context.pending += 1;
    1427. 

  1427. 	} else {
    1428. 

  1428. 		COLLECT_MS(ctag->attributes[ctag->n_attributes - 1].value,
    1429. 

  1429. 				cptr, len);
    1430. 

  1430. 		tokeniser->context.pending += len;
    1431. 

  1431. 	}
    1432. 

  1432. 

  1433. 	return HUBBUB_OK;
    1434. 

  1434. }
    1435. 

  1435. 

  1436. hubbub_error hubbub_tokeniser_handle_attribute_value_sq(
    1437. 

  1437. 		hubbub_tokeniser *tokeniser)
    1438. 

  1438. {
    1439. 

  1439. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1440. 

  1440. 

  1441. 	size_t len;
    1442. 

  1442. 	const uint8_t *cptr;
    1443. 

  1443. 	parserutils_error error;
    1444. 

  1444. 	uint8_t c;
    1445. 

  1445. 

  1446. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1447. 

  1447. 			tokeniser->context.pending, &cptr, &len);
    1448. 

  1448. 

  1449. 	if (error != PARSERUTILS_OK) {
    1450. 

  1450. 		if (error == PARSERUTILS_EOF) {
    1451. 

  1451. 			tokeniser->state = STATE_DATA;
    1452. 

  1452. 			return emit_current_tag(tokeniser);
    1453. 

  1453. 		} else {
    1454. 

  1454. 			return hubbub_error_from_parserutils_error(error);
    1455. 

  1455. 		}
    1456. 

  1456. 	}
    1457. 

  1457. 

  1458. 	c = *cptr;
    1459. 

  1459. 

  1460. 	if (c == '\'') {
    1461. 

  1461. 		tokeniser->context.pending += len;
    1462. 

  1462. 		tokeniser->state = STATE_AFTER_ATTRIBUTE_VALUE_Q;
    1463. 

  1463. 	} else if (c == '&') {
    1464. 

  1464. 		tokeniser->context.prev_state = tokeniser->state;
    1465. 

  1465. 		tokeniser->state = STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE;
    1466. 

  1466. 		tokeniser->context.allowed_char = '\'';
    1467. 

  1467. 		/* Don't eat the '&'; it'll be handled by entity consumption */
    1468. 

  1468. 	} else if (c == '\0') {
    1469. 

  1469. 		COLLECT_MS(ctag->attributes[ctag->n_attributes - 1].value,
    1470. 

  1470. 				u_fffd, sizeof(u_fffd));
    1471. 

  1471. 		tokeniser->context.pending += len;
    1472. 

  1472. 	} else if (c == '\r') {
    1473. 

  1473. 		error = parserutils_inputstream_peek(
    1474. 

  1474. 				tokeniser->input,
    1475. 

  1475. 				tokeniser->context.pending + len,
    1476. 

  1476. 				&cptr,
    1477. 

  1477. 				&len);
    1478. 

  1478. 

  1479. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    1480. 

  1480. 			return hubbub_error_from_parserutils_error(error);
    1481. 

  1481. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    1482. 

  1482. 			COLLECT_MS(ctag->attributes[
    1483. 

  1483. 					ctag->n_attributes - 1].value,
    1484. 

  1484. 					&lf, sizeof(lf));
    1485. 

  1485. 		}
    1486. 

  1486. 

  1487. 		/* Consume \r */
    1488. 

  1488. 		tokeniser->context.pending += 1;
    1489. 

  1489. 	} else {
    1490. 

  1490. 		COLLECT_MS(ctag->attributes[ctag->n_attributes - 1].value,
    1491. 

  1491. 				cptr, len);
    1492. 

  1492. 		tokeniser->context.pending += len;
    1493. 

  1493. 	}
    1494. 

  1494. 

  1495. 	return HUBBUB_OK;
    1496. 

  1496. }
    1497. 

  1497. 

  1498. hubbub_error hubbub_tokeniser_handle_attribute_value_uq(
    1499. 

  1499. 		hubbub_tokeniser *tokeniser)
    1500. 

  1500. {
    1501. 

  1501. 	hubbub_tag *ctag = &tokeniser->context.current_tag;
    1502. 

  1502. 	uint8_t c;
    1503. 

  1503. 

  1504. 	size_t len;
    1505. 

  1505. 	const uint8_t *cptr;
    1506. 

  1506. 	parserutils_error error;
    1507. 

  1507. 

  1508. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1509. 

  1509. 			tokeniser->context.pending, &cptr, &len);
    1510. 

  1510. 

  1511. 	if (error != PARSERUTILS_OK) {
    1512. 

  1512. 		if (error == PARSERUTILS_EOF) {
    1513. 

  1513. 			tokeniser->state = STATE_DATA;
    1514. 

  1514. 			return emit_current_tag(tokeniser);
    1515. 

  1515. 		} else {
    1516. 

  1516. 			return hubbub_error_from_parserutils_error(error);
    1517. 

  1517. 		}
    1518. 

  1518. 	}
    1519. 

  1519. 

  1520. 	c = *cptr;
    1521. 

  1521. 

  1522. 	assert(c == '&' ||
    1523. 

  1523. 		ctag->attributes[ctag->n_attributes - 1].value.len >= 1);
    1524. 

  1524. 

  1525. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1526. 

  1526. 		tokeniser->context.pending += len;
    1527. 

  1527. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_NAME;
    1528. 

  1528. 	} else if (c == '&') {
    1529. 

  1529. 		tokeniser->context.prev_state = tokeniser->state;
    1530. 

  1530. 		tokeniser->state = STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE;
    1531. 

  1531. 		/* Don't eat the '&'; it'll be handled by entity consumption */
    1532. 

  1532. 	} else if (c == '>') {
    1533. 

  1533. 		tokeniser->context.pending += len;
    1534. 

  1534. 		tokeniser->state = STATE_DATA;
    1535. 

  1535. 		return emit_current_tag(tokeniser);
    1536. 

  1536. 	} else if (c == '\0') {
    1537. 

  1537. 		COLLECT(ctag->attributes[ctag->n_attributes - 1].value,
    1538. 

  1538. 				u_fffd, sizeof(u_fffd));
    1539. 

  1539. 		tokeniser->context.pending += len;
    1540. 

  1540. 	} else {
    1541. 

  1541. 		if (c == '"' || c == '\'' || c == '=') {
    1542. 

  1542. 			/** \todo parse error */
    1543. 

  1543. 		}
    1544. 

  1544. 

  1545. 		COLLECT(ctag->attributes[ctag->n_attributes - 1].value,
    1546. 

  1546. 				cptr, len);
    1547. 

  1547. 		tokeniser->context.pending += len;
    1548. 

  1548. 	}
    1549. 

  1549. 

  1550. 	return HUBBUB_OK;
    1551. 

  1551. }
    1552. 

  1552. 

  1553. hubbub_error hubbub_tokeniser_handle_character_reference_in_attribute_value(
    1554. 

  1554. 		hubbub_tokeniser *tokeniser)
    1555. 

  1555. {
    1556. 

  1556. 	if (tokeniser->context.match_entity.complete == false) {
    1557. 

  1557. 		return hubbub_tokeniser_consume_character_reference(tokeniser,
    1558. 

  1558. 				tokeniser->context.pending);
    1559. 

  1559. 	} else {
    1560. 

  1560. 		hubbub_tag *ctag = &tokeniser->context.current_tag;
    1561. 

  1561. 		hubbub_attribute *attr = &ctag->attributes[
    1562. 

  1562. 				ctag->n_attributes - 1];
    1563. 

  1563. 

  1564. 		uint8_t utf8[6];
    1565. 

  1565. 		uint8_t *utf8ptr = utf8;
    1566. 

  1566. 		size_t len = sizeof(utf8);
    1567. 

  1567. 

  1568. 		if (tokeniser->context.match_entity.codepoint) {
    1569. 

  1569. 			parserutils_charset_utf8_from_ucs4(
    1570. 

  1570. 				tokeniser->context.match_entity.codepoint,
    1571. 

  1571. 				&utf8ptr, &len);
    1572. 

  1572. 

  1573. 			COLLECT_MS(attr->value, utf8, sizeof(utf8) - len);
    1574. 

  1574. 

  1575. 			/* +1 for the ampersand */
    1576. 

  1576. 			tokeniser->context.pending +=
    1577. 

  1577. 					tokeniser->context.match_entity.length
    1578. 

  1578. 					+ 1;
    1579. 

  1579. 		} else {
    1580. 

  1580. 			size_t len = 0;
    1581. 

  1581. 			const uint8_t *cptr = NULL;
    1582. 

  1582. 			parserutils_error error;
    1583. 

  1583. 

  1584. 			error = parserutils_inputstream_peek(
    1585. 

  1585. 					tokeniser->input,
    1586. 

  1586. 					tokeniser->context.pending, 
    1587. 

  1587. 					&cptr,
    1588. 

  1588. 					&len);
    1589. 

  1589. 

  1590. 			assert(error == PARSERUTILS_OK);
    1591. 

  1591. 

  1592. 			/* Insert the ampersand */
    1593. 

  1593. 			COLLECT_MS(attr->value, cptr, len);
    1594. 

  1594. 			tokeniser->context.pending += len;
    1595. 

  1595. 		}
    1596. 

  1596. 

  1597. 		/* Reset for next time */
    1598. 

  1598. 		tokeniser->context.match_entity.complete = false;
    1599. 

  1599. 

  1600. 		/* And back to the previous state */
    1601. 

  1601. 		tokeniser->state = tokeniser->context.prev_state;
    1602. 

  1602. 	}
    1603. 

  1603. 

  1604. 	return HUBBUB_OK;
    1605. 

  1605. }
    1606. 

  1606. 

  1607. /* always switches state */
    1608. 

  1608. hubbub_error hubbub_tokeniser_handle_after_attribute_value_q(
    1609. 

  1609. 		hubbub_tokeniser *tokeniser)
    1610. 

  1610. {
    1611. 

  1611. 	size_t len;
    1612. 

  1612. 	const uint8_t *cptr;
    1613. 

  1613. 	parserutils_error error;
    1614. 

  1614. 	uint8_t c;
    1615. 

  1615. 

  1616. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1617. 

  1617. 			tokeniser->context.pending, &cptr, &len);
    1618. 

  1618. 

  1619. 	if (error != PARSERUTILS_OK) {
    1620. 

  1620. 		if (error == PARSERUTILS_EOF) {
    1621. 

  1621. 			tokeniser->state = STATE_DATA;
    1622. 

  1622. 			return emit_current_tag(tokeniser);
    1623. 

  1623. 		} else {
    1624. 

  1624. 			return hubbub_error_from_parserutils_error(error);
    1625. 

  1625. 		}
    1626. 

  1626. 	}
    1627. 

  1627. 

  1628. 	c = *cptr;
    1629. 

  1629. 

  1630. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    1631. 

  1631. 		tokeniser->context.pending += len;
    1632. 

  1632. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_NAME;
    1633. 

  1633. 	} else if (c == '>') {
    1634. 

  1634. 		tokeniser->context.pending += len;
    1635. 

  1635. 

  1636. 		tokeniser->state = STATE_DATA;
    1637. 

  1637. 		return emit_current_tag(tokeniser);
    1638. 

  1638. 	} else if (c == '/') {
    1639. 

  1639. 		tokeniser->context.pending += len;
    1640. 

  1640. 		tokeniser->state = STATE_SELF_CLOSING_START_TAG;
    1641. 

  1641. 	} else {
    1642. 

  1642. 		/** \todo parse error */
    1643. 

  1643. 		/* Reprocess character in before attribute name state */
    1644. 

  1644. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_NAME;
    1645. 

  1645. 	}
    1646. 

  1646. 

  1647. 	return HUBBUB_OK;
    1648. 

  1648. }
    1649. 

  1649. 

  1650. hubbub_error hubbub_tokeniser_handle_self_closing_start_tag(
    1651. 

  1651. 		hubbub_tokeniser *tokeniser)
    1652. 

  1652. {
    1653. 

  1653. 	size_t len;
    1654. 

  1654. 	const uint8_t *cptr;
    1655. 

  1655. 	parserutils_error error;
    1656. 

  1656. 	uint8_t c;
    1657. 

  1657. 

  1658. 	error = parserutils_inputstream_peek(tokeniser->input,
    1659. 

  1659. 			tokeniser->context.pending, &cptr, &len);
    1660. 

  1660. 

  1661. 	if (error != PARSERUTILS_OK) {
    1662. 

  1662. 		if (error == PARSERUTILS_EOF) {
    1663. 

  1663. 			tokeniser->state = STATE_DATA;
    1664. 

  1664. 			return emit_current_tag(tokeniser);
    1665. 

  1665. 		} else {
    1666. 

  1666. 			return hubbub_error_from_parserutils_error(error);
    1667. 

  1667. 		}
    1668. 

  1668. 	}
    1669. 

  1669. 

  1670. 	c = *cptr;
    1671. 

  1671. 

  1672. 	if (c == '>') {
    1673. 

  1673. 		tokeniser->context.pending += len;
    1674. 

  1674. 		tokeniser->state = STATE_DATA;
    1675. 

  1675. 

  1676. 		tokeniser->context.current_tag.self_closing = true;
    1677. 

  1677. 		return emit_current_tag(tokeniser);
    1678. 

  1678. 	} else {
    1679. 

  1679. 		/* Reprocess character in before attribute name state */
    1680. 

  1680. 		tokeniser->state = STATE_BEFORE_ATTRIBUTE_NAME;
    1681. 

  1681. 	}
    1682. 

  1682. 

  1683. 	return HUBBUB_OK;
    1684. 

  1684. }
    1685. 

  1685. 

  1686. /* this state expects tokeniser->context.chars to be empty on first entry */
    1687. 

  1687. hubbub_error hubbub_tokeniser_handle_bogus_comment(hubbub_tokeniser *tokeniser)
    1688. 

  1688. {
    1689. 

  1689. 	size_t len;
    1690. 

  1690. 	const uint8_t *cptr;
    1691. 

  1691. 	parserutils_error error;
    1692. 

  1692. 	uint8_t c;
    1693. 

  1693. 

  1694. 	error = parserutils_inputstream_peek(tokeniser->input,
    1695. 

  1695. 			tokeniser->context.pending, &cptr, &len);
    1696. 

  1696. 

  1697. 	if (error != PARSERUTILS_OK) {
    1698. 

  1698. 		if (error == PARSERUTILS_EOF) {
    1699. 

  1699. 			tokeniser->state = STATE_DATA;
    1700. 

  1700. 			return emit_current_comment(tokeniser);
    1701. 

  1701. 		} else {
    1702. 

  1702. 			return hubbub_error_from_parserutils_error(error);
    1703. 

  1703. 		}
    1704. 

  1704. 	}
    1705. 

  1705. 

  1706. 	c = *cptr;
    1707. 

  1707. 

  1708. 	if (c == '>') {
    1709. 

  1709. 		tokeniser->context.pending += len;
    1710. 

  1710. 		tokeniser->state = STATE_DATA;
    1711. 

  1711. 		return emit_current_comment(tokeniser);
    1712. 

  1712. 	} else if (c == '\0') {
    1713. 

  1713. 		error = parserutils_buffer_append(tokeniser->buffer,
    1714. 

  1714. 				u_fffd, sizeof(u_fffd));
    1715. 

  1715. 		if (error != PARSERUTILS_OK)
    1716. 

  1716. 			return hubbub_error_from_parserutils_error(error);
    1717. 

  1717. 

  1718. 		tokeniser->context.pending += len;
    1719. 

  1719. 	} else if (c == '\r') {
    1720. 

  1720. 		error = parserutils_inputstream_peek(
    1721. 

  1721. 				tokeniser->input,
    1722. 

  1722. 				tokeniser->context.pending,
    1723. 

  1723. 				&cptr,
    1724. 

  1724. 				&len);
    1725. 

  1725. 

  1726. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    1727. 

  1727. 			return hubbub_error_from_parserutils_error(error);
    1728. 

  1728. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    1729. 

  1729. 			error = parserutils_buffer_append(tokeniser->buffer,
    1730. 

  1730. 					&lf, sizeof(lf));
    1731. 

  1731. 			if (error != PARSERUTILS_OK) {
    1732. 

  1732. 				return hubbub_error_from_parserutils_error(
    1733. 

  1733. 						error);
    1734. 

  1734. 			}
    1735. 

  1735. 		}
    1736. 

  1736. 		tokeniser->context.pending += len;
    1737. 

  1737. 	} else {
    1738. 

  1738. 		error = parserutils_buffer_append(tokeniser->buffer,
    1739. 

  1739. 				(uint8_t *) cptr, len);
    1740. 

  1740. 		if (error != PARSERUTILS_OK)
    1741. 

  1741. 			return hubbub_error_from_parserutils_error(error);
    1742. 

  1742. 

  1743. 		tokeniser->context.pending += len;
    1744. 

  1744. 	}
    1745. 

  1745. 

  1746. 	return HUBBUB_OK;
    1747. 

  1747. }
    1748. 

  1748. 

  1749. /* this state always switches to another state straight away */
    1750. 

  1750. hubbub_error hubbub_tokeniser_handle_markup_declaration_open(
    1751. 

  1751. 		hubbub_tokeniser *tokeniser)
    1752. 

  1752. {
    1753. 

  1753. 	size_t len;
    1754. 

  1754. 	const uint8_t *cptr;
    1755. 

  1755. 	parserutils_error error;
    1756. 

  1756. 	uint8_t c;
    1757. 

  1757. 

  1758. 	assert(tokeniser->context.pending == 0);
    1759. 

  1759. 

  1760. 	error = parserutils_inputstream_peek(tokeniser->input, 0, &cptr, &len);
    1761. 

  1761. 

  1762. 	if (error != PARSERUTILS_OK) {
    1763. 

  1763. 		if (error == PARSERUTILS_EOF) {
    1764. 

  1764. 			tokeniser->state = STATE_BOGUS_COMMENT;
    1765. 

  1765. 			return HUBBUB_OK;
    1766. 

  1766. 		} else {
    1767. 

  1767. 			return hubbub_error_from_parserutils_error(error);
    1768. 

  1768. 		}
    1769. 

  1769. 	}
    1770. 

  1770. 

  1771. 	c = *cptr;
    1772. 

  1772. 

  1773. 	if (c == '-') {
    1774. 

  1774. 		tokeniser->context.pending = len;
    1775. 

  1775. 		tokeniser->state = STATE_MATCH_COMMENT;
    1776. 

  1776. 	} else if ((c & ~0x20) == 'D') {
    1777. 

  1777. 		tokeniser->context.pending = len;
    1778. 

  1778. 		tokeniser->context.match_doctype.count = len;
    1779. 

  1779. 		tokeniser->state = STATE_MATCH_DOCTYPE;
    1780. 

  1780. 	} else if (tokeniser->process_cdata_section == true && c == '[') {
    1781. 

  1781. 		tokeniser->context.pending = len;
    1782. 

  1782. 		tokeniser->context.match_cdata.count = len;
    1783. 

  1783. 		tokeniser->state = STATE_MATCH_CDATA;
    1784. 

  1784. 	} else {
    1785. 

  1785. 		tokeniser->state = STATE_BOGUS_COMMENT;
    1786. 

  1786. 	}
    1787. 

  1787. 

  1788. 	return HUBBUB_OK;
    1789. 

  1789. }
    1790. 

  1790. 

  1791. 

  1792. hubbub_error hubbub_tokeniser_handle_match_comment(hubbub_tokeniser *tokeniser)
    1793. 

  1793. {
    1794. 

  1794. 	size_t len;
    1795. 

  1795. 	const uint8_t *cptr;
    1796. 

  1796. 	parserutils_error error;
    1797. 

  1797. 

  1798. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1799. 

  1799. 			tokeniser->context.pending, &cptr, &len);
    1800. 

  1800. 

  1801. 	if (error != PARSERUTILS_OK) {
    1802. 

  1802. 		if (error == PARSERUTILS_EOF) {
    1803. 

  1803. 			tokeniser->context.pending =
    1804. 

  1804. 				tokeniser->context.current_comment.len = 0;
    1805. 

  1805. 			tokeniser->state = STATE_BOGUS_COMMENT;
    1806. 

  1806. 			return HUBBUB_OK;
    1807. 

  1807. 		} else {
    1808. 

  1808. 			return hubbub_error_from_parserutils_error(error);
    1809. 

  1809. 		}
    1810. 

  1810. 	}
    1811. 

  1811. 

  1812. 	tokeniser->context.pending = tokeniser->context.current_comment.len = 0;
    1813. 

  1813. 

  1814. 	if (*cptr == '-') {
    1815. 

  1815. 		parserutils_inputstream_advance(tokeniser->input, SLEN("--"));
    1816. 

  1816. 		tokeniser->state = STATE_COMMENT_START;
    1817. 

  1817. 	} else {
    1818. 

  1818. 		tokeniser->state = STATE_BOGUS_COMMENT;
    1819. 

  1819. 	}
    1820. 

  1820. 

  1821. 	return HUBBUB_OK;
    1822. 

  1822. }
    1823. 

  1823. 

  1824. 

  1825. hubbub_error hubbub_tokeniser_handle_comment(hubbub_tokeniser *tokeniser)
    1826. 

  1826. {
    1827. 

  1827. 	size_t len;
    1828. 

  1828. 	const uint8_t *cptr;
    1829. 

  1829. 	parserutils_error error;
    1830. 

  1830. 	uint8_t c;
    1831. 

  1831. 

  1832. 	error = parserutils_inputstream_peek(tokeniser->input, 
    1833. 

  1833. 			tokeniser->context.pending, &cptr, &len);
    1834. 

  1834. 

  1835. 	if (error != PARSERUTILS_OK) {
    1836. 

  1836. 		if (error == PARSERUTILS_EOF) {
    1837. 

  1837. 			tokeniser->state = STATE_DATA;
    1838. 

  1838. 			return emit_current_comment(tokeniser);
    1839. 

  1839. 		} else {
    1840. 

  1840. 			return hubbub_error_from_parserutils_error(error);
    1841. 

  1841. 		}
    1842. 

  1842. 	}
    1843. 

  1843. 

  1844. 	c = *cptr;
    1845. 

  1845. 

  1846. 	if (c == '>' && (tokeniser->state == STATE_COMMENT_START_DASH ||
    1847. 

  1847. 			tokeniser->state == STATE_COMMENT_START ||
    1848. 

  1848. 			tokeniser->state == STATE_COMMENT_END)) {
    1849. 

  1849. 		tokeniser->context.pending += len;
    1850. 

  1850. 

  1851. 		/** \todo parse error if state != COMMENT_END */
    1852. 

  1852. 		tokeniser->state = STATE_DATA;
    1853. 

  1853. 		return emit_current_comment(tokeniser);
    1854. 

  1854. 	} else if (c == '-') {
    1855. 

  1855. 		if (tokeniser->state == STATE_COMMENT_START) {
    1856. 

  1856. 			tokeniser->state = STATE_COMMENT_START_DASH;
    1857. 

  1857. 		} else if (tokeniser->state == STATE_COMMENT_START_DASH) {
    1858. 

  1858. 			tokeniser->state = STATE_COMMENT_END;
    1859. 

  1859. 		} else if (tokeniser->state == STATE_COMMENT) {
    1860. 

  1860. 			tokeniser->state = STATE_COMMENT_END_DASH;
    1861. 

  1861. 		} else if (tokeniser->state == STATE_COMMENT_END_DASH) {
    1862. 

  1862. 			tokeniser->state = STATE_COMMENT_END;
    1863. 

  1863. 		} else if (tokeniser->state == STATE_COMMENT_END) {
    1864. 

  1864. 			error = parserutils_buffer_append(tokeniser->buffer,
    1865. 

  1865. 					(uint8_t *) "-", SLEN("-"));
    1866. 

  1866. 			if (error != PARSERUTILS_OK) {
    1867. 

  1867. 				return hubbub_error_from_parserutils_error(
    1868. 

  1868. 						error);
    1869. 

  1869. 			}
    1870. 

  1870. 		}
    1871. 

  1871. 

  1872. 		tokeniser->context.pending += len;
    1873. 

  1873. 	} else {
    1874. 

  1874. 		if (tokeniser->state == STATE_COMMENT_START_DASH ||
    1875. 

  1875. 				tokeniser->state == STATE_COMMENT_END_DASH) {
    1876. 

  1876. 			error = parserutils_buffer_append(tokeniser->buffer,
    1877. 

  1877. 					(uint8_t *) "-", SLEN("-"));
    1878. 

  1878. 			if (error != PARSERUTILS_OK) {
    1879. 

  1879. 				return hubbub_error_from_parserutils_error(
    1880. 

  1880. 						error);
    1881. 

  1881. 			}
    1882. 

  1882. 		} else if (tokeniser->state == STATE_COMMENT_END) {
    1883. 

  1883. 			error = parserutils_buffer_append(tokeniser->buffer,
    1884. 

  1884. 					(uint8_t *) "--", SLEN("--"));
    1885. 

  1885. 			if (error != PARSERUTILS_OK) {
    1886. 

  1886. 				return hubbub_error_from_parserutils_error(
    1887. 

  1887. 						error);
    1888. 

  1888. 			}
    1889. 

  1889. 		}
    1890. 

  1890. 

  1891. 		if (c == '\0') {
    1892. 

  1892. 			error = parserutils_buffer_append(tokeniser->buffer,
    1893. 

  1893. 					u_fffd, sizeof(u_fffd));
    1894. 

  1894. 			if (error != PARSERUTILS_OK) {
    1895. 

  1895. 				return hubbub_error_from_parserutils_error(
    1896. 

  1896. 						error);
    1897. 

  1897. 			}
    1898. 

  1898. 		} else if (c == '\r') {
    1899. 

  1899. 			size_t next_len;
    1900. 

  1900. 			error = parserutils_inputstream_peek(
    1901. 

  1901. 					tokeniser->input,
    1902. 

  1902. 					tokeniser->context.pending + len,
    1903. 

  1903. 					&cptr,
    1904. 

  1904. 					&next_len);
    1905. 

  1905. 			if (error != PARSERUTILS_OK && 
    1906. 

  1906. 					error != PARSERUTILS_EOF) {
    1907. 

  1907. 				return hubbub_error_from_parserutils_error(
    1908. 

  1908. 						error);
    1909. 

  1909. 			} else if (error != PARSERUTILS_EOF && *cptr != '\n') {
    1910. 

  1910. 				error = parserutils_buffer_append(
    1911. 

  1911. 						tokeniser->buffer,
    1912. 

  1912. 						&lf, sizeof(lf));
    1913. 

  1913. 				if (error != PARSERUTILS_OK) {
    1914. 

  1914. 					return hubbub_error_from_parserutils_error(
    1915. 

  1915. 							error);
    1916. 

  1916. 				}
    1917. 

  1917. 			}
    1918. 

  1918. 		} else {
    1919. 

  1919. 			error = parserutils_buffer_append(tokeniser->buffer, 
    1920. 

  1920. 					cptr, len);
    1921. 

  1921. 			if (error != PARSERUTILS_OK) {
    1922. 

  1922. 				return hubbub_error_from_parserutils_error(
    1923. 

  1923. 						error);
    1924. 

  1924. 			}
    1925. 

  1925. 		}
    1926. 

  1926. 

  1927. 		tokeniser->context.pending += len;
    1928. 

  1928. 		tokeniser->state = STATE_COMMENT;
    1929. 

  1929. 	}
    1930. 

  1930. 

  1931. 	return HUBBUB_OK;
    1932. 

  1932. }
    1933. 

  1933. 

  1934. 

  1935. 

  1936. 

  1937. #define DOCTYPE		"DOCTYPE"
    1938. 

  1938. #define DOCTYPE_LEN	(SLEN(DOCTYPE) - 1)
    1939. 

  1939. 

  1940. hubbub_error hubbub_tokeniser_handle_match_doctype(hubbub_tokeniser *tokeniser)
    1941. 

  1941. {
    1942. 

  1942. 	size_t len;
    1943. 

  1943. 	const uint8_t *cptr;
    1944. 

  1944. 	parserutils_error error;
    1945. 

  1945. 	uint8_t c;
    1946. 

  1946. 

  1947. 	error = parserutils_inputstream_peek(tokeniser->input,
    1948. 

  1948. 			tokeniser->context.match_doctype.count, &cptr, &len);
    1949. 

  1949. 

  1950. 	if (error != PARSERUTILS_OK) {
    1951. 

  1951. 		if (error == PARSERUTILS_EOF) {
    1952. 

  1952. 			tokeniser->context.current_comment.len =
    1953. 

  1953. 					tokeniser->context.pending = 0;
    1954. 

  1954. 			tokeniser->state = STATE_BOGUS_COMMENT;
    1955. 

  1955. 			return HUBBUB_OK;
    1956. 

  1956. 		} else {
    1957. 

  1957. 			return hubbub_error_from_parserutils_error(error);
    1958. 

  1958. 		}
    1959. 

  1959. 	}
    1960. 

  1960. 

  1961. 	c = *cptr;
    1962. 

  1962. 

  1963. 	assert(tokeniser->context.match_doctype.count <= DOCTYPE_LEN);
    1964. 

  1964. 

  1965. 	if (DOCTYPE[tokeniser->context.match_doctype.count] != (c & ~0x20)) {
    1966. 

  1966. 		tokeniser->context.current_comment.len =
    1967. 

  1967. 				tokeniser->context.pending = 0;
    1968. 

  1968. 		tokeniser->state = STATE_BOGUS_COMMENT;
    1969. 

  1969. 		return HUBBUB_OK;
    1970. 

  1970. 	}
    1971. 

  1971. 

  1972. 	tokeniser->context.pending += len;
    1973. 

  1973. 

  1974. 	if (tokeniser->context.match_doctype.count == DOCTYPE_LEN) {
    1975. 

  1975. 		/* Skip over the DOCTYPE bit */
    1976. 

  1976. 		parserutils_inputstream_advance(tokeniser->input,
    1977. 

  1977. 				tokeniser->context.pending);
    1978. 

  1978. 

  1979. 		memset(&tokeniser->context.current_doctype, 0,
    1980. 

  1980. 				sizeof tokeniser->context.current_doctype);
    1981. 

  1981. 		tokeniser->context.current_doctype.public_missing = true;
    1982. 

  1982. 		tokeniser->context.current_doctype.system_missing = true;
    1983. 

  1983. 		tokeniser->context.pending = 0;
    1984. 

  1984. 

  1985. 		tokeniser->state = STATE_DOCTYPE;
    1986. 

  1986. 	}
    1987. 

  1987. 

  1988. 	tokeniser->context.match_doctype.count++;
    1989. 

  1989. 

  1990. 	return HUBBUB_OK;
    1991. 

  1991. }
    1992. 

  1992. 

  1993. #undef DOCTYPE
    1994. 

  1994. #undef DOCTYPE_LEN
    1995. 

  1995. 

  1996. hubbub_error hubbub_tokeniser_handle_doctype(hubbub_tokeniser *tokeniser)
    1997. 

  1997. {
    1998. 

  1998. 	size_t len;
    1999. 

  1999. 	const uint8_t *cptr;
    2000. 

  2000. 	parserutils_error error;
    2001. 

  2001. 	uint8_t c;
    2002. 

  2002. 

  2003. 	error = parserutils_inputstream_peek(tokeniser->input,
    2004. 

  2004. 			tokeniser->context.pending, &cptr, &len);
    2005. 

  2005. 

  2006. 	if (error != PARSERUTILS_OK) {
    2007. 

  2007. 		if (error == PARSERUTILS_EOF) {
    2008. 

  2008. 			tokeniser->state = STATE_BEFORE_DOCTYPE_NAME;
    2009. 

  2009. 			return HUBBUB_OK;
    2010. 

  2010. 		} else {
    2011. 

  2011. 			return hubbub_error_from_parserutils_error(error);
    2012. 

  2012. 		}
    2013. 

  2013. 	}
    2014. 

  2014. 

  2015. 	c = *cptr;
    2016. 

  2016. 

  2017. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2018. 

  2018. 		tokeniser->context.pending += len;
    2019. 

  2019. 	}
    2020. 

  2020. 

  2021. 	tokeniser->state = STATE_BEFORE_DOCTYPE_NAME;
    2022. 

  2022. 

  2023. 	return HUBBUB_OK;
    2024. 

  2024. }
    2025. 

  2025. 

  2026. hubbub_error hubbub_tokeniser_handle_before_doctype_name(
    2027. 

  2027. 		hubbub_tokeniser *tokeniser)
    2028. 

  2028. {
    2029. 

  2029. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2030. 

  2030. 	size_t len;
    2031. 

  2031. 	const uint8_t *cptr;
    2032. 

  2032. 	parserutils_error error;
    2033. 

  2033. 	uint8_t c;
    2034. 

  2034. 

  2035. 	error = parserutils_inputstream_peek(tokeniser->input,
    2036. 

  2036. 			tokeniser->context.pending, &cptr, &len);
    2037. 

  2037. 

  2038. 	if (error != PARSERUTILS_OK) {
    2039. 

  2039. 		if (error == PARSERUTILS_EOF) {
    2040. 

  2040. 			/** \todo parse error */
    2041. 

  2041. 			/* Emit current doctype, force-quirks on */
    2042. 

  2042. 			tokeniser->state = STATE_DATA;
    2043. 

  2043. 			return emit_current_doctype(tokeniser, true);
    2044. 

  2044. 		} else {
    2045. 

  2045. 			return hubbub_error_from_parserutils_error(error);
    2046. 

  2046. 		}
    2047. 

  2047. 	}
    2048. 

  2048. 

  2049. 	c = *cptr;
    2050. 

  2050. 

  2051. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2052. 

  2052. 		/* pass over in silence */
    2053. 

  2053. 		tokeniser->context.pending += len;
    2054. 

  2054. 	} else if (c == '>') {
    2055. 

  2055. 		/** \todo parse error */
    2056. 

  2056. 		tokeniser->context.pending += len;
    2057. 

  2057. 		tokeniser->state = STATE_DATA;
    2058. 

  2058. 		return emit_current_doctype(tokeniser, true);
    2059. 

  2059. 	} else {
    2060. 

  2060. 		if (c == '\0') {
    2061. 

  2061. 			START_BUF(cdoc->name, u_fffd, sizeof(u_fffd));
    2062. 

  2062. 		} else if ('A' <= c && c <= 'Z') {
    2063. 

  2063. 			uint8_t lc = c + 0x20;
    2064. 

  2064. 

  2065. 			START_BUF(cdoc->name, &lc, len);
    2066. 

  2066. 		} else {
    2067. 

  2067. 			START_BUF(cdoc->name, cptr, len);
    2068. 

  2068. 		}
    2069. 

  2069. 

  2070. 		tokeniser->context.pending += len;
    2071. 

  2071. 		tokeniser->state = STATE_DOCTYPE_NAME;
    2072. 

  2072. 	}
    2073. 

  2073. 

  2074. 	return HUBBUB_OK;
    2075. 

  2075. }
    2076. 

  2076. 

  2077. hubbub_error hubbub_tokeniser_handle_doctype_name(hubbub_tokeniser *tokeniser)
    2078. 

  2078. {
    2079. 

  2079. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2080. 

  2080. 	size_t len;
    2081. 

  2081. 	const uint8_t *cptr;
    2082. 

  2082. 	parserutils_error error;
    2083. 

  2083. 	uint8_t c;
    2084. 

  2084. 

  2085. 	error = parserutils_inputstream_peek(tokeniser->input,
    2086. 

  2086. 			tokeniser->context.pending, &cptr, &len);
    2087. 

  2087. 

  2088. 	if (error != PARSERUTILS_OK) {
    2089. 

  2089. 		if (error == PARSERUTILS_EOF) {
    2090. 

  2090. 			tokeniser->state = STATE_DATA;
    2091. 

  2091. 			return emit_current_doctype(tokeniser, true);
    2092. 

  2092. 		} else {
    2093. 

  2093. 			return hubbub_error_from_parserutils_error(error);
    2094. 

  2094. 		}
    2095. 

  2095. 	}
    2096. 

  2096. 

  2097. 	c = *cptr;
    2098. 

  2098. 

  2099. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2100. 

  2100. 		tokeniser->context.pending += len;
    2101. 

  2101. 		tokeniser->state = STATE_AFTER_DOCTYPE_NAME;
    2102. 

  2102. 	} else if (c == '>') {
    2103. 

  2103. 		tokeniser->context.pending += len;
    2104. 

  2104. 		tokeniser->state = STATE_DATA;
    2105. 

  2105. 		return emit_current_doctype(tokeniser, false);
    2106. 

  2106. 	} else if (c == '\0') {
    2107. 

  2107. 		COLLECT(cdoc->name, u_fffd, sizeof(u_fffd));
    2108. 

  2108. 		tokeniser->context.pending += len;
    2109. 

  2109. 	} else if ('A' <= c && c <= 'Z') {
    2110. 

  2110. 		uint8_t lc = c + 0x20;
    2111. 

  2111. 		COLLECT(cdoc->name, &lc, len);
    2112. 

  2112. 		tokeniser->context.pending += len;
    2113. 

  2113. 	} else {
    2114. 

  2114. 		COLLECT(cdoc->name, cptr, len);
    2115. 

  2115. 		tokeniser->context.pending += len;
    2116. 

  2116. 	}
    2117. 

  2117. 

  2118. 	return HUBBUB_OK;
    2119. 

  2119. }
    2120. 

  2120. 

  2121. hubbub_error hubbub_tokeniser_handle_after_doctype_name(
    2122. 

  2122. 		hubbub_tokeniser *tokeniser)
    2123. 

  2123. {
    2124. 

  2124. 	size_t len;
    2125. 

  2125. 	const uint8_t *cptr;
    2126. 

  2126. 	parserutils_error error;
    2127. 

  2127. 	uint8_t c;
    2128. 

  2128. 

  2129. 	error = parserutils_inputstream_peek(tokeniser->input,
    2130. 

  2130. 			tokeniser->context.pending, &cptr, &len);
    2131. 

  2131. 

  2132. 	if (error != PARSERUTILS_OK) {
    2133. 

  2133. 		if (error == PARSERUTILS_EOF) {
    2134. 

  2134. 			tokeniser->state = STATE_DATA;
    2135. 

  2135. 			return emit_current_doctype(tokeniser, true);
    2136. 

  2136. 		} else {
    2137. 

  2137. 			return hubbub_error_from_parserutils_error(error);
    2138. 

  2138. 		}
    2139. 

  2139. 	}
    2140. 

  2140. 

  2141. 	c = *cptr;
    2142. 

  2142. 	tokeniser->context.pending += len;
    2143. 

  2143. 

  2144. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2145. 

  2145. 		/* pass over in silence */
    2146. 

  2146. 	} else if (c == '>') {
    2147. 

  2147. 		tokeniser->state = STATE_DATA;
    2148. 

  2148. 		return emit_current_doctype(tokeniser, false);
    2149. 

  2149. 	} else if ((c & ~0x20) == 'P') {
    2150. 

  2150. 		tokeniser->context.match_doctype.count = 1;
    2151. 

  2151. 		tokeniser->state = STATE_MATCH_PUBLIC;
    2152. 

  2152. 	} else if ((c & ~0x20) == 'S') {
    2153. 

  2153. 		tokeniser->context.match_doctype.count = 1;
    2154. 

  2154. 		tokeniser->state = STATE_MATCH_SYSTEM;
    2155. 

  2155. 	} else {
    2156. 

  2156. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2157. 

  2157. 		tokeniser->context.current_doctype.force_quirks = true;
    2158. 

  2158. 	}
    2159. 

  2159. 

  2160. 	return HUBBUB_OK;
    2161. 

  2161. }
    2162. 

  2162. 

  2163. #define PUBLIC		"PUBLIC"
    2164. 

  2164. #define PUBLIC_LEN	(SLEN(PUBLIC) - 1)
    2165. 

  2165. 

  2166. hubbub_error hubbub_tokeniser_handle_match_public(hubbub_tokeniser *tokeniser)
    2167. 

  2167. {
    2168. 

  2168. 	size_t len;
    2169. 

  2169. 	const uint8_t *cptr;
    2170. 

  2170. 	parserutils_error error;
    2171. 

  2171. 	uint8_t c;
    2172. 

  2172. 

  2173. 	error = parserutils_inputstream_peek(tokeniser->input,
    2174. 

  2174. 			tokeniser->context.pending, &cptr, &len);
    2175. 

  2175. 

  2176. 	if (error != PARSERUTILS_OK) {
    2177. 

  2177. 		if (error == PARSERUTILS_EOF) {
    2178. 

  2178. 			tokeniser->context.current_doctype.force_quirks = true;
    2179. 

  2179. 			tokeniser->state = STATE_BOGUS_DOCTYPE;
    2180. 

  2180. 			return HUBBUB_OK;
    2181. 

  2181. 		} else {
    2182. 

  2182. 			return hubbub_error_from_parserutils_error(error);
    2183. 

  2183. 		}
    2184. 

  2184. 	}
    2185. 

  2185. 

  2186. 	c = *cptr;
    2187. 

  2187. 

  2188. 	assert(tokeniser->context.match_doctype.count <= PUBLIC_LEN);
    2189. 

  2189. 

  2190. 	if (PUBLIC[tokeniser->context.match_doctype.count] != (c & ~0x20)) {
    2191. 

  2191. 		tokeniser->context.current_doctype.force_quirks = true;
    2192. 

  2192. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2193. 

  2193. 		return HUBBUB_OK;
    2194. 

  2194. 	}
    2195. 

  2195. 

  2196. 	tokeniser->context.pending += len;
    2197. 

  2197. 

  2198. 	if (tokeniser->context.match_doctype.count == PUBLIC_LEN) {
    2199. 

  2199. 		tokeniser->state = STATE_BEFORE_DOCTYPE_PUBLIC;
    2200. 

  2200. 	}
    2201. 

  2201. 

  2202. 	tokeniser->context.match_doctype.count++;
    2203. 

  2203. 

  2204. 	return HUBBUB_OK;
    2205. 

  2205. }
    2206. 

  2206. 

  2207. #undef PUBLIC
    2208. 

  2208. #undef PUBLIC_LEN
    2209. 

  2209. 

  2210. hubbub_error hubbub_tokeniser_handle_before_doctype_public(
    2211. 

  2211. 		hubbub_tokeniser *tokeniser)
    2212. 

  2212. {
    2213. 

  2213. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2214. 

  2214. 	size_t len;
    2215. 

  2215. 	const uint8_t *cptr;
    2216. 

  2216. 	parserutils_error error;
    2217. 

  2217. 	uint8_t c;
    2218. 

  2218. 

  2219. 	error = parserutils_inputstream_peek(tokeniser->input,
    2220. 

  2220. 			tokeniser->context.pending, &cptr, &len);
    2221. 

  2221. 

  2222. 	if (error != PARSERUTILS_OK) {
    2223. 

  2223. 		if (error == PARSERUTILS_EOF) {
    2224. 

  2224. 			tokeniser->state = STATE_DATA;
    2225. 

  2225. 			return emit_current_doctype(tokeniser, true);
    2226. 

  2226. 		} else {
    2227. 

  2227. 			return hubbub_error_from_parserutils_error(error);
    2228. 

  2228. 		}
    2229. 

  2229. 	}
    2230. 

  2230. 

  2231. 	c = *cptr;
    2232. 

  2232. 	tokeniser->context.pending += len;
    2233. 

  2233. 

  2234. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2235. 

  2235. 		/* pass over in silence */
    2236. 

  2236. 	} else if (c == '"') {
    2237. 

  2237. 		cdoc->public_missing = false;
    2238. 

  2238. 		cdoc->public_id.len = 0;
    2239. 

  2239. 		tokeniser->state = STATE_DOCTYPE_PUBLIC_DQ;
    2240. 

  2240. 	} else if (c == '\'') {
    2241. 

  2241. 		cdoc->public_missing = false;
    2242. 

  2242. 		cdoc->public_id.len = 0;
    2243. 

  2243. 		tokeniser->state = STATE_DOCTYPE_PUBLIC_SQ;
    2244. 

  2244. 	} else if (c == '>') {
    2245. 

  2245. 		tokeniser->state = STATE_DATA;
    2246. 

  2246. 		return emit_current_doctype(tokeniser, true);
    2247. 

  2247. 	} else {
    2248. 

  2248. 		cdoc->force_quirks = true;
    2249. 

  2249. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2250. 

  2250. 	}
    2251. 

  2251. 

  2252. 	return HUBBUB_OK;
    2253. 

  2253. }
    2254. 

  2254. 

  2255. hubbub_error hubbub_tokeniser_handle_doctype_public_dq(
    2256. 

  2256. 		hubbub_tokeniser *tokeniser)
    2257. 

  2257. {
    2258. 

  2258. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2259. 

  2259. 	size_t len;
    2260. 

  2260. 	const uint8_t *cptr;
    2261. 

  2261. 	parserutils_error error;
    2262. 

  2262. 	uint8_t c;
    2263. 

  2263. 

  2264. 	error = parserutils_inputstream_peek(tokeniser->input,
    2265. 

  2265. 			tokeniser->context.pending, &cptr, &len);
    2266. 

  2266. 

  2267. 	if (error != PARSERUTILS_OK) {
    2268. 

  2268. 		if (error == PARSERUTILS_EOF) {
    2269. 

  2269. 			tokeniser->state = STATE_DATA;
    2270. 

  2270. 			return emit_current_doctype(tokeniser, true);
    2271. 

  2271. 		} else {
    2272. 

  2272. 			return hubbub_error_from_parserutils_error(error);
    2273. 

  2273. 		}
    2274. 

  2274. 	}
    2275. 

  2275. 

  2276. 	c = *cptr;
    2277. 

  2277. 

  2278. 	if (c == '"') {
    2279. 

  2279. 		tokeniser->context.pending += len;
    2280. 

  2280. 		tokeniser->state = STATE_AFTER_DOCTYPE_PUBLIC;
    2281. 

  2281. 	} else if (c == '>') {
    2282. 

  2282. 		tokeniser->context.pending += len;
    2283. 

  2283. 		tokeniser->state = STATE_DATA;
    2284. 

  2284. 		return emit_current_doctype(tokeniser, true);
    2285. 

  2285. 	} else if (c == '\0') {
    2286. 

  2286. 		COLLECT_MS(cdoc->public_id, u_fffd, sizeof(u_fffd));
    2287. 

  2287. 		tokeniser->context.pending += len;
    2288. 

  2288. 	} else if (c == '\r') {
    2289. 

  2289. 		error = parserutils_inputstream_peek(
    2290. 

  2290. 				tokeniser->input,
    2291. 

  2291. 				tokeniser->context.pending,
    2292. 

  2292. 				&cptr,
    2293. 

  2293. 				&len);
    2294. 

  2294. 

  2295. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2296. 

  2296. 			return hubbub_error_from_parserutils_error(error);
    2297. 

  2297. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    2298. 

  2298. 			COLLECT_MS(cdoc->public_id, &lf, sizeof(lf));
    2299. 

  2299. 		}
    2300. 

  2300. 

  2301. 		/* Collect '\r' */
    2302. 

  2302. 		tokeniser->context.pending += 1;
    2303. 

  2303. 	} else {
    2304. 

  2304. 		COLLECT_MS(cdoc->public_id, cptr, len);
    2305. 

  2305. 

  2306. 		tokeniser->context.pending += len;
    2307. 

  2307. 	}
    2308. 

  2308. 

  2309. 	return HUBBUB_OK;
    2310. 

  2310. }
    2311. 

  2311. 

  2312. hubbub_error hubbub_tokeniser_handle_doctype_public_sq(
    2313. 

  2313. 		hubbub_tokeniser *tokeniser)
    2314. 

  2314. {
    2315. 

  2315. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2316. 

  2316. 	size_t len;
    2317. 

  2317. 	const uint8_t *cptr;
    2318. 

  2318. 	parserutils_error error;
    2319. 

  2319. 	uint8_t c;
    2320. 

  2320. 

  2321. 	error = parserutils_inputstream_peek(tokeniser->input,
    2322. 

  2322. 			tokeniser->context.pending, &cptr, &len);
    2323. 

  2323. 

  2324. 	if (error != PARSERUTILS_OK) {
    2325. 

  2325. 		if (error == PARSERUTILS_EOF) {
    2326. 

  2326. 			tokeniser->state = STATE_DATA;
    2327. 

  2327. 			return emit_current_doctype(tokeniser, true);
    2328. 

  2328. 		} else {
    2329. 

  2329. 			return hubbub_error_from_parserutils_error(error);
    2330. 

  2330. 		}
    2331. 

  2331. 	}
    2332. 

  2332. 

  2333. 	c = *cptr;
    2334. 

  2334. 

  2335. 	if (c == '\'') {
    2336. 

  2336. 		tokeniser->context.pending += len;
    2337. 

  2337. 		tokeniser->state = STATE_AFTER_DOCTYPE_PUBLIC;
    2338. 

  2338. 	} else if (c == '>') {
    2339. 

  2339. 		tokeniser->context.pending += len;
    2340. 

  2340. 		tokeniser->state = STATE_DATA;
    2341. 

  2341. 		return emit_current_doctype(tokeniser, true);
    2342. 

  2342. 	} else if (c == '\0') {
    2343. 

  2343. 		COLLECT_MS(cdoc->public_id, u_fffd, sizeof(u_fffd));
    2344. 

  2344. 		tokeniser->context.pending += len;
    2345. 

  2345. 	} else if (c == '\r') {
    2346. 

  2346. 		error = parserutils_inputstream_peek(
    2347. 

  2347. 				tokeniser->input,
    2348. 

  2348. 				tokeniser->context.pending,
    2349. 

  2349. 				&cptr,
    2350. 

  2350. 				&len);
    2351. 

  2351. 

  2352. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2353. 

  2353. 			return hubbub_error_from_parserutils_error(error);
    2354. 

  2354. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    2355. 

  2355. 			COLLECT_MS(cdoc->public_id, &lf, sizeof(lf));
    2356. 

  2356. 		}
    2357. 

  2357. 	
    2358. 

  2358. 		/* Collect '\r' */
    2359. 

  2359. 		tokeniser->context.pending += 1;
    2360. 

  2360. 	} else {
    2361. 

  2361. 		COLLECT_MS(cdoc->public_id, cptr, len);
    2362. 

  2362. 		tokeniser->context.pending += len;
    2363. 

  2363. 	}
    2364. 

  2364. 

  2365. 	return HUBBUB_OK;
    2366. 

  2366. }
    2367. 

  2367. 

  2368. 

  2369. hubbub_error hubbub_tokeniser_handle_after_doctype_public(
    2370. 

  2370. 		hubbub_tokeniser *tokeniser)
    2371. 

  2371. {
    2372. 

  2372. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2373. 

  2373. 	size_t len;
    2374. 

  2374. 	const uint8_t *cptr;
    2375. 

  2375. 	parserutils_error error;
    2376. 

  2376. 	uint8_t c;
    2377. 

  2377. 

  2378. 	error = parserutils_inputstream_peek(tokeniser->input,
    2379. 

  2379. 			tokeniser->context.pending, &cptr, &len);
    2380. 

  2380. 

  2381. 	if (error != PARSERUTILS_OK) {
    2382. 

  2382. 		if (error == PARSERUTILS_EOF) {
    2383. 

  2383. 			tokeniser->state = STATE_DATA;
    2384. 

  2384. 			return emit_current_doctype(tokeniser, true);
    2385. 

  2385. 		} else {
    2386. 

  2386. 			return hubbub_error_from_parserutils_error(error);
    2387. 

  2387. 		}
    2388. 

  2388. 	}
    2389. 

  2389. 

  2390. 	c = *cptr;
    2391. 

  2391. 	tokeniser->context.pending += len;
    2392. 

  2392. 

  2393. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2394. 

  2394. 		/* pass over in silence */
    2395. 

  2395. 	} else if (c == '"') {
    2396. 

  2396. 		cdoc->system_missing = false;
    2397. 

  2397. 		cdoc->system_id.len = 0;
    2398. 

  2398. 

  2399. 		tokeniser->state = STATE_DOCTYPE_SYSTEM_DQ;
    2400. 

  2400. 	} else if (c == '\'') {
    2401. 

  2401. 		cdoc->system_missing = false;
    2402. 

  2402. 		cdoc->system_id.len = 0;
    2403. 

  2403. 

  2404. 		tokeniser->state = STATE_DOCTYPE_SYSTEM_SQ;
    2405. 

  2405. 	} else if (c == '>') {
    2406. 

  2406. 		tokeniser->state = STATE_DATA;
    2407. 

  2407. 		return emit_current_doctype(tokeniser, false);
    2408. 

  2408. 	} else {
    2409. 

  2409. 		cdoc->force_quirks = true;
    2410. 

  2410. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2411. 

  2411. 	}
    2412. 

  2412. 

  2413. 	return HUBBUB_OK;
    2414. 

  2414. }
    2415. 

  2415. 

  2416. 

  2417. 

  2418. #define SYSTEM		"SYSTEM"
    2419. 

  2419. #define SYSTEM_LEN	(SLEN(SYSTEM) - 1)
    2420. 

  2420. 

  2421. hubbub_error hubbub_tokeniser_handle_match_system(hubbub_tokeniser *tokeniser)
    2422. 

  2422. {
    2423. 

  2423. 	size_t len;
    2424. 

  2424. 	const uint8_t *cptr;
    2425. 

  2425. 	parserutils_error error;
    2426. 

  2426. 	uint8_t c;
    2427. 

  2427. 

  2428. 	error = parserutils_inputstream_peek(tokeniser->input,
    2429. 

  2429. 			tokeniser->context.pending, &cptr, &len);
    2430. 

  2430. 

  2431. 	if (error != PARSERUTILS_OK){
    2432. 

  2432. 		if (error == PARSERUTILS_EOF) {
    2433. 

  2433. 			tokeniser->context.current_doctype.force_quirks = true;
    2434. 

  2434. 			tokeniser->state = STATE_BOGUS_DOCTYPE;
    2435. 

  2435. 			return HUBBUB_OK;
    2436. 

  2436. 		} else {
    2437. 

  2437. 			return hubbub_error_from_parserutils_error(error);
    2438. 

  2438. 		}
    2439. 

  2439. 	}
    2440. 

  2440. 

  2441. 	c = *cptr;
    2442. 

  2442. 

  2443. 	assert(tokeniser->context.match_doctype.count <= SYSTEM_LEN);
    2444. 

  2444. 

  2445. 	if (SYSTEM[tokeniser->context.match_doctype.count] != (c & ~0x20)) {
    2446. 

  2446. 		tokeniser->context.current_doctype.force_quirks = true;
    2447. 

  2447. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2448. 

  2448. 		return HUBBUB_OK;
    2449. 

  2449. 	}
    2450. 

  2450. 

  2451. 	tokeniser->context.pending += len;
    2452. 

  2452. 

  2453. 	if (tokeniser->context.match_doctype.count == SYSTEM_LEN) {
    2454. 

  2454. 		tokeniser->state = STATE_BEFORE_DOCTYPE_SYSTEM;
    2455. 

  2455. 	}
    2456. 

  2456. 

  2457. 	tokeniser->context.match_doctype.count++;
    2458. 

  2458. 

  2459. 	return HUBBUB_OK;
    2460. 

  2460. }
    2461. 

  2461. 

  2462. #undef SYSTEM
    2463. 

  2463. #undef SYSTEM_LEN
    2464. 

  2464. 

  2465. hubbub_error hubbub_tokeniser_handle_before_doctype_system(
    2466. 

  2466. 		hubbub_tokeniser *tokeniser)
    2467. 

  2467. {
    2468. 

  2468. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2469. 

  2469. 	size_t len;
    2470. 

  2470. 	const uint8_t *cptr;
    2471. 

  2471. 	parserutils_error error;
    2472. 

  2472. 	uint8_t c;
    2473. 

  2473. 

  2474. 	error = parserutils_inputstream_peek(tokeniser->input,
    2475. 

  2475. 			tokeniser->context.pending, &cptr, &len);
    2476. 

  2476. 

  2477. 	if (error != PARSERUTILS_OK) {
    2478. 

  2478. 		if (error == PARSERUTILS_EOF) {
    2479. 

  2479. 			tokeniser->state = STATE_DATA;
    2480. 

  2480. 			return emit_current_doctype(tokeniser, true);
    2481. 

  2481. 		} else {
    2482. 

  2482. 			return hubbub_error_from_parserutils_error(error);
    2483. 

  2483. 		}
    2484. 

  2484. 	}
    2485. 

  2485. 

  2486. 	c = *cptr;
    2487. 

  2487. 	tokeniser->context.pending += len;
    2488. 

  2488. 

  2489. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2490. 

  2490. 		/* pass over */
    2491. 

  2491. 	} else if (c == '"') {
    2492. 

  2492. 		cdoc->system_missing = false;
    2493. 

  2493. 		cdoc->system_id.len = 0;
    2494. 

  2494. 

  2495. 		tokeniser->state = STATE_DOCTYPE_SYSTEM_DQ;
    2496. 

  2496. 	} else if (c == '\'') {
    2497. 

  2497. 		cdoc->system_missing = false;
    2498. 

  2498. 		cdoc->system_id.len = 0;
    2499. 

  2499. 

  2500. 		tokeniser->state = STATE_DOCTYPE_SYSTEM_SQ;
    2501. 

  2501. 	} else if (c == '>') {
    2502. 

  2502. 		tokeniser->state = STATE_DATA;
    2503. 

  2503. 		return emit_current_doctype(tokeniser, true);
    2504. 

  2504. 	} else {
    2505. 

  2505. 		cdoc->force_quirks = true;
    2506. 

  2506. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2507. 

  2507. 	}
    2508. 

  2508. 

  2509. 	return HUBBUB_OK;
    2510. 

  2510. }
    2511. 

  2511. 

  2512. hubbub_error hubbub_tokeniser_handle_doctype_system_dq(
    2513. 

  2513. 		hubbub_tokeniser *tokeniser)
    2514. 

  2514. {
    2515. 

  2515. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2516. 

  2516. 	size_t len;
    2517. 

  2517. 	const uint8_t *cptr;
    2518. 

  2518. 	parserutils_error error;
    2519. 

  2519. 	uint8_t c;
    2520. 

  2520. 

  2521. 	error = parserutils_inputstream_peek(tokeniser->input,
    2522. 

  2522. 			tokeniser->context.pending, &cptr, &len);
    2523. 

  2523. 

  2524. 	if (error != PARSERUTILS_OK) {
    2525. 

  2525. 		if (error == PARSERUTILS_EOF) {
    2526. 

  2526. 			tokeniser->state = STATE_DATA;
    2527. 

  2527. 			return emit_current_doctype(tokeniser, true);
    2528. 

  2528. 		} else {
    2529. 

  2529. 			return hubbub_error_from_parserutils_error(error);
    2530. 

  2530. 		}
    2531. 

  2531. 	}
    2532. 

  2532. 

  2533. 	c = *cptr;
    2534. 

  2534. 

  2535. 	if (c == '"') {
    2536. 

  2536. 		tokeniser->context.pending += len;
    2537. 

  2537. 		tokeniser->state = STATE_AFTER_DOCTYPE_SYSTEM;
    2538. 

  2538. 	} else if (c == '>') {
    2539. 

  2539. 		tokeniser->context.pending += len;
    2540. 

  2540. 		tokeniser->state = STATE_DATA;
    2541. 

  2541. 		return emit_current_doctype(tokeniser, true);
    2542. 

  2542. 	} else if (c == '\0') {
    2543. 

  2543. 		COLLECT_MS(cdoc->system_id, u_fffd, sizeof(u_fffd));
    2544. 

  2544. 		tokeniser->context.pending += len;
    2545. 

  2545. 	} else if (c == '\r') {
    2546. 

  2546. 		error = parserutils_inputstream_peek(
    2547. 

  2547. 				tokeniser->input,
    2548. 

  2548. 				tokeniser->context.pending,
    2549. 

  2549. 				&cptr,
    2550. 

  2550. 				&len);
    2551. 

  2551. 

  2552. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2553. 

  2553. 			return hubbub_error_from_parserutils_error(error);
    2554. 

  2554. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    2555. 

  2555. 			COLLECT_MS(cdoc->system_id, &lf, sizeof(lf));
    2556. 

  2556. 		}
    2557. 

  2557. 

  2558. 		/* Collect '\r' */
    2559. 

  2559. 		tokeniser->context.pending += 1;
    2560. 

  2560. 	} else {
    2561. 

  2561. 		COLLECT_MS(cdoc->system_id, cptr, len);
    2562. 

  2562. 		tokeniser->context.pending += len;
    2563. 

  2563. 	}
    2564. 

  2564. 

  2565. 	return HUBBUB_OK;
    2566. 

  2566. }
    2567. 

  2567. 

  2568. hubbub_error hubbub_tokeniser_handle_doctype_system_sq(
    2569. 

  2569. 		hubbub_tokeniser *tokeniser)
    2570. 

  2570. {
    2571. 

  2571. 	hubbub_doctype *cdoc = &tokeniser->context.current_doctype;
    2572. 

  2572. 	size_t len;
    2573. 

  2573. 	const uint8_t *cptr;
    2574. 

  2574. 	parserutils_error error;
    2575. 

  2575. 	uint8_t c;
    2576. 

  2576. 

  2577. 	error = parserutils_inputstream_peek(tokeniser->input,
    2578. 

  2578. 			tokeniser->context.pending, &cptr, &len);
    2579. 

  2579. 

  2580. 	if (error != PARSERUTILS_OK) {
    2581. 

  2581. 		if (error == PARSERUTILS_EOF) {
    2582. 

  2582. 			tokeniser->state = STATE_DATA;
    2583. 

  2583. 			return emit_current_doctype(tokeniser, true);
    2584. 

  2584. 		} else {
    2585. 

  2585. 			return hubbub_error_from_parserutils_error(error);
    2586. 

  2586. 		}
    2587. 

  2587. 	}
    2588. 

  2588. 

  2589. 	c = *cptr;
    2590. 

  2590. 

  2591. 	if (c == '\'') {
    2592. 

  2592. 		tokeniser->context.pending += len;
    2593. 

  2593. 		tokeniser->state = STATE_AFTER_DOCTYPE_SYSTEM;
    2594. 

  2594. 	} else if (c == '>') {
    2595. 

  2595. 		tokeniser->context.pending += len;
    2596. 

  2596. 		tokeniser->state = STATE_DATA;
    2597. 

  2597. 		return emit_current_doctype(tokeniser, true);
    2598. 

  2598. 	} else if (c == '\0') {
    2599. 

  2599. 		COLLECT_MS(cdoc->system_id, u_fffd, sizeof(u_fffd));
    2600. 

  2600. 		tokeniser->context.pending += len;
    2601. 

  2601. 	} else if (c == '\r') {
    2602. 

  2602. 		error = parserutils_inputstream_peek(
    2603. 

  2603. 				tokeniser->input,
    2604. 

  2604. 				tokeniser->context.pending,
    2605. 

  2605. 				&cptr,
    2606. 

  2606. 				&len);
    2607. 

  2607. 

  2608. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2609. 

  2609. 			return hubbub_error_from_parserutils_error(error);
    2610. 

  2610. 		} else if (error == PARSERUTILS_EOF || *cptr != '\n') {
    2611. 

  2611. 			COLLECT_MS(cdoc->system_id, &lf, sizeof(lf));
    2612. 

  2612. 		}
    2613. 

  2613. 

  2614. 		/* Collect '\r' */
    2615. 

  2615. 		tokeniser->context.pending += 1;
    2616. 

  2616. 	} else {
    2617. 

  2617. 		COLLECT_MS(cdoc->system_id, cptr, len);
    2618. 

  2618. 		tokeniser->context.pending += len;
    2619. 

  2619. 	}
    2620. 

  2620. 

  2621. 	return HUBBUB_OK;
    2622. 

  2622. }
    2623. 

  2623. 

  2624. hubbub_error hubbub_tokeniser_handle_after_doctype_system(
    2625. 

  2625. 		hubbub_tokeniser *tokeniser)
    2626. 

  2626. {
    2627. 

  2627. 	size_t len;
    2628. 

  2628. 	const uint8_t *cptr;
    2629. 

  2629. 	parserutils_error error;
    2630. 

  2630. 	uint8_t c;
    2631. 

  2631. 

  2632. 	error = parserutils_inputstream_peek(tokeniser->input,
    2633. 

  2633. 			tokeniser->context.pending, &cptr, &len);
    2634. 

  2634. 

  2635. 	if (error != PARSERUTILS_OK) {
    2636. 

  2636. 		if (error == PARSERUTILS_EOF) {
    2637. 

  2637. 			tokeniser->state = STATE_DATA;
    2638. 

  2638. 			return emit_current_doctype(tokeniser, true);
    2639. 

  2639. 		} else {
    2640. 

  2640. 			return hubbub_error_from_parserutils_error(error);
    2641. 

  2641. 		}
    2642. 

  2642. 	}
    2643. 

  2643. 

  2644. 	c = *cptr;
    2645. 

  2645. 	tokeniser->context.pending += len;
    2646. 

  2646. 

  2647. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' || c == '\r') {
    2648. 

  2648. 		/* pass over in silence */
    2649. 

  2649. 	} else if (c == '>') {
    2650. 

  2650. 		tokeniser->state = STATE_DATA;
    2651. 

  2651. 		return emit_current_doctype(tokeniser, false);
    2652. 

  2652. 	} else {
    2653. 

  2653. 		tokeniser->state = STATE_BOGUS_DOCTYPE;
    2654. 

  2654. 	}
    2655. 

  2655. 

  2656. 	return HUBBUB_OK;
    2657. 

  2657. }
    2658. 

  2658. 

  2659. 

  2660. hubbub_error hubbub_tokeniser_handle_bogus_doctype(hubbub_tokeniser *tokeniser)
    2661. 

  2661. {
    2662. 

  2662. 	size_t len;
    2663. 

  2663. 	const uint8_t *cptr;
    2664. 

  2664. 	parserutils_error error;
    2665. 

  2665. 	uint8_t c;
    2666. 

  2666. 

  2667. 	error = parserutils_inputstream_peek(tokeniser->input,
    2668. 

  2668. 			tokeniser->context.pending, &cptr, &len);
    2669. 

  2669. 

  2670. 	if (error != PARSERUTILS_OK) {
    2671. 

  2671. 		if (error == PARSERUTILS_EOF) {
    2672. 

  2672. 			tokeniser->state = STATE_DATA;
    2673. 

  2673. 			return emit_current_doctype(tokeniser, false);
    2674. 

  2674. 		} else {
    2675. 

  2675. 			return hubbub_error_from_parserutils_error(error);
    2676. 

  2676. 		}
    2677. 

  2677. 	}
    2678. 

  2678. 

  2679. 	c = *cptr;
    2680. 

  2680. 	tokeniser->context.pending += len;
    2681. 

  2681. 

  2682. 	if (c == '>') {
    2683. 

  2683. 		tokeniser->state = STATE_DATA;
    2684. 

  2684. 		return emit_current_doctype(tokeniser, false);
    2685. 

  2685. 	}
    2686. 

  2686. 

  2687. 	return HUBBUB_OK;
    2688. 

  2688. }
    2689. 

  2689. 

  2690. 

  2691. 

  2692. #define CDATA		"[CDATA["
    2693. 

  2693. #define CDATA_LEN	(SLEN(CDATA) - 1)
    2694. 

  2694. 

  2695. hubbub_error hubbub_tokeniser_handle_match_cdata(hubbub_tokeniser *tokeniser)
    2696. 

  2696. {
    2697. 

  2697. 	size_t len;
    2698. 

  2698. 	const uint8_t *cptr;
    2699. 

  2699. 	parserutils_error error;
    2700. 

  2700. 	uint8_t c;
    2701. 

  2701. 

  2702. 	error = parserutils_inputstream_peek(tokeniser->input,
    2703. 

  2703. 			tokeniser->context.pending, &cptr, &len);
    2704. 

  2704. 

  2705. 	if (error != PARSERUTILS_OK) {
    2706. 

  2706. 		if (error == PARSERUTILS_EOF) {
    2707. 

  2707. 			tokeniser->context.current_comment.len =
    2708. 

  2708. 					tokeniser->context.pending = 0;
    2709. 

  2709. 			tokeniser->state = STATE_BOGUS_COMMENT;
    2710. 

  2710. 			return HUBBUB_OK;
    2711. 

  2711. 		} else {
    2712. 

  2712. 			return hubbub_error_from_parserutils_error(error);
    2713. 

  2713. 		}
    2714. 

  2714. 	}
    2715. 

  2715. 

  2716. 	c = *cptr;
    2717. 

  2717. 

  2718. 	assert(tokeniser->context.match_cdata.count <= CDATA_LEN);
    2719. 

  2719. 

  2720. 	if (CDATA[tokeniser->context.match_cdata.count] != (c & ~0x20)) {
    2721. 

  2721. 		tokeniser->context.current_comment.len =
    2722. 

  2722. 				tokeniser->context.pending =
    2723. 

  2723. 				0;
    2724. 

  2724. 		tokeniser->state = STATE_BOGUS_COMMENT;
    2725. 

  2725. 		return HUBBUB_OK;
    2726. 

  2726. 	}
    2727. 

  2727. 

  2728. 	tokeniser->context.pending += len;
    2729. 

  2729. 

  2730. 	if (tokeniser->context.match_cdata.count == CDATA_LEN) {
    2731. 

  2731. 		parserutils_inputstream_advance(tokeniser->input,
    2732. 

  2732. 				tokeniser->context.match_cdata.count + len);
    2733. 

  2733. 		tokeniser->context.pending = 0;
    2734. 

  2734. 		tokeniser->context.match_cdata.end = 0;
    2735. 

  2735. 		tokeniser->state = STATE_CDATA_BLOCK;
    2736. 

  2736. 	}
    2737. 

  2737. 

  2738. 	tokeniser->context.match_cdata.count += len;
    2739. 

  2739. 

  2740. 	return HUBBUB_OK;
    2741. 

  2741. }
    2742. 

  2742. 

  2743. #undef CDATA
    2744. 

  2744. #undef CDATA_LEN
    2745. 

  2745. 

  2746. 

  2747. hubbub_error hubbub_tokeniser_handle_cdata_block(hubbub_tokeniser *tokeniser)
    2748. 

  2748. {
    2749. 

  2749. 	size_t len;
    2750. 

  2750. 	const uint8_t *cptr;
    2751. 

  2751. 	parserutils_error error;
    2752. 

  2752. 	uint8_t c;
    2753. 

  2753. 

  2754. 	error = parserutils_inputstream_peek(tokeniser->input,
    2755. 

  2755. 			tokeniser->context.pending, &cptr, &len);
    2756. 

  2756. 

  2757. 	if (error != PARSERUTILS_OK) {
    2758. 

  2758. 		if (error == PARSERUTILS_EOF) {
    2759. 

  2759. 			tokeniser->state = STATE_DATA;
    2760. 

  2760. 			return emit_current_chars(tokeniser);
    2761. 

  2761. 		} else {
    2762. 

  2762. 			return hubbub_error_from_parserutils_error(error);
    2763. 

  2763. 		}
    2764. 

  2764. 	}
    2765. 

  2765. 

  2766. 	c = *cptr;
    2767. 

  2767. 

  2768. 	if (c == ']' && (tokeniser->context.match_cdata.end == 0 ||
    2769. 

  2769. 			tokeniser->context.match_cdata.end == 1)) {
    2770. 

  2770. 		tokeniser->context.pending += len;
    2771. 

  2771. 		tokeniser->context.match_cdata.end += len;
    2772. 

  2772. 	} else if (c == '>' && tokeniser->context.match_cdata.end == 2) {
    2773. 

  2773. 		/* Remove the previous two "]]" */
    2774. 

  2774. 		tokeniser->context.pending -= 2;
    2775. 

  2775. 

  2776. 		/* Emit any pending characters */
    2777. 

  2777. 		emit_current_chars(tokeniser);
    2778. 

  2778. 

  2779. 		/* Now move past the "]]>" bit */
    2780. 

  2780. 		parserutils_inputstream_advance(tokeniser->input, SLEN("]]>"));
    2781. 

  2781. 

  2782. 		tokeniser->state = STATE_DATA;
    2783. 

  2783. 	} else if (c == '\0') {
    2784. 

  2784. 		if (tokeniser->context.pending > 0) {
    2785. 

  2785. 			/* Emit any pending characters */
    2786. 

  2786. 			emit_current_chars(tokeniser);
    2787. 

  2787. 		}
    2788. 

  2788. 

  2789. 		/* Perform NUL-byte replacement */
    2790. 

  2790. 		emit_character_token(tokeniser, &u_fffd_str);
    2791. 

  2791. 

  2792. 		parserutils_inputstream_advance(tokeniser->input, len);
    2793. 

  2793. 		tokeniser->context.match_cdata.end = 0;
    2794. 

  2794. 	} else if (c == '\r') {
    2795. 

  2795. 		error = parserutils_inputstream_peek(
    2796. 

  2796. 				tokeniser->input,
    2797. 

  2797. 				tokeniser->context.pending + len,
    2798. 

  2798. 				&cptr,
    2799. 

  2799. 				&len);
    2800. 

  2800. 

  2801. 		if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2802. 

  2802. 			return hubbub_error_from_parserutils_error(error);
    2803. 

  2803. 		}
    2804. 

  2804. 

  2805. 		if (tokeniser->context.pending > 0) {
    2806. 

  2806. 			/* Emit any pending characters */
    2807. 

  2807. 			emit_current_chars(tokeniser);
    2808. 

  2808. 		}
    2809. 

  2809. 

  2810. 		if (error == PARSERUTILS_EOF || *cptr != '\n') {
    2811. 

  2811. 			/* Emit newline */
    2812. 

  2812. 			emit_character_token(tokeniser, &lf_str);
    2813. 

  2813. 		}
    2814. 

  2814. 

  2815. 		/* Advance over \r */
    2816. 

  2816. 		parserutils_inputstream_advance(tokeniser->input, 1);
    2817. 

  2817. 		tokeniser->context.match_cdata.end = 0;
    2818. 

  2818. 	} else {
    2819. 

  2819. 		tokeniser->context.pending += len;
    2820. 

  2820. 		tokeniser->context.match_cdata.end = 0;
    2821. 

  2821. 	}
    2822. 

  2822. 

  2823. 	return HUBBUB_OK;
    2824. 

  2824. }
    2825. 

  2825. 

  2826. 

  2827. hubbub_error hubbub_tokeniser_consume_character_reference(
    2828. 

  2828. 		hubbub_tokeniser *tokeniser, size_t pos)
    2829. 

  2829. {
    2830. 

  2830. 	uint32_t allowed_char = tokeniser->context.allowed_char;
    2831. 

  2831. 

  2832. 	size_t len;
    2833. 

  2833. 	const uint8_t *cptr;
    2834. 

  2834. 	parserutils_error error;
    2835. 

  2835. 	uint8_t c;
    2836. 

  2836. 	size_t off;
    2837. 

  2837. 

  2838. 	error = parserutils_inputstream_peek(tokeniser->input, pos, 
    2839. 

  2839. 			&cptr, &len);
    2840. 

  2840. 

  2841. 	/* We should always start on an ampersand */
    2842. 

  2842. 	assert(error == PARSERUTILS_OK);
    2843. 

  2843. 	assert(len == 1 && *cptr == '&');
    2844. 

  2844. 

  2845. 	off = pos + len;
    2846. 

  2846. 

  2847. 	/* Look at the character after the ampersand */
    2848. 

  2848. 	error = parserutils_inputstream_peek(tokeniser->input, off, 
    2849. 

  2849. 			&cptr, &len);
    2850. 

  2850. 

  2851. 	if (error != PARSERUTILS_OK) {
    2852. 

  2852. 		if (error == PARSERUTILS_EOF) {
    2853. 

  2853. 			tokeniser->context.match_entity.complete = true;
    2854. 

  2854. 			tokeniser->context.match_entity.codepoint = 0;
    2855. 

  2855. 			return HUBBUB_OK;
    2856. 

  2856. 		} else {
    2857. 

  2857. 			return hubbub_error_from_parserutils_error(error);
    2858. 

  2858. 		}
    2859. 

  2859. 	}
    2860. 

  2860. 

  2861. 	c = *cptr;
    2862. 

  2862. 

  2863. 	/* Set things up */
    2864. 

  2864. 	tokeniser->context.match_entity.offset = off;
    2865. 

  2865. 	tokeniser->context.match_entity.poss_length = 0;
    2866. 

  2866. 	tokeniser->context.match_entity.length = 0;
    2867. 

  2867. 	tokeniser->context.match_entity.base = 0;
    2868. 

  2868. 	tokeniser->context.match_entity.codepoint = 0;
    2869. 

  2869. 	tokeniser->context.match_entity.had_data = false;
    2870. 

  2870. 	tokeniser->context.match_entity.return_state = tokeniser->state;
    2871. 

  2871. 	tokeniser->context.match_entity.complete = false;
    2872. 

  2872. 	tokeniser->context.match_entity.overflow = false;
    2873. 

  2873. 	tokeniser->context.match_entity.context = NULL;
    2874. 

  2874. 	tokeniser->context.match_entity.prev_len = len;
    2875. 

  2875. 

  2876. 	/* Reset allowed character for future calls */
    2877. 

  2877. 	tokeniser->context.allowed_char = '\0';
    2878. 

  2878. 

  2879. 	if (c == '\t' || c == '\n' || c == '\f' || c == ' ' ||
    2880. 

  2880. 			c == '<' || c == '&' ||
    2881. 

  2881. 			(allowed_char && c == allowed_char)) {
    2882. 

  2882. 		tokeniser->context.match_entity.complete = true;
    2883. 

  2883. 		tokeniser->context.match_entity.codepoint = 0;
    2884. 

  2884. 	} else if (c == '#') {
    2885. 

  2885. 		tokeniser->context.match_entity.length += len;
    2886. 

  2886. 		tokeniser->state = STATE_NUMBERED_ENTITY;
    2887. 

  2887. 	} else {
    2888. 

  2888. 		tokeniser->state = STATE_NAMED_ENTITY;
    2889. 

  2889. 	}
    2890. 

  2890. 

  2891. 	return HUBBUB_OK;
    2892. 

  2892. }
    2893. 

  2893. 

  2894. 

  2895. hubbub_error hubbub_tokeniser_handle_numbered_entity(
    2896. 

  2896. 		hubbub_tokeniser *tokeniser)
    2897. 

  2897. {
    2898. 

  2898. 	hubbub_tokeniser_context *ctx = &tokeniser->context;
    2899. 

  2899. 

  2900. 	size_t len;
    2901. 

  2901. 	const uint8_t *cptr;
    2902. 

  2902. 	parserutils_error error;
    2903. 

  2903. 

  2904. 	error = parserutils_inputstream_peek(tokeniser->input,
    2905. 

  2905. 			ctx->match_entity.offset + ctx->match_entity.length,
    2906. 

  2906. 			&cptr, &len);
    2907. 

  2907. 

  2908. 	if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2909. 

  2909. 		return hubbub_error_from_parserutils_error(error);
    2910. 

  2910. 	}
    2911. 

  2911. 

  2912. 	if (error != PARSERUTILS_EOF && ctx->match_entity.base == 0) {
    2913. 

  2913. 		uint8_t c = *cptr;
    2914. 

  2914. 		if ((c & ~0x20) == 'X') {
    2915. 

  2915. 			ctx->match_entity.base = 16;
    2916. 

  2916. 			ctx->match_entity.length += len;
    2917. 

  2917. 		} else {
    2918. 

  2918. 			ctx->match_entity.base = 10;
    2919. 

  2919. 		}
    2920. 

  2920. 	}
    2921. 

  2921. 

  2922. 	while ((error = parserutils_inputstream_peek(tokeniser->input,
    2923. 

  2923. 			ctx->match_entity.offset + ctx->match_entity.length,
    2924. 

  2924. 			&cptr, &len)) == PARSERUTILS_OK) {
    2925. 

  2925. 		uint8_t c = *cptr;
    2926. 

  2926. 

  2927. 		if (ctx->match_entity.base == 10 &&
    2928. 

  2928. 				('0' <= c && c <= '9')) {
    2929. 

  2929. 			ctx->match_entity.had_data = true;
    2930. 

  2930. 			ctx->match_entity.codepoint =
    2931. 

  2931. 				ctx->match_entity.codepoint * 10 + (c - '0');
    2932. 

  2932. 

  2933. 			ctx->match_entity.length += len;
    2934. 

  2934. 		} else if (ctx->match_entity.base == 16 &&
    2935. 

  2935. 				(('0' <= c && c <= '9') ||
    2936. 

  2936. 				('A' <= (c & ~0x20) &&
    2937. 

  2937. 						(c & ~0x20) <= 'F'))) {
    2938. 

  2938. 			ctx->match_entity.had_data = true;
    2939. 

  2939. 			ctx->match_entity.codepoint *= 16;
    2940. 

  2940. 

  2941. 			if ('0' <= c && c <= '9') {
    2942. 

  2942. 				ctx->match_entity.codepoint += (c - '0');
    2943. 

  2943. 			} else {
    2944. 

  2944. 				ctx->match_entity.codepoint +=
    2945. 

  2945. 						((c & ~0x20) - 'A' + 10);
    2946. 

  2946. 			}
    2947. 

  2947. 

  2948. 			ctx->match_entity.length += len;
    2949. 

  2949. 		} else {
    2950. 

  2950. 			break;
    2951. 

  2951. 		}
    2952. 

  2952. 

  2953. 		if (ctx->match_entity.codepoint >= 0x10FFFF) {
    2954. 

  2954. 			ctx->match_entity.overflow = true;
    2955. 

  2955. 		}
    2956. 

  2956. 	}
    2957. 

  2957. 

  2958. 	if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    2959. 

  2959. 		return hubbub_error_from_parserutils_error(error);
    2960. 

  2960. 	}
    2961. 

  2961. 

  2962. 	/* Eat trailing semicolon, if any */
    2963. 

  2963. 	if (error != PARSERUTILS_EOF && *cptr == ';') {
    2964. 

  2964. 		ctx->match_entity.length += len;
    2965. 

  2965. 	}
    2966. 

  2966. 

  2967. 	/* Had data, so calculate final codepoint */
    2968. 

  2968. 	if (ctx->match_entity.had_data) {
    2969. 

  2969. 		uint32_t cp = ctx->match_entity.codepoint;
    2970. 

  2970. 

  2971. 		if (0x80 <= cp && cp <= 0x9F) {
    2972. 

  2972. 			cp = cp1252Table[cp - 0x80];
    2973. 

  2973. 		} else if (cp == 0x0D) {
    2974. 

  2974. 			cp = 0x000A;
    2975. 

  2975. 		} else if (ctx->match_entity.overflow || 
    2976. 

  2976. 				cp <= 0x0008 || cp == 0x000B ||
    2977. 

  2977. 				(0x000E <= cp && cp <= 0x001F) ||
    2978. 

  2978. 				(0x007F <= cp && cp <= 0x009F) ||
    2979. 

  2979. 				(0xD800 <= cp && cp <= 0xDFFF) ||
    2980. 

  2980. 				(0xFDD0 <= cp && cp <= 0xFDEF) ||
    2981. 

  2981. 				(cp & 0xFFFE) == 0xFFFE) {
    2982. 

  2982. 			/* the check for cp > 0x10FFFF per spec is performed
    2983. 

  2983. 			 * in the loop above to avoid overflow */
    2984. 

  2984. 			cp = 0xFFFD;
    2985. 

  2985. 		}
    2986. 

  2986. 

  2987. 		ctx->match_entity.codepoint = cp;
    2988. 

  2988. 	}
    2989. 

  2989. 

  2990. 	/* Flag completion */
    2991. 

  2991. 	ctx->match_entity.complete = true;
    2992. 

  2992. 

  2993. 	/* And back to the state we were entered in */
    2994. 

  2994. 	tokeniser->state = ctx->match_entity.return_state;
    2995. 

  2995. 

  2996. 	return HUBBUB_OK;
    2997. 

  2997. }
    2998. 

  2998. 

  2999. hubbub_error hubbub_tokeniser_handle_named_entity(hubbub_tokeniser *tokeniser)
    3000. 

  3000. {
    3001. 

  3001. 	hubbub_tokeniser_context *ctx = &tokeniser->context;
    3002. 

  3002. 

  3003. 	size_t len;
    3004. 

  3004. 	const uint8_t *cptr;
    3005. 

  3005. 	parserutils_error error;
    3006. 

  3006. 

  3007. 	while ((error = parserutils_inputstream_peek(tokeniser->input,
    3008. 

  3008. 			ctx->match_entity.offset +
    3009. 

  3009. 					ctx->match_entity.poss_length,
    3010. 

  3010. 			&cptr, &len)) == PARSERUTILS_OK) {
    3011. 

  3011. 		uint32_t cp;
    3012. 

  3012. 

  3013. 		uint8_t c = *cptr;
    3014. 

  3014. 		hubbub_error error;
    3015. 

  3015. 

  3016. 		if (c > 0x7F) {
    3017. 

  3017. 			/* Entity names are ASCII only */
    3018. 

  3018. 			break;
    3019. 

  3019. 		}
    3020. 

  3020. 

  3021. 		error = hubbub_entities_search_step(c, &cp,
    3022. 

  3022. 				&ctx->match_entity.context);
    3023. 

  3023. 		if (error == HUBBUB_OK) {
    3024. 

  3024. 			/* Had a match - store it for later */
    3025. 

  3025. 			ctx->match_entity.codepoint = cp;
    3026. 

  3026. 

  3027. 			ctx->match_entity.length =
    3028. 

  3028. 					ctx->match_entity.poss_length + len;
    3029. 

  3029. 			ctx->match_entity.poss_length =
    3030. 

  3030. 					ctx->match_entity.length;
    3031. 

  3031. 		} else if (error == HUBBUB_INVALID) {
    3032. 

  3032. 			/* No further matches - use last found */
    3033. 

  3033. 			break;
    3034. 

  3034. 		} else {
    3035. 

  3035. 			/* Need more data */
    3036. 

  3036. 			ctx->match_entity.poss_length += len;
    3037. 

  3037. 		}
    3038. 

  3038. 	}
    3039. 

  3039. 

  3040. 	if (error != PARSERUTILS_OK && error != PARSERUTILS_EOF) {
    3041. 

  3041. 		return hubbub_error_from_parserutils_error(error);
    3042. 

  3042. 	}
    3043. 

  3043. 

  3044. 	if (ctx->match_entity.length > 0) {
    3045. 

  3045. 		uint8_t c;
    3046. 

  3046. 		error = parserutils_inputstream_peek(tokeniser->input,
    3047. 

  3047. 				ctx->match_entity.offset + 
    3048. 

  3048. 					ctx->match_entity.length - 1,
    3049. 

  3049. 				&cptr, &len);
    3050. 

  3050. 		/* We're re-reading a character we've already read after. 
    3051. 

  3051. 		 * Therefore, there's no way that an error may occur as 
    3052. 

  3052. 		 * a result. */
    3053. 

  3053. 		assert(error == PARSERUTILS_OK);
    3054. 

  3054. 

  3055. 		c = *cptr;
    3056. 

  3056. 

  3057. 		if ((tokeniser->context.match_entity.return_state ==
    3058. 

  3058. 				STATE_CHARACTER_REFERENCE_IN_ATTRIBUTE_VALUE) &&
    3059. 

  3059. 				c != ';') {
    3060. 

  3060. 			error = parserutils_inputstream_peek(tokeniser->input,
    3061. 

  3061. 					ctx->match_entity.offset +
    3062. 

  3062. 						ctx->match_entity.length,
    3063. 

  3063. 					&cptr, &len);
    3064. 

  3064. 			/* We must have attempted to read one more character 
    3065. 

  3065. 			 * than was present in the entity name, as that is the 
    3066. 

  3066. 			 * only way to break out of the loop above. If that 
    3067. 

  3067. 			 * failed, then any non-EOF case will have been handled
    3068. 

  3068. 			 * by the if statement after the loop thus it cannot 
    3069. 

  3069. 			 * occur here. */
    3070. 

  3070. 			assert(error == PARSERUTILS_OK || 
    3071. 

  3071. 					error == PARSERUTILS_EOF);
    3072. 

  3072. 

  3073. 			if (error == PARSERUTILS_EOF) {
    3074. 

  3074. 				ctx->match_entity.codepoint = 0;
    3075. 

  3075. 			}
    3076. 

  3076. 

  3077. 			c = *cptr;
    3078. 

  3078. 			if ((0x0030 <= c && c <= 0x0039) ||
    3079. 

  3079. 					(0x0041 <= c && c <= 0x005A) ||
    3080. 

  3080. 					(0x0061 <= c && c <= 0x007A)) {
    3081. 

  3081. 				ctx->match_entity.codepoint = 0;
    3082. 

  3082. 			}
    3083. 

  3083. 		}
    3084. 

  3084. 	}
    3085. 

  3085. 

  3086. 	/* Flag completion */
    3087. 

  3087. 	ctx->match_entity.complete = true;
    3088. 

  3088. 

  3089. 	/* And back to the state from whence we came */
    3090. 

  3090. 	tokeniser->state = ctx->match_entity.return_state;
    3091. 

  3091. 

  3092. 	return HUBBUB_OK;
    3093. 

  3093. }
    3094. 

  3094. 

  3095. 

  3096. 

  3097. /*** Token emitting bits ***/
    3098. 

  3098. 

  3099. /**
    3100. 

  3100.  * Emit a character token.
    3101. 

  3101.  *
    3102. 

  3102.  * \param tokeniser	Tokeniser instance
    3103. 

  3103.  * \param chars		Pointer to hubbub_string to emit
    3104. 

  3104.  * \return	true
    3105. 

  3105.  */
    3106. 

  3106. hubbub_error emit_character_token(hubbub_tokeniser *tokeniser,
    3107. 

  3107. 		const hubbub_string *chars)
    3108. 

  3108. {
    3109. 

  3109. 	hubbub_token token;
    3110. 

  3110. 

  3111. 	token.type = HUBBUB_TOKEN_CHARACTER;
    3112. 

  3112. 	token.data.character = *chars;
    3113. 

  3113. 

  3114. 	return hubbub_tokeniser_emit_token(tokeniser, &token);
    3115. 

  3115. }
    3116. 

  3116. 

  3117. /**
    3118. 

  3118.  * Emit the current pending characters being stored in the tokeniser context.
    3119. 

  3119.  *
    3120. 

  3120.  * \param tokeniser	Tokeniser instance
    3121. 

  3121.  * \return	true
    3122. 

  3122.  */
    3123. 

  3123. hubbub_error emit_current_chars(hubbub_tokeniser *tokeniser)
    3124. 

  3124. {
    3125. 

  3125. 	hubbub_token token;
    3126. 

  3126. 	size_t len;
    3127. 

  3127. 	const uint8_t *cptr = NULL;
    3128. 

  3128. 	parserutils_error error;
    3129. 

  3129. 

  3130. 	/* Calling this with nothing to output is a probable bug */
    3131. 

  3131. 	assert(tokeniser->context.pending > 0);
    3132. 

  3132. 

  3133. 	error = parserutils_inputstream_peek(tokeniser->input, 0, &cptr, &len);
    3134. 

  3134. 

  3135. 	assert(error == PARSERUTILS_OK);
    3136. 

  3136. 

  3137. 	token.type = HUBBUB_TOKEN_CHARACTER;
    3138. 

  3138. 	token.data.character.ptr = cptr;
    3139. 

  3139. 	token.data.character.len = tokeniser->context.pending;
    3140. 

  3140. 

  3141. 	return hubbub_tokeniser_emit_token(tokeniser, &token);
    3142. 

  3142. }
    3143. 

  3143. 

  3144. /**
    3145. 

  3145.  * Emit the current tag token being stored in the tokeniser context.
    3146. 

  3146.  *
    3147. 

  3147.  * \param tokeniser	Tokeniser instance
    3148. 

  3148.  * \return	true
    3149. 

  3149.  */
    3150. 

  3150. hubbub_error emit_current_tag(hubbub_tokeniser *tokeniser)
    3151. 

  3151. {
    3152. 

  3152. 	hubbub_error err;
    3153. 

  3153. 	hubbub_token token;
    3154. 

  3154. 	uint32_t n_attributes;
    3155. 

  3155. 	hubbub_attribute *attrs;
    3156. 

  3156. 	uint8_t *ptr;
    3157. 

  3157. 	uint32_t i, j;
    3158. 

  3158. 

  3159. 	/* Emit current tag */
    3160. 

  3160. 	token.type = tokeniser->context.current_tag_type;
    3161. 

  3161. 	token.data.tag = tokeniser->context.current_tag;
    3162. 

  3162.  	token.data.tag.ns = HUBBUB_NS_HTML;
    3163. 

  3163. 

  3164. 

  3165. 	n_attributes = token.data.tag.n_attributes;
    3166. 

  3166. 	attrs = token.data.tag.attributes;
    3167. 

  3167. 

  3168. 	/* Set pointers correctly... */
    3169. 

  3169. 	ptr = tokeniser->buffer->data;
    3170. 

  3170. 	token.data.tag.name.ptr = tokeniser->buffer->data;
    3171. 

  3171. 	ptr += token.data.tag.name.len;
    3172. 

  3172. 

  3173. 	for (i = 0; i < n_attributes; i++) {
    3174. 

  3174. 		attrs[i].name.ptr = ptr;
    3175. 

  3175. 		ptr += attrs[i].name.len;
    3176. 

  3176. 		attrs[i].value.ptr = ptr;
    3177. 

  3177. 		ptr += attrs[i].value.len;
    3178. 

  3178. 	}
    3179. 

  3179. 

  3180. 

  3181. 	/* Discard duplicate attributes */
    3182. 

  3182. 	for (i = 0; i < n_attributes; i++) {
    3183. 

  3183. 		for (j = 0; j < n_attributes; j++) {
    3184. 

  3184. 			uint32_t move;
    3185. 

  3185. 

  3186. 			if (j == i ||
    3187. 

  3187. 				attrs[i].name.len !=
    3188. 

  3188. 						attrs[j].name.len ||
    3189. 

  3189. 				strncmp((char *) attrs[i].name.ptr,
    3190. 

  3190. 					(char *) attrs[j].name.ptr,
    3191. 

  3191. 					attrs[i].name.len) != 0) {
    3192. 

  3192. 				/* Attributes don't match */
    3193. 

  3193. 				continue;
    3194. 

  3194. 			}
    3195. 

  3195. 

  3196. 			assert(i < j);
    3197. 

  3197. 

  3198. 			/* Calculate amount to move */
    3199. 

  3199. 			move = (n_attributes - 1 - j) *
    3200. 

  3200. 					sizeof(hubbub_attribute);
    3201. 

  3201. 

  3202. 			if (move > 0) {
    3203. 

  3203. 				memmove(&attrs[j],&attrs[j+1], move);
    3204. 

  3204. 			}
    3205. 

  3205. 

  3206. 			/* We've deleted an item, so we need to 
    3207. 

  3207. 			 * reprocess this index */
    3208. 

  3208. 			j--;
    3209. 

  3209. 

  3210. 			/* And reduce the number of attributes */
    3211. 

  3211. 			n_attributes--;
    3212. 

  3212. 		}
    3213. 

  3213. 	}
    3214. 

  3214. 

  3215. 	token.data.tag.n_attributes = n_attributes;
    3216. 

  3216. 

  3217. 	err = hubbub_tokeniser_emit_token(tokeniser, &token);
    3218. 

  3218. 

  3219. 	if (token.type == HUBBUB_TOKEN_START_TAG) {
    3220. 

  3220. 		/* Save start tag name for R?CDATA */
    3221. 

  3221. 		if (token.data.tag.name.len <
    3222. 

  3222. 			sizeof(tokeniser->context.last_start_tag_name)) {
    3223. 

  3223. 			strncpy((char *) tokeniser->context.last_start_tag_name,
    3224. 

  3224. 				(const char *) token.data.tag.name.ptr,
    3225. 

  3225. 				token.data.tag.name.len);
    3226. 

  3226. 			tokeniser->context.last_start_tag_len =
    3227. 

  3227. 					token.data.tag.name.len;
    3228. 

  3228. 		} else {
    3229. 

  3229. 			tokeniser->context.last_start_tag_name[0] = '\0';
    3230. 

  3230. 			tokeniser->context.last_start_tag_len = 0;
    3231. 

  3231. 		}
    3232. 

  3232. 	} else /* if (token->type == HUBBUB_TOKEN_END_TAG) */ {
    3233. 

  3233. 		/* Reset content model after R?CDATA elements */
    3234. 

  3234. 		tokeniser->content_model = HUBBUB_CONTENT_MODEL_PCDATA;
    3235. 

  3235. 	}
    3236. 

  3236. 

  3237. 	/* Reset the self-closing flag */
    3238. 

  3238. 	tokeniser->context.current_tag.self_closing = false;
    3239. 

  3239. 

  3240. 	return err;
    3241. 

  3241. }
    3242. 

  3242. 

  3243. /**
    3244. 

  3244.  * Emit the current comment token being stored in the tokeniser context.
    3245. 

  3245.  *
    3246. 

  3246.  * \param tokeniser	Tokeniser instance
    3247. 

  3247.  * \return	true
    3248. 

  3248.  */
    3249. 

  3249. hubbub_error emit_current_comment(hubbub_tokeniser *tokeniser)
    3250. 

  3250. {
    3251. 

  3251. 	hubbub_token token;
    3252. 

  3252. 

  3253. 	token.type = HUBBUB_TOKEN_COMMENT;
    3254. 

  3254. 	token.data.comment.ptr = tokeniser->buffer->data;
    3255. 

  3255. 	token.data.comment.len = tokeniser->buffer->length;
    3256. 

  3256. 

  3257. 	return hubbub_tokeniser_emit_token(tokeniser, &token);
    3258. 

  3258. }
    3259. 

  3259. 

  3260. /**
    3261. 

  3261.  * Emit the current doctype token being stored in the tokeniser context.
    3262. 

  3262.  *
    3263. 

  3263.  * \param tokeniser	Tokeniser instance
    3264. 

  3264.  * \param force_quirks	Force quirks mode on this document
    3265. 

  3265.  * \return	true
    3266. 

  3266.  */
    3267. 

  3267. hubbub_error emit_current_doctype(hubbub_tokeniser *tokeniser,
    3268. 

  3268. 		bool force_quirks)
    3269. 

  3269. {
    3270. 

  3270. 	hubbub_token token;
    3271. 

  3271. 

  3272. 	/* Emit doctype */
    3273. 

  3273. 	token.type = HUBBUB_TOKEN_DOCTYPE;
    3274. 

  3274. 	token.data.doctype = tokeniser->context.current_doctype;
    3275. 

  3275. 	if (force_quirks == true)
    3276. 

  3276. 		token.data.doctype.force_quirks = true;
    3277. 

  3277. 

  3278. 	/* Set pointers correctly */
    3279. 

  3279. 	token.data.doctype.name.ptr = tokeniser->buffer->data;
    3280. 

  3280. 

  3281. 	if (token.data.doctype.public_missing == false) {
    3282. 

  3282. 		token.data.doctype.public_id.ptr = tokeniser->buffer->data + 
    3283. 

  3283. 				token.data.doctype.name.len;
    3284. 

  3284. 	}
    3285. 

  3285. 

  3286. 	if (token.data.doctype.system_missing == false) {
    3287. 

  3287. 		token.data.doctype.system_id.ptr = tokeniser->buffer->data +
    3288. 

  3288. 				token.data.doctype.name.len +
    3289. 

  3289. 				token.data.doctype.public_id.len;
    3290. 

  3290. 	}
    3291. 

  3291. 

  3292. 	return hubbub_tokeniser_emit_token(tokeniser, &token);
    3293. 

  3293. }
    3294. 

  3294. 

  3295. /**
    3296. 

  3296.  * Emit a token, performing sanity checks if necessary
    3297. 

  3297.  *
    3298. 

  3298.  * \param tokeniser  Tokeniser instance
    3299. 

  3299.  * \param token      Token to emit
    3300. 

  3300.  */
    3301. 

  3301. hubbub_error hubbub_tokeniser_emit_token(hubbub_tokeniser *tokeniser,
    3302. 

  3302. 		hubbub_token *token)
    3303. 

  3303. {
    3304. 

  3304. 	hubbub_error err = HUBBUB_OK;
    3305. 

  3305. 

  3306. 	assert(tokeniser != NULL);
    3307. 

  3307. 	assert(token != NULL);
    3308. 

  3308. 

  3309. #ifndef NDEBUG
    3310. 

  3310. 	/* Sanity checks */
    3311. 

  3311. 	switch (token->type) {
    3312. 

  3312. 	case HUBBUB_TOKEN_DOCTYPE:
    3313. 

  3313. 		assert(memchr(token->data.doctype.name.ptr, 0xff, 
    3314. 

  3314. 				token->data.doctype.name.len) == NULL);
    3315. 

  3315. 		if (token->data.doctype.public_missing == false)
    3316. 

  3316. 			assert(memchr(token->data.doctype.public_id.ptr, 0xff,
    3317. 

  3317. 				token->data.doctype.public_id.len) == NULL);
    3318. 

  3318. 		if (token->data.doctype.system_missing == false)
    3319. 

  3319. 			assert(memchr(token->data.doctype.system_id.ptr, 0xff,
    3320. 

  3320. 				token->data.doctype.system_id.len) == NULL);
    3321. 

  3321. 		break;
    3322. 

  3322. 	case HUBBUB_TOKEN_START_TAG:
    3323. 

  3323. 	case HUBBUB_TOKEN_END_TAG:
    3324. 

  3324. 	{
    3325. 

  3325. 		uint32_t i;
    3326. 

  3326. 		assert(memchr(token->data.tag.name.ptr, 0xff, 
    3327. 

  3327. 				token->data.tag.name.len) == NULL);
    3328. 

  3328. 		for (i = 0; i < token->data.tag.n_attributes; i++) {
    3329. 

  3329. 			hubbub_attribute *attr = &token->data.tag.attributes[i];
    3330. 

  3330. 

  3331. 			assert(memchr(attr->name.ptr, 0xff, attr->name.len) == 
    3332. 

  3332. 					NULL);
    3333. 

  3333. 			assert(memchr(attr->value.ptr, 0xff, attr->value.len) ==
    3334. 

  3334. 					NULL);
    3335. 

  3335. 		}
    3336. 

  3336. 	}
    3337. 

  3337. 		break;
    3338. 

  3338. 	case HUBBUB_TOKEN_COMMENT:
    3339. 

  3339. 		assert(memchr(token->data.comment.ptr, 0xff, 
    3340. 

  3340. 				token->data.comment.len) == NULL);
    3341. 

  3341. 		break;
    3342. 

  3342. 	case HUBBUB_TOKEN_CHARACTER:
    3343. 

  3343. 		assert(memchr(token->data.character.ptr, 0xff,
    3344. 

  3344. 				token->data.character.len) == NULL);
    3345. 

  3345. 		break;
    3346. 

  3346. 	case HUBBUB_TOKEN_EOF:
    3347. 

  3347. 		break;
    3348. 

  3348. 	}
    3349. 

  3349. #endif
    3350. 

  3350. 

  3351. 	/* Emit the token */
    3352. 

  3352. 	if (tokeniser->token_handler) {
    3353. 

  3353. 		err = tokeniser->token_handler(token, tokeniser->token_pw);
    3354. 

  3354. 	}
    3355. 

  3355. 

  3356. 	/* Discard current buffer */
    3357. 

  3357. 	if (tokeniser->buffer->length) {
    3358. 

  3358. 		parserutils_buffer_discard(tokeniser->buffer, 0,
    3359. 

  3359. 				tokeniser->buffer->length);
    3360. 

  3360. 	}
    3361. 

  3361. 

  3362. 	/* Advance the pointer */
    3363. 

  3363. 	if (tokeniser->context.pending) {
    3364. 

  3364. 		parserutils_inputstream_advance(tokeniser->input,
    3365. 

  3365. 				tokeniser->context.pending);
    3366. 

  3366. 		tokeniser->context.pending = 0;
    3367. 

  3367. 	}
    3368. 

  3368. 

  3369. 	return err;
    3370. 

  3370. }
    3371. 

  3371.