1 /*
2 +----------------------------------------------------------------------+
3 | Copyright (c) The PHP Group |
4 +----------------------------------------------------------------------+
5 | This source file is subject to version 3.01 of the PHP license, |
6 | that is bundled with this package in the file LICENSE, and is |
7 | available through the world-wide-web at the following url: |
8 | https://www.php.net/license/3_01.txt |
9 | If you did not receive a copy of the PHP license and are unable to |
10 | obtain it through the world-wide-web, please send a note to |
11 | license@php.net so we can mail you a copy immediately. |
12 +----------------------------------------------------------------------+
13 | Authors: Niels Dossche <nielsdos@php.net> |
14 +----------------------------------------------------------------------+
15 */
16
17 #ifdef HAVE_CONFIG_H
18 #include <config.h>
19 #endif
20
21 #include "php.h"
22 #if defined(HAVE_LIBXML) && defined(HAVE_DOM)
23 #include "php_dom.h"
24 #include "infra.h"
25 #include "html5_parser.h"
26 #include "html5_serializer.h"
27 #include "namespace_compat.h"
28 #include "private_data.h"
29 #include "dom_properties.h"
30 #include <Zend/zend_smart_string.h>
31 #include <lexbor/html/encoding.h>
32 #include <lexbor/encoding/encoding.h>
33
34 /* Implementation defined, but as HTML5 defaults in all other cases to UTF-8, we'll do the same. */
35 #define DOM_FALLBACK_ENCODING_ID LXB_ENCODING_UTF_8
36
37 typedef struct dom_line_column_cache {
38 size_t last_line;
39 size_t last_column;
40 size_t last_offset;
41 } dom_line_column_cache;
42
43 typedef struct dom_lexbor_libxml2_bridge_application_data {
44 const char *input_name;
45 const lxb_codepoint_t *current_input_codepoints;
46 const char *current_input_characters;
47 size_t current_input_length;
48 size_t current_total_offset;
49 dom_line_column_cache cache_tokenizer;
50 bool html_no_implied;
51 } dom_lexbor_libxml2_bridge_application_data;
52
53 typedef struct dom_character_encoding_data {
54 const lxb_encoding_data_t *encoding_data;
55 size_t bom_shift;
56 } dom_character_encoding_data;
57
58 typedef zend_result (*dom_write_output)(void*, const char *, size_t);
59
60 typedef struct dom_output_ctx {
61 const lxb_encoding_data_t *encoding_data;
62 const lxb_encoding_data_t *decoding_data;
63 lxb_encoding_encode_t *encode;
64 lxb_encoding_decode_t *decode;
65 lxb_codepoint_t *codepoints;
66 lxb_char_t *encoding_output;
67 void *output_data;
68 dom_write_output write_output;
69 } dom_output_ctx;
70
71 typedef struct dom_decoding_encoding_ctx {
72 /* We can skip some conversion if the input and output encoding are both UTF-8,
73 * we only have to validate and substitute replacement characters */
74 bool fast_path; /* Put first, near the encode & decode structures, for cache locality */
75 lxb_encoding_encode_t encode;
76 lxb_encoding_decode_t decode;
77 const lxb_encoding_data_t *encode_data;
78 const lxb_encoding_data_t *decode_data;
79 lxb_char_t encoding_output[4096];
80 lxb_codepoint_t codepoints[4096];
81 } dom_decoding_encoding_ctx;
82
83 /* https://dom.spec.whatwg.org/#dom-document-implementation */
dom_modern_document_implementation_read(dom_object * obj,zval * retval)84 zend_result dom_modern_document_implementation_read(dom_object *obj, zval *retval)
85 {
86 const uint32_t PROP_INDEX = 0;
87
88 #if ZEND_DEBUG
89 zend_string *implementation_str = ZSTR_INIT_LITERAL("implementation", false);
90 const zend_property_info *prop_info = zend_get_property_info(dom_abstract_base_document_class_entry, implementation_str, 0);
91 zend_string_release_ex(implementation_str, false);
92 ZEND_ASSERT(OBJ_PROP_TO_NUM(prop_info->offset) == PROP_INDEX);
93 #endif
94
95 zval *cached_implementation = OBJ_PROP_NUM(&obj->std, PROP_INDEX);
96 if (Z_ISUNDEF_P(cached_implementation)) {
97 php_dom_create_implementation(cached_implementation, true);
98 }
99
100 ZVAL_OBJ_COPY(retval, Z_OBJ_P(cached_implementation));
101
102 return SUCCESS;
103 }
104
dom_decoding_encoding_ctx_init(dom_decoding_encoding_ctx * ctx)105 static void dom_decoding_encoding_ctx_init(dom_decoding_encoding_ctx *ctx)
106 {
107 ctx->encode_data = lxb_encoding_data(LXB_ENCODING_UTF_8);
108 ctx->decode_data = NULL;
109 /* Set fast path on by default so that the decoder finishing is skipped if this was never initialised properly. */
110 ctx->fast_path = true;
111 (void) lxb_encoding_encode_init(
112 &ctx->encode,
113 ctx->encode_data,
114 ctx->encoding_output,
115 sizeof(ctx->encoding_output) / sizeof(*ctx->encoding_output)
116 );
117 (void) lxb_encoding_encode_replace_set(&ctx->encode, LXB_ENCODING_REPLACEMENT_BYTES, LXB_ENCODING_REPLACEMENT_SIZE);
118 }
119
dom_lexbor_tokenizer_error_code_to_string(lxb_html_tokenizer_error_id_t id)120 static const char *dom_lexbor_tokenizer_error_code_to_string(lxb_html_tokenizer_error_id_t id)
121 {
122 switch (id) {
123 case LXB_HTML_TOKENIZER_ERROR_ABCLOFEMCO: return "abrupt-closing-of-empty-comment";
124 case LXB_HTML_TOKENIZER_ERROR_ABDOPUID: return "abrupt-doctype-public-identifier";
125 case LXB_HTML_TOKENIZER_ERROR_ABDOSYID: return "abrupt-doctype-system-identifier";
126 case LXB_HTML_TOKENIZER_ERROR_ABOFDIINNUCHRE: return "absence-of-digits-in-numeric-character-reference";
127 case LXB_HTML_TOKENIZER_ERROR_CDINHTCO: return "cdata-in-html-content";
128 case LXB_HTML_TOKENIZER_ERROR_CHREOUUNRA: return "character-reference-outside-unicode-range";
129 case LXB_HTML_TOKENIZER_ERROR_COCHININST: return "control-character-in-input-stream";
130 case LXB_HTML_TOKENIZER_ERROR_COCHRE: return "control-character-reference";
131 case LXB_HTML_TOKENIZER_ERROR_ENTAWIAT: return "end-tag-with-attributes";
132 case LXB_HTML_TOKENIZER_ERROR_DUAT: return "duplicate-attribute";
133 case LXB_HTML_TOKENIZER_ERROR_ENTAWITRSO: return "end-tag-with-trailing-solidus";
134 case LXB_HTML_TOKENIZER_ERROR_EOBETANA: return "eof-before-tag-name";
135 case LXB_HTML_TOKENIZER_ERROR_EOINCD: return "eof-in-cdata";
136 case LXB_HTML_TOKENIZER_ERROR_EOINCO: return "eof-in-comment";
137 case LXB_HTML_TOKENIZER_ERROR_EOINDO: return "eof-in-doctype";
138 case LXB_HTML_TOKENIZER_ERROR_EOINSCHTCOLITE: return "eof-in-script-html-comment-like-text";
139 case LXB_HTML_TOKENIZER_ERROR_EOINTA: return "eof-in-tag";
140 case LXB_HTML_TOKENIZER_ERROR_INCLCO: return "incorrectly-closed-comment";
141 case LXB_HTML_TOKENIZER_ERROR_INOPCO: return "incorrectly-opened-comment";
142 case LXB_HTML_TOKENIZER_ERROR_INCHSEAFDONA: return "invalid-character-sequence-after-doctype-name";
143 case LXB_HTML_TOKENIZER_ERROR_INFICHOFTANA: return "invalid-first-character-of-tag-name";
144 case LXB_HTML_TOKENIZER_ERROR_MIATVA: return "missing-attribute-value";
145 case LXB_HTML_TOKENIZER_ERROR_MIDONA: return "missing-doctype-name";
146 case LXB_HTML_TOKENIZER_ERROR_MIDOPUID: return "missing-doctype-public-identifier";
147 case LXB_HTML_TOKENIZER_ERROR_MIDOSYID: return "missing-doctype-system-identifier";
148 case LXB_HTML_TOKENIZER_ERROR_MIENTANA: return "missing-end-tag-name";
149 case LXB_HTML_TOKENIZER_ERROR_MIQUBEDOPUID: return "missing-quote-before-doctype-public-identifier";
150 case LXB_HTML_TOKENIZER_ERROR_MIQUBEDOSYID: return "missing-quote-before-doctype-system-identifier";
151 case LXB_HTML_TOKENIZER_ERROR_MISEAFCHRE: return "missing-semicolon-after-character-reference";
152 case LXB_HTML_TOKENIZER_ERROR_MIWHAFDOPUKE: return "missing-whitespace-after-doctype-public-keyword";
153 case LXB_HTML_TOKENIZER_ERROR_MIWHAFDOSYKE: return "missing-whitespace-after-doctype-system-keyword";
154 case LXB_HTML_TOKENIZER_ERROR_MIWHBEDONA: return "missing-whitespace-before-doctype-name";
155 case LXB_HTML_TOKENIZER_ERROR_MIWHBEAT: return "missing-whitespace-between-attributes";
156 case LXB_HTML_TOKENIZER_ERROR_MIWHBEDOPUANSYID: return "missing-whitespace-between-doctype-public-and-system-identifiers";
157 case LXB_HTML_TOKENIZER_ERROR_NECO: return "nested-comment";
158 case LXB_HTML_TOKENIZER_ERROR_NOCHRE: return "noncharacter-character-reference";
159 case LXB_HTML_TOKENIZER_ERROR_NOININST: return "noncharacter-in-input-stream";
160 case LXB_HTML_TOKENIZER_ERROR_NOVOHTELSTTAWITRSO: return "non-void-html-element-start-tag-with-trailing-solidus";
161 case LXB_HTML_TOKENIZER_ERROR_NUCHRE: return "null-character-reference";
162 case LXB_HTML_TOKENIZER_ERROR_SUCHRE: return "surrogate-character-reference";
163 case LXB_HTML_TOKENIZER_ERROR_SUININST: return "surrogate-in-input-stream";
164 case LXB_HTML_TOKENIZER_ERROR_UNCHAFDOSYID: return "unexpected-character-after-doctype-system-identifier";
165 case LXB_HTML_TOKENIZER_ERROR_UNCHINATNA: return "unexpected-character-in-attribute-name";
166 case LXB_HTML_TOKENIZER_ERROR_UNCHINUNATVA: return "unexpected-character-in-unquoted-attribute-value";
167 case LXB_HTML_TOKENIZER_ERROR_UNEQSIBEATNA: return "unexpected-equals-sign-before-attribute-name";
168 case LXB_HTML_TOKENIZER_ERROR_UNNUCH: return "unexpected-null-character";
169 case LXB_HTML_TOKENIZER_ERROR_UNQUMAINOFTANA: return "unexpected-question-mark-instead-of-tag-name";
170 case LXB_HTML_TOKENIZER_ERROR_UNSOINTA: return "unexpected-solidus-in-tag";
171 case LXB_HTML_TOKENIZER_ERROR_UNNACHRE: return "unknown-named-character-reference";
172 default: return "unknown error";
173 }
174 }
175
dom_lexbor_tree_error_code_to_string(lxb_html_tree_error_id_t id)176 static const char *dom_lexbor_tree_error_code_to_string(lxb_html_tree_error_id_t id)
177 {
178 switch (id) {
179 case LXB_HTML_RULES_ERROR_UNTO: return "unexpected-token";
180 case LXB_HTML_RULES_ERROR_UNCLTO: return "unexpected-closed-token";
181 case LXB_HTML_RULES_ERROR_NUCH: return "null-character";
182 case LXB_HTML_RULES_ERROR_UNCHTO: return "unexpected-character-token";
183 case LXB_HTML_RULES_ERROR_UNTOININMO: return "unexpected-token-in-initial-mode";
184 case LXB_HTML_RULES_ERROR_BADOTOININMO: return "bad-doctype-token-in-initial-mode";
185 case LXB_HTML_RULES_ERROR_DOTOINBEHTMO: return "doctype-token-in-before-html-mode";
186 case LXB_HTML_RULES_ERROR_UNCLTOINBEHTMO: return "unexpected-closed-token-in-before-html-mode";
187 case LXB_HTML_RULES_ERROR_DOTOINBEHEMO: return "doctype-token-in-before-head-mode";
188 case LXB_HTML_RULES_ERROR_UNCLTOINBEHEMO: return "unexpected-closed_token-in-before-head-mode";
189 case LXB_HTML_RULES_ERROR_DOTOINHEMO: return "doctype-token-in-head-mode";
190 case LXB_HTML_RULES_ERROR_NOVOHTELSTTAWITRSO: return "non-void-html-element-start-tag-with-trailing-solidus";
191 case LXB_HTML_RULES_ERROR_HETOINHEMO: return "head-token-in-head-mode";
192 case LXB_HTML_RULES_ERROR_UNCLTOINHEMO: return "unexpected-closed-token-in-head-mode";
193 case LXB_HTML_RULES_ERROR_TECLTOWIOPINHEMO: return "template-closed-token-without-opening-in-head-mode";
194 case LXB_HTML_RULES_ERROR_TEELISNOCUINHEMO: return "template-element-is-not-current-in-head-mode";
195 case LXB_HTML_RULES_ERROR_DOTOINHENOMO: return "doctype-token-in-head-noscript-mode";
196 case LXB_HTML_RULES_ERROR_DOTOAFHEMO: return "doctype-token-after-head-mode";
197 case LXB_HTML_RULES_ERROR_HETOAFHEMO: return "head-token-after-head-mode";
198 case LXB_HTML_RULES_ERROR_DOTOINBOMO: return "doctype-token-in-body-mode";
199 case LXB_HTML_RULES_ERROR_BAENOPELISWR: return "bad-ending-open-elements-is-wrong";
200 case LXB_HTML_RULES_ERROR_OPELISWR: return "open-elements-is-wrong";
201 case LXB_HTML_RULES_ERROR_UNELINOPELST: return "unexpected-element-in-open-elements-stack";
202 case LXB_HTML_RULES_ERROR_MIELINOPELST: return "missing-element-in-open-elements-stack";
203 case LXB_HTML_RULES_ERROR_NOBOELINSC: return "no-body-element-in-scope";
204 case LXB_HTML_RULES_ERROR_MIELINSC: return "missing-element-in-scope";
205 case LXB_HTML_RULES_ERROR_UNELINSC: return "unexpected-element-in-scope";
206 case LXB_HTML_RULES_ERROR_UNELINACFOST: return "unexpected-element-in-active-formatting-stack";
207 case LXB_HTML_RULES_ERROR_UNENOFFI: return "unexpected-end-of-file";
208 case LXB_HTML_RULES_ERROR_CHINTATE: return "characters-in-table-text";
209 case LXB_HTML_RULES_ERROR_DOTOINTAMO: return "doctype-token-in-table-mode";
210 case LXB_HTML_RULES_ERROR_DOTOINSEMO: return "doctype-token-in-select-mode";
211 case LXB_HTML_RULES_ERROR_DOTOAFBOMO: return "doctype-token-after-body-mode";
212 case LXB_HTML_RULES_ERROR_DOTOINFRMO: return "doctype-token-in-frameset-mode";
213 case LXB_HTML_RULES_ERROR_DOTOAFFRMO: return "doctype-token-after-frameset-mode";
214 case LXB_HTML_RULES_ERROR_DOTOFOCOMO: return "doctype-token-foreign-content-mode";
215 default: return "unknown error";
216 }
217 }
218
dom_lexbor_libxml2_bridge_status_code_to_string(lexbor_libxml2_bridge_status status)219 static const char *dom_lexbor_libxml2_bridge_status_code_to_string(lexbor_libxml2_bridge_status status)
220 {
221 switch (status) {
222 case LEXBOR_LIBXML2_BRIDGE_STATUS_CANNOT_INIT: return "cannot initialize data structures";
223 case LEXBOR_LIBXML2_BRIDGE_STATUS_FATAL_PARSE: return "fatal error in parsing";
224 case LEXBOR_LIBXML2_BRIDGE_STATUS_OVERFLOW: return "string length overflow";
225 case LEXBOR_LIBXML2_BRIDGE_STATUS_OOM: return "out of memory";
226 default: return "unknown error";
227 }
228 }
229
dom_reset_line_column_cache(dom_line_column_cache * cache)230 static void dom_reset_line_column_cache(dom_line_column_cache *cache)
231 {
232 cache->last_line = 1;
233 cache->last_column = 1;
234 cache->last_offset = 0;
235 }
236
dom_find_line_and_column_using_cache(const dom_lexbor_libxml2_bridge_application_data * application_data,dom_line_column_cache * cache,size_t offset)237 static void dom_find_line_and_column_using_cache(
238 const dom_lexbor_libxml2_bridge_application_data *application_data,
239 dom_line_column_cache *cache,
240 size_t offset
241 )
242 {
243 offset -= application_data->current_total_offset;
244 if (offset > application_data->current_input_length) {
245 /* Possible with empty input, also just good for general safety */
246 offset = application_data->current_input_length;
247 }
248
249 size_t last_column = cache->last_column;
250 size_t last_line = cache->last_line;
251 size_t last_offset = cache->last_offset;
252
253 /* Either unicode or UTF-8 data */
254 if (application_data->current_input_codepoints != NULL) {
255 while (last_offset < offset) {
256 if (application_data->current_input_codepoints[last_offset] == 0x000A /* Unicode codepoint for line feed */) {
257 last_line++;
258 last_column = 1;
259 } else {
260 last_column++;
261 }
262 last_offset++;
263 }
264 } else {
265 while (last_offset < offset) {
266 const lxb_char_t current = application_data->current_input_characters[last_offset];
267 if (current == '\n') {
268 last_line++;
269 last_column = 1;
270 last_offset++;
271 } else {
272 /* See Lexbor tokenizer patch
273 * Note for future self: branchlessly computing the length and jumping by the length would be nice,
274 * however it takes so many instructions to do so that it is slower than this naive method. */
275 if ((current & 0b11000000) != 0b10000000) {
276 last_column++;
277 }
278 last_offset++;
279 }
280 }
281 }
282
283 cache->last_column = last_column;
284 cache->last_line = last_line;
285 cache->last_offset = last_offset;
286 }
287
dom_lexbor_libxml2_bridge_tokenizer_error_reporter(void * application_data_voidptr,lxb_html_tokenizer_error_t * error,size_t offset)288 static void dom_lexbor_libxml2_bridge_tokenizer_error_reporter(
289 void *application_data_voidptr,
290 lxb_html_tokenizer_error_t *error,
291 size_t offset
292 )
293 {
294 dom_lexbor_libxml2_bridge_application_data *application_data = application_data_voidptr;
295 dom_find_line_and_column_using_cache(application_data, &application_data->cache_tokenizer, offset);
296 php_libxml_pretend_ctx_error_ex(application_data->input_name, application_data->cache_tokenizer.last_line, application_data->cache_tokenizer.last_column, "tokenizer error %s in %s, line: %zu, column: %zu\n", dom_lexbor_tokenizer_error_code_to_string(error->id), application_data->input_name, application_data->cache_tokenizer.last_line, application_data->cache_tokenizer.last_column);
297 }
298
dom_lexbor_libxml2_bridge_tree_error_reporter(void * application_data_voidptr,lxb_html_tree_error_t * error,size_t line,size_t column,size_t len)299 static void dom_lexbor_libxml2_bridge_tree_error_reporter(
300 void *application_data_voidptr,
301 lxb_html_tree_error_t *error,
302 size_t line,
303 size_t column,
304 size_t len
305 )
306 {
307 dom_lexbor_libxml2_bridge_application_data *application_data = application_data_voidptr;
308
309 if (line == 1 && application_data->html_no_implied && error->id == LXB_HTML_RULES_ERROR_UNTOININMO) {
310 /* For no implied mode, we want to mimick libxml's behaviour of not reporting an error for a lacking doctype. */
311 return;
312 }
313
314 if (len <= 1) {
315 /* Possible with EOF, or single-character tokens, don't use a range in the error display in this case */
316 php_libxml_pretend_ctx_error_ex(
317 application_data->input_name,
318 line,
319 column,
320 "tree error %s in %s, line: %zu, column: %zu\n",
321 dom_lexbor_tree_error_code_to_string(error->id),
322 application_data->input_name,
323 line,
324 column
325 );
326 } else {
327 php_libxml_pretend_ctx_error_ex(
328 application_data->input_name,
329 line,
330 column,
331 "tree error %s in %s, line: %zu, column: %zu-%zu\n",
332 dom_lexbor_tree_error_code_to_string(error->id),
333 application_data->input_name,
334 line,
335 column,
336 column + len - 1
337 );
338 }
339 }
340
dom_search_child(xmlNodePtr parent,const char * searching_for)341 static xmlNodePtr dom_search_child(xmlNodePtr parent, const char *searching_for)
342 {
343 xmlNodePtr node = parent->children;
344 while (node != NULL) {
345 if (node->type == XML_ELEMENT_NODE && strcmp((const char *) node->name, searching_for) == 0) {
346 return node;
347 }
348 node = node->next;
349 }
350 return NULL;
351 }
352
dom_place_remove_element_and_hoist_children(xmlNodePtr parent,const char * searching_for)353 static void dom_place_remove_element_and_hoist_children(xmlNodePtr parent, const char *searching_for)
354 {
355 xmlNodePtr node = dom_search_child(parent, searching_for);
356 if (node != NULL) {
357 xmlUnlinkNode(node);
358
359 xmlNodePtr child = node->children;
360 while (child != NULL) {
361 xmlUnlinkNode(child);
362 xmlAddChild(parent, child);
363 child = node->children;
364 }
365
366 xmlFreeNode(node);
367 }
368 }
369
dom_post_process_html5_loading(xmlDocPtr lxml_doc,zend_long options,const lexbor_libxml2_bridge_extracted_observations * observations)370 static void dom_post_process_html5_loading(
371 xmlDocPtr lxml_doc,
372 zend_long options,
373 const lexbor_libxml2_bridge_extracted_observations *observations
374 )
375 {
376 if (options & HTML_PARSE_NOIMPLIED) {
377 xmlNodePtr html_node = dom_search_child((xmlNodePtr) lxml_doc, "html");
378 if (!observations->has_explicit_head_tag) {
379 dom_place_remove_element_and_hoist_children(html_node, "head");
380 }
381 if (!observations->has_explicit_body_tag) {
382 dom_place_remove_element_and_hoist_children(html_node, "body");
383 }
384 if (!observations->has_explicit_html_tag) {
385 dom_place_remove_element_and_hoist_children((xmlNodePtr) lxml_doc, "html");
386 }
387 }
388 }
389
390 /* https://html.spec.whatwg.org/multipage/parsing.html#determining-the-character-encoding */
dom_determine_encoding(const char * source,size_t source_len)391 static dom_character_encoding_data dom_determine_encoding(const char *source, size_t source_len)
392 {
393 dom_character_encoding_data result;
394
395 /* BOM sniffing */
396 if (source_len >= 3 && source[0] == '\xEF' && source[1] == '\xBB' && source[2] == '\xBF') {
397 result.encoding_data = lxb_encoding_data(LXB_ENCODING_UTF_8);
398 result.bom_shift = 3;
399 return result;
400 } else if (source_len >= 2) {
401 if (source[0] == '\xFE' && source[1] == '\xFF') {
402 result.encoding_data = lxb_encoding_data(LXB_ENCODING_UTF_16BE);
403 result.bom_shift = 2;
404 return result;
405 } else if (source[0] == '\xFF' && source[1] == '\xFE') {
406 result.encoding_data = lxb_encoding_data(LXB_ENCODING_UTF_16LE);
407 result.bom_shift = 2;
408 return result;
409 }
410 }
411
412 /* Perform prescan */
413 lxb_html_encoding_t encoding;
414 lxb_status_t status = lxb_html_encoding_init(&encoding);
415 if (status != LXB_STATUS_OK) {
416 goto fallback_uninit;
417 }
418 /* This is the "wait either for 1024 bytes or 500ms" part */
419 if (source_len > 1024) {
420 source_len = 1024;
421 }
422 status = lxb_html_encoding_determine(&encoding, (const lxb_char_t *) source, (const lxb_char_t *) source + source_len);
423 if (status != LXB_STATUS_OK) {
424 goto fallback;
425 }
426 lxb_html_encoding_entry_t *entry = lxb_html_encoding_meta_entry(&encoding, 0);
427 if (entry == NULL) {
428 goto fallback;
429 }
430 result.encoding_data = lxb_encoding_data_by_pre_name(entry->name, entry->end - entry->name);
431 if (!result.encoding_data) {
432 goto fallback;
433 }
434 result.bom_shift = 0;
435 lxb_html_encoding_destroy(&encoding, false);
436 return result;
437
438 fallback:
439 lxb_html_encoding_destroy(&encoding, false);
440 fallback_uninit:
441 result.encoding_data = lxb_encoding_data(DOM_FALLBACK_ENCODING_ID);
442 result.bom_shift = 0;
443 return result;
444 }
445
dom_setup_parser_encoding_manually(const lxb_char_t * buf_start,const lxb_encoding_data_t * encoding_data,dom_decoding_encoding_ctx * decoding_encoding_ctx,dom_lexbor_libxml2_bridge_application_data * application_data)446 static void dom_setup_parser_encoding_manually(const lxb_char_t *buf_start, const lxb_encoding_data_t *encoding_data, dom_decoding_encoding_ctx *decoding_encoding_ctx, dom_lexbor_libxml2_bridge_application_data *application_data)
447 {
448 static const lxb_codepoint_t replacement_codepoint = LXB_ENCODING_REPLACEMENT_CODEPOINT;
449
450 decoding_encoding_ctx->decode_data = encoding_data;
451
452 (void) lxb_encoding_decode_init(
453 &decoding_encoding_ctx->decode,
454 decoding_encoding_ctx->decode_data,
455 decoding_encoding_ctx->codepoints,
456 sizeof(decoding_encoding_ctx->codepoints) / sizeof(*decoding_encoding_ctx->codepoints)
457 );
458 (void) lxb_encoding_decode_replace_set(
459 &decoding_encoding_ctx->decode,
460 &replacement_codepoint,
461 LXB_ENCODING_REPLACEMENT_BUFFER_LEN
462 );
463 /* Note: encode_data is for UTF-8 */
464 decoding_encoding_ctx->fast_path = decoding_encoding_ctx->decode_data == decoding_encoding_ctx->encode_data;
465
466 if (decoding_encoding_ctx->fast_path) {
467 application_data->current_input_codepoints = NULL;
468 application_data->current_input_characters = (const char *) buf_start;
469 } else {
470 application_data->current_input_codepoints = decoding_encoding_ctx->codepoints;
471 application_data->current_input_characters = NULL;
472 }
473 }
474
dom_setup_parser_encoding_implicitly(const lxb_char_t ** buf_ref,size_t * read,dom_decoding_encoding_ctx * decoding_encoding_ctx,dom_lexbor_libxml2_bridge_application_data * application_data)475 static void dom_setup_parser_encoding_implicitly(
476 const lxb_char_t **buf_ref,
477 size_t *read,
478 dom_decoding_encoding_ctx *decoding_encoding_ctx,
479 dom_lexbor_libxml2_bridge_application_data *application_data
480 )
481 {
482 const char *buf_start = (const char *) *buf_ref;
483 dom_character_encoding_data dom_encoding_data = dom_determine_encoding(buf_start, *read);
484 *buf_ref += dom_encoding_data.bom_shift;
485 *read -= dom_encoding_data.bom_shift;
486 dom_setup_parser_encoding_manually((const lxb_char_t *) buf_start, dom_encoding_data.encoding_data, decoding_encoding_ctx, application_data);
487 }
488
dom_process_parse_chunk(lexbor_libxml2_bridge_parse_context * ctx,lxb_html_document_t * document,lxb_html_parser_t * parser,size_t encoded_length,const lxb_char_t * encoding_output,size_t input_buffer_length,size_t * tokenizer_error_offset,size_t * tree_error_offset)489 static bool dom_process_parse_chunk(
490 lexbor_libxml2_bridge_parse_context *ctx,
491 lxb_html_document_t *document,
492 lxb_html_parser_t *parser,
493 size_t encoded_length,
494 const lxb_char_t *encoding_output,
495 size_t input_buffer_length,
496 size_t *tokenizer_error_offset,
497 size_t *tree_error_offset
498 )
499 {
500 dom_lexbor_libxml2_bridge_application_data *application_data = ctx->application_data;
501 application_data->current_input_length = input_buffer_length;
502 lexbor_status_t lexbor_status = lxb_html_document_parse_chunk(document, encoding_output, encoded_length);
503 if (UNEXPECTED(lexbor_status != LXB_STATUS_OK)) {
504 return false;
505 }
506 if (ctx->tokenizer_error_reporter || ctx->tree_error_reporter) {
507 lexbor_libxml2_bridge_report_errors(ctx, parser, encoding_output, application_data->current_total_offset, tokenizer_error_offset, tree_error_offset);
508 dom_find_line_and_column_using_cache(application_data, &application_data->cache_tokenizer, application_data->current_total_offset + input_buffer_length);
509 }
510 application_data->current_total_offset += input_buffer_length;
511 application_data->cache_tokenizer.last_offset = 0;
512 return true;
513 }
514
dom_decode_encode_fast_path(lexbor_libxml2_bridge_parse_context * ctx,lxb_html_document_t * document,lxb_html_parser_t * parser,const lxb_char_t ** buf_ref_ref,const lxb_char_t * buf_end,dom_decoding_encoding_ctx * decoding_encoding_ctx,size_t * tokenizer_error_offset,size_t * tree_error_offset)515 static bool dom_decode_encode_fast_path(
516 lexbor_libxml2_bridge_parse_context *ctx,
517 lxb_html_document_t *document,
518 lxb_html_parser_t *parser,
519 const lxb_char_t **buf_ref_ref,
520 const lxb_char_t *buf_end,
521 dom_decoding_encoding_ctx *decoding_encoding_ctx,
522 size_t *tokenizer_error_offset,
523 size_t *tree_error_offset
524 )
525 {
526 const lxb_char_t *buf_ref = *buf_ref_ref;
527 const lxb_char_t *last_output = buf_ref;
528 while (buf_ref != buf_end) {
529 /* Fast path converts non-validated UTF-8 -> validated UTF-8 */
530 if (decoding_encoding_ctx->decode.u.utf_8.need == 0 && *buf_ref < 0x80) {
531 /* Fast path within the fast path: try to skip non-mb bytes in bulk if we are not in a state where we
532 * need more UTF-8 bytes to complete a sequence.
533 * It might be tempting to use SIMD here, but it turns out that this is less efficient because
534 * we need to process the same byte multiple times sometimes when mixing ASCII with multibyte. */
535 buf_ref++;
536 continue;
537 }
538 const lxb_char_t *buf_ref_backup = buf_ref;
539 lxb_codepoint_t codepoint = lxb_encoding_decode_utf_8_single(&decoding_encoding_ctx->decode, &buf_ref, buf_end);
540 if (UNEXPECTED(codepoint > LXB_ENCODING_MAX_CODEPOINT)) {
541 size_t skip = buf_ref - buf_ref_backup; /* Skip invalid data, it's replaced by the UTF-8 replacement bytes */
542 if (!dom_process_parse_chunk(
543 ctx,
544 document,
545 parser,
546 buf_ref - last_output - skip,
547 last_output,
548 buf_ref - last_output,
549 tokenizer_error_offset,
550 tree_error_offset
551 )) {
552 goto fail_oom;
553 }
554 if (!dom_process_parse_chunk(
555 ctx,
556 document,
557 parser,
558 LXB_ENCODING_REPLACEMENT_SIZE,
559 LXB_ENCODING_REPLACEMENT_BYTES,
560 0,
561 tokenizer_error_offset,
562 tree_error_offset
563 )) {
564 goto fail_oom;
565 }
566 last_output = buf_ref;
567 }
568 }
569 if (buf_ref != last_output
570 && !dom_process_parse_chunk(
571 ctx,
572 document,
573 parser,
574 buf_ref - last_output,
575 last_output,
576 buf_ref - last_output,
577 tokenizer_error_offset,
578 tree_error_offset
579 )) {
580 goto fail_oom;
581 }
582 *buf_ref_ref = buf_ref;
583 return true;
584 fail_oom:
585 *buf_ref_ref = buf_ref;
586 return false;
587 }
588
dom_decode_encode_slow_path(lexbor_libxml2_bridge_parse_context * ctx,lxb_html_document_t * document,lxb_html_parser_t * parser,const lxb_char_t ** buf_ref_ref,const lxb_char_t * buf_end,dom_decoding_encoding_ctx * decoding_encoding_ctx,size_t * tokenizer_error_offset,size_t * tree_error_offset)589 static bool dom_decode_encode_slow_path(
590 lexbor_libxml2_bridge_parse_context *ctx,
591 lxb_html_document_t *document,
592 lxb_html_parser_t *parser,
593 const lxb_char_t **buf_ref_ref,
594 const lxb_char_t *buf_end,
595 dom_decoding_encoding_ctx *decoding_encoding_ctx,
596 size_t *tokenizer_error_offset,
597 size_t *tree_error_offset
598 )
599 {
600 const lxb_char_t *buf_ref = *buf_ref_ref;
601 lexbor_status_t decode_status, encode_status;
602 do {
603 decode_status = decoding_encoding_ctx->decode_data->decode(&decoding_encoding_ctx->decode, &buf_ref, buf_end);
604
605 const lxb_codepoint_t *codepoints_ref = (const lxb_codepoint_t *) decoding_encoding_ctx->codepoints;
606 size_t decoding_buffer_used = lxb_encoding_decode_buf_used(&decoding_encoding_ctx->decode);
607 const lxb_codepoint_t *codepoints_end = decoding_encoding_ctx->codepoints + decoding_buffer_used;
608 do {
609 encode_status = decoding_encoding_ctx->encode_data->encode(&decoding_encoding_ctx->encode, &codepoints_ref, codepoints_end);
610 ZEND_ASSERT(encode_status != LXB_STATUS_ERROR && "parameters and replacements should be valid");
611 if (!dom_process_parse_chunk(
612 ctx,
613 document,
614 parser,
615 lxb_encoding_encode_buf_used(&decoding_encoding_ctx->encode),
616 decoding_encoding_ctx->encoding_output,
617 decoding_buffer_used,
618 tokenizer_error_offset,
619 tree_error_offset
620 )) {
621 goto fail_oom;
622 }
623 lxb_encoding_encode_buf_used_set(&decoding_encoding_ctx->encode, 0);
624 } while (encode_status == LXB_STATUS_SMALL_BUFFER);
625 lxb_encoding_decode_buf_used_set(&decoding_encoding_ctx->decode, 0);
626 } while (decode_status == LXB_STATUS_SMALL_BUFFER);
627 *buf_ref_ref = buf_ref;
628 return true;
629 fail_oom:
630 *buf_ref_ref = buf_ref;
631 return false;
632 }
633
dom_parse_decode_encode_step(lexbor_libxml2_bridge_parse_context * ctx,lxb_html_document_t * document,lxb_html_parser_t * parser,const lxb_char_t ** buf_ref_ref,const lxb_char_t * buf_end,dom_decoding_encoding_ctx * decoding_encoding_ctx,size_t * tokenizer_error_offset,size_t * tree_error_offset)634 static bool dom_parse_decode_encode_step(
635 lexbor_libxml2_bridge_parse_context *ctx,
636 lxb_html_document_t *document,
637 lxb_html_parser_t *parser,
638 const lxb_char_t **buf_ref_ref,
639 const lxb_char_t *buf_end,
640 dom_decoding_encoding_ctx *decoding_encoding_ctx,
641 size_t *tokenizer_error_offset,
642 size_t *tree_error_offset
643 )
644 {
645 if (decoding_encoding_ctx->fast_path) {
646 return dom_decode_encode_fast_path(
647 ctx,
648 document,
649 parser,
650 buf_ref_ref,
651 buf_end,
652 decoding_encoding_ctx,
653 tokenizer_error_offset,
654 tree_error_offset
655 );
656 } else {
657 return dom_decode_encode_slow_path(
658 ctx,
659 document,
660 parser,
661 buf_ref_ref,
662 buf_end,
663 decoding_encoding_ctx,
664 tokenizer_error_offset,
665 tree_error_offset
666 );
667 }
668 }
669
dom_parse_decode_encode_finish(lexbor_libxml2_bridge_parse_context * ctx,lxb_html_document_t * document,lxb_html_parser_t * parser,dom_decoding_encoding_ctx * decoding_encoding_ctx,size_t * tokenizer_error_offset,size_t * tree_error_offset)670 static bool dom_parse_decode_encode_finish(
671 lexbor_libxml2_bridge_parse_context *ctx,
672 lxb_html_document_t *document,
673 lxb_html_parser_t *parser,
674 dom_decoding_encoding_ctx *decoding_encoding_ctx,
675 size_t *tokenizer_error_offset,
676 size_t *tree_error_offset
677 )
678 {
679 if (!decoding_encoding_ctx->fast_path) {
680 /* Fast path handles codepoints one by one, so this part is not applicable in that case */
681 (void) lxb_encoding_decode_finish(&decoding_encoding_ctx->decode);
682 size_t decoding_buffer_size = lxb_encoding_decode_buf_used(&decoding_encoding_ctx->decode);
683 if (decoding_buffer_size > 0) {
684 const lxb_codepoint_t *codepoints_ref = (const lxb_codepoint_t *) decoding_encoding_ctx->codepoints;
685 const lxb_codepoint_t *codepoints_end = codepoints_ref + decoding_buffer_size;
686 (void) decoding_encoding_ctx->encode_data->encode(&decoding_encoding_ctx->encode, &codepoints_ref, codepoints_end);
687 if (!dom_process_parse_chunk(
688 ctx,
689 document,
690 parser,
691 lxb_encoding_encode_buf_used(&decoding_encoding_ctx->encode),
692 decoding_encoding_ctx->encoding_output,
693 decoding_buffer_size,
694 tokenizer_error_offset,
695 tree_error_offset
696 )) {
697 return false;
698 }
699 }
700 }
701 (void) lxb_encoding_encode_finish(&decoding_encoding_ctx->encode);
702 if (lxb_encoding_encode_buf_used(&decoding_encoding_ctx->encode)
703 && !dom_process_parse_chunk(
704 ctx,
705 document,
706 parser,
707 lxb_encoding_encode_buf_used(&decoding_encoding_ctx->encode),
708 decoding_encoding_ctx->encoding_output,
709 lxb_encoding_decode_buf_used(&decoding_encoding_ctx->decode),
710 tokenizer_error_offset,
711 tree_error_offset
712 )) {
713 return false;
714 }
715 return true;
716 }
717
check_options_validity(uint32_t arg_num,zend_long options)718 static bool check_options_validity(uint32_t arg_num, zend_long options)
719 {
720 const zend_long VALID_OPTIONS = XML_PARSE_NOERROR | XML_PARSE_COMPACT | HTML_PARSE_NOIMPLIED | DOM_HTML_NO_DEFAULT_NS;
721 if ((options & ~VALID_OPTIONS) != 0) {
722 zend_argument_value_error(arg_num, "contains invalid flags (allowed flags: "
723 "LIBXML_NOERROR, "
724 "LIBXML_COMPACT, "
725 "LIBXML_HTML_NOIMPLIED, "
726 "Dom\\NO_DEFAULT_NS)");
727 return false;
728 }
729 return true;
730 }
731
PHP_METHOD(Dom_HTMLDocument,createEmpty)732 PHP_METHOD(Dom_HTMLDocument, createEmpty)
733 {
734 const char *encoding = "UTF-8";
735 size_t encoding_len = strlen("UTF-8");
736 if (zend_parse_parameters(ZEND_NUM_ARGS(), "|p", &encoding, &encoding_len) == FAILURE) {
737 RETURN_THROWS();
738 }
739
740 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name((const lxb_char_t *) encoding, encoding_len);
741
742 if (encoding_data == NULL) {
743 zend_argument_value_error(1, "must be a valid document encoding");
744 RETURN_THROWS();
745 }
746
747 xmlDocPtr lxml_doc = php_dom_create_html_doc();
748 if (UNEXPECTED(lxml_doc == NULL)) {
749 goto oom;
750 }
751
752 lxml_doc->encoding = xmlStrdup((const xmlChar *) encoding);
753
754 dom_object *intern = php_dom_instantiate_object_helper(
755 return_value,
756 dom_html_document_class_entry,
757 (xmlNodePtr) lxml_doc,
758 NULL
759 );
760 dom_set_xml_class(intern->document);
761 intern->document->private_data = php_dom_libxml_private_data_header(php_dom_private_data_create());
762 return;
763
764 oom:
765 php_dom_throw_error(INVALID_STATE_ERR, true);
766 RETURN_THROWS();
767 }
768
769 /* Only bother to register error handling when the error reports can become observable. */
dom_should_register_error_handlers(zend_long options)770 static bool dom_should_register_error_handlers(zend_long options)
771 {
772 if (options & XML_PARSE_NOERROR) {
773 return false;
774 }
775
776 return php_libxml_uses_internal_errors() || ((EG(error_reporting) | EG(user_error_handler_error_reporting)) & E_WARNING);
777 }
778
PHP_METHOD(Dom_HTMLDocument,createFromString)779 PHP_METHOD(Dom_HTMLDocument, createFromString)
780 {
781 const char *source, *override_encoding = NULL;
782 size_t source_len, override_encoding_len;
783 zend_long options = 0;
784 if (zend_parse_parameters(
785 ZEND_NUM_ARGS(),
786 "s|lp!",
787 &source,
788 &source_len,
789 &options,
790 &override_encoding,
791 &override_encoding_len
792 ) == FAILURE) {
793 RETURN_THROWS();
794 }
795
796 if (!check_options_validity(2, options)) {
797 RETURN_THROWS();
798 }
799
800 dom_lexbor_libxml2_bridge_application_data application_data;
801 application_data.input_name = "Entity";
802 application_data.current_total_offset = 0;
803 application_data.html_no_implied = options & HTML_PARSE_NOIMPLIED;
804 dom_reset_line_column_cache(&application_data.cache_tokenizer);
805 lexbor_libxml2_bridge_parse_context ctx;
806 lexbor_libxml2_bridge_parse_context_init(&ctx);
807 if (dom_should_register_error_handlers(options)) {
808 lexbor_libxml2_bridge_parse_set_error_callbacks(
809 &ctx,
810 dom_lexbor_libxml2_bridge_tokenizer_error_reporter,
811 dom_lexbor_libxml2_bridge_tree_error_reporter
812 );
813 }
814 ctx.application_data = &application_data;
815
816 size_t tokenizer_error_offset = 0;
817 size_t tree_error_offset = 0;
818
819 /* Setup everything encoding & decoding related */
820 const lxb_char_t *buf_ref = (const lxb_char_t *) source;
821 dom_decoding_encoding_ctx decoding_encoding_ctx;
822 dom_decoding_encoding_ctx_init(&decoding_encoding_ctx);
823 if (override_encoding != NULL) {
824 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name(
825 (const lxb_char_t *) override_encoding,
826 override_encoding_len
827 );
828 if (!encoding_data) {
829 zend_argument_value_error(3, "must be a valid document encoding");
830 RETURN_THROWS();
831 }
832 dom_setup_parser_encoding_manually(buf_ref, encoding_data, &decoding_encoding_ctx, &application_data);
833 } else {
834 dom_setup_parser_encoding_implicitly(&buf_ref, &source_len, &decoding_encoding_ctx, &application_data);
835 }
836
837 lxb_html_document_t *document = lxb_html_document_create();
838 if (UNEXPECTED(document == NULL)) {
839 goto fail_oom;
840 }
841
842 lxb_status_t lexbor_status = lxb_html_document_parse_chunk_begin(document);
843 if (UNEXPECTED(lexbor_status != LXB_STATUS_OK)) {
844 goto fail_oom;
845 }
846
847 lxb_html_parser_t *parser = document->dom_document.parser;
848
849 while (source_len > 0) {
850 size_t chunk_size = source_len;
851 const size_t MAX_CHUNK_SIZE = sizeof(decoding_encoding_ctx.encoding_output) / sizeof(*decoding_encoding_ctx.encoding_output);
852 if (chunk_size > MAX_CHUNK_SIZE) {
853 chunk_size = MAX_CHUNK_SIZE;
854 }
855 source_len -= chunk_size;
856
857 const lxb_char_t *buf_end = buf_ref + chunk_size;
858 bool result = dom_parse_decode_encode_step(
859 &ctx,
860 document,
861 parser,
862 &buf_ref,
863 buf_end,
864 &decoding_encoding_ctx,
865 &tokenizer_error_offset,
866 &tree_error_offset
867 );
868 if (!result) {
869 goto fail_oom;
870 }
871 }
872
873 if (!dom_parse_decode_encode_finish(&ctx, document, parser, &decoding_encoding_ctx, &tokenizer_error_offset, &tree_error_offset)) {
874 goto fail_oom;
875 }
876
877 lexbor_status = lxb_html_document_parse_chunk_end(document);
878 if (lexbor_status != LXB_STATUS_OK) {
879 goto fail_oom;
880 }
881
882 php_dom_private_data *private_data = php_dom_private_data_create();
883
884 xmlDocPtr lxml_doc;
885 lexbor_libxml2_bridge_status bridge_status = lexbor_libxml2_bridge_convert_document(
886 document,
887 &lxml_doc,
888 options & XML_PARSE_COMPACT,
889 !(options & DOM_HTML_NO_DEFAULT_NS),
890 private_data
891 );
892 lexbor_libxml2_bridge_copy_observations(parser->tree, &ctx.observations);
893 if (UNEXPECTED(bridge_status != LEXBOR_LIBXML2_BRIDGE_STATUS_OK)) {
894 php_dom_private_data_destroy(private_data);
895 php_libxml_ctx_error(
896 NULL,
897 "%s in %s",
898 dom_lexbor_libxml2_bridge_status_code_to_string(bridge_status),
899 application_data.input_name
900 );
901 lxb_html_document_destroy(document);
902 RETURN_FALSE;
903 }
904 lxb_html_document_destroy(document);
905
906 dom_post_process_html5_loading(lxml_doc, options, &ctx.observations);
907
908 if (decoding_encoding_ctx.decode_data) {
909 lxml_doc->encoding = xmlStrdup((const xmlChar *) decoding_encoding_ctx.decode_data->name);
910 } else {
911 lxml_doc->encoding = xmlStrdup((const xmlChar *) "UTF-8");
912 }
913
914 dom_object *intern = php_dom_instantiate_object_helper(
915 return_value,
916 dom_html_document_class_entry,
917 (xmlNodePtr) lxml_doc,
918 NULL
919 );
920 dom_set_xml_class(intern->document);
921 intern->document->quirks_mode = ctx.observations.quirks_mode;
922 intern->document->private_data = php_dom_libxml_private_data_header(private_data);
923 return;
924
925 fail_oom:
926 lxb_html_document_destroy(document);
927 php_dom_throw_error(INVALID_STATE_ERR, true);
928 RETURN_THROWS();
929 }
930
PHP_METHOD(Dom_HTMLDocument,createFromFile)931 PHP_METHOD(Dom_HTMLDocument, createFromFile)
932 {
933 const char *filename, *override_encoding = NULL;
934 php_dom_private_data *private_data = NULL;
935 size_t filename_len, override_encoding_len;
936 zend_long options = 0;
937 php_stream *stream = NULL;
938 if (zend_parse_parameters(
939 ZEND_NUM_ARGS(),
940 "p|lp!",
941 &filename,
942 &filename_len,
943 &options,
944 &override_encoding,
945 &override_encoding_len
946 ) == FAILURE) {
947 RETURN_THROWS();
948 }
949
950 /* See php_libxml_streams_IO_open_wrapper(), apparently this caused issues in the past. */
951 if (strstr(filename, "%00")) {
952 zend_argument_value_error(1, "must not contain percent-encoded NUL bytes");
953 RETURN_THROWS();
954 }
955
956 if (!check_options_validity(2, options)) {
957 RETURN_THROWS();
958 }
959
960 dom_lexbor_libxml2_bridge_application_data application_data;
961 application_data.input_name = filename;
962 application_data.current_total_offset = 0;
963 application_data.html_no_implied = options & HTML_PARSE_NOIMPLIED;
964 dom_reset_line_column_cache(&application_data.cache_tokenizer);
965 lexbor_libxml2_bridge_parse_context ctx;
966 lexbor_libxml2_bridge_parse_context_init(&ctx);
967 if (dom_should_register_error_handlers(options)) {
968 lexbor_libxml2_bridge_parse_set_error_callbacks(
969 &ctx,
970 dom_lexbor_libxml2_bridge_tokenizer_error_reporter,
971 dom_lexbor_libxml2_bridge_tree_error_reporter
972 );
973 }
974 ctx.application_data = &application_data;
975
976 char buf[4096];
977
978 /* Setup everything encoding & decoding related */
979 dom_decoding_encoding_ctx decoding_encoding_ctx;
980 dom_decoding_encoding_ctx_init(&decoding_encoding_ctx);
981 bool should_determine_encoding_implicitly = true; /* First read => determine encoding implicitly */
982 if (override_encoding != NULL) {
983 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name(
984 (const lxb_char_t *) override_encoding,
985 override_encoding_len
986 );
987 if (!encoding_data) {
988 zend_argument_value_error(3, "must be a valid document encoding");
989 RETURN_THROWS();
990 }
991 should_determine_encoding_implicitly = false;
992 dom_setup_parser_encoding_manually((const lxb_char_t *) buf, encoding_data, &decoding_encoding_ctx, &application_data);
993 }
994
995 zend_string *opened_path = NULL;
996 stream = php_stream_open_wrapper_ex(filename, "rb", REPORT_ERRORS, &opened_path, php_libxml_get_stream_context());
997 if (!stream) {
998 if (!EG(exception)) {
999 zend_throw_exception_ex(NULL, 0, "Cannot open file '%s'", filename);
1000 }
1001 RETURN_THROWS();
1002 }
1003
1004 /* MIME sniff */
1005 if (should_determine_encoding_implicitly) {
1006 zend_string *charset = php_libxml_sniff_charset_from_stream(stream);
1007 if (charset != NULL) {
1008 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name(
1009 (const lxb_char_t *) ZSTR_VAL(charset),
1010 ZSTR_LEN(charset)
1011 );
1012 if (encoding_data != NULL) {
1013 should_determine_encoding_implicitly = false;
1014 dom_setup_parser_encoding_manually(
1015 (const lxb_char_t *) buf,
1016 encoding_data,
1017 &decoding_encoding_ctx,
1018 &application_data
1019 );
1020 }
1021 zend_string_release_ex(charset, false);
1022 }
1023 }
1024
1025 lxb_html_document_t *document = lxb_html_document_create();
1026 if (UNEXPECTED(document == NULL)) {
1027 goto fail_oom;
1028 }
1029
1030 lxb_status_t lexbor_status = lxb_html_document_parse_chunk_begin(document);
1031 if (UNEXPECTED(lexbor_status != LXB_STATUS_OK)) {
1032 goto fail_oom;
1033 }
1034
1035 size_t tokenizer_error_offset = 0;
1036 size_t tree_error_offset = 0;
1037 ssize_t read;
1038 lxb_html_parser_t *parser = document->dom_document.parser;
1039
1040 while ((read = php_stream_read(stream, buf, sizeof(buf))) > 0) {
1041 const lxb_char_t *buf_ref = (const lxb_char_t *) buf;
1042
1043 if (should_determine_encoding_implicitly) {
1044 should_determine_encoding_implicitly = false;
1045 dom_setup_parser_encoding_implicitly(&buf_ref, (size_t *) &read, &decoding_encoding_ctx, &application_data);
1046 }
1047
1048 const lxb_char_t *buf_end = buf_ref + read;
1049 bool result = dom_parse_decode_encode_step(
1050 &ctx,
1051 document,
1052 parser,
1053 &buf_ref,
1054 buf_end,
1055 &decoding_encoding_ctx,
1056 &tokenizer_error_offset,
1057 &tree_error_offset
1058 );
1059 if (!result) {
1060 goto fail_oom;
1061 }
1062 }
1063
1064 if (!dom_parse_decode_encode_finish(&ctx, document, parser, &decoding_encoding_ctx, &tokenizer_error_offset, &tree_error_offset)) {
1065 goto fail_oom;
1066 }
1067
1068 lexbor_status = lxb_html_document_parse_chunk_end(document);
1069 if (lexbor_status != LXB_STATUS_OK) {
1070 goto fail_oom;
1071 }
1072
1073 private_data = php_dom_private_data_create();
1074
1075 xmlDocPtr lxml_doc;
1076 lexbor_libxml2_bridge_status bridge_status = lexbor_libxml2_bridge_convert_document(
1077 document,
1078 &lxml_doc,
1079 options & XML_PARSE_COMPACT,
1080 !(options & DOM_HTML_NO_DEFAULT_NS),
1081 private_data
1082 );
1083 lexbor_libxml2_bridge_copy_observations(parser->tree, &ctx.observations);
1084 if (UNEXPECTED(bridge_status != LEXBOR_LIBXML2_BRIDGE_STATUS_OK)) {
1085 php_libxml_ctx_error(NULL, "%s in %s", dom_lexbor_libxml2_bridge_status_code_to_string(bridge_status), filename);
1086 RETVAL_FALSE;
1087 goto fail_general;
1088 }
1089 lxb_html_document_destroy(document);
1090
1091 dom_post_process_html5_loading(lxml_doc, options, &ctx.observations);
1092
1093 if (decoding_encoding_ctx.decode_data) {
1094 lxml_doc->encoding = xmlStrdup((const xmlChar *) decoding_encoding_ctx.decode_data->name);
1095 } else {
1096 lxml_doc->encoding = xmlStrdup((const xmlChar *) "UTF-8");
1097 }
1098
1099 if (stream->wrapper == &php_plain_files_wrapper && opened_path != NULL) {
1100 xmlChar *converted = xmlPathToURI((const xmlChar *) ZSTR_VAL(opened_path));
1101 if (UNEXPECTED(!converted)) {
1102 goto fail_oom;
1103 }
1104 /* Check for "file:/" instead of "file://" because of libxml2 quirk */
1105 if (strncmp((const char *) converted, "file:/", sizeof("file:/") - 1) != 0) {
1106 xmlChar *buffer = xmlStrdup((const xmlChar *) "file://");
1107 if (UNEXPECTED(!buffer)) {
1108 xmlFree(converted);
1109 goto fail_oom;
1110 }
1111 xmlChar *new_buffer = xmlStrcat(buffer, converted);
1112 if (UNEXPECTED(!new_buffer)) {
1113 xmlFree(buffer);
1114 xmlFree(converted);
1115 goto fail_oom;
1116 }
1117 xmlFree(converted);
1118 lxml_doc->URL = new_buffer;
1119 } else {
1120 #ifdef PHP_WIN32
1121 converted = php_dom_libxml_fix_file_path(converted);
1122 #endif
1123 lxml_doc->URL = converted;
1124 }
1125 } else {
1126 lxml_doc->URL = xmlStrdup((const xmlChar *) filename);
1127 }
1128
1129 if (opened_path != NULL) {
1130 zend_string_release_ex(opened_path, false);
1131 }
1132 php_stream_close(stream);
1133 stream = NULL;
1134
1135 dom_object *intern = php_dom_instantiate_object_helper(
1136 return_value,
1137 dom_html_document_class_entry,
1138 (xmlNodePtr) lxml_doc,
1139 NULL
1140 );
1141 dom_set_xml_class(intern->document);
1142 intern->document->quirks_mode = ctx.observations.quirks_mode;
1143 intern->document->private_data = php_dom_libxml_private_data_header(private_data);
1144 return;
1145
1146 fail_oom:
1147 php_dom_throw_error(INVALID_STATE_ERR, true);
1148 fail_general:
1149 if (private_data != NULL) {
1150 php_dom_private_data_destroy(private_data);
1151 }
1152 lxb_html_document_destroy(document);
1153 php_stream_close(stream);
1154 if (opened_path != NULL) {
1155 zend_string_release_ex(opened_path, false);
1156 }
1157 }
1158
dom_write_output_smart_str(void * ctx,const char * buf,size_t size)1159 static zend_result dom_write_output_smart_str(void *ctx, const char *buf, size_t size)
1160 {
1161 smart_str_appendl((smart_str *) ctx, buf, size);
1162 return SUCCESS;
1163 }
1164
dom_write_output_stream(void * application_data,const char * buf,size_t len)1165 static zend_result dom_write_output_stream(void *application_data, const char *buf, size_t len)
1166 {
1167 php_stream *stream = (php_stream *) application_data;
1168 if (UNEXPECTED(php_stream_write(stream, buf, len) < 0)) {
1169 return FAILURE;
1170 }
1171 return SUCCESS;
1172 }
1173
dom_saveHTML_write_string_len(void * application_data,const char * buf,size_t len)1174 static zend_result dom_saveHTML_write_string_len(void *application_data, const char *buf, size_t len)
1175 {
1176 dom_output_ctx *output = (dom_output_ctx *) application_data;
1177 lxb_status_t decode_status, encode_status;
1178 const lxb_char_t *buf_ref = (const lxb_char_t *) buf;
1179 const lxb_char_t *buf_end = buf_ref + len;
1180
1181 do {
1182 decode_status = output->decoding_data->decode(output->decode, &buf_ref, buf_end);
1183
1184 const lxb_codepoint_t *codepoints_ref = output->codepoints;
1185 const lxb_codepoint_t *codepoints_end = codepoints_ref + lxb_encoding_decode_buf_used(output->decode);
1186 do {
1187 encode_status = output->encoding_data->encode(output->encode, &codepoints_ref, codepoints_end);
1188 if (UNEXPECTED(output->write_output(
1189 output->output_data,
1190 (const char *) output->encoding_output,
1191 lxb_encoding_encode_buf_used(output->encode)
1192 ) != SUCCESS)) {
1193 return FAILURE;
1194 }
1195 lxb_encoding_encode_buf_used_set(output->encode, 0);
1196 } while (encode_status == LXB_STATUS_SMALL_BUFFER);
1197 lxb_encoding_decode_buf_used_set(output->decode, 0);
1198 } while (decode_status == LXB_STATUS_SMALL_BUFFER);
1199
1200 return SUCCESS;
1201 }
1202
dom_saveHTML_write_string(void * application_data,const char * buf)1203 static zend_result dom_saveHTML_write_string(void *application_data, const char *buf)
1204 {
1205 return dom_saveHTML_write_string_len(application_data, buf, strlen(buf));
1206 }
1207
dom_common_save(dom_output_ctx * output_ctx,dom_object * intern,const xmlDoc * docp,const xmlNode * node)1208 static zend_result dom_common_save(dom_output_ctx *output_ctx, dom_object *intern, const xmlDoc *docp, const xmlNode *node)
1209 {
1210 /* Initialize everything related to encoding & decoding */
1211 const lxb_encoding_data_t *decoding_data = lxb_encoding_data(LXB_ENCODING_UTF_8);
1212 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name(
1213 (const lxb_char_t *) docp->encoding,
1214 strlen((const char *) docp->encoding)
1215 );
1216 lxb_encoding_encode_t encode;
1217 lxb_encoding_decode_t decode;
1218 lxb_char_t encoding_output[4096];
1219 lxb_codepoint_t codepoints[4096];
1220 (void) lxb_encoding_encode_init(&encode, encoding_data, encoding_output, sizeof(encoding_output) / sizeof(*encoding_output));
1221 (void) lxb_encoding_decode_init(&decode, decoding_data, codepoints, sizeof(codepoints) / sizeof(*codepoints));
1222 if (encoding_data->encoding == LXB_ENCODING_UTF_8) {
1223 lxb_encoding_encode_replace_set(&encode, LXB_ENCODING_REPLACEMENT_BYTES, LXB_ENCODING_REPLACEMENT_SIZE);
1224 } else {
1225 /* Fallback if there is no replacement by default */
1226 lxb_encoding_encode_replace_set(&encode, (const lxb_char_t *) "?", 1);
1227 }
1228 lxb_encoding_decode_replace_set(&decode, LXB_ENCODING_REPLACEMENT_BUFFER, LXB_ENCODING_REPLACEMENT_BUFFER_LEN);
1229
1230 output_ctx->encoding_data = encoding_data;
1231 output_ctx->decoding_data = decoding_data;
1232 output_ctx->encode = &encode;
1233 output_ctx->decode = &decode;
1234 output_ctx->codepoints = codepoints;
1235 output_ctx->encoding_output = encoding_output;
1236
1237 dom_html5_serialize_context ctx;
1238 ctx.write_string_len = dom_saveHTML_write_string_len;
1239 ctx.write_string = dom_saveHTML_write_string;
1240 ctx.application_data = output_ctx;
1241 ctx.private_data = php_dom_get_private_data(intern);
1242 if (UNEXPECTED(dom_html5_serialize_outer(&ctx, node) != SUCCESS)) {
1243 return FAILURE;
1244 }
1245
1246 (void) lxb_encoding_decode_finish(&decode);
1247 if (lxb_encoding_decode_buf_used(&decode)) {
1248 const lxb_codepoint_t *codepoints_ref = (const lxb_codepoint_t *) codepoints;
1249 (void) encoding_data->encode(&encode, &codepoints_ref, codepoints_ref + lxb_encoding_decode_buf_used(&decode));
1250 if (UNEXPECTED(output_ctx->write_output(
1251 output_ctx->output_data,
1252 (const char *) encoding_output,
1253 lxb_encoding_encode_buf_used(&encode)) != SUCCESS
1254 )) {
1255 return FAILURE;
1256 }
1257 }
1258 (void) lxb_encoding_encode_finish(&encode);
1259 if (lxb_encoding_encode_buf_used(&encode)) {
1260 if (UNEXPECTED(output_ctx->write_output(
1261 output_ctx->output_data,
1262 (const char *) encoding_output,
1263 lxb_encoding_encode_buf_used(&encode)) != SUCCESS
1264 )) {
1265 return FAILURE;
1266 }
1267 }
1268
1269 return SUCCESS;
1270 }
1271
PHP_METHOD(Dom_HTMLDocument,saveHtmlFile)1272 PHP_METHOD(Dom_HTMLDocument, saveHtmlFile)
1273 {
1274 zval *id;
1275 xmlDoc *docp;
1276 size_t file_len;
1277 dom_object *intern;
1278 char *file;
1279
1280 id = ZEND_THIS;
1281 if (zend_parse_parameters(ZEND_NUM_ARGS(), "p", &file, &file_len) == FAILURE) {
1282 RETURN_THROWS();
1283 }
1284
1285 if (file_len == 0) {
1286 zend_argument_must_not_be_empty_error(1);
1287 RETURN_THROWS();
1288 }
1289
1290 php_stream *stream = php_stream_open_wrapper_ex(file, "wb", REPORT_ERRORS, /* opened_path */ NULL, php_libxml_get_stream_context());
1291 if (!stream) {
1292 RETURN_FALSE;
1293 }
1294
1295 DOM_GET_OBJ(docp, id, xmlDocPtr, intern);
1296
1297 dom_output_ctx output_ctx;
1298 output_ctx.output_data = stream;
1299 output_ctx.write_output = dom_write_output_stream;
1300 if (UNEXPECTED(dom_common_save(&output_ctx, intern, docp, (const xmlNode *) docp) != SUCCESS)) {
1301 php_stream_close(stream);
1302 RETURN_FALSE;
1303 }
1304
1305 zend_long bytes = php_stream_tell(stream);
1306 php_stream_close(stream);
1307
1308 RETURN_LONG(bytes);
1309 }
1310
PHP_METHOD(Dom_HTMLDocument,saveHtml)1311 PHP_METHOD(Dom_HTMLDocument, saveHtml)
1312 {
1313 zval *nodep = NULL;
1314 const xmlDoc *docp;
1315 const xmlNode *node;
1316 dom_object *intern, *nodeobj;
1317
1318 if (zend_parse_parameters(ZEND_NUM_ARGS(), "|O!", &nodep, dom_modern_node_class_entry) == FAILURE) {
1319 RETURN_THROWS();
1320 }
1321
1322 DOM_GET_OBJ(docp, ZEND_THIS, xmlDocPtr, intern);
1323
1324 if (nodep != NULL) {
1325 DOM_GET_OBJ(node, nodep, xmlNodePtr, nodeobj);
1326 if (node->doc != docp) {
1327 php_dom_throw_error(WRONG_DOCUMENT_ERR, true);
1328 RETURN_THROWS();
1329 }
1330 } else {
1331 node = (const xmlNode *) docp;
1332 }
1333
1334 smart_str buf = {0};
1335 dom_output_ctx output_ctx;
1336 output_ctx.output_data = &buf;
1337 output_ctx.write_output = dom_write_output_smart_str;
1338 /* Can't fail because dom_write_output_smart_str() can't fail. */
1339 zend_result result = dom_common_save(&output_ctx, intern, docp, node);
1340 ZEND_ASSERT(result == SUCCESS);
1341
1342 RETURN_STR(smart_str_extract(&buf));
1343 }
1344
dom_html_document_encoding_write(dom_object * obj,zval * newval)1345 zend_result dom_html_document_encoding_write(dom_object *obj, zval *newval)
1346 {
1347 DOM_PROP_NODE(xmlDocPtr, docp, obj);
1348
1349 /* Typed property, can only be IS_STRING. */
1350 ZEND_ASSERT(Z_TYPE_P(newval) == IS_STRING);
1351
1352 zend_string *str = Z_STR_P(newval);
1353 const lxb_encoding_data_t *encoding_data = lxb_encoding_data_by_name((const lxb_char_t *) ZSTR_VAL(str), ZSTR_LEN(str));
1354
1355 if (encoding_data != NULL) {
1356 xmlFree(BAD_CAST docp->encoding);
1357 docp->encoding = xmlStrdup((const xmlChar *) encoding_data->name);
1358 } else {
1359 zend_value_error("Invalid document encoding");
1360 return FAILURE;
1361 }
1362
1363 return SUCCESS;
1364 }
1365
dom_html_document_element_read_raw(const xmlDoc * docp,bool (* accept)(const xmlChar *))1366 static xmlNodePtr dom_html_document_element_read_raw(const xmlDoc *docp, bool (*accept)(const xmlChar *))
1367 {
1368 const xmlNode *root = xmlDocGetRootElement(docp);
1369 if (root == NULL || !(php_dom_ns_is_fast(root, php_dom_ns_is_html_magic_token) && xmlStrEqual(root->name, BAD_CAST "html"))) {
1370 return NULL;
1371 }
1372
1373 xmlNodePtr cur = root->children;
1374 while (cur != NULL) {
1375 if (cur->type == XML_ELEMENT_NODE && php_dom_ns_is_fast(cur, php_dom_ns_is_html_magic_token) && accept(cur->name)) {
1376 return cur;
1377 }
1378 cur = cur->next;
1379 }
1380
1381 return NULL;
1382 }
1383
dom_html_document_element_read_helper(dom_object * obj,zval * retval,bool (* accept)(const xmlChar *))1384 zend_result dom_html_document_element_read_helper(dom_object *obj, zval *retval, bool (*accept)(const xmlChar *))
1385 {
1386 DOM_PROP_NODE(const xmlDoc *, docp, obj);
1387
1388 const xmlNode *element = dom_html_document_element_read_raw(docp, accept);
1389 php_dom_create_nullable_object((xmlNodePtr) element, retval, obj);
1390
1391 return SUCCESS;
1392 }
1393
dom_accept_body_name(const xmlChar * name)1394 static bool dom_accept_body_name(const xmlChar *name)
1395 {
1396 return xmlStrEqual(name, BAD_CAST "body") || xmlStrEqual(name, BAD_CAST "frameset");
1397 }
1398
dom_accept_head_name(const xmlChar * name)1399 static bool dom_accept_head_name(const xmlChar *name)
1400 {
1401 return xmlStrEqual(name, BAD_CAST "head");
1402 }
1403
1404 /* https://html.spec.whatwg.org/#dom-document-body */
dom_html_document_body_read(dom_object * obj,zval * retval)1405 zend_result dom_html_document_body_read(dom_object *obj, zval *retval)
1406 {
1407 return dom_html_document_element_read_helper(obj, retval, dom_accept_body_name);
1408 }
1409
1410 /* https://html.spec.whatwg.org/#dom-document-head */
dom_html_document_head_read(dom_object * obj,zval * retval)1411 zend_result dom_html_document_head_read(dom_object *obj, zval *retval)
1412 {
1413 return dom_html_document_element_read_helper(obj, retval, dom_accept_head_name);
1414 }
1415
1416 /* https://html.spec.whatwg.org/#dom-document-body */
dom_html_document_body_write(dom_object * obj,zval * newval)1417 zend_result dom_html_document_body_write(dom_object *obj, zval *newval)
1418 {
1419 DOM_PROP_NODE(xmlDocPtr, docp, obj);
1420
1421 /* 1. If the new value is not a body or frameset element, then throw a "HierarchyRequestError" DOMException. */
1422 if (Z_TYPE_P(newval) != IS_NULL) {
1423 dom_object *newval_intern = Z_DOMOBJ_P(newval);
1424 if (newval_intern->ptr != NULL) {
1425 xmlNodePtr newval_node = ((php_libxml_node_ptr *) newval_intern->ptr)->node;
1426 /* Note: because this property has type HTMLElement, we know the namespace is correct. */
1427 if (dom_accept_body_name(newval_node->name)) {
1428 /* 2. If the new value is the same as the body element, return. */
1429 const xmlNode *current_body_element = dom_html_document_element_read_raw(docp, dom_accept_body_name);
1430 if (current_body_element == newval_node) {
1431 return SUCCESS;
1432 }
1433
1434 /* 3. If the body element is not null, then replace the body element with the new value within the body element's parent and return. */
1435 if (current_body_element != NULL) {
1436 php_dom_adopt_node(newval_node, obj, docp);
1437 xmlNodePtr old = xmlReplaceNode((xmlNodePtr) current_body_element, newval_node);
1438 if (old != NULL && old->_private == NULL) {
1439 php_libxml_node_free_resource(old);
1440 }
1441 return SUCCESS;
1442 }
1443
1444 /* 4. If there is no document element, throw a "HierarchyRequestError" DOMException. */
1445 xmlNodePtr root = xmlDocGetRootElement(docp);
1446 if (root == NULL) {
1447 php_dom_throw_error_with_message(HIERARCHY_REQUEST_ERR, "A body can only be set if there is a document element", true);
1448 return FAILURE;
1449 }
1450
1451 /* 5. Append the new value to the document element. */
1452 php_dom_adopt_node(newval_node, obj, docp);
1453 xmlAddChild(root, newval_node);
1454 return SUCCESS;
1455 }
1456 }
1457 }
1458
1459 php_dom_throw_error_with_message(HIERARCHY_REQUEST_ERR, "The new body must either be a body or a frameset tag", true);
1460 return FAILURE;
1461 }
1462
1463 /* https://dom.spec.whatwg.org/#concept-child-text-content */
dom_get_child_text_content(const xmlNode * node)1464 static zend_string *dom_get_child_text_content(const xmlNode *node)
1465 {
1466 smart_str content = {0};
1467
1468 const xmlNode *text = node->children;
1469 while (text != NULL) {
1470 if ((text->type == XML_TEXT_NODE || text->type == XML_CDATA_SECTION_NODE) && text->content != NULL) {
1471 smart_str_appends(&content, (const char *) text->content);
1472 }
1473 text = text->next;
1474 }
1475
1476 return smart_str_extract(&content);
1477 }
1478
1479 /* https://html.spec.whatwg.org/#the-title-element-2 */
dom_get_title_element(const xmlDoc * doc)1480 static xmlNodePtr dom_get_title_element(const xmlDoc *doc)
1481 {
1482 xmlNodePtr node = doc->children;
1483
1484 while (node != NULL) {
1485 if (node->type == XML_ELEMENT_NODE) {
1486 if (php_dom_ns_is_fast(node, php_dom_ns_is_html_magic_token) && xmlStrEqual(node->name, BAD_CAST "title")) {
1487 break;
1488 }
1489 }
1490
1491 node = php_dom_next_in_tree_order(node, NULL);
1492 }
1493
1494 return node;
1495 }
1496
1497 /* The subtle difference is that this is about the direct title descendant of the svg element,
1498 * whereas the html variant of this function is about the first in-tree title element. */
dom_get_svg_title_element(xmlNodePtr svg)1499 static xmlNodePtr dom_get_svg_title_element(xmlNodePtr svg)
1500 {
1501 xmlNodePtr cur = svg->children;
1502
1503 while (cur != NULL) {
1504 if (cur->type == XML_ELEMENT_NODE
1505 && php_dom_ns_is_fast(cur, php_dom_ns_is_svg_magic_token) && xmlStrEqual(cur->name, BAD_CAST "title")) {
1506 break;
1507 }
1508 cur = cur->next;
1509 }
1510
1511 return cur;
1512 }
1513
1514 /* https://html.spec.whatwg.org/#document.title */
dom_html_document_title_read(dom_object * obj,zval * retval)1515 zend_result dom_html_document_title_read(dom_object *obj, zval *retval)
1516 {
1517 DOM_PROP_NODE(const xmlDoc *, docp, obj);
1518 xmlNodePtr root = xmlDocGetRootElement(docp);
1519
1520 if (root == NULL) {
1521 ZVAL_EMPTY_STRING(retval);
1522 return SUCCESS;
1523 }
1524
1525 zend_string *value = zend_empty_string;
1526
1527 /* 1. If the document element is an SVG svg element,
1528 * then let value be the child text content of the first SVG title element that is a child of the document element. */
1529 if (php_dom_ns_is_fast(root, php_dom_ns_is_svg_magic_token) && xmlStrEqual(root->name, BAD_CAST "svg")) {
1530 const xmlNode *title = dom_get_svg_title_element(root);
1531 if (title != NULL) {
1532 value = dom_get_child_text_content(title);
1533 }
1534 } else {
1535 /* 2. Otherwise, let value be the child text content of the title element,
1536 * or the empty string if the title element is null. */
1537 const xmlNode *title = dom_get_title_element(docp);
1538 if (title != NULL) {
1539 value = dom_get_child_text_content(title);
1540 }
1541 }
1542
1543 /* 3. Strip and collapse ASCII whitespace in value. */
1544 value = dom_strip_and_collapse_ascii_whitespace(value);
1545
1546 /* 4. Return value. */
1547 ZVAL_STR(retval, value);
1548
1549 return SUCCESS;
1550 }
1551
dom_string_replace_all(xmlDocPtr docp,xmlNodePtr element,zval * zv)1552 static void dom_string_replace_all(xmlDocPtr docp, xmlNodePtr element, zval *zv)
1553 {
1554 dom_remove_all_children(element);
1555 xmlNode *text = xmlNewDocText(docp, BAD_CAST Z_STRVAL_P(zv));
1556 xmlAddChild(element, text);
1557 }
1558
1559 /* https://html.spec.whatwg.org/#document.title */
dom_html_document_title_write(dom_object * obj,zval * newval)1560 zend_result dom_html_document_title_write(dom_object *obj, zval *newval)
1561 {
1562 DOM_PROP_NODE(xmlDocPtr, docp, obj);
1563 xmlNodePtr root = xmlDocGetRootElement(docp);
1564
1565 if (root == NULL) {
1566 return SUCCESS;
1567 }
1568
1569 /* If the document element is an SVG svg element */
1570 if (php_dom_ns_is_fast(root, php_dom_ns_is_svg_magic_token) && xmlStrEqual(root->name, BAD_CAST "svg")) {
1571 /* 1. If there is an SVG title element that is a child of the document element, let element be the first such element. */
1572 xmlNodePtr element = dom_get_svg_title_element(root);
1573
1574 /* 2. Otherwise: */
1575 if (element == NULL) {
1576 /* 2.1. Let element be the result of creating an element given the document element's node document,
1577 * title, and the SVG namespace. */
1578
1579 /* Annoyingly, we must create it in the svg namespace _without_ prefix... */
1580 xmlNsPtr ns = root->ns;
1581 if (ns->prefix != NULL) {
1582 /* Slow path... */
1583 php_dom_libxml_ns_mapper *ns_mapper = php_dom_get_ns_mapper(obj);
1584 zend_string *href = ZSTR_INIT_LITERAL(DOM_SVG_NS_URI, false);
1585 ns = php_dom_libxml_ns_mapper_get_ns(ns_mapper, zend_empty_string, href);
1586 zend_string_release_ex(href, false);
1587 }
1588
1589 element = xmlNewDocNode(docp, ns, BAD_CAST "title", NULL);
1590 if (UNEXPECTED(element == NULL)) {
1591 php_dom_throw_error(INVALID_STATE_ERR, true);
1592 return FAILURE;
1593 }
1594
1595 /* 2.2. Insert element as the first child of the document element. */
1596 if (root->children == NULL) {
1597 root->last = element;
1598 } else {
1599 element->next = root->children;
1600 root->children->prev = element;
1601 }
1602 root->children = element;
1603 element->parent = root;
1604 }
1605
1606 /* 3. String replace all with the given value within element. */
1607 dom_string_replace_all(docp, element, newval);
1608 }
1609 /* If the document element is in the HTML namespace */
1610 else if (php_dom_ns_is_fast(root, php_dom_ns_is_html_magic_token)) {
1611 /* 1. If the title element is null and the head element is null, then return. */
1612 xmlNodePtr title = dom_get_title_element(docp);
1613 xmlNodePtr head = dom_html_document_element_read_raw(docp, dom_accept_head_name);
1614 if (title == NULL && head == NULL) {
1615 return SUCCESS;
1616 }
1617
1618 /* 2. If the title element is non-null, let element be the title element. */
1619 xmlNodePtr element = title;
1620
1621 /* 3. Otherwise: */
1622 if (element == NULL) {
1623 /* 3.1. Let element be the result of creating an element given the document element's node document, title,
1624 * and the HTML namespace. */
1625 php_dom_libxml_ns_mapper *ns_mapper = php_dom_get_ns_mapper(obj);
1626 element = xmlNewDocNode(docp, php_dom_libxml_ns_mapper_ensure_html_ns(ns_mapper), BAD_CAST "title", NULL);
1627 if (UNEXPECTED(element == NULL)) {
1628 php_dom_throw_error(INVALID_STATE_ERR, true);
1629 return FAILURE;
1630 }
1631
1632 /* 3.2. Append element to the head element. */
1633 xmlAddChild(head, element);
1634 }
1635
1636 /* 4. String replace all with the given value within element. */
1637 dom_string_replace_all(docp, element, newval);
1638 }
1639
1640 return SUCCESS;
1641 }
1642
1643 #if ZEND_DEBUG
PHP_METHOD(Dom_HTMLDocument,debugGetTemplateCount)1644 PHP_METHOD(Dom_HTMLDocument, debugGetTemplateCount)
1645 {
1646 xmlDocPtr doc;
1647 dom_object *intern;
1648
1649 ZEND_PARSE_PARAMETERS_NONE();
1650
1651 DOM_GET_OBJ(doc, ZEND_THIS, xmlDocPtr, intern);
1652 ZEND_IGNORE_VALUE(doc);
1653
1654 RETURN_LONG((zend_long) php_dom_get_template_count((const php_dom_private_data *) intern->document->private_data));
1655 }
1656 #endif
1657
1658 #endif /* HAVE_LIBXML && HAVE_DOM */
1659