1 /* Licensed to the Apache Software Foundation (ASF) under one or more
2 * contributor license agreements. See the NOTICE file distributed with
3 * this work for additional information regarding copyright ownership.
4 * The ASF licenses this file to You under the Apache License, Version 2.0
5 * (the "License"); you may not use this file except in compliance with
6 * the License. You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17 #include "apr.h"
18 #include "apr_strings.h"
19
20 #define APR_WANT_STDIO /* for sprintf() */
21 #define APR_WANT_STRFUNC
22 #include "apr_want.h"
23
24 #include "apr_xml.h"
25
26 #include "apu_config.h"
27
28 #if defined(HAVE_XMLPARSE_XMLPARSE_H)
29 #include <xmlparse/xmlparse.h>
30 #elif defined(HAVE_XMLTOK_XMLPARSE_H)
31 #include <xmltok/xmlparse.h>
32 #elif defined(HAVE_XML_XMLPARSE_H)
33 #include <xml/xmlparse.h>
34 #else
35 #include <expat.h>
36 #endif
37
38 #define DEBUG_CR "\r\n"
39
40 static const char APR_KW_xmlns[] = { 0x78, 0x6D, 0x6C, 0x6E, 0x73, '\0' };
41 static const char APR_KW_xmlns_lang[] = { 0x78, 0x6D, 0x6C, 0x3A, 0x6C, 0x61, 0x6E, 0x67, '\0' };
42 static const char APR_KW_DAV[] = { 0x44, 0x41, 0x56, 0x3A, '\0' };
43
44 /* errors related to namespace processing */
45 #define APR_XML_NS_ERROR_UNKNOWN_PREFIX (-1000)
46 #define APR_XML_NS_ERROR_INVALID_DECL (-1001)
47
48 /* test for a namespace prefix that begins with [Xx][Mm][Ll] */
49 #define APR_XML_NS_IS_RESERVED(name) \
50 ( (name[0] == 0x58 || name[0] == 0x78) && \
51 (name[1] == 0x4D || name[1] == 0x6D) && \
52 (name[2] == 0x4C || name[2] == 0x6C) )
53
54
55 /* the real (internal) definition of the parser context */
56 struct apr_xml_parser {
57 apr_xml_doc *doc; /* the doc we're parsing */
58 apr_pool_t *p; /* the pool we allocate from */
59 apr_xml_elem *cur_elem; /* current element */
60
61 int error; /* an error has occurred */
62 #define APR_XML_ERROR_EXPAT 1
63 #define APR_XML_ERROR_PARSE_DONE 2
64 /* also: public APR_XML_NS_ERROR_* values (if any) */
65
66 XML_Parser xp; /* the actual (Expat) XML parser */
67 enum XML_Error xp_err; /* stored Expat error code */
68 };
69
70 /* struct for scoping namespace declarations */
71 typedef struct apr_xml_ns_scope {
72 const char *prefix; /* prefix used for this ns */
73 int ns; /* index into namespace table */
74 int emptyURI; /* the namespace URI is the empty string */
75 struct apr_xml_ns_scope *next; /* next scoped namespace */
76 } apr_xml_ns_scope;
77
78
79 /* return namespace table index for a given prefix */
find_prefix(apr_xml_parser * parser,const char * prefix)80 static int find_prefix(apr_xml_parser *parser, const char *prefix)
81 {
82 apr_xml_elem *elem = parser->cur_elem;
83
84 /*
85 ** Walk up the tree, looking for a namespace scope that defines this
86 ** prefix.
87 */
88 for (; elem; elem = elem->parent) {
89 apr_xml_ns_scope *ns_scope = elem->ns_scope;
90
91 for (ns_scope = elem->ns_scope; ns_scope; ns_scope = ns_scope->next) {
92 if (strcmp(prefix, ns_scope->prefix) == 0) {
93 if (ns_scope->emptyURI) {
94 /*
95 ** It is possible to set the default namespace to an
96 ** empty URI string; this resets the default namespace
97 ** to mean "no namespace." We just found the prefix
98 ** refers to an empty URI, so return "no namespace."
99 */
100 return APR_XML_NS_NONE;
101 }
102
103 return ns_scope->ns;
104 }
105 }
106 }
107
108 /*
109 * If the prefix is empty (""), this means that a prefix was not
110 * specified in the element/attribute. The search that was performed
111 * just above did not locate a default namespace URI (which is stored
112 * into ns_scope with an empty prefix). This means the element/attribute
113 * has "no namespace". We have a reserved value for this.
114 */
115 if (*prefix == '\0') {
116 return APR_XML_NS_NONE;
117 }
118
119 /* not found */
120 return APR_XML_NS_ERROR_UNKNOWN_PREFIX;
121 }
122
start_handler(void * userdata,const char * name,const char ** attrs)123 static void start_handler(void *userdata, const char *name, const char **attrs)
124 {
125 apr_xml_parser *parser = userdata;
126 apr_xml_elem *elem;
127 apr_xml_attr *attr;
128 apr_xml_attr *prev;
129 char *colon;
130 const char *quoted;
131 char *elem_name;
132
133 /* punt once we find an error */
134 if (parser->error)
135 return;
136
137 elem = apr_pcalloc(parser->p, sizeof(*elem));
138
139 /* prep the element */
140 elem->name = elem_name = apr_pstrdup(parser->p, name);
141
142 /* fill in the attributes (note: ends up in reverse order) */
143 while (*attrs) {
144 attr = apr_palloc(parser->p, sizeof(*attr));
145 attr->name = apr_pstrdup(parser->p, *attrs++);
146 attr->value = apr_pstrdup(parser->p, *attrs++);
147 attr->next = elem->attr;
148 elem->attr = attr;
149 }
150
151 /* hook the element into the tree */
152 if (parser->cur_elem == NULL) {
153 /* no current element; this also becomes the root */
154 parser->cur_elem = parser->doc->root = elem;
155 }
156 else {
157 /* this element appeared within the current elem */
158 elem->parent = parser->cur_elem;
159
160 /* set up the child/sibling links */
161 if (elem->parent->last_child == NULL) {
162 /* no first child either */
163 elem->parent->first_child = elem->parent->last_child = elem;
164 }
165 else {
166 /* hook onto the end of the parent's children */
167 elem->parent->last_child->next = elem;
168 elem->parent->last_child = elem;
169 }
170
171 /* this element is now the current element */
172 parser->cur_elem = elem;
173 }
174
175 /* scan the attributes for namespace declarations */
176 for (prev = NULL, attr = elem->attr;
177 attr;
178 attr = attr->next) {
179 if (strncmp(attr->name, APR_KW_xmlns, 5) == 0) {
180 const char *prefix = &attr->name[5];
181 apr_xml_ns_scope *ns_scope;
182
183 /* test for xmlns:foo= form and xmlns= form */
184 if (*prefix == 0x3A) {
185 /* a namespace prefix declaration must have a
186 non-empty value. */
187 if (attr->value[0] == '\0') {
188 parser->error = APR_XML_NS_ERROR_INVALID_DECL;
189 return;
190 }
191 ++prefix;
192 }
193 else if (*prefix != '\0') {
194 /* advance "prev" since "attr" is still present */
195 prev = attr;
196 continue;
197 }
198
199 /* quote the URI before we ever start working with it */
200 quoted = apr_xml_quote_string(parser->p, attr->value, 1);
201
202 /* build and insert the new scope */
203 ns_scope = apr_pcalloc(parser->p, sizeof(*ns_scope));
204 ns_scope->prefix = prefix;
205 ns_scope->ns = apr_xml_insert_uri(parser->doc->namespaces, quoted);
206 ns_scope->emptyURI = *quoted == '\0';
207 ns_scope->next = elem->ns_scope;
208 elem->ns_scope = ns_scope;
209
210 /* remove this attribute from the element */
211 if (prev == NULL)
212 elem->attr = attr->next;
213 else
214 prev->next = attr->next;
215
216 /* Note: prev will not be advanced since we just removed "attr" */
217 }
218 else if (strcmp(attr->name, APR_KW_xmlns_lang) == 0) {
219 /* save away the language (in quoted form) */
220 elem->lang = apr_xml_quote_string(parser->p, attr->value, 1);
221
222 /* remove this attribute from the element */
223 if (prev == NULL)
224 elem->attr = attr->next;
225 else
226 prev->next = attr->next;
227
228 /* Note: prev will not be advanced since we just removed "attr" */
229 }
230 else {
231 /* advance "prev" since "attr" is still present */
232 prev = attr;
233 }
234 }
235
236 /*
237 ** If an xml:lang attribute didn't exist (lang==NULL), then copy the
238 ** language from the parent element (if present).
239 **
240 ** NOTE: elem_size() *depends* upon this pointer equality.
241 */
242 if (elem->lang == NULL && elem->parent != NULL)
243 elem->lang = elem->parent->lang;
244
245 /* adjust the element's namespace */
246 colon = strchr(elem_name, 0x3A);
247 if (colon == NULL) {
248 /*
249 * The element is using the default namespace, which will always
250 * be found. Either it will be "no namespace", or a default
251 * namespace URI has been specified at some point.
252 */
253 elem->ns = find_prefix(parser, "");
254 }
255 else if (APR_XML_NS_IS_RESERVED(elem->name)) {
256 elem->ns = APR_XML_NS_NONE;
257 }
258 else {
259 *colon = '\0';
260 elem->ns = find_prefix(parser, elem->name);
261 elem->name = colon + 1;
262
263 if (APR_XML_NS_IS_ERROR(elem->ns)) {
264 parser->error = elem->ns;
265 return;
266 }
267 }
268
269 /* adjust all remaining attributes' namespaces */
270 for (attr = elem->attr; attr; attr = attr->next) {
271 /*
272 * apr_xml_attr defines this as "const" but we dup'd it, so we
273 * know that we can change it. a bit hacky, but the existing
274 * structure def is best.
275 */
276 char *attr_name = (char *)attr->name;
277
278 colon = strchr(attr_name, 0x3A);
279 if (colon == NULL) {
280 /*
281 * Attributes do NOT use the default namespace. Therefore,
282 * we place them into the "no namespace" category.
283 */
284 attr->ns = APR_XML_NS_NONE;
285 }
286 else if (APR_XML_NS_IS_RESERVED(attr->name)) {
287 attr->ns = APR_XML_NS_NONE;
288 }
289 else {
290 *colon = '\0';
291 attr->ns = find_prefix(parser, attr->name);
292 attr->name = colon + 1;
293
294 if (APR_XML_NS_IS_ERROR(attr->ns)) {
295 parser->error = attr->ns;
296 return;
297 }
298 }
299 }
300 }
301
end_handler(void * userdata,const char * name)302 static void end_handler(void *userdata, const char *name)
303 {
304 apr_xml_parser *parser = userdata;
305
306 /* punt once we find an error */
307 if (parser->error)
308 return;
309
310 /* pop up one level */
311 parser->cur_elem = parser->cur_elem->parent;
312 }
313
cdata_handler(void * userdata,const char * data,int len)314 static void cdata_handler(void *userdata, const char *data, int len)
315 {
316 apr_xml_parser *parser = userdata;
317 apr_xml_elem *elem;
318 apr_text_header *hdr;
319 const char *s;
320
321 /* punt once we find an error */
322 if (parser->error)
323 return;
324
325 elem = parser->cur_elem;
326 s = apr_pstrndup(parser->p, data, len);
327
328 if (elem->last_child == NULL) {
329 /* no children yet. this cdata follows the start tag */
330 hdr = &elem->first_cdata;
331 }
332 else {
333 /* child elements exist. this cdata follows the last child. */
334 hdr = &elem->last_child->following_cdata;
335 }
336
337 apr_text_append(parser->p, hdr, s);
338 }
339
cleanup_parser(void * ctx)340 static apr_status_t cleanup_parser(void *ctx)
341 {
342 apr_xml_parser *parser = ctx;
343
344 XML_ParserFree(parser->xp);
345 parser->xp = NULL;
346
347 return APR_SUCCESS;
348 }
349
350 #if XML_MAJOR_VERSION > 1
351 /* Stop the parser if an entity declaration is hit. */
entity_declaration(void * userData,const XML_Char * entityName,int is_parameter_entity,const XML_Char * value,int value_length,const XML_Char * base,const XML_Char * systemId,const XML_Char * publicId,const XML_Char * notationName)352 static void entity_declaration(void *userData, const XML_Char *entityName,
353 int is_parameter_entity, const XML_Char *value,
354 int value_length, const XML_Char *base,
355 const XML_Char *systemId, const XML_Char *publicId,
356 const XML_Char *notationName)
357 {
358 apr_xml_parser *parser = userData;
359
360 XML_StopParser(parser->xp, XML_FALSE);
361 }
362 #else
363 /* A noop default_handler. */
default_handler(void * userData,const XML_Char * s,int len)364 static void default_handler(void *userData, const XML_Char *s, int len)
365 {
366 }
367 #endif
368
apr_xml_parser_create(apr_pool_t * pool)369 APU_DECLARE(apr_xml_parser *) apr_xml_parser_create(apr_pool_t *pool)
370 {
371 apr_xml_parser *parser = apr_pcalloc(pool, sizeof(*parser));
372
373 parser->p = pool;
374 parser->doc = apr_pcalloc(pool, sizeof(*parser->doc));
375
376 parser->doc->namespaces = apr_array_make(pool, 5, sizeof(const char *));
377
378 /* ### is there a way to avoid hard-coding this? */
379 apr_xml_insert_uri(parser->doc->namespaces, APR_KW_DAV);
380
381 parser->xp = XML_ParserCreate(NULL);
382 if (parser->xp == NULL) {
383 (*apr_pool_abort_get(pool))(APR_ENOMEM);
384 return NULL;
385 }
386
387 apr_pool_cleanup_register(pool, parser, cleanup_parser,
388 apr_pool_cleanup_null);
389
390 XML_SetUserData(parser->xp, parser);
391 XML_SetElementHandler(parser->xp, start_handler, end_handler);
392 XML_SetCharacterDataHandler(parser->xp, cdata_handler);
393
394 /* Prevent the "billion laughs" attack against expat by disabling
395 * internal entity expansion. With 2.x, forcibly stop the parser
396 * if an entity is declared - this is safer and a more obvious
397 * failure mode. With older versions, installing a noop
398 * DefaultHandler means that internal entities will be expanded as
399 * the empty string, which is also sufficient to prevent the
400 * attack. */
401 #if XML_MAJOR_VERSION > 1
402 XML_SetEntityDeclHandler(parser->xp, entity_declaration);
403 #else
404 XML_SetDefaultHandler(parser->xp, default_handler);
405 #endif
406
407 return parser;
408 }
409
do_parse(apr_xml_parser * parser,const char * data,apr_size_t len,int is_final)410 static apr_status_t do_parse(apr_xml_parser *parser,
411 const char *data, apr_size_t len,
412 int is_final)
413 {
414 if (parser->xp == NULL) {
415 parser->error = APR_XML_ERROR_PARSE_DONE;
416 }
417 else {
418 int rv = XML_Parse(parser->xp, data, (int)len, is_final);
419
420 if (rv == 0) {
421 parser->error = APR_XML_ERROR_EXPAT;
422 parser->xp_err = XML_GetErrorCode(parser->xp);
423 }
424 }
425
426 /* ### better error code? */
427 return parser->error ? APR_EGENERAL : APR_SUCCESS;
428 }
429
apr_xml_parser_feed(apr_xml_parser * parser,const char * data,apr_size_t len)430 APU_DECLARE(apr_status_t) apr_xml_parser_feed(apr_xml_parser *parser,
431 const char *data,
432 apr_size_t len)
433 {
434 return do_parse(parser, data, len, 0 /* is_final */);
435 }
436
apr_xml_parser_done(apr_xml_parser * parser,apr_xml_doc ** pdoc)437 APU_DECLARE(apr_status_t) apr_xml_parser_done(apr_xml_parser *parser,
438 apr_xml_doc **pdoc)
439 {
440 char end;
441 apr_status_t status = do_parse(parser, &end, 0, 1 /* is_final */);
442
443 /* get rid of the parser */
444 (void) apr_pool_cleanup_run(parser->p, parser, cleanup_parser);
445
446 if (status)
447 return status;
448
449 if (pdoc != NULL)
450 *pdoc = parser->doc;
451 return APR_SUCCESS;
452 }
453
apr_xml_parser_geterror(apr_xml_parser * parser,char * errbuf,apr_size_t errbufsize)454 APU_DECLARE(char *) apr_xml_parser_geterror(apr_xml_parser *parser,
455 char *errbuf,
456 apr_size_t errbufsize)
457 {
458 int error = parser->error;
459 const char *msg;
460
461 /* clear our record of an error */
462 parser->error = 0;
463
464 switch (error) {
465 case 0:
466 msg = "No error.";
467 break;
468
469 case APR_XML_NS_ERROR_UNKNOWN_PREFIX:
470 msg = "An undefined namespace prefix was used.";
471 break;
472
473 case APR_XML_NS_ERROR_INVALID_DECL:
474 msg = "A namespace prefix was defined with an empty URI.";
475 break;
476
477 case APR_XML_ERROR_EXPAT:
478 (void) apr_snprintf(errbuf, errbufsize,
479 "XML parser error code: %s (%d)",
480 XML_ErrorString(parser->xp_err), parser->xp_err);
481 return errbuf;
482
483 case APR_XML_ERROR_PARSE_DONE:
484 msg = "The parser is not active.";
485 break;
486
487 default:
488 msg = "There was an unknown error within the XML body.";
489 break;
490 }
491
492 (void) apr_cpystrn(errbuf, msg, errbufsize);
493 return errbuf;
494 }
495
apr_xml_parse_file(apr_pool_t * p,apr_xml_parser ** parser,apr_xml_doc ** ppdoc,apr_file_t * xmlfd,apr_size_t buffer_length)496 APU_DECLARE(apr_status_t) apr_xml_parse_file(apr_pool_t *p,
497 apr_xml_parser **parser,
498 apr_xml_doc **ppdoc,
499 apr_file_t *xmlfd,
500 apr_size_t buffer_length)
501 {
502 apr_status_t rv;
503 char *buffer;
504 apr_size_t length;
505
506 *parser = apr_xml_parser_create(p);
507 if (*parser == NULL) {
508 /* FIXME: returning an error code would be nice,
509 * but we dont get one ;( */
510 return APR_EGENERAL;
511 }
512 buffer = apr_palloc(p, buffer_length);
513 length = buffer_length;
514
515 rv = apr_file_read(xmlfd, buffer, &length);
516
517 while (rv == APR_SUCCESS) {
518 rv = apr_xml_parser_feed(*parser, buffer, length);
519 if (rv != APR_SUCCESS) {
520 return rv;
521 }
522
523 length = buffer_length;
524 rv = apr_file_read(xmlfd, buffer, &length);
525 }
526 if (rv != APR_EOF) {
527 return rv;
528 }
529 rv = apr_xml_parser_done(*parser, ppdoc);
530 *parser = NULL;
531 return rv;
532 }
533
apr_text_append(apr_pool_t * p,apr_text_header * hdr,const char * text)534 APU_DECLARE(void) apr_text_append(apr_pool_t * p, apr_text_header *hdr,
535 const char *text)
536 {
537 apr_text *t = apr_palloc(p, sizeof(*t));
538
539 t->text = text;
540 t->next = NULL;
541
542 if (hdr->first == NULL) {
543 /* no text elements yet */
544 hdr->first = hdr->last = t;
545 }
546 else {
547 /* append to the last text element */
548 hdr->last->next = t;
549 hdr->last = t;
550 }
551 }
552
553
554 /* ---------------------------------------------------------------
555 **
556 ** XML UTILITY FUNCTIONS
557 */
558
559 /*
560 ** apr_xml_quote_string: quote an XML string
561 **
562 ** Replace '<', '>', and '&' with '<', '>', and '&'.
563 ** If quotes is true, then replace '"' with '"'.
564 **
565 ** quotes is typically set to true for XML strings that will occur within
566 ** double quotes -- attribute values.
567 */
apr_xml_quote_string(apr_pool_t * p,const char * s,int quotes)568 APU_DECLARE(const char *) apr_xml_quote_string(apr_pool_t *p, const char *s,
569 int quotes)
570 {
571 const char *scan;
572 apr_size_t len = 0;
573 apr_size_t extra = 0;
574 char *qstr;
575 char *qscan;
576 char c;
577
578 for (scan = s; (c = *scan) != '\0'; ++scan, ++len) {
579 if (c == '<' || c == '>')
580 extra += 3; /* < or > */
581 else if (c == '&')
582 extra += 4; /* & */
583 else if (quotes && c == '"')
584 extra += 5; /* " */
585 }
586
587 /* nothing to do? */
588 if (extra == 0)
589 return s;
590
591 qstr = apr_palloc(p, len + extra + 1);
592 for (scan = s, qscan = qstr; (c = *scan) != '\0'; ++scan) {
593 if (c == '<') {
594 *qscan++ = '&';
595 *qscan++ = 'l';
596 *qscan++ = 't';
597 *qscan++ = ';';
598 }
599 else if (c == '>') {
600 *qscan++ = '&';
601 *qscan++ = 'g';
602 *qscan++ = 't';
603 *qscan++ = ';';
604 }
605 else if (c == '&') {
606 *qscan++ = '&';
607 *qscan++ = 'a';
608 *qscan++ = 'm';
609 *qscan++ = 'p';
610 *qscan++ = ';';
611 }
612 else if (quotes && c == '"') {
613 *qscan++ = '&';
614 *qscan++ = 'q';
615 *qscan++ = 'u';
616 *qscan++ = 'o';
617 *qscan++ = 't';
618 *qscan++ = ';';
619 }
620 else {
621 *qscan++ = c;
622 }
623 }
624
625 *qscan = '\0';
626 return qstr;
627 }
628
629 /* how many characters for the given integer? */
630 #define APR_XML_NS_LEN(ns) ((ns) < 10 ? 1 : (ns) < 100 ? 2 : (ns) < 1000 ? 3 : \
631 (ns) < 10000 ? 4 : (ns) < 100000 ? 5 : \
632 (ns) < 1000000 ? 6 : (ns) < 10000000 ? 7 : \
633 (ns) < 100000000 ? 8 : (ns) < 1000000000 ? 9 : 10)
634
text_size(const apr_text * t)635 static apr_size_t text_size(const apr_text *t)
636 {
637 apr_size_t size = 0;
638
639 for (; t; t = t->next)
640 size += strlen(t->text);
641 return size;
642 }
643
elem_size(const apr_xml_elem * elem,int style,apr_array_header_t * namespaces,int * ns_map)644 static apr_size_t elem_size(const apr_xml_elem *elem, int style,
645 apr_array_header_t *namespaces, int *ns_map)
646 {
647 apr_size_t size;
648
649 if (style == APR_XML_X2T_FULL || style == APR_XML_X2T_FULL_NS_LANG) {
650 const apr_xml_attr *attr;
651
652 size = 0;
653
654 if (style == APR_XML_X2T_FULL_NS_LANG) {
655 int i;
656
657 /*
658 ** The outer element will contain xmlns:ns%d="%s" attributes
659 ** and an xml:lang attribute, if applicable.
660 */
661
662 for (i = namespaces->nelts; i--;) {
663 /* compute size of: ' xmlns:ns%d="%s"' */
664 size += (9 + APR_XML_NS_LEN(i) + 2 +
665 strlen(APR_XML_GET_URI_ITEM(namespaces, i)) + 1);
666 }
667
668 if (elem->lang != NULL) {
669 /* compute size of: ' xml:lang="%s"' */
670 size += 11 + strlen(elem->lang) + 1;
671 }
672 }
673
674 if (elem->ns == APR_XML_NS_NONE) {
675 /* compute size of: <%s> */
676 size += 1 + strlen(elem->name) + 1;
677 }
678 else {
679 int ns = ns_map ? ns_map[elem->ns] : elem->ns;
680
681 /* compute size of: <ns%d:%s> */
682 size += 3 + APR_XML_NS_LEN(ns) + 1 + strlen(elem->name) + 1;
683 }
684
685 if (APR_XML_ELEM_IS_EMPTY(elem)) {
686 /* insert a closing "/" */
687 size += 1;
688 }
689 else {
690 /*
691 * two of above plus "/":
692 * <ns%d:%s> ... </ns%d:%s>
693 * OR <%s> ... </%s>
694 */
695 size = 2 * size + 1;
696 }
697
698 for (attr = elem->attr; attr; attr = attr->next) {
699 if (attr->ns == APR_XML_NS_NONE) {
700 /* compute size of: ' %s="%s"' */
701 size += 1 + strlen(attr->name) + 2 + strlen(attr->value) + 1;
702 }
703 else {
704 /* compute size of: ' ns%d:%s="%s"' */
705 int ns = ns_map ? ns_map[attr->ns] : attr->ns;
706 size += 3 + APR_XML_NS_LEN(ns) + 1 + strlen(attr->name) + 2 + strlen(attr->value) + 1;
707 }
708 }
709
710 /*
711 ** If the element has an xml:lang value that is *different* from
712 ** its parent, then add the thing in: ' xml:lang="%s"'.
713 **
714 ** NOTE: we take advantage of the pointer equality established by
715 ** the parsing for "inheriting" the xml:lang values from parents.
716 */
717 if (elem->lang != NULL &&
718 (elem->parent == NULL || elem->lang != elem->parent->lang)) {
719 size += 11 + strlen(elem->lang) + 1;
720 }
721 }
722 else if (style == APR_XML_X2T_LANG_INNER) {
723 /*
724 * This style prepends the xml:lang value plus a null terminator.
725 * If a lang value is not present, then we insert a null term.
726 */
727 size = elem->lang ? strlen(elem->lang) + 1 : 1;
728 }
729 else
730 size = 0;
731
732 size += text_size(elem->first_cdata.first);
733
734 for (elem = elem->first_child; elem; elem = elem->next) {
735 /* the size of the child element plus the CDATA that follows it */
736 size += (elem_size(elem, APR_XML_X2T_FULL, NULL, ns_map) +
737 text_size(elem->following_cdata.first));
738 }
739
740 return size;
741 }
742
write_text(char * s,const apr_text * t)743 static char *write_text(char *s, const apr_text *t)
744 {
745 for (; t; t = t->next) {
746 apr_size_t len = strlen(t->text);
747 memcpy(s, t->text, len);
748 s += len;
749 }
750 return s;
751 }
752
write_elem(char * s,const apr_xml_elem * elem,int style,apr_array_header_t * namespaces,int * ns_map)753 static char *write_elem(char *s, const apr_xml_elem *elem, int style,
754 apr_array_header_t *namespaces, int *ns_map)
755 {
756 const apr_xml_elem *child;
757 apr_size_t len;
758 int ns;
759
760 if (style == APR_XML_X2T_FULL || style == APR_XML_X2T_FULL_NS_LANG) {
761 int empty = APR_XML_ELEM_IS_EMPTY(elem);
762 const apr_xml_attr *attr;
763
764 if (elem->ns == APR_XML_NS_NONE) {
765 len = sprintf(s, "<%s", elem->name);
766 }
767 else {
768 ns = ns_map ? ns_map[elem->ns] : elem->ns;
769 len = sprintf(s, "<ns%d:%s", ns, elem->name);
770 }
771 s += len;
772
773 for (attr = elem->attr; attr; attr = attr->next) {
774 if (attr->ns == APR_XML_NS_NONE)
775 len = sprintf(s, " %s=\"%s\"", attr->name, attr->value);
776 else {
777 ns = ns_map ? ns_map[attr->ns] : attr->ns;
778 len = sprintf(s, " ns%d:%s=\"%s\"", ns, attr->name, attr->value);
779 }
780 s += len;
781 }
782
783 /* add the xml:lang value if necessary */
784 if (elem->lang != NULL &&
785 (style == APR_XML_X2T_FULL_NS_LANG ||
786 elem->parent == NULL ||
787 elem->lang != elem->parent->lang)) {
788 len = sprintf(s, " xml:lang=\"%s\"", elem->lang);
789 s += len;
790 }
791
792 /* add namespace definitions, if required */
793 if (style == APR_XML_X2T_FULL_NS_LANG) {
794 int i;
795
796 for (i = namespaces->nelts; i--;) {
797 len = sprintf(s, " xmlns:ns%d=\"%s\"", i,
798 APR_XML_GET_URI_ITEM(namespaces, i));
799 s += len;
800 }
801 }
802
803 /* no more to do. close it up and go. */
804 if (empty) {
805 *s++ = '/';
806 *s++ = '>';
807 return s;
808 }
809
810 /* just close it */
811 *s++ = '>';
812 }
813 else if (style == APR_XML_X2T_LANG_INNER) {
814 /* prepend the xml:lang value */
815 if (elem->lang != NULL) {
816 len = strlen(elem->lang);
817 memcpy(s, elem->lang, len);
818 s += len;
819 }
820 *s++ = '\0';
821 }
822
823 s = write_text(s, elem->first_cdata.first);
824
825 for (child = elem->first_child; child; child = child->next) {
826 s = write_elem(s, child, APR_XML_X2T_FULL, NULL, ns_map);
827 s = write_text(s, child->following_cdata.first);
828 }
829
830 if (style == APR_XML_X2T_FULL || style == APR_XML_X2T_FULL_NS_LANG) {
831 if (elem->ns == APR_XML_NS_NONE) {
832 len = sprintf(s, "</%s>", elem->name);
833 }
834 else {
835 ns = ns_map ? ns_map[elem->ns] : elem->ns;
836 len = sprintf(s, "</ns%d:%s>", ns, elem->name);
837 }
838 s += len;
839 }
840
841 return s;
842 }
843
apr_xml_quote_elem(apr_pool_t * p,apr_xml_elem * elem)844 APU_DECLARE(void) apr_xml_quote_elem(apr_pool_t *p, apr_xml_elem *elem)
845 {
846 apr_text *scan_txt;
847 apr_xml_attr *scan_attr;
848 apr_xml_elem *scan_elem;
849
850 /* convert the element's text */
851 for (scan_txt = elem->first_cdata.first;
852 scan_txt != NULL;
853 scan_txt = scan_txt->next) {
854 scan_txt->text = apr_xml_quote_string(p, scan_txt->text, 0);
855 }
856 for (scan_txt = elem->following_cdata.first;
857 scan_txt != NULL;
858 scan_txt = scan_txt->next) {
859 scan_txt->text = apr_xml_quote_string(p, scan_txt->text, 0);
860 }
861
862 /* convert the attribute values */
863 for (scan_attr = elem->attr;
864 scan_attr != NULL;
865 scan_attr = scan_attr->next) {
866 scan_attr->value = apr_xml_quote_string(p, scan_attr->value, 1);
867 }
868
869 /* convert the child elements */
870 for (scan_elem = elem->first_child;
871 scan_elem != NULL;
872 scan_elem = scan_elem->next) {
873 apr_xml_quote_elem(p, scan_elem);
874 }
875 }
876
877 /* convert an element to a text string */
apr_xml_to_text(apr_pool_t * p,const apr_xml_elem * elem,int style,apr_array_header_t * namespaces,int * ns_map,const char ** pbuf,apr_size_t * psize)878 APU_DECLARE(void) apr_xml_to_text(apr_pool_t * p, const apr_xml_elem *elem,
879 int style, apr_array_header_t *namespaces,
880 int *ns_map, const char **pbuf,
881 apr_size_t *psize)
882 {
883 /* get the exact size, plus a null terminator */
884 apr_size_t size = elem_size(elem, style, namespaces, ns_map) + 1;
885 char *s = apr_palloc(p, size);
886
887 (void) write_elem(s, elem, style, namespaces, ns_map);
888 s[size - 1] = '\0';
889
890 *pbuf = s;
891 if (psize)
892 *psize = size;
893 }
894
apr_xml_empty_elem(apr_pool_t * p,const apr_xml_elem * elem)895 APU_DECLARE(const char *) apr_xml_empty_elem(apr_pool_t * p,
896 const apr_xml_elem *elem)
897 {
898 if (elem->ns == APR_XML_NS_NONE) {
899 /*
900 * The prefix (xml...) is already within the prop name, or
901 * the element simply has no prefix.
902 */
903 return apr_psprintf(p, "<%s/>" DEBUG_CR, elem->name);
904 }
905
906 return apr_psprintf(p, "<ns%d:%s/>" DEBUG_CR, elem->ns, elem->name);
907 }
908
909 /* return the URI's (existing) index, or insert it and return a new index */
apr_xml_insert_uri(apr_array_header_t * uri_array,const char * uri)910 APU_DECLARE(int) apr_xml_insert_uri(apr_array_header_t *uri_array,
911 const char *uri)
912 {
913 int i;
914 const char **pelt;
915
916 /* never insert an empty URI; this index is always APR_XML_NS_NONE */
917 if (*uri == '\0')
918 return APR_XML_NS_NONE;
919
920 for (i = uri_array->nelts; i--;) {
921 if (strcmp(uri, APR_XML_GET_URI_ITEM(uri_array, i)) == 0)
922 return i;
923 }
924
925 pelt = apr_array_push(uri_array);
926 *pelt = uri; /* assume uri is const or in a pool */
927 return uri_array->nelts - 1;
928 }
929
930 /* convert the element to EBCDIC */
931 #if APR_CHARSET_EBCDIC
apr_xml_parser_convert_elem(apr_xml_elem * e,apr_xlate_t * convset)932 static apr_status_t apr_xml_parser_convert_elem(apr_xml_elem *e,
933 apr_xlate_t *convset)
934 {
935 apr_xml_attr *a;
936 apr_xml_elem *ec;
937 apr_text *t;
938 apr_size_t inbytes_left, outbytes_left;
939 apr_status_t status;
940
941 inbytes_left = outbytes_left = strlen(e->name);
942 status = apr_xlate_conv_buffer(convset, e->name, &inbytes_left, (char *) e->name, &outbytes_left);
943 if (status) {
944 return status;
945 }
946
947 for (t = e->first_cdata.first; t != NULL; t = t->next) {
948 inbytes_left = outbytes_left = strlen(t->text);
949 status = apr_xlate_conv_buffer(convset, t->text, &inbytes_left, (char *) t->text, &outbytes_left);
950 if (status) {
951 return status;
952 }
953 }
954
955 for (t = e->following_cdata.first; t != NULL; t = t->next) {
956 inbytes_left = outbytes_left = strlen(t->text);
957 status = apr_xlate_conv_buffer(convset, t->text, &inbytes_left, (char *) t->text, &outbytes_left);
958 if (status) {
959 return status;
960 }
961 }
962
963 for (a = e->attr; a != NULL; a = a->next) {
964 inbytes_left = outbytes_left = strlen(a->name);
965 status = apr_xlate_conv_buffer(convset, a->name, &inbytes_left, (char *) a->name, &outbytes_left);
966 if (status) {
967 return status;
968 }
969 inbytes_left = outbytes_left = strlen(a->value);
970 status = apr_xlate_conv_buffer(convset, a->value, &inbytes_left, (char *) a->value, &outbytes_left);
971 if (status) {
972 return status;
973 }
974 }
975
976 for (ec = e->first_child; ec != NULL; ec = ec->next) {
977 status = apr_xml_parser_convert_elem(ec, convset);
978 if (status) {
979 return status;
980 }
981 }
982 return APR_SUCCESS;
983 }
984
985 /* convert the whole document to EBCDIC */
apr_xml_parser_convert_doc(apr_pool_t * pool,apr_xml_doc * pdoc,apr_xlate_t * convset)986 APU_DECLARE(apr_status_t) apr_xml_parser_convert_doc(apr_pool_t *pool,
987 apr_xml_doc *pdoc,
988 apr_xlate_t *convset)
989 {
990 apr_status_t status;
991 /* Don't convert the namespaces: they are constant! */
992 if (pdoc->namespaces != NULL) {
993 int i;
994 apr_array_header_t *namespaces;
995 namespaces = apr_array_make(pool, pdoc->namespaces->nelts, sizeof(const char *));
996 if (namespaces == NULL)
997 return APR_ENOMEM;
998 for (i = 0; i < pdoc->namespaces->nelts; i++) {
999 apr_size_t inbytes_left, outbytes_left;
1000 char *ptr = (char *) APR_XML_GET_URI_ITEM(pdoc->namespaces, i);
1001 ptr = apr_pstrdup(pool, ptr);
1002 if ( ptr == NULL)
1003 return APR_ENOMEM;
1004 inbytes_left = outbytes_left = strlen(ptr);
1005 status = apr_xlate_conv_buffer(convset, ptr, &inbytes_left, ptr, &outbytes_left);
1006 if (status) {
1007 return status;
1008 }
1009 apr_xml_insert_uri(namespaces, ptr);
1010 }
1011 pdoc->namespaces = namespaces;
1012 }
1013 return apr_xml_parser_convert_elem(pdoc->root, convset);
1014 }
1015 #endif
1016