1
/* $Id: xml.c 3553 2011-05-05 06:14:19Z nanang $ */
3
* Copyright (C) 2008-2011 Teluu Inc. (http://www.teluu.com)
4
* Copyright (C) 2003-2008 Benny Prijono <benny@prijono.org>
6
* This program is free software; you can redistribute it and/or modify
7
* it under the terms of the GNU General Public License as published by
8
* the Free Software Foundation; either version 2 of the License, or
9
* (at your option) any later version.
11
* This program is distributed in the hope that it will be useful,
12
* but WITHOUT ANY WARRANTY; without even the implied warranty of
13
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14
* GNU General Public License for more details.
16
* You should have received a copy of the GNU General Public License
17
* along with this program; if not, write to the Free Software
18
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20
#include <pjlib-util/xml.h>
21
#include <pjlib-util/scanner.h>
22
#include <pj/except.h>
24
#include <pj/string.h>
28
#define EX_SYNTAX_ERROR 12
29
#define THIS_FILE "xml.c"
31
static void on_syntax_error(struct pj_scanner *scanner)
33
PJ_UNUSED_ARG(scanner);
34
PJ_THROW(EX_SYNTAX_ERROR);
37
static pj_xml_node *alloc_node( pj_pool_t *pool )
41
node = PJ_POOL_ZALLOC_T(pool, pj_xml_node);
42
pj_list_init( &node->attr_head );
43
pj_list_init( &node->node_head );
48
static pj_xml_attr *alloc_attr( pj_pool_t *pool )
50
return PJ_POOL_ZALLOC_T(pool, pj_xml_attr);
53
/* This is a recursive function! */
54
static pj_xml_node *xml_parse_node( pj_pool_t *pool, pj_scanner *scanner)
61
if (*scanner->curptr != '<')
62
on_syntax_error(scanner);
64
/* Handle Processing Instructino (PI) construct (i.e. "<?") */
65
if (*scanner->curptr == '<' && *(scanner->curptr+1) == '?') {
66
pj_scan_advance_n(scanner, 2, PJ_FALSE);
69
pj_scan_get_until_ch(scanner, '?', &dummy);
70
if (*scanner->curptr=='?' && *(scanner->curptr+1)=='>') {
71
pj_scan_advance_n(scanner, 2, PJ_TRUE);
74
pj_scan_advance_n(scanner, 1, PJ_FALSE);
77
return xml_parse_node(pool, scanner);
80
/* Handle comments construct (i.e. "<!") */
81
if (pj_scan_strcmp(scanner, "<!", 2) == 0) {
82
pj_scan_advance_n(scanner, 2, PJ_FALSE);
85
pj_scan_get_until_ch(scanner, '>', &dummy);
86
if (pj_scan_strcmp(scanner, ">", 1) == 0) {
87
pj_scan_advance_n(scanner, 1, PJ_TRUE);
90
pj_scan_advance_n(scanner, 1, PJ_FALSE);
93
return xml_parse_node(pool, scanner);
97
node = alloc_node(pool);
100
pj_scan_get_char(scanner);
103
pj_scan_get_until_chr( scanner, " />\t", &node->name);
105
/* Get attributes. */
106
while (*scanner->curptr != '>' && *scanner->curptr != '/') {
107
pj_xml_attr *attr = alloc_attr(pool);
109
pj_scan_get_until_chr( scanner, "=> \t", &attr->name);
110
if (*scanner->curptr == '=') {
111
pj_scan_get_char( scanner );
112
pj_scan_get_quotes(scanner, "\"'", "\"'", 2, &attr->value);
113
/* remove quote characters */
115
attr->value.slen -= 2;
118
pj_list_push_back( &node->attr_head, attr );
121
if (*scanner->curptr == '/') {
122
pj_scan_get_char(scanner);
123
if (pj_scan_get_char(scanner) != '>')
124
on_syntax_error(scanner);
128
/* Enclosing bracket. */
129
if (pj_scan_get_char(scanner) != '>')
130
on_syntax_error(scanner);
133
while (*scanner->curptr == '<' && *(scanner->curptr+1) != '/') {
134
pj_xml_node *sub_node = xml_parse_node(pool, scanner);
135
pj_list_push_back( &node->node_head, sub_node );
139
if (!pj_scan_is_eof(scanner) && *scanner->curptr != '<') {
140
pj_scan_get_until_ch(scanner, '<', &node->content);
143
/* Enclosing node. */
144
if (pj_scan_get_char(scanner) != '<' || pj_scan_get_char(scanner) != '/')
145
on_syntax_error(scanner);
147
pj_scan_get_until_chr(scanner, " \t>", &end_name);
150
if (pj_stricmp(&node->name, &end_name) != 0)
151
on_syntax_error(scanner);
154
if (pj_scan_get_char(scanner) != '>')
155
on_syntax_error(scanner);
160
PJ_DEF(pj_xml_node*) pj_xml_parse( pj_pool_t *pool, char *msg, pj_size_t len)
162
pj_xml_node *node = NULL;
166
if (!msg || !len || !pool)
169
pj_scan_init( &scanner, msg, len,
170
PJ_SCAN_AUTOSKIP_WS|PJ_SCAN_AUTOSKIP_NEWLINE,
173
node = xml_parse_node(pool, &scanner);
176
PJ_LOG(4,(THIS_FILE, "Syntax error parsing XML in line %d column %d",
177
scanner.line, pj_scan_get_col(&scanner)));
180
pj_scan_fini( &scanner );
184
/* This is a recursive function. */
185
static int xml_print_node( const pj_xml_node *node, int indent,
186
char *buf, pj_size_t len )
191
pj_xml_node *sub_node;
193
#define SIZE_LEFT() ((int)(len - (p-buf)))
198
if (SIZE_LEFT() < node->name.slen + indent + 5)
200
for (i=0; i<indent; ++i)
203
pj_memcpy(p, node->name.ptr, node->name.slen);
204
p += node->name.slen;
206
/* Print attributes. */
207
attr = node->attr_head.next;
208
while (attr != &node->attr_head) {
210
if (SIZE_LEFT() < attr->name.slen + attr->value.slen + 4)
215
/* Attribute name. */
216
pj_memcpy(p, attr->name.ptr, attr->name.slen);
217
p += attr->name.slen;
219
/* Attribute value. */
220
if (attr->value.slen) {
223
pj_memcpy(p, attr->value.ptr, attr->value.slen);
224
p += attr->value.slen;
231
/* Check for empty node. */
232
if (node->content.slen==0 &&
233
node->node_head.next==(pj_xml_node*)&node->node_head)
242
if (SIZE_LEFT() < 1) return -1;
245
/* Print sub nodes. */
246
sub_node = node->node_head.next;
247
while (sub_node != (pj_xml_node*)&node->node_head) {
250
if (SIZE_LEFT() < indent + 3)
255
printed = xml_print_node(sub_node, indent + 1, p, SIZE_LEFT());
260
sub_node = sub_node->next;
264
if (node->content.slen) {
265
if (SIZE_LEFT() < node->content.slen) return -1;
266
pj_memcpy(p, node->content.ptr, node->content.slen);
267
p += node->content.slen;
270
/* Enclosing node. */
271
if (node->node_head.next != (pj_xml_node*)&node->node_head) {
272
if (SIZE_LEFT() < node->name.slen + 5 + indent)
276
for (i=0; i<indent; ++i)
279
if (SIZE_LEFT() < node->name.slen + 3)
284
pj_memcpy(p, node->name.ptr, node->name.slen);
285
p += node->name.slen;
293
PJ_DEF(int) pj_xml_print(const pj_xml_node *node, char *buf, pj_size_t len,
294
pj_bool_t include_prolog)
299
if (!node || !buf || !len)
302
if (include_prolog) {
303
pj_str_t prolog = {"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n", 39};
304
if ((int)len < prolog.slen)
306
pj_memcpy(buf, prolog.ptr, prolog.slen);
307
prolog_len = prolog.slen;
310
printed = xml_print_node(node, 0, buf+prolog_len, len-prolog_len) + prolog_len;
311
if (printed > 0 && len-printed >= 1) {
312
buf[printed++] = '\n';
317
PJ_DEF(pj_xml_node*) pj_xml_node_new(pj_pool_t *pool, const pj_str_t *name)
319
pj_xml_node *node = alloc_node(pool);
320
pj_strdup(pool, &node->name, name);
324
PJ_DEF(pj_xml_attr*) pj_xml_attr_new( pj_pool_t *pool, const pj_str_t *name,
325
const pj_str_t *value)
327
pj_xml_attr *attr = alloc_attr(pool);
328
pj_strdup( pool, &attr->name, name);
329
pj_strdup( pool, &attr->value, value);
333
PJ_DEF(void) pj_xml_add_node( pj_xml_node *parent, pj_xml_node *node )
335
pj_list_push_back(&parent->node_head, node);
338
PJ_DEF(void) pj_xml_add_attr( pj_xml_node *node, pj_xml_attr *attr )
340
pj_list_push_back(&node->attr_head, attr);
343
PJ_DEF(pj_xml_node*) pj_xml_find_node(const pj_xml_node *parent,
344
const pj_str_t *name)
346
const pj_xml_node *node = parent->node_head.next;
350
while (node != (void*)&parent->node_head) {
351
if (pj_stricmp(&node->name, name) == 0)
352
return (pj_xml_node*)node;
358
PJ_DEF(pj_xml_node*) pj_xml_find_node_rec(const pj_xml_node *parent,
359
const pj_str_t *name)
361
const pj_xml_node *node = parent->node_head.next;
365
while (node != (void*)&parent->node_head) {
367
if (pj_stricmp(&node->name, name) == 0)
368
return (pj_xml_node*)node;
369
found = pj_xml_find_node_rec(node, name);
371
return (pj_xml_node*)found;
377
PJ_DEF(pj_xml_node*) pj_xml_find_next_node( const pj_xml_node *parent,
378
const pj_xml_node *node,
379
const pj_str_t *name)
384
while (node != (void*)&parent->node_head) {
385
if (pj_stricmp(&node->name, name) == 0)
386
return (pj_xml_node*)node;
393
PJ_DEF(pj_xml_attr*) pj_xml_find_attr( const pj_xml_node *node,
394
const pj_str_t *name,
395
const pj_str_t *value)
397
const pj_xml_attr *attr = node->attr_head.next;
398
while (attr != (void*)&node->attr_head) {
399
if (pj_stricmp(&attr->name, name)==0) {
401
if (pj_stricmp(&attr->value, value)==0)
402
return (pj_xml_attr*)attr;
404
return (pj_xml_attr*)attr;
414
PJ_DEF(pj_xml_node*) pj_xml_find( const pj_xml_node *parent,
415
const pj_str_t *name,
417
pj_bool_t (*match)(const pj_xml_node *,
420
const pj_xml_node *node = (const pj_xml_node *)parent->node_head.next;
425
while (node != (const pj_xml_node*) &parent->node_head) {
427
if (pj_stricmp(&node->name, name)!=0) {
433
if (match(node, data))
434
return (pj_xml_node*)node;
436
return (pj_xml_node*)node;
444
PJ_DEF(pj_xml_node*) pj_xml_find_rec( const pj_xml_node *parent,
445
const pj_str_t *name,
447
pj_bool_t (*match)(const pj_xml_node*,
450
const pj_xml_node *node = (const pj_xml_node *)parent->node_head.next;
455
while (node != (const pj_xml_node*) &parent->node_head) {
459
if (pj_stricmp(&node->name, name)==0) {
461
if (match(node, data))
462
return (pj_xml_node*)node;
464
return (pj_xml_node*)node;
469
if (match(node, data))
470
return (pj_xml_node*)node;
473
found = pj_xml_find_rec(node, name, data, match);
482
PJ_DEF(pj_xml_node*) pj_xml_clone( pj_pool_t *pool, const pj_xml_node *rhs)
485
const pj_xml_attr *r_attr;
486
const pj_xml_node *child;
488
node = alloc_node(pool);
490
pj_strdup(pool, &node->name, &rhs->name);
491
pj_strdup(pool, &node->content, &rhs->content);
493
/* Clone all attributes */
494
r_attr = rhs->attr_head.next;
495
while (r_attr != &rhs->attr_head) {
499
attr = alloc_attr(pool);
500
pj_strdup(pool, &attr->name, &r_attr->name);
501
pj_strdup(pool, &attr->value, &r_attr->value);
503
pj_list_push_back(&node->attr_head, attr);
505
r_attr = r_attr->next;
508
/* Clone all child nodes. */
509
child = rhs->node_head.next;
510
while (child != (pj_xml_node*) &rhs->node_head) {
511
pj_xml_node *new_child;
513
new_child = pj_xml_clone(pool, child);
514
pj_list_push_back(&node->node_head, new_child);