Commit | Line | Data |
---|---|---|
fd79ecee DJ |
1 | /* Helper routines for parsing XML using Expat. |
2 | ||
e2882c85 | 3 | Copyright (C) 2006-2018 Free Software Foundation, Inc. |
fd79ecee DJ |
4 | |
5 | This file is part of GDB. | |
6 | ||
7 | This program is free software; you can redistribute it and/or modify | |
8 | it under the terms of the GNU General Public License as published by | |
a9762ec7 | 9 | the Free Software Foundation; either version 3 of the License, or |
fd79ecee DJ |
10 | (at your option) any later version. |
11 | ||
12 | This program is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
a9762ec7 | 18 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
fd79ecee DJ |
19 | |
20 | #include "defs.h" | |
e776119f | 21 | #include "gdbcmd.h" |
05a4558a | 22 | #include "xml-support.h" |
614c279d | 23 | #include "filestuff.h" |
05a4558a | 24 | #include "safe-ctype.h" |
a7fc9b61 | 25 | #include <vector> |
bd8a901f | 26 | #include <string> |
e776119f DJ |
27 | |
28 | /* Debugging flag. */ | |
29 | static int debug_xml; | |
fd79ecee DJ |
30 | |
31 | /* The contents of this file are only useful if XML support is | |
32 | available. */ | |
33 | #ifdef HAVE_LIBEXPAT | |
34 | ||
dbc981de | 35 | #include "gdb_expat.h" |
e776119f | 36 | |
108546a0 DJ |
37 | /* The maximum depth of <xi:include> nesting. No need to be miserly, |
38 | we just want to avoid running out of stack on loops. */ | |
39 | #define MAX_XINCLUDE_DEPTH 30 | |
40 | ||
123dc839 DJ |
41 | /* Simplified XML parser infrastructure. */ |
42 | ||
e776119f DJ |
43 | /* A parsing level -- used to keep track of the current element |
44 | nesting. */ | |
45 | struct scope_level | |
46 | { | |
a7fc9b61 PA |
47 | explicit scope_level (const gdb_xml_element *elements_ = NULL) |
48 | : elements (elements_), | |
49 | element (NULL), | |
bd8a901f | 50 | seen (0) |
a7fc9b61 PA |
51 | {} |
52 | ||
e776119f DJ |
53 | /* Elements we allow at this level. */ |
54 | const struct gdb_xml_element *elements; | |
55 | ||
56 | /* The element which we are within. */ | |
57 | const struct gdb_xml_element *element; | |
58 | ||
59 | /* Mask of which elements we've seen at this level (used for | |
60 | optional and repeatable checking). */ | |
61 | unsigned int seen; | |
62 | ||
bd8a901f PA |
63 | /* Body text accumulation. */ |
64 | std::string body; | |
e776119f | 65 | }; |
e776119f DJ |
66 | |
67 | /* The parser itself, and our additional state. */ | |
68 | struct gdb_xml_parser | |
69 | { | |
010151c9 PA |
70 | gdb_xml_parser (const char *name, |
71 | const gdb_xml_element *elements, | |
72 | void *user_data); | |
73 | ~gdb_xml_parser(); | |
74 | ||
4895cde2 PA |
75 | /* Associate DTD_NAME, which must be the name of a compiled-in DTD, |
76 | with the parser. */ | |
77 | void use_dtd (const char *dtd_name); | |
78 | ||
79 | /* Return the name of the expected / default DTD, if specified. */ | |
80 | const char *dtd_name () | |
81 | { return m_dtd_name; } | |
82 | ||
83 | /* Invoke the parser on BUFFER. BUFFER is the data to parse, which | |
84 | should be NUL-terminated. | |
85 | ||
86 | The return value is 0 for success or -1 for error. It may throw, | |
87 | but only if something unexpected goes wrong during parsing; parse | |
88 | errors will be caught, warned about, and reported as failure. */ | |
89 | int parse (const char *buffer); | |
90 | ||
91 | /* Issue a debugging message. */ | |
92 | void vdebug (const char *format, va_list ap) | |
93 | ATTRIBUTE_PRINTF (2, 0); | |
94 | ||
95 | /* Issue an error message, and stop parsing. */ | |
96 | void verror (const char *format, va_list ap) | |
97 | ATTRIBUTE_NORETURN ATTRIBUTE_PRINTF (2, 0); | |
98 | ||
99 | void body_text (const XML_Char *text, int length); | |
100 | void start_element (const XML_Char *name, const XML_Char **attrs); | |
101 | void end_element (const XML_Char *name); | |
102 | ||
103 | /* Return the name of this parser. */ | |
104 | const char *name () | |
105 | { return m_name; } | |
106 | ||
107 | /* Return the user's callback data, for handlers. */ | |
108 | void *user_data () | |
109 | { return m_user_data; }; | |
110 | ||
111 | /* Are we the special <xi:include> parser? */ | |
112 | void set_is_xinclude (bool is_xinclude) | |
113 | { m_is_xinclude = is_xinclude; } | |
114 | ||
115 | /* A thrown error, if any. */ | |
116 | void set_error (gdb_exception error) | |
117 | { | |
118 | m_error = error; | |
119 | #ifdef HAVE_XML_STOPPARSER | |
120 | XML_StopParser (m_expat_parser, XML_FALSE); | |
121 | #endif | |
122 | } | |
123 | ||
124 | /* Return the underlying expat parser. */ | |
125 | XML_Parser expat_parser () | |
126 | { return m_expat_parser; } | |
127 | ||
128 | private: | |
129 | /* The underlying expat parser. */ | |
130 | XML_Parser m_expat_parser; | |
131 | ||
132 | /* Name of this parser. */ | |
133 | const char *m_name; | |
e776119f | 134 | |
4895cde2 PA |
135 | /* The user's callback data, for handlers. */ |
136 | void *m_user_data; | |
e776119f | 137 | |
a7fc9b61 | 138 | /* Scoping stack. */ |
4895cde2 | 139 | std::vector<scope_level> m_scopes; |
e776119f | 140 | |
4895cde2 PA |
141 | /* A thrown error, if any. */ |
142 | struct gdb_exception m_error; | |
108546a0 | 143 | |
4895cde2 PA |
144 | /* The line of the thrown error, or 0. */ |
145 | int m_last_line; | |
146 | ||
147 | /* The name of the expected / default DTD, if specified. */ | |
148 | const char *m_dtd_name; | |
149 | ||
150 | /* Are we the special <xi:include> parser? */ | |
151 | bool m_is_xinclude; | |
e776119f DJ |
152 | }; |
153 | ||
154 | /* Process some body text. We accumulate the text for later use; it's | |
155 | wrong to do anything with it immediately, because a single block of | |
156 | text might be broken up into multiple calls to this function. */ | |
157 | ||
4895cde2 PA |
158 | void |
159 | gdb_xml_parser::body_text (const XML_Char *text, int length) | |
160 | { | |
161 | if (m_error.reason < 0) | |
162 | return; | |
163 | ||
164 | scope_level &scope = m_scopes.back (); | |
165 | scope.body.append (text, length); | |
166 | } | |
167 | ||
e776119f DJ |
168 | static void |
169 | gdb_xml_body_text (void *data, const XML_Char *text, int length) | |
170 | { | |
19ba03f4 | 171 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data; |
e776119f | 172 | |
4895cde2 | 173 | parser->body_text (text, length); |
e776119f | 174 | } |
fd79ecee | 175 | |
e776119f | 176 | /* Issue a debugging message from one of PARSER's handlers. */ |
fd79ecee | 177 | |
e776119f | 178 | void |
4895cde2 | 179 | gdb_xml_parser::vdebug (const char *format, va_list ap) |
e776119f | 180 | { |
4895cde2 | 181 | int line = XML_GetCurrentLineNumber (m_expat_parser); |
e776119f DJ |
182 | char *message; |
183 | ||
e776119f DJ |
184 | message = xstrvprintf (format, ap); |
185 | if (line) | |
186 | fprintf_unfiltered (gdb_stderr, "%s (line %d): %s\n", | |
4895cde2 | 187 | m_name, line, message); |
e776119f DJ |
188 | else |
189 | fprintf_unfiltered (gdb_stderr, "%s: %s\n", | |
4895cde2 | 190 | m_name, message); |
e776119f DJ |
191 | xfree (message); |
192 | } | |
193 | ||
4895cde2 PA |
194 | void |
195 | gdb_xml_debug (struct gdb_xml_parser *parser, const char *format, ...) | |
196 | { | |
197 | if (!debug_xml) | |
198 | return; | |
199 | ||
200 | va_list ap; | |
201 | va_start (ap, format); | |
202 | parser->vdebug (format, ap); | |
203 | va_end (ap); | |
204 | } | |
205 | ||
e776119f DJ |
206 | /* Issue an error message from one of PARSER's handlers, and stop |
207 | parsing. */ | |
208 | ||
4895cde2 PA |
209 | void |
210 | gdb_xml_parser::verror (const char *format, va_list ap) | |
211 | { | |
212 | int line = XML_GetCurrentLineNumber (m_expat_parser); | |
213 | ||
214 | m_last_line = line; | |
215 | throw_verror (XML_PARSE_ERROR, format, ap); | |
216 | } | |
217 | ||
e776119f DJ |
218 | void |
219 | gdb_xml_error (struct gdb_xml_parser *parser, const char *format, ...) | |
220 | { | |
e776119f | 221 | va_list ap; |
e776119f | 222 | va_start (ap, format); |
4895cde2 PA |
223 | parser->verror (format, ap); |
224 | va_end (ap); | |
e776119f DJ |
225 | } |
226 | ||
3d2c1d41 PA |
227 | /* Find the attribute named NAME in the set of parsed attributes |
228 | ATTRIBUTES. Returns NULL if not found. */ | |
229 | ||
230 | struct gdb_xml_value * | |
4d0fdd9b SM |
231 | xml_find_attribute (std::vector<gdb_xml_value> &attributes, |
232 | const char *name) | |
3d2c1d41 | 233 | { |
4d0fdd9b SM |
234 | for (gdb_xml_value &value : attributes) |
235 | if (strcmp (value.name, name) == 0) | |
236 | return &value; | |
3d2c1d41 PA |
237 | |
238 | return NULL; | |
239 | } | |
240 | ||
4895cde2 PA |
241 | /* Handle the start of an element. NAME is the element, and ATTRS are |
242 | the names and values of this element's attributes. */ | |
e776119f | 243 | |
4895cde2 PA |
244 | void |
245 | gdb_xml_parser::start_element (const XML_Char *name, | |
246 | const XML_Char **attrs) | |
fd79ecee | 247 | { |
4895cde2 PA |
248 | if (m_error.reason < 0) |
249 | return; | |
250 | ||
e776119f DJ |
251 | const struct gdb_xml_element *element; |
252 | const struct gdb_xml_attribute *attribute; | |
e776119f | 253 | unsigned int seen; |
e776119f | 254 | |
e776119f DJ |
255 | /* Push an error scope. If we return or throw an exception before |
256 | filling this in, it will tell us to ignore children of this | |
a7fc9b61 PA |
257 | element. Note we don't take a reference to the element yet |
258 | because further below we'll process the element which may recurse | |
259 | back here and push more elements to the vector. When the | |
260 | recursion unrolls all such elements will have been popped back | |
261 | already, but if one of those pushes reallocates the vector, | |
262 | previous element references will be invalidated. */ | |
4895cde2 | 263 | m_scopes.emplace_back (); |
a7fc9b61 PA |
264 | |
265 | /* Get a reference to the current scope. */ | |
4895cde2 | 266 | scope_level &scope = m_scopes[m_scopes.size () - 2]; |
e776119f | 267 | |
4895cde2 | 268 | gdb_xml_debug (this, _("Entering element <%s>"), name); |
e776119f DJ |
269 | |
270 | /* Find this element in the list of the current scope's allowed | |
271 | children. Record that we've seen it. */ | |
272 | ||
273 | seen = 1; | |
a7fc9b61 | 274 | for (element = scope.elements; element && element->name; |
e776119f DJ |
275 | element++, seen <<= 1) |
276 | if (strcmp (element->name, name) == 0) | |
277 | break; | |
278 | ||
279 | if (element == NULL || element->name == NULL) | |
280 | { | |
108546a0 DJ |
281 | /* If we're working on XInclude, <xi:include> can be the child |
282 | of absolutely anything. Copy the previous scope's element | |
283 | list into the new scope even if there was no match. */ | |
4895cde2 | 284 | if (m_is_xinclude) |
108546a0 | 285 | { |
4895cde2 | 286 | XML_DefaultCurrent (m_expat_parser); |
108546a0 | 287 | |
4895cde2 | 288 | scope_level &unknown_scope = m_scopes.back (); |
a7fc9b61 | 289 | unknown_scope.elements = scope.elements; |
108546a0 DJ |
290 | return; |
291 | } | |
292 | ||
4895cde2 | 293 | gdb_xml_debug (this, _("Element <%s> unknown"), name); |
e776119f DJ |
294 | return; |
295 | } | |
296 | ||
a7fc9b61 | 297 | if (!(element->flags & GDB_XML_EF_REPEATABLE) && (seen & scope.seen)) |
4895cde2 | 298 | gdb_xml_error (this, _("Element <%s> only expected once"), name); |
e776119f | 299 | |
a7fc9b61 | 300 | scope.seen |= seen; |
e776119f | 301 | |
4d0fdd9b | 302 | std::vector<gdb_xml_value> attributes; |
108546a0 | 303 | |
e776119f DJ |
304 | for (attribute = element->attributes; |
305 | attribute != NULL && attribute->name != NULL; | |
306 | attribute++) | |
307 | { | |
308 | const char *val = NULL; | |
309 | const XML_Char **p; | |
310 | void *parsed_value; | |
e776119f DJ |
311 | |
312 | for (p = attrs; *p != NULL; p += 2) | |
313 | if (!strcmp (attribute->name, p[0])) | |
314 | { | |
315 | val = p[1]; | |
316 | break; | |
317 | } | |
318 | ||
319 | if (*p != NULL && val == NULL) | |
320 | { | |
4895cde2 | 321 | gdb_xml_debug (this, _("Attribute \"%s\" missing a value"), |
e776119f DJ |
322 | attribute->name); |
323 | continue; | |
324 | } | |
325 | ||
326 | if (*p == NULL && !(attribute->flags & GDB_XML_AF_OPTIONAL)) | |
327 | { | |
4895cde2 | 328 | gdb_xml_error (this, _("Required attribute \"%s\" of " |
e776119f DJ |
329 | "<%s> not specified"), |
330 | attribute->name, element->name); | |
331 | continue; | |
332 | } | |
333 | ||
334 | if (*p == NULL) | |
335 | continue; | |
336 | ||
4895cde2 | 337 | gdb_xml_debug (this, _("Parsing attribute %s=\"%s\""), |
e776119f DJ |
338 | attribute->name, val); |
339 | ||
340 | if (attribute->handler) | |
4895cde2 | 341 | parsed_value = attribute->handler (this, attribute, val); |
e776119f DJ |
342 | else |
343 | parsed_value = xstrdup (val); | |
344 | ||
4d0fdd9b | 345 | attributes.emplace_back (attribute->name, parsed_value); |
e776119f DJ |
346 | } |
347 | ||
348 | /* Check for unrecognized attributes. */ | |
349 | if (debug_xml) | |
fd79ecee | 350 | { |
e776119f DJ |
351 | const XML_Char **p; |
352 | ||
353 | for (p = attrs; *p != NULL; p += 2) | |
354 | { | |
355 | for (attribute = element->attributes; | |
356 | attribute != NULL && attribute->name != NULL; | |
357 | attribute++) | |
358 | if (strcmp (attribute->name, *p) == 0) | |
359 | break; | |
360 | ||
361 | if (attribute == NULL || attribute->name == NULL) | |
4895cde2 | 362 | gdb_xml_debug (this, _("Ignoring unknown attribute %s"), *p); |
e776119f DJ |
363 | } |
364 | } | |
365 | ||
366 | /* Call the element handler if there is one. */ | |
367 | if (element->start_handler) | |
4895cde2 | 368 | element->start_handler (this, element, m_user_data, attributes); |
e776119f | 369 | |
a7fc9b61 PA |
370 | /* Fill in a new scope level. Note that we must delay getting a |
371 | back reference till here because above we might have recursed, | |
372 | which may have reallocated the vector which invalidates | |
373 | iterators/pointers/references. */ | |
4895cde2 | 374 | scope_level &new_scope = m_scopes.back (); |
a7fc9b61 PA |
375 | new_scope.element = element; |
376 | new_scope.elements = element->children; | |
e776119f DJ |
377 | } |
378 | ||
379 | /* Wrapper for gdb_xml_start_element, to prevent throwing exceptions | |
380 | through expat. */ | |
381 | ||
382 | static void | |
383 | gdb_xml_start_element_wrapper (void *data, const XML_Char *name, | |
384 | const XML_Char **attrs) | |
385 | { | |
19ba03f4 | 386 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data; |
e776119f | 387 | |
492d29ea | 388 | TRY |
e776119f | 389 | { |
4895cde2 | 390 | parser->start_element (name, attrs); |
e776119f | 391 | } |
492d29ea | 392 | CATCH (ex, RETURN_MASK_ALL) |
e776119f | 393 | { |
4895cde2 | 394 | parser->set_error (ex); |
fd79ecee | 395 | } |
492d29ea | 396 | END_CATCH |
e776119f DJ |
397 | } |
398 | ||
4895cde2 | 399 | /* Handle the end of an element. NAME is the current element. */ |
e776119f | 400 | |
4895cde2 PA |
401 | void |
402 | gdb_xml_parser::end_element (const XML_Char *name) | |
e776119f | 403 | { |
4895cde2 PA |
404 | if (m_error.reason < 0) |
405 | return; | |
406 | ||
407 | struct scope_level *scope = &m_scopes.back (); | |
e776119f DJ |
408 | const struct gdb_xml_element *element; |
409 | unsigned int seen; | |
e776119f | 410 | |
4895cde2 | 411 | gdb_xml_debug (this, _("Leaving element <%s>"), name); |
e776119f DJ |
412 | |
413 | for (element = scope->elements, seen = 1; | |
414 | element != NULL && element->name != NULL; | |
415 | element++, seen <<= 1) | |
416 | if ((scope->seen & seen) == 0 | |
417 | && (element->flags & GDB_XML_EF_OPTIONAL) == 0) | |
4895cde2 | 418 | gdb_xml_error (this, _("Required element <%s> is missing"), |
e776119f DJ |
419 | element->name); |
420 | ||
581e13c1 | 421 | /* Call the element processor. */ |
108546a0 | 422 | if (scope->element != NULL && scope->element->end_handler) |
e776119f | 423 | { |
bd8a901f | 424 | const char *body; |
e776119f | 425 | |
bd8a901f PA |
426 | if (scope->body.empty ()) |
427 | body = ""; | |
108546a0 DJ |
428 | else |
429 | { | |
430 | int length; | |
e776119f | 431 | |
bd8a901f PA |
432 | length = scope->body.size (); |
433 | body = scope->body.c_str (); | |
e776119f | 434 | |
108546a0 | 435 | /* Strip leading and trailing whitespace. */ |
bd8a901f PA |
436 | while (length > 0 && ISSPACE (body[length - 1])) |
437 | length--; | |
438 | scope->body.erase (length); | |
108546a0 DJ |
439 | while (*body && ISSPACE (*body)) |
440 | body++; | |
441 | } | |
442 | ||
4895cde2 PA |
443 | scope->element->end_handler (this, scope->element, |
444 | m_user_data, body); | |
108546a0 DJ |
445 | } |
446 | else if (scope->element == NULL) | |
4895cde2 | 447 | XML_DefaultCurrent (m_expat_parser); |
e776119f DJ |
448 | |
449 | /* Pop the scope level. */ | |
4895cde2 | 450 | m_scopes.pop_back (); |
e776119f DJ |
451 | } |
452 | ||
453 | /* Wrapper for gdb_xml_end_element, to prevent throwing exceptions | |
454 | through expat. */ | |
455 | ||
456 | static void | |
457 | gdb_xml_end_element_wrapper (void *data, const XML_Char *name) | |
458 | { | |
19ba03f4 | 459 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data; |
e776119f | 460 | |
492d29ea | 461 | TRY |
e776119f | 462 | { |
4895cde2 | 463 | parser->end_element (name); |
e776119f | 464 | } |
492d29ea | 465 | CATCH (ex, RETURN_MASK_ALL) |
e776119f | 466 | { |
4895cde2 | 467 | parser->set_error (ex); |
e776119f | 468 | } |
492d29ea | 469 | END_CATCH |
e776119f DJ |
470 | } |
471 | ||
472 | /* Free a parser and all its associated state. */ | |
473 | ||
010151c9 | 474 | gdb_xml_parser::~gdb_xml_parser () |
e776119f | 475 | { |
4895cde2 | 476 | XML_ParserFree (m_expat_parser); |
e776119f DJ |
477 | } |
478 | ||
010151c9 PA |
479 | /* Initialize a parser. */ |
480 | ||
4895cde2 | 481 | gdb_xml_parser::gdb_xml_parser (const char *name, |
010151c9 | 482 | const gdb_xml_element *elements, |
4895cde2 PA |
483 | void *user_data) |
484 | : m_name (name), | |
485 | m_user_data (user_data), | |
486 | m_error (exception_none), | |
487 | m_last_line (0), | |
488 | m_dtd_name (NULL), | |
489 | m_is_xinclude (false) | |
490 | { | |
491 | m_expat_parser = XML_ParserCreateNS (NULL, '!'); | |
492 | if (m_expat_parser == NULL) | |
010151c9 | 493 | malloc_failure (0); |
e776119f | 494 | |
4895cde2 | 495 | XML_SetUserData (m_expat_parser, this); |
e776119f DJ |
496 | |
497 | /* Set the callbacks. */ | |
4895cde2 | 498 | XML_SetElementHandler (m_expat_parser, gdb_xml_start_element_wrapper, |
e776119f | 499 | gdb_xml_end_element_wrapper); |
4895cde2 | 500 | XML_SetCharacterDataHandler (m_expat_parser, gdb_xml_body_text); |
e776119f DJ |
501 | |
502 | /* Initialize the outer scope. */ | |
4895cde2 | 503 | m_scopes.emplace_back (elements); |
efc0eabd PA |
504 | } |
505 | ||
108546a0 DJ |
506 | /* External entity handler. The only external entities we support |
507 | are those compiled into GDB (we do not fetch entities from the | |
508 | target). */ | |
509 | ||
510 | static int XMLCALL | |
511 | gdb_xml_fetch_external_entity (XML_Parser expat_parser, | |
512 | const XML_Char *context, | |
513 | const XML_Char *base, | |
514 | const XML_Char *systemId, | |
515 | const XML_Char *publicId) | |
516 | { | |
108546a0 DJ |
517 | XML_Parser entity_parser; |
518 | const char *text; | |
519 | enum XML_Status status; | |
520 | ||
521 | if (systemId == NULL) | |
522 | { | |
4895cde2 PA |
523 | gdb_xml_parser *parser |
524 | = (gdb_xml_parser *) XML_GetUserData (expat_parser); | |
525 | ||
526 | text = fetch_xml_builtin (parser->dtd_name ()); | |
108546a0 | 527 | if (text == NULL) |
9b20d036 MS |
528 | internal_error (__FILE__, __LINE__, |
529 | _("could not locate built-in DTD %s"), | |
4895cde2 | 530 | parser->dtd_name ()); |
108546a0 DJ |
531 | } |
532 | else | |
533 | { | |
534 | text = fetch_xml_builtin (systemId); | |
535 | if (text == NULL) | |
536 | return XML_STATUS_ERROR; | |
537 | } | |
538 | ||
4895cde2 PA |
539 | entity_parser = XML_ExternalEntityParserCreate (expat_parser, |
540 | context, NULL); | |
108546a0 DJ |
541 | |
542 | /* Don't use our handlers for the contents of the DTD. Just let expat | |
543 | process it. */ | |
544 | XML_SetElementHandler (entity_parser, NULL, NULL); | |
545 | XML_SetDoctypeDeclHandler (entity_parser, NULL, NULL); | |
546 | XML_SetXmlDeclHandler (entity_parser, NULL); | |
547 | XML_SetDefaultHandler (entity_parser, NULL); | |
548 | XML_SetUserData (entity_parser, NULL); | |
549 | ||
550 | status = XML_Parse (entity_parser, text, strlen (text), 1); | |
551 | ||
552 | XML_ParserFree (entity_parser); | |
553 | return status; | |
554 | } | |
555 | ||
108546a0 | 556 | void |
4895cde2 | 557 | gdb_xml_parser::use_dtd (const char *dtd_name) |
108546a0 DJ |
558 | { |
559 | enum XML_Error err; | |
560 | ||
4895cde2 | 561 | m_dtd_name = dtd_name; |
108546a0 | 562 | |
4895cde2 | 563 | XML_SetParamEntityParsing (m_expat_parser, |
108546a0 | 564 | XML_PARAM_ENTITY_PARSING_UNLESS_STANDALONE); |
4895cde2 | 565 | XML_SetExternalEntityRefHandler (m_expat_parser, |
108546a0 DJ |
566 | gdb_xml_fetch_external_entity); |
567 | ||
568 | /* Even if no DTD is provided, use the built-in DTD anyway. */ | |
4895cde2 | 569 | err = XML_UseForeignDTD (m_expat_parser, XML_TRUE); |
108546a0 DJ |
570 | if (err != XML_ERROR_NONE) |
571 | internal_error (__FILE__, __LINE__, | |
9b20d036 MS |
572 | _("XML_UseForeignDTD failed: %s"), |
573 | XML_ErrorString (err)); | |
108546a0 DJ |
574 | } |
575 | ||
e776119f DJ |
576 | /* Invoke PARSER on BUFFER. BUFFER is the data to parse, which |
577 | should be NUL-terminated. | |
578 | ||
579 | The return value is 0 for success or -1 for error. It may throw, | |
580 | but only if something unexpected goes wrong during parsing; parse | |
581 | errors will be caught, warned about, and reported as failure. */ | |
582 | ||
583 | int | |
4895cde2 | 584 | gdb_xml_parser::parse (const char *buffer) |
e776119f DJ |
585 | { |
586 | enum XML_Status status; | |
587 | const char *error_string; | |
588 | ||
4895cde2 | 589 | gdb_xml_debug (this, _("Starting:\n%s"), buffer); |
de584861 | 590 | |
4895cde2 | 591 | status = XML_Parse (m_expat_parser, buffer, strlen (buffer), 1); |
e776119f | 592 | |
4895cde2 | 593 | if (status == XML_STATUS_OK && m_error.reason == 0) |
e776119f DJ |
594 | return 0; |
595 | ||
4895cde2 PA |
596 | if (m_error.reason == RETURN_ERROR |
597 | && m_error.error == XML_PARSE_ERROR) | |
e776119f | 598 | { |
4895cde2 PA |
599 | gdb_assert (m_error.message != NULL); |
600 | error_string = m_error.message; | |
e776119f DJ |
601 | } |
602 | else if (status == XML_STATUS_ERROR) | |
603 | { | |
4895cde2 | 604 | enum XML_Error err = XML_GetErrorCode (m_expat_parser); |
a109c7c1 | 605 | |
e776119f DJ |
606 | error_string = XML_ErrorString (err); |
607 | } | |
608 | else | |
609 | { | |
4895cde2 PA |
610 | gdb_assert (m_error.reason < 0); |
611 | throw_exception (m_error); | |
e776119f DJ |
612 | } |
613 | ||
4895cde2 PA |
614 | if (m_last_line != 0) |
615 | warning (_("while parsing %s (at line %d): %s"), m_name, | |
616 | m_last_line, error_string); | |
e776119f | 617 | else |
4895cde2 | 618 | warning (_("while parsing %s: %s"), m_name, error_string); |
e776119f DJ |
619 | |
620 | return -1; | |
fd79ecee DJ |
621 | } |
622 | ||
efc0eabd PA |
623 | int |
624 | gdb_xml_parse_quick (const char *name, const char *dtd_name, | |
625 | const struct gdb_xml_element *elements, | |
626 | const char *document, void *user_data) | |
627 | { | |
010151c9 | 628 | gdb_xml_parser parser (name, elements, user_data); |
efc0eabd | 629 | if (dtd_name != NULL) |
4895cde2 PA |
630 | parser.use_dtd (dtd_name); |
631 | return parser.parse (document); | |
efc0eabd PA |
632 | } |
633 | ||
fd79ecee DJ |
634 | /* Parse a field VALSTR that we expect to contain an integer value. |
635 | The integer is returned in *VALP. The string is parsed with an | |
636 | equivalent to strtoul. | |
637 | ||
638 | Returns 0 for success, -1 for error. */ | |
639 | ||
640 | static int | |
641 | xml_parse_unsigned_integer (const char *valstr, ULONGEST *valp) | |
642 | { | |
643 | const char *endptr; | |
644 | ULONGEST result; | |
645 | ||
646 | if (*valstr == '\0') | |
647 | return -1; | |
648 | ||
649 | result = strtoulst (valstr, &endptr, 0); | |
650 | if (*endptr != '\0') | |
651 | return -1; | |
652 | ||
653 | *valp = result; | |
654 | return 0; | |
655 | } | |
656 | ||
e776119f DJ |
657 | /* Parse an integer string into a ULONGEST and return it, or call |
658 | gdb_xml_error if it could not be parsed. */ | |
fd79ecee DJ |
659 | |
660 | ULONGEST | |
e776119f | 661 | gdb_xml_parse_ulongest (struct gdb_xml_parser *parser, const char *value) |
fd79ecee DJ |
662 | { |
663 | ULONGEST result; | |
fd79ecee DJ |
664 | |
665 | if (xml_parse_unsigned_integer (value, &result) != 0) | |
e776119f DJ |
666 | gdb_xml_error (parser, _("Can't convert \"%s\" to an integer"), value); |
667 | ||
fd79ecee DJ |
668 | return result; |
669 | } | |
670 | ||
e776119f DJ |
671 | /* Parse an integer attribute into a ULONGEST. */ |
672 | ||
673 | void * | |
674 | gdb_xml_parse_attr_ulongest (struct gdb_xml_parser *parser, | |
675 | const struct gdb_xml_attribute *attribute, | |
676 | const char *value) | |
677 | { | |
678 | ULONGEST result; | |
679 | void *ret; | |
fd79ecee | 680 | |
e776119f DJ |
681 | if (xml_parse_unsigned_integer (value, &result) != 0) |
682 | gdb_xml_error (parser, _("Can't convert %s=\"%s\" to an integer"), | |
683 | attribute->name, value); | |
fd79ecee | 684 | |
8d749320 | 685 | ret = XNEW (ULONGEST); |
e776119f DJ |
686 | memcpy (ret, &result, sizeof (result)); |
687 | return ret; | |
688 | } | |
689 | ||
123dc839 DJ |
690 | /* A handler_data for yes/no boolean values. */ |
691 | ||
692 | const struct gdb_xml_enum gdb_xml_enums_boolean[] = { | |
693 | { "yes", 1 }, | |
694 | { "no", 0 }, | |
695 | { NULL, 0 } | |
696 | }; | |
697 | ||
e776119f DJ |
698 | /* Map NAME to VALUE. A struct gdb_xml_enum * should be saved as the |
699 | value of handler_data when using gdb_xml_parse_attr_enum to parse a | |
700 | fixed list of possible strings. The list is terminated by an entry | |
701 | with NAME == NULL. */ | |
702 | ||
703 | void * | |
704 | gdb_xml_parse_attr_enum (struct gdb_xml_parser *parser, | |
705 | const struct gdb_xml_attribute *attribute, | |
706 | const char *value) | |
fd79ecee | 707 | { |
19ba03f4 SM |
708 | const struct gdb_xml_enum *enums |
709 | = (const struct gdb_xml_enum *) attribute->handler_data; | |
e776119f | 710 | void *ret; |
fd79ecee | 711 | |
19ba03f4 SM |
712 | for (enums = (const struct gdb_xml_enum *) attribute->handler_data; |
713 | enums->name != NULL; enums++) | |
123dc839 | 714 | if (strcasecmp (enums->name, value) == 0) |
e776119f DJ |
715 | break; |
716 | ||
717 | if (enums->name == NULL) | |
718 | gdb_xml_error (parser, _("Unknown attribute value %s=\"%s\""), | |
719 | attribute->name, value); | |
720 | ||
721 | ret = xmalloc (sizeof (enums->value)); | |
722 | memcpy (ret, &enums->value, sizeof (enums->value)); | |
723 | return ret; | |
fd79ecee | 724 | } |
108546a0 DJ |
725 | \f |
726 | ||
727 | /* XInclude processing. This is done as a separate step from actually | |
728 | parsing the document, so that we can produce a single combined XML | |
729 | document - e.g. to hand to a front end or to simplify comparing two | |
730 | documents. We make extensive use of XML_DefaultCurrent, to pass | |
731 | input text directly into the output without reformatting or | |
732 | requoting it. | |
733 | ||
734 | We output the DOCTYPE declaration for the first document unchanged, | |
735 | if present, and discard DOCTYPEs from included documents. Only the | |
736 | one we pass through here is used when we feed the result back to | |
737 | expat. The XInclude standard explicitly does not discuss | |
738 | validation of the result; we choose to apply the same DTD applied | |
739 | to the outermost document. | |
740 | ||
741 | We can not simply include the external DTD subset in the document | |
742 | as an internal subset, because <!IGNORE> and <!INCLUDE> are valid | |
743 | only in external subsets. But if we do not pass the DTD into the | |
744 | output at all, default values will not be filled in. | |
745 | ||
746 | We don't pass through any <?xml> declaration because we generate | |
747 | UTF-8, not whatever the input encoding was. */ | |
748 | ||
749 | struct xinclude_parsing_data | |
750 | { | |
bd8a901f PA |
751 | xinclude_parsing_data (std::string &output_, |
752 | xml_fetch_another fetcher_, void *fetcher_baton_, | |
010151c9 | 753 | int include_depth_) |
bd8a901f PA |
754 | : output (output_), |
755 | skip_depth (0), | |
010151c9 PA |
756 | include_depth (include_depth_), |
757 | fetcher (fetcher_), | |
758 | fetcher_baton (fetcher_baton_) | |
bd8a901f | 759 | {} |
010151c9 | 760 | |
bd8a901f PA |
761 | /* Where the output goes. */ |
762 | std::string &output; | |
108546a0 DJ |
763 | |
764 | /* A count indicating whether we are in an element whose | |
765 | children should not be copied to the output, and if so, | |
766 | how deep we are nested. This is used for anything inside | |
767 | an xi:include, and for the DTD. */ | |
768 | int skip_depth; | |
769 | ||
770 | /* The number of <xi:include> elements currently being processed, | |
771 | to detect loops. */ | |
772 | int include_depth; | |
773 | ||
774 | /* A function to call to obtain additional features, and its | |
775 | baton. */ | |
776 | xml_fetch_another fetcher; | |
777 | void *fetcher_baton; | |
778 | }; | |
779 | ||
780 | static void | |
781 | xinclude_start_include (struct gdb_xml_parser *parser, | |
782 | const struct gdb_xml_element *element, | |
4d0fdd9b SM |
783 | void *user_data, |
784 | std::vector<gdb_xml_value> &attributes) | |
108546a0 | 785 | { |
19ba03f4 SM |
786 | struct xinclude_parsing_data *data |
787 | = (struct xinclude_parsing_data *) user_data; | |
4d0fdd9b | 788 | char *href = (char *) xml_find_attribute (attributes, "href")->value.get (); |
108546a0 DJ |
789 | |
790 | gdb_xml_debug (parser, _("Processing XInclude of \"%s\""), href); | |
791 | ||
792 | if (data->include_depth > MAX_XINCLUDE_DEPTH) | |
793 | gdb_xml_error (parser, _("Maximum XInclude depth (%d) exceeded"), | |
794 | MAX_XINCLUDE_DEPTH); | |
795 | ||
9018be22 SM |
796 | gdb::optional<gdb::char_vector> text |
797 | = data->fetcher (href, data->fetcher_baton); | |
798 | if (!text) | |
108546a0 | 799 | gdb_xml_error (parser, _("Could not load XML document \"%s\""), href); |
108546a0 | 800 | |
4895cde2 | 801 | if (!xml_process_xincludes (data->output, parser->name (), |
9018be22 | 802 | text->data (), data->fetcher, |
bd8a901f PA |
803 | data->fetcher_baton, |
804 | data->include_depth + 1)) | |
108546a0 DJ |
805 | gdb_xml_error (parser, _("Parsing \"%s\" failed"), href); |
806 | ||
108546a0 DJ |
807 | data->skip_depth++; |
808 | } | |
809 | ||
810 | static void | |
811 | xinclude_end_include (struct gdb_xml_parser *parser, | |
812 | const struct gdb_xml_element *element, | |
813 | void *user_data, const char *body_text) | |
814 | { | |
19ba03f4 SM |
815 | struct xinclude_parsing_data *data |
816 | = (struct xinclude_parsing_data *) user_data; | |
108546a0 DJ |
817 | |
818 | data->skip_depth--; | |
819 | } | |
820 | ||
821 | static void XMLCALL | |
822 | xml_xinclude_default (void *data_, const XML_Char *s, int len) | |
823 | { | |
19ba03f4 | 824 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data_; |
4895cde2 | 825 | xinclude_parsing_data *data = (xinclude_parsing_data *) parser->user_data (); |
108546a0 DJ |
826 | |
827 | /* If we are inside of e.g. xi:include or the DTD, don't save this | |
828 | string. */ | |
829 | if (data->skip_depth) | |
830 | return; | |
831 | ||
832 | /* Otherwise just add it to the end of the document we're building | |
833 | up. */ | |
bd8a901f | 834 | data->output.append (s, len); |
108546a0 DJ |
835 | } |
836 | ||
837 | static void XMLCALL | |
838 | xml_xinclude_start_doctype (void *data_, const XML_Char *doctypeName, | |
839 | const XML_Char *sysid, const XML_Char *pubid, | |
840 | int has_internal_subset) | |
841 | { | |
19ba03f4 | 842 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data_; |
4895cde2 | 843 | xinclude_parsing_data *data = (xinclude_parsing_data *) parser->user_data (); |
108546a0 DJ |
844 | |
845 | /* Don't print out the doctype, or the contents of the DTD internal | |
846 | subset, if any. */ | |
847 | data->skip_depth++; | |
848 | } | |
849 | ||
850 | static void XMLCALL | |
851 | xml_xinclude_end_doctype (void *data_) | |
852 | { | |
19ba03f4 | 853 | struct gdb_xml_parser *parser = (struct gdb_xml_parser *) data_; |
4895cde2 | 854 | xinclude_parsing_data *data = (xinclude_parsing_data *) parser->user_data (); |
108546a0 DJ |
855 | |
856 | data->skip_depth--; | |
857 | } | |
858 | ||
859 | static void XMLCALL | |
860 | xml_xinclude_xml_decl (void *data_, const XML_Char *version, | |
861 | const XML_Char *encoding, int standalone) | |
862 | { | |
863 | /* Do nothing - this function prevents the default handler from | |
864 | being called, thus suppressing the XML declaration from the | |
865 | output. */ | |
866 | } | |
867 | ||
108546a0 DJ |
868 | const struct gdb_xml_attribute xinclude_attributes[] = { |
869 | { "href", GDB_XML_AF_NONE, NULL, NULL }, | |
870 | { NULL, GDB_XML_AF_NONE, NULL, NULL } | |
871 | }; | |
872 | ||
873 | const struct gdb_xml_element xinclude_elements[] = { | |
874 | { "http://www.w3.org/2001/XInclude!include", xinclude_attributes, NULL, | |
875 | GDB_XML_EF_OPTIONAL | GDB_XML_EF_REPEATABLE, | |
876 | xinclude_start_include, xinclude_end_include }, | |
877 | { NULL, NULL, NULL, GDB_XML_EF_NONE, NULL, NULL } | |
878 | }; | |
879 | ||
880 | /* The main entry point for <xi:include> processing. */ | |
881 | ||
bd8a901f PA |
882 | bool |
883 | xml_process_xincludes (std::string &result, | |
884 | const char *name, const char *text, | |
108546a0 DJ |
885 | xml_fetch_another fetcher, void *fetcher_baton, |
886 | int depth) | |
887 | { | |
bd8a901f | 888 | xinclude_parsing_data data (result, fetcher, fetcher_baton, depth); |
108546a0 | 889 | |
010151c9 | 890 | gdb_xml_parser parser (name, xinclude_elements, &data); |
4895cde2 | 891 | parser.set_is_xinclude (true); |
108546a0 | 892 | |
4895cde2 PA |
893 | XML_SetCharacterDataHandler (parser.expat_parser (), NULL); |
894 | XML_SetDefaultHandler (parser.expat_parser (), xml_xinclude_default); | |
108546a0 DJ |
895 | |
896 | /* Always discard the XML version declarations; the only important | |
897 | thing this provides is encoding, and our result will have been | |
898 | converted to UTF-8. */ | |
4895cde2 | 899 | XML_SetXmlDeclHandler (parser.expat_parser (), xml_xinclude_xml_decl); |
108546a0 DJ |
900 | |
901 | if (depth > 0) | |
902 | /* Discard the doctype for included documents. */ | |
4895cde2 | 903 | XML_SetDoctypeDeclHandler (parser.expat_parser (), |
108546a0 DJ |
904 | xml_xinclude_start_doctype, |
905 | xml_xinclude_end_doctype); | |
906 | ||
4895cde2 | 907 | parser.use_dtd ("xinclude.dtd"); |
108546a0 | 908 | |
4895cde2 | 909 | if (parser.parse (text) == 0) |
108546a0 | 910 | { |
108546a0 | 911 | if (depth == 0) |
010151c9 | 912 | gdb_xml_debug (&parser, _("XInclude processing succeeded.")); |
bd8a901f | 913 | return true; |
108546a0 | 914 | } |
108546a0 | 915 | |
bd8a901f | 916 | return false; |
108546a0 | 917 | } |
05a4558a | 918 | #endif /* HAVE_LIBEXPAT */ |
108546a0 DJ |
919 | \f |
920 | ||
921 | /* Return an XML document which was compiled into GDB, from | |
922 | the given FILENAME, or NULL if the file was not compiled in. */ | |
923 | ||
924 | const char * | |
925 | fetch_xml_builtin (const char *filename) | |
926 | { | |
927 | const char *(*p)[2]; | |
928 | ||
929 | for (p = xml_builtin; (*p)[0]; p++) | |
930 | if (strcmp ((*p)[0], filename) == 0) | |
931 | return (*p)[1]; | |
932 | ||
933 | return NULL; | |
934 | } | |
fd79ecee | 935 | |
05a4558a DJ |
936 | /* A to_xfer_partial helper function which reads XML files which were |
937 | compiled into GDB. The target may call this function from its own | |
938 | to_xfer_partial handler, after converting object and annex to the | |
939 | appropriate filename. */ | |
940 | ||
941 | LONGEST | |
942 | xml_builtin_xfer_partial (const char *filename, | |
943 | gdb_byte *readbuf, const gdb_byte *writebuf, | |
944 | ULONGEST offset, LONGEST len) | |
945 | { | |
946 | const char *buf; | |
947 | LONGEST len_avail; | |
948 | ||
949 | gdb_assert (readbuf != NULL && writebuf == NULL); | |
950 | gdb_assert (filename != NULL); | |
951 | ||
952 | buf = fetch_xml_builtin (filename); | |
953 | if (buf == NULL) | |
954 | return -1; | |
955 | ||
956 | len_avail = strlen (buf); | |
957 | if (offset >= len_avail) | |
958 | return 0; | |
959 | ||
960 | if (len > len_avail - offset) | |
961 | len = len_avail - offset; | |
962 | memcpy (readbuf, buf + offset, len); | |
963 | return len; | |
964 | } | |
965 | \f | |
fd79ecee DJ |
966 | |
967 | static void | |
e776119f DJ |
968 | show_debug_xml (struct ui_file *file, int from_tty, |
969 | struct cmd_list_element *c, const char *value) | |
fd79ecee | 970 | { |
e776119f | 971 | fprintf_filtered (file, _("XML debugging is %s.\n"), value); |
fd79ecee DJ |
972 | } |
973 | ||
9018be22 | 974 | gdb::optional<gdb::char_vector> |
a96d9b2e SDJ |
975 | xml_fetch_content_from_file (const char *filename, void *baton) |
976 | { | |
19ba03f4 | 977 | const char *dirname = (const char *) baton; |
d419f42d | 978 | gdb_file_up file; |
a96d9b2e SDJ |
979 | |
980 | if (dirname && *dirname) | |
981 | { | |
982 | char *fullname = concat (dirname, "/", filename, (char *) NULL); | |
a109c7c1 | 983 | |
a96d9b2e | 984 | if (fullname == NULL) |
d26e3629 | 985 | malloc_failure (0); |
614c279d | 986 | file = gdb_fopen_cloexec (fullname, FOPEN_RT); |
a96d9b2e SDJ |
987 | xfree (fullname); |
988 | } | |
989 | else | |
614c279d | 990 | file = gdb_fopen_cloexec (filename, FOPEN_RT); |
a96d9b2e SDJ |
991 | |
992 | if (file == NULL) | |
9018be22 | 993 | return {}; |
a96d9b2e | 994 | |
2edf834e | 995 | /* Read in the whole file. */ |
a96d9b2e | 996 | |
2edf834e | 997 | size_t len; |
a96d9b2e | 998 | |
2edf834e PA |
999 | if (fseek (file.get (), 0, SEEK_END) == -1) |
1000 | perror_with_name (_("seek to end of file")); | |
1001 | len = ftell (file.get ()); | |
1002 | rewind (file.get ()); | |
a96d9b2e | 1003 | |
9018be22 | 1004 | gdb::char_vector text (len + 1); |
a96d9b2e | 1005 | |
9018be22 | 1006 | if (fread (text.data (), 1, len, file.get ()) != len |
63929e84 | 1007 | || ferror (file.get ())) |
2edf834e PA |
1008 | { |
1009 | warning (_("Read error from \"%s\""), filename); | |
9018be22 | 1010 | return {}; |
a96d9b2e SDJ |
1011 | } |
1012 | ||
9018be22 | 1013 | text.back () = '\0'; |
a96d9b2e SDJ |
1014 | return text; |
1015 | } | |
1016 | ||
fd79ecee | 1017 | void |
e776119f | 1018 | _initialize_xml_support (void) |
fd79ecee | 1019 | { |
e776119f DJ |
1020 | add_setshow_boolean_cmd ("xml", class_maintenance, &debug_xml, |
1021 | _("Set XML parser debugging."), | |
1022 | _("Show XML parser debugging."), | |
1023 | _("When set, debugging messages for XML parsers " | |
1024 | "are displayed."), | |
1025 | NULL, show_debug_xml, | |
1026 | &setdebuglist, &showdebuglist); | |
fd79ecee | 1027 | } |