eglib/src/gmarkup.c

   1 /*
   2  * gmakrup.c: Minimal XML markup reader.
   3  *
   4  * Unlike the GLib one, this can not be restarted with more text
   5  * as the Mono use does not require it.
   6  *
   7  * Actually, with further thought, I think that this could be made
   8  * to restart very easily.  The pos == end condition would mean
   9  * "return to caller" and only at end parse this would be a fatal
  10  * error.
  11  *
  12  * Not that it matters to Mono, but it is very simple to change, there
  13  * is a tricky situation: there are a few places where we check p+n
  14  * in the source, and that would have to change to be progressive, instead
  15  * of depending on the string to be complete at that point, so we would
  16  * have to introduce extra states to cope with that.
  17  *
  18  * Author:
  19  *   Miguel de Icaza (miguel@novell.com)
  20  *
  21  * (C) 2006 Novell, Inc.
  22  *
  23  * Permission is hereby granted, free of charge, to any person obtaining
  24  * a copy of this software and associated documentation files (the
  25  * "Software"), to deal in the Software without restriction, including
  26  * without limitation the rights to use, copy, modify, merge, publish,
  27  * distribute, sublicense, and/or sell copies of the Software, and to
  28  * permit persons to whom the Software is furnished to do so, subject to
  29  * the following conditions:
  30  *
  31  * The above copyright notice and this permission notice shall be
  32  * included in all copies or substantial portions of the Software.
  33  *
  34  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  35  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  36  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  37  * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
  38  * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
  39  * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
  40  * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  41  */
  42 #include <stdio.h>
  43 #include <ctype.h>
  44 #include <glib.h>
  45
  46 #define set_error(msg, ...) do { if (error != NULL) *error = g_error_new (GINT_TO_POINTER (1), 1, msg, __VA_ARGS__); } while (0);
  47
  48 typedef enum {
  49         START,
  50         START_ELEMENT,
  51         TEXT,
  52         FLUSH_TEXT,
  53         CLOSING_ELEMENT,
  54         COMMENT,
  55         SKIP_XML_DECLARATION
  56 } ParseState;
  57
  58 struct _GMarkupParseContext {
  59         GMarkupParser  parser;
  60         gpointer       user_data;
  61         GDestroyNotify user_data_dnotify;
  62         ParseState     state;
  63
  64         /* Stores the name of the current element, so we can issue the end_element */
  65         GSList         *level;
  66
  67         GString        *text;
  68 };
  69
  70 GMarkupParseContext *
  71 g_markup_parse_context_new (const GMarkupParser *parser,
  72                             GMarkupParseFlags flags,
  73                             gpointer user_data,
  74                             GDestroyNotify user_data_dnotify)
  75 {
  76         GMarkupParseContext *context = g_new0 (GMarkupParseContext, 1);
  77
  78         context->parser = *parser;
  79         context->user_data = user_data;
  80         context->user_data_dnotify = user_data_dnotify;
  81
  82         return context;
  83 }
  84
  85 void
  86 g_markup_parse_context_free (GMarkupParseContext *context)
  87 {
  88         GSList *l;
  89
  90         g_return_if_fail (context != NULL);
  91
  92         if (context->user_data_dnotify != NULL)
  93                 (context->user_data_dnotify) (context->user_data);
  94
  95         if (context->text != NULL)
  96                 g_string_free (context->text, TRUE);
  97         for (l = context->level; l; l = l->next)
  98                 g_free (l->data);
  99         g_slist_free (context->level);
 100         g_free (context);
 101 }
 102
 103 static gboolean
 104 my_isspace (char c)
 105 {
 106         if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\v')
 107                 return TRUE;
 108         return FALSE;
 109 }
 110
 111 static gboolean
 112 my_isalnum (char c)
 113 {
 114         if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'))
 115                 return TRUE;
 116         if (c >= '0' && c <= '9')
 117                 return TRUE;
 118
 119         return FALSE;
 120 }
 121
 122 static gboolean
 123 my_isalpha (char c)
 124 {
 125         if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'))
 126                 return TRUE;
 127         return FALSE;
 128 }
 129
 130 static const char *
 131 skip_space (const char *p, const char *end)
 132 {
 133         for (; p < end && my_isspace (*p); p++)
 134                 ;
 135         return p;
 136 }
 137
 138 static const char *
 139 parse_value (const char *p, const char *end, char **value, GError **error)
 140 {
 141         const char *start;
 142         int l;
 143
 144         if (*p != '"'){
 145                 set_error ("%s", "Expected the attribute value to start with a quote");
 146                 return end;
 147         }
 148         start = ++p;
 149         for (; p < end && *p != '"'; p++)
 150                 ;
 151         if (p == end)
 152                 return end;
 153         l = (int)(p - start);
 154         p++;
 155         *value = g_malloc (l + 1);
 156         if (*value == NULL)
 157                 return end;
 158         strncpy (*value, start, l);
 159         (*value) [l] = 0;
 160         return p;
 161 }
 162
 163 static const char *
 164 parse_name (const char *p, const char *end, char **value)
 165 {
 166         const char *start = p;
 167         int l;
 168
 169         for (; p < end && my_isalnum (*p); p++)
 170                 ;
 171         if (p == end)
 172                 return end;
 173
 174         l = (int)(p - start);
 175         *value = g_malloc (l + 1);
 176         if (*value == NULL)
 177                 return end;
 178         strncpy (*value, start, l);
 179         (*value) [l] = 0;
 180         return p;
 181 }
 182
 183 static const char *
 184 parse_attributes (const char *p, const char *end, char ***names, char ***values, GError **error, int *full_stop, int state)
 185 {
 186         int nnames = 0;
 187
 188         while (TRUE){
 189                 p = skip_space (p, end);
 190                 if (p == end)
 191                         return end;
 192
 193                 if (*p == '>'){
 194                         *full_stop = 0;
 195                         return p;
 196                 }
 197                 if (state == SKIP_XML_DECLARATION && *p == '?' && ((p+1) < end) && *(p+1) == '>'){
 198                         *full_stop = 0;
 199                         return p+1;
 200                 }
 201
 202                 if (*p == '/' && ((p+1) < end && *(p+1) == '>')){
 203                         *full_stop = 1;
 204                         return p+1;
 205                 } else {
 206                         char *name, *value;
 207
 208                         p = parse_name (p, end, &name);
 209                         if (p == end)
 210                                 return p;
 211
 212                         p = skip_space (p, end);
 213                         if (p == end){
 214                                 g_free (name);
 215                                 return p;
 216                         }
 217                         if (*p != '='){
 218                                 set_error ("Expected an = after the attribute name `%s'", name);
 219                                 g_free (name);
 220                                 return end;
 221                         }
 222                         p++;
 223                         p = skip_space (p, end);
 224                         if (p == end){
 225                                 g_free (name);
 226                                 return end;
 227                         }
 228
 229                         p = parse_value (p, end, &value, error);
 230                         if (p == end){
 231                                 g_free (name);
 232                                 return p;
 233                         }
 234
 235                         ++nnames;
 236                         *names = g_realloc (*names, sizeof (char **) * (nnames+1));
 237                         *values = g_realloc (*values, sizeof (char **) * (nnames+1));
 238                         (*names) [nnames-1] = name;
 239                         (*values) [nnames-1] = value;
 240                         (*names) [nnames] = NULL;
 241                         (*values) [nnames] = NULL;
 242                 }
 243         }
 244 }
 245
 246 static void
 247 destroy_parse_state (GMarkupParseContext *context)
 248 {
 249         GSList *p;
 250
 251         for (p = context->level; p != NULL; p = p->next)
 252                 g_free (p->data);
 253
 254         g_slist_free (context->level);
 255         if (context->text != NULL)
 256                 g_string_free (context->text, TRUE);
 257         context->text = NULL;
 258         context->level = NULL;
 259 }
 260
 261 gboolean
 262 g_markup_parse_context_parse (GMarkupParseContext *context,
 263                               const gchar *text, gssize text_len,
 264                               GError **error)
 265 {
 266         const char *p,  *end;
 267
 268         g_return_val_if_fail (context != NULL, FALSE);
 269         g_return_val_if_fail (text != NULL, FALSE);
 270         g_return_val_if_fail (text_len >= 0, FALSE);
 271
 272         end = text + text_len;
 273
 274         for (p = text; p < end; p++){
 275                 char c = *p;
 276
 277                 switch (context->state){
 278                 case START:
 279                         if (c == ' ' || c == '\t' || c == '\f' || c == '\n')
 280                                 continue;
 281                         if (c == '<'){
 282                                 if (p+1 < end && p [1] == '?'){
 283                                         context->state = SKIP_XML_DECLARATION;
 284                                         p++;
 285                                 } else
 286                                         context->state = START_ELEMENT;
 287                                 continue;
 288                         }
 289                         set_error ("%s", "Expected < to start the document");
 290                         goto fail;
 291
 292                 case SKIP_XML_DECLARATION:
 293                 case START_ELEMENT: {
 294                         const char *element_start = p, *element_end;
 295                         char *ename = NULL;
 296                         int full_stop = 0, l;
 297                         gchar **names = NULL, **values = NULL;
 298
 299                         for (; p < end && my_isspace (*p); p++)
 300                                 ;
 301                         if (p == end){
 302                                 set_error ("%s", "Unfinished element");
 303                                 goto fail;
 304                         }
 305
 306                         if (*p == '!' && (p+2 < end) && (p [1] == '-') && (p [2] == '-')){
 307                                 context->state = COMMENT;
 308                                 p += 2;
 309                                 break;
 310                         }
 311
 312                         if (!my_isalpha (*p)){
 313                                 set_error ("%s", "Expected an element name");
 314                                 goto fail;
 315                         }
 316
 317                         for (++p; p < end && (my_isalnum (*p) || (*p == '.')); p++)
 318                                 ;
 319                         if (p == end){
 320                                 set_error ("%s", "Expected an element");
 321                                 goto fail;
 322                         }
 323                         element_end = p;
 324
 325                         for (; p < end && my_isspace (*p); p++)
 326                                 ;
 327                         if (p == end){
 328                                 set_error ("%s", "Unfinished element");
 329                                 goto fail;
 330                         }
 331                         p = parse_attributes (p, end, &names, &values, error, &full_stop, context->state);
 332                         if (p == end){
 333                                 if (names != NULL) {
 334                                         g_strfreev (names);
 335                                         g_strfreev (values);
 336                                 }
 337                                 /* Only set the error if parse_attributes did not */
 338                                 if (error != NULL && *error == NULL)
 339                                         set_error ("%s", "Unfinished sequence");
 340                                 goto fail;
 341                         }
 342                         l = (int)(element_end - element_start);
 343                         ename = g_malloc (l + 1);
 344                         if (ename == NULL)
 345                                 goto fail;
 346                         strncpy (ename, element_start, l);
 347                         ename [l] = 0;
 348
 349                         if (context->state == START_ELEMENT)
 350                                 if (context->parser.start_element != NULL)
 351                                         context->parser.start_element (context, ename,
 352                                                                        (const gchar **) names,
 353                                                                        (const gchar **) values,
 354                                                                        context->user_data, error);
 355
 356                         if (names != NULL){
 357                                 g_strfreev (names);
 358                                 g_strfreev (values);
 359                         }
 360
 361                         if (error != NULL && *error != NULL){
 362                                 g_free (ename);
 363                                 goto fail;
 364                         }
 365
 366                         if (full_stop){
 367                                 if (context->parser.end_element != NULL &&  context->state == START_ELEMENT){
 368                                         context->parser.end_element (context, ename, context->user_data, error);
 369                                         if (error != NULL && *error != NULL){
 370                                                 free (ename);
 371                                                 goto fail;
 372                                         }
 373                                 }
 374                                 g_free (ename);
 375                         } else {
 376                                 context->level = g_slist_prepend (context->level, ename);
 377                         }
 378
 379                         context->state = TEXT;
 380                         break;
 381                 } /* case START_ELEMENT */
 382
 383                 case TEXT: {
 384                         if (c == '<'){
 385                                 context->state = FLUSH_TEXT;
 386                                 break;
 387                         }
 388                         if (context->parser.text != NULL){
 389                                 if (context->text == NULL)
 390                                         context->text = g_string_new ("");
 391                                 g_string_append_c (context->text, c);
 392                         }
 393                         break;
 394                 }
 395
 396                 case COMMENT:
 397                         if (*p != '-')
 398                                 break;
 399                         if (p+2 < end && (p [1] == '-') && (p [2] == '>')){
 400                                 context->state = TEXT;
 401                                 p += 2;
 402                                 break;
 403                         }
 404                         break;
 405
 406                 case FLUSH_TEXT:
 407                         if (context->parser.text != NULL){
 408                                 context->parser.text (context, context->text->str, context->text->len,
 409                                                       context->user_data, error);
 410                                 if (error != NULL && *error != NULL)
 411                                         goto fail;
 412                         }
 413
 414                         if (c == '/')
 415                                 context->state = CLOSING_ELEMENT;
 416                         else {
 417                                 p--;
 418                                 context->state = START_ELEMENT;
 419                         }
 420                         break;
 421
 422                 case CLOSING_ELEMENT: {
 423                         GSList *current = context->level;
 424                         char *text;
 425
 426                         if (context->level == NULL){
 427                                 set_error ("%s", "Too many closing tags, not enough open tags");
 428                                 goto fail;
 429                         }
 430
 431                         text = current->data;
 432                         if (context->parser.end_element != NULL){
 433                                 context->parser.end_element (context, text, context->user_data, error);
 434                                 if (error != NULL && *error != NULL){
 435                                         g_free (text);
 436                                         goto fail;
 437                                 }
 438                         }
 439                         g_free (text);
 440
 441                         while (p < end && *p != '>')
 442                                 p++;
 443
 444                         context->level = context->level->next;
 445                         g_slist_free_1 (current);
 446                         context->state = TEXT;
 447                         break;
 448                 } /* case CLOSING_ELEMENT */
 449
 450                 } /* switch */
 451         }
 452
 453
 454         return TRUE;
 455  fail:
 456         if (context->parser.error && error != NULL && *error)
 457                 context->parser.error (context, *error, context->user_data);
 458
 459         destroy_parse_state (context);
 460         return FALSE;
 461 }
 462
 463 gboolean
 464 g_markup_parse_context_end_parse (GMarkupParseContext *context, GError **error)
 465 {
 466         g_return_val_if_fail (context != NULL, FALSE);
 467
 468         /*
 469          * In our case, we always signal errors during parse, not at the end
 470          * see the notes at the top of this file for details on how this
 471          * could be moved here
 472          */
 473         return TRUE;
 474 }