src/html.c

   1 /*
   2  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2005 Hiroyuki Yamamoto
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License
  16  * along with this program; if not, write to the Free Software
  17  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  18  */
  19
  20 #include <glib.h>
  21 #include <stdio.h>
  22 #include <string.h>
  23 #include <ctype.h>
  24
  25 #include "html.h"
  26 #include "codeconv.h"
  27 #include "utils.h"
  28
  29 #define HTMLBUFSIZE     8192
  30 #define HR_STR          "------------------------------------------------"
  31
  32 typedef struct _HTMLSymbol      HTMLSymbol;
  33
  34 struct _HTMLSymbol
  35 {
  36         gchar *const key;
  37         gchar *const val;
  38 };
  39
  40 static HTMLSymbol symbol_list[] = {
  41         {"&lt;"    , "<"},
  42         {"&gt;"    , ">"},
  43         {"&amp;"   , "&"},
  44         {"&quot;"  , "\""},
  45         {"&nbsp;"  , " "},
  46         {"&trade;" , "(TM)"},
  47
  48         {"&#153;", "(TM)"},
  49 };
  50
  51 static HTMLSymbol ascii_symbol_list[] = {
  52         {"&iexcl;" , "^!"},
  53         {"&brvbar;", "|"},
  54         {"&copy;"  , "(C)"},
  55         {"&laquo;" , "<<"},
  56         {"&reg;"   , "(R)"},
  57
  58         {"&sup2;"  , "^2"},
  59         {"&sup3;"  , "^3"},
  60         {"&acute;" , "'"},
  61         {"&cedil;" , ","},
  62         {"&sup1;"  , "^1"},
  63         {"&raquo;" , ">>"},
  64         {"&frac14;", "1/4"},
  65         {"&frac12;", "1/2"},
  66         {"&frac34;", "3/4"},
  67         {"&iquest;", "^?"},
  68
  69         {"&Agrave;", "A`"},
  70         {"&Aacute;", "A'"},
  71         {"&Acirc;" , "A^"},
  72         {"&Atilde;", "A~"},
  73         {"&AElig;" , "AE"},
  74         {"&Egrave;", "E`"},
  75         {"&Eacute;", "E'"},
  76         {"&Ecirc;" , "E^"},
  77         {"&Igrave;", "I`"},
  78         {"&Iacute;", "I'"},
  79         {"&Icirc;" , "I^"},
  80
  81         {"&Ntilde;", "N~"},
  82         {"&Ograve;", "O`"},
  83         {"&Oacute;", "O'"},
  84         {"&Ocirc;" , "O^"},
  85         {"&Otilde;", "O~"},
  86         {"&Ugrave;", "U`"},
  87         {"&Uacute;", "U'"},
  88         {"&Ucirc;" , "U^"},
  89         {"&Yacute;", "Y'"},
  90
  91         {"&agrave;", "a`"},
  92         {"&aacute;", "a'"},
  93         {"&acirc;" , "a^"},
  94         {"&atilde;", "a~"},
  95         {"&aelig;" , "ae"},
  96         {"&egrave;", "e`"},
  97         {"&eacute;", "e'"},
  98         {"&ecirc;" , "e^"},
  99         {"&igrave;", "i`"},
 100         {"&iacute;", "i'"},
 101         {"&icirc;" , "i^"},
 102
 103         {"&ntilde;", "n~"},
 104         {"&ograve;", "o`"},
 105         {"&oacute;", "o'"},
 106         {"&ocirc;" , "o^"},
 107         {"&otilde;", "o~"},
 108         {"&ugrave;", "u`"},
 109         {"&uacute;", "u'"},
 110         {"&ucirc;" , "u^"},
 111         {"&yacute;", "y'"},
 112 };
 113
 114 static GHashTable *default_symbol_table;
 115
 116 static HTMLState html_read_line         (HTMLParser     *parser);
 117 static void html_append_char            (HTMLParser     *parser,
 118                                          gchar           ch);
 119 static void html_append_str             (HTMLParser     *parser,
 120                                          const gchar    *str,
 121                                          gint            len);
 122 static HTMLState html_parse_tag         (HTMLParser     *parser);
 123 static void html_parse_special          (HTMLParser     *parser);
 124 static void html_get_parenthesis        (HTMLParser     *parser,
 125                                          gchar          *buf,
 126                                          gint            len);
 127
 128
 129 HTMLParser *html_parser_new(FILE *fp, CodeConverter *conv)
 130 {
 131         HTMLParser *parser;
 132
 133         g_return_val_if_fail(fp != NULL, NULL);
 134         g_return_val_if_fail(conv != NULL, NULL);
 135
 136         parser = g_new0(HTMLParser, 1);
 137         parser->fp = fp;
 138         parser->conv = conv;
 139         parser->str = g_string_new(NULL);
 140         parser->buf = g_string_new(NULL);
 141         parser->bufp = parser->buf->str;
 142         parser->state = HTML_NORMAL;
 143         parser->href = NULL;
 144         parser->newline = TRUE;
 145         parser->empty_line = TRUE;
 146         parser->space = FALSE;
 147         parser->pre = FALSE;
 148
 149 #define SYMBOL_TABLE_ADD(table, list) \
 150 { \
 151         gint i; \
 152  \
 153         for (i = 0; i < sizeof(list) / sizeof(list[0]); i++) \
 154                 g_hash_table_insert(table, list[i].key, list[i].val); \
 155 }
 156
 157         if (!default_symbol_table) {
 158                 default_symbol_table =
 159                         g_hash_table_new(g_str_hash, g_str_equal);
 160                 SYMBOL_TABLE_ADD(default_symbol_table, symbol_list);
 161                 SYMBOL_TABLE_ADD(default_symbol_table, ascii_symbol_list);
 162         }
 163
 164 #undef SYMBOL_TABLE_ADD
 165
 166         parser->symbol_table = default_symbol_table;
 167
 168         return parser;
 169 }
 170
 171 void html_parser_destroy(HTMLParser *parser)
 172 {
 173         g_string_free(parser->str, TRUE);
 174         g_string_free(parser->buf, TRUE);
 175         g_free(parser->href);
 176         g_free(parser);
 177 }
 178
 179 gchar *html_parse(HTMLParser *parser)
 180 {
 181         parser->state = HTML_NORMAL;
 182         g_string_truncate(parser->str, 0);
 183
 184         if (*parser->bufp == '\0') {
 185                 g_string_truncate(parser->buf, 0);
 186                 parser->bufp = parser->buf->str;
 187                 if (html_read_line(parser) == HTML_EOF)
 188                         return NULL;
 189         }
 190
 191         while (*parser->bufp != '\0') {
 192                 switch (*parser->bufp) {
 193                 case '<': {
 194                         HTMLState st;
 195                         st = html_parse_tag(parser);
 196                         /* when we see an href, we need to flush the str
 197                          * buffer.  Then collect all the chars until we
 198                          * see the end anchor tag
 199                          */
 200                         if (HTML_HREF_BEG == st || HTML_HREF == st)
 201                                 return parser->str->str;
 202                         }
 203                         break;
 204                 case '&':
 205                         html_parse_special(parser);
 206                         break;
 207                 case ' ':
 208                 case '\t':
 209                 case '\r':
 210                 case '\n':
 211                         if (parser->bufp[0] == '\r' && parser->bufp[1] == '\n')
 212                                 parser->bufp++;
 213
 214                         if (!parser->pre) {
 215                                 if (!parser->newline)
 216                                         parser->space = TRUE;
 217
 218                                 parser->bufp++;
 219                                 break;
 220                         }
 221                         /* fallthrough */
 222                 default:
 223                         html_append_char(parser, *parser->bufp++);
 224                 }
 225         }
 226
 227         return parser->str->str;
 228 }
 229
 230 static HTMLState html_read_line(HTMLParser *parser)
 231 {
 232         gchar buf[HTMLBUFSIZE];
 233         gchar buf2[HTMLBUFSIZE];
 234         gint index;
 235
 236         if (fgets(buf, sizeof(buf), parser->fp) == NULL) {
 237                 parser->state = HTML_EOF;
 238                 return HTML_EOF;
 239         }
 240
 241         if (conv_convert(parser->conv, buf2, sizeof(buf2), buf) < 0) {
 242                 index = parser->bufp - parser->buf->str;
 243
 244                 conv_utf8todisp(buf2, sizeof(buf2), buf);
 245                 g_string_append(parser->buf, buf2);
 246
 247                 parser->bufp = parser->buf->str + index;
 248
 249                 return HTML_CONV_FAILED;
 250         }
 251
 252         index = parser->bufp - parser->buf->str;
 253
 254         g_string_append(parser->buf, buf2);
 255
 256         parser->bufp = parser->buf->str + index;
 257
 258         return HTML_NORMAL;
 259 }
 260
 261 static void html_append_char(HTMLParser *parser, gchar ch)
 262 {
 263         GString *str = parser->str;
 264
 265         if (!parser->pre && parser->space) {
 266                 g_string_append_c(str, ' ');
 267                 parser->space = FALSE;
 268         }
 269
 270         g_string_append_c(str, ch);
 271
 272         parser->empty_line = FALSE;
 273         if (ch == '\n') {
 274                 parser->newline = TRUE;
 275                 if (str->len > 1 && str->str[str->len - 2] == '\n')
 276                         parser->empty_line = TRUE;
 277         } else
 278                 parser->newline = FALSE;
 279 }
 280
 281 static void html_append_str(HTMLParser *parser, const gchar *str, gint len)
 282 {
 283         GString *string = parser->str;
 284
 285         if (!parser->pre && parser->space) {
 286                 g_string_append_c(string, ' ');
 287                 parser->space = FALSE;
 288         }
 289
 290         if (len == 0) return;
 291         if (len < 0)
 292                 g_string_append(string, str);
 293         else {
 294                 gchar *s;
 295                 Xstrndup_a(s, str, len, return);
 296                 g_string_append(string, s);
 297         }
 298
 299         parser->empty_line = FALSE;
 300         if (string->len > 0 && string->str[string->len - 1] == '\n') {
 301                 parser->newline = TRUE;
 302                 if (string->len > 1 && string->str[string->len - 2] == '\n')
 303                         parser->empty_line = TRUE;
 304         } else
 305                 parser->newline = FALSE;
 306 }
 307
 308 static HTMLTag *html_get_tag(const gchar *str)
 309 {
 310         HTMLTag *tag;
 311         gchar *tmp;
 312         guchar *tmpp;
 313
 314         g_return_val_if_fail(str != NULL, NULL);
 315
 316         if (*str == '\0' || *str == '!') return NULL;
 317
 318         Xstrdup_a(tmp, str, return NULL);
 319
 320         tag = g_new0(HTMLTag, 1);
 321
 322         for (tmpp = tmp; *tmpp != '\0' && !g_ascii_isspace(*tmpp); tmpp++)
 323                 ;
 324
 325         if (*tmpp == '\0') {
 326                 g_strdown(tmp);
 327                 tag->name = g_strdup(tmp);
 328                 return tag;
 329         } else {
 330                 *tmpp++ = '\0';
 331                 g_strdown(tmp);
 332                 tag->name = g_strdup(tmp);
 333         }
 334
 335         while (*tmpp != '\0') {
 336                 HTMLAttr *attr;
 337                 gchar *attr_name;
 338                 gchar *attr_value;
 339                 gchar *p;
 340                 gchar quote;
 341
 342                 while (g_ascii_isspace(*tmpp)) tmpp++;
 343                 attr_name = tmpp;
 344
 345                 while (*tmpp != '\0' && !g_ascii_isspace(*tmpp) &&
 346                        *tmpp != '=')
 347                         tmpp++;
 348                 if (*tmpp != '\0' && *tmpp != '=') {
 349                         *tmpp++ = '\0';
 350                         while (g_ascii_isspace(*tmpp)) tmpp++;
 351                 }
 352
 353                 if (*tmpp == '=') {
 354                         *tmpp++ = '\0';
 355                         while (g_ascii_isspace(*tmpp)) tmpp++;
 356
 357                         if (*tmpp == '"' || *tmpp == '\'') {
 358                                 /* name="value" */
 359                                 quote = *tmpp;
 360                                 tmpp++;
 361                                 attr_value = tmpp;
 362                                 if ((p = strchr(attr_value, quote)) == NULL) {
 363                                         g_warning("html_get_tag(): syntax error in tag: '%s'\n", str);
 364                                         return tag;
 365                                 }
 366                                 tmpp = p;
 367                                 *tmpp++ = '\0';
 368                                 while (g_ascii_isspace(*tmpp)) tmpp++;
 369                         } else {
 370                                 /* name=value */
 371                                 attr_value = tmpp;
 372                                 while (*tmpp != '\0' && !g_ascii_isspace(*tmpp)) tmpp++;
 373                                 if (*tmpp != '\0')
 374                                         *tmpp++ = '\0';
 375                         }
 376                 } else
 377                         attr_value = "";
 378
 379                 g_strchomp(attr_name);
 380                 g_strdown(attr_name);
 381                 attr = g_new(HTMLAttr, 1);
 382                 attr->name = g_strdup(attr_name);
 383                 attr->value = g_strdup(attr_value);
 384                 tag->attr = g_list_append(tag->attr, attr);
 385         }
 386
 387         return tag;
 388 }
 389
 390 static void html_free_tag(HTMLTag *tag)
 391 {
 392         if (!tag) return;
 393
 394         g_free(tag->name);
 395         while (tag->attr != NULL) {
 396                 HTMLAttr *attr = (HTMLAttr *)tag->attr->data;
 397                 g_free(attr->name);
 398                 g_free(attr->value);
 399                 g_free(attr);
 400                 tag->attr = g_list_remove(tag->attr, tag->attr->data);
 401         }
 402         g_free(tag);
 403 }
 404
 405 static HTMLState html_parse_tag(HTMLParser *parser)
 406 {
 407         gchar buf[HTMLBUFSIZE];
 408         HTMLTag *tag;
 409
 410         html_get_parenthesis(parser, buf, sizeof(buf));
 411
 412         tag = html_get_tag(buf);
 413
 414         parser->state = HTML_UNKNOWN;
 415         if (!tag) return HTML_UNKNOWN;
 416
 417         if (!strcmp(tag->name, "br")) {
 418                 parser->space = FALSE;
 419                 html_append_char(parser, '\n');
 420                 parser->state = HTML_BR;
 421         } else if (!strcmp(tag->name, "a")) {
 422                 if (tag->attr && tag->attr->data &&
 423                     !strcmp(((HTMLAttr *)tag->attr->data)->name, "href")) {
 424                         g_free(parser->href);
 425                         parser->href =
 426                                 g_strdup(((HTMLAttr *)tag->attr->data)->value);
 427                         parser->state = HTML_HREF_BEG;
 428                 }
 429         } else if (!strcmp(tag->name, "/a")) {
 430                 parser->state = HTML_HREF;
 431         } else if (!strcmp(tag->name, "p")) {
 432                 parser->space = FALSE;
 433                 if (!parser->empty_line) {
 434                         parser->space = FALSE;
 435                         if (!parser->newline) html_append_char(parser, '\n');
 436                         html_append_char(parser, '\n');
 437                 }
 438                 parser->state = HTML_PAR;
 439         } else if (!strcmp(tag->name, "pre")) {
 440                 parser->pre = TRUE;
 441                 parser->state = HTML_PRE;
 442         } else if (!strcmp(tag->name, "/pre")) {
 443                 parser->pre = FALSE;
 444                 parser->state = HTML_NORMAL;
 445         } else if (!strcmp(tag->name, "hr")) {
 446                 if (!parser->newline) {
 447                         parser->space = FALSE;
 448                         html_append_char(parser, '\n');
 449                 }
 450                 html_append_str(parser, HR_STR "\n", -1);
 451                 parser->state = HTML_HR;
 452         } else if (!strcmp(tag->name, "div")    ||
 453                    !strcmp(tag->name, "ul")     ||
 454                    !strcmp(tag->name, "li")     ||
 455                    !strcmp(tag->name, "table")  ||
 456                    !strcmp(tag->name, "tr")     ||
 457                    (tag->name[0] == 'h' && g_ascii_isdigit(tag->name[1]))) {
 458                 if (!parser->newline) {
 459                         parser->space = FALSE;
 460                         html_append_char(parser, '\n');
 461                 }
 462                 parser->state = HTML_NORMAL;
 463         } else if (!strcmp(tag->name, "/table") ||
 464                    (tag->name[0] == '/' &&
 465                     tag->name[1] == 'h' &&
 466                     g_ascii_isdigit(tag->name[1]))) {
 467                 if (!parser->empty_line) {
 468                         parser->space = FALSE;
 469                         if (!parser->newline) html_append_char(parser, '\n');
 470                         html_append_char(parser, '\n');
 471                 }
 472                 parser->state = HTML_NORMAL;
 473         } else if (!strcmp(tag->name, "/div")   ||
 474                    !strcmp(tag->name, "/ul")    ||
 475                    !strcmp(tag->name, "/li")) {
 476                 if (!parser->newline) {
 477                         parser->space = FALSE;
 478                         html_append_char(parser, '\n');
 479                 }
 480                 parser->state = HTML_NORMAL;
 481                         }
 482
 483         html_free_tag(tag);
 484
 485         return parser->state;
 486 }
 487
 488 static void html_parse_special(HTMLParser *parser)
 489 {
 490         gchar symbol_name[9];
 491         gint n;
 492         const gchar *val;
 493
 494         parser->state = HTML_UNKNOWN;
 495         g_return_if_fail(*parser->bufp == '&');
 496
 497         /* &foo; */
 498         for (n = 0; parser->bufp[n] != '\0' && parser->bufp[n] != ';'; n++)
 499                 ;
 500         if (n > 7 || parser->bufp[n] != ';') {
 501                 /* output literal `&' */
 502                 html_append_char(parser, *parser->bufp++);
 503                 parser->state = HTML_NORMAL;
 504                 return;
 505         }
 506         strncpy2(symbol_name, parser->bufp, n + 2);
 507         parser->bufp += n + 1;
 508
 509         if ((val = g_hash_table_lookup(parser->symbol_table, symbol_name))
 510             != NULL) {
 511                 html_append_str(parser, val, -1);
 512                 parser->state = HTML_NORMAL;
 513                 return;
 514         } else if (symbol_name[1] == '#' && g_ascii_isdigit(symbol_name[2])) {
 515                 gint ch;
 516
 517                 ch = atoi(symbol_name + 2);
 518                 if ((ch > 0 && ch <= 127) ||
 519                     (ch >= 128 && ch <= 255 &&
 520                      parser->conv->charset == C_ISO_8859_1)) {
 521                         html_append_char(parser, ch);
 522                         parser->state = HTML_NORMAL;
 523                         return;
 524                 } else {
 525                         char *symb = NULL;
 526                         switch (ch) {
 527                         /* http://www.w3schools.com/html/html_entitiesref.asp */
 528                         case 338:       /* capital ligature OE  &OElig;  */
 529                                 symb = "OE";
 530                                 break;
 531                         case 339:       /* small ligature OE    &oelig;  */
 532                                 symb = "oe";
 533                                 break;
 534                         case 352:       /* capital S w/caron    &Scaron; */
 535                         case 353:       /* small S w/caron      &scaron; */
 536                         case 376:       /* cap Y w/ diaeres     &Yuml;   */
 537                                 break;
 538                         case 710:       /* circumflex accent    &circ;   */
 539                                 symb = "^";
 540                                 break;
 541                         case 732:       /* small tilde          &tilde;  */
 542                                 symb = "~";
 543                                 break;
 544                         case 8194:      /* en space             &ensp;   */
 545                         case 8195:      /* em space             &emsp;   */
 546                         case 8201:      /* thin space           &thinsp; */
 547                                 symb = " ";
 548                                 break;
 549                         case 8204:      /* zero width non-joiner &zwnj;  */
 550                         case 8205:      /* zero width joiner    &zwj;   */
 551                         case 8206:      /* l-t-r mark           &lrm;   */
 552                         case 8207:      /* r-t-l mark           &rlm     */
 553                                 break;
 554                         case 8211:      /* en dash              &ndash;  */
 555                                 symb = "-";
 556                                 break;
 557                         case 8212:      /* em dash              &mdash;  */
 558                                 symb = "--";
 559                                 break;
 560                         case 8216:      /* l single quot mark   &lsquo;  */
 561                                 symb = "`";
 562                                 break;
 563                         case 8217:      /* r single quot mark   &rsquo;  */
 564                                 symb = "'";
 565                                 break;
 566                         case 8218:      /* single low-9 quot    &sbquo;  */
 567                                 symb = ",";
 568                                 break;
 569                         case 8220:      /* l double quot mark   &ldquo;  */
 570                                 symb = "``";
 571                                 break;
 572                         case 8221:      /* r double quot mark   &rdquo;  */
 573                                 symb = "''";
 574                                 break;
 575                         case 8222:      /* double low-9 quot    &bdquo;  */
 576                                 symb = ",,";
 577                                 break;
 578                         case 8224:      /* dagger               &dagger; */
 579                         case 8225:      /* double dagger        &Dagger; */
 580                                 break;
 581                         case 8230:      /* horizontal ellipsis  &hellip; */
 582                                 symb = "...";
 583                                 break;
 584                         case 8240:      /* per mile             &permil; */
 585                                 symb = "\%o";
 586                                 break;
 587                         case 8249:      /* l-pointing angle quot &lsaquo; */
 588                                 symb = "<";
 589                                 break;
 590                         case 8250:      /* r-pointing angle quot &rsaquo; */
 591                                 symb = ">";
 592                                 break;
 593                         case 8364:      /* euro                 &euro;   */
 594                                 symb = "EUR";
 595                                 break;
 596                         case 8482:      /* trademark            &trade;  */
 597                                 symb  = "(TM)";
 598                                 break;
 599                         default:
 600                                 break;
 601                         }
 602                         if (symb) {
 603                                 html_append_str(parser, symb, -1);
 604                                 parser->state = HTML_NORMAL;
 605                                 return;
 606                         }
 607                 }
 608         }
 609
 610         html_append_str(parser, symbol_name, -1);
 611 }
 612
 613 static void html_get_parenthesis(HTMLParser *parser, gchar *buf, gint len)
 614 {
 615         gchar *p;
 616
 617         buf[0] = '\0';
 618         g_return_if_fail(*parser->bufp == '<');
 619
 620         /* ignore comment / CSS / script stuff */
 621         if (!strncmp(parser->bufp, "<!--", 4)) {
 622                 parser->bufp += 4;
 623                 while ((p = strstr(parser->bufp, "-->")) == NULL)
 624                         if (html_read_line(parser) == HTML_EOF) return;
 625                 parser->bufp = p + 3;
 626                 return;
 627         }
 628         if (!g_ascii_strncasecmp(parser->bufp, "<style", 6)) {
 629                 parser->bufp += 6;
 630                 while ((p = strcasestr(parser->bufp, "</style>")) == NULL)
 631                         if (html_read_line(parser) == HTML_EOF) return;
 632                 parser->bufp = p + 8;
 633                 return;
 634         }
 635         if (!g_ascii_strncasecmp(parser->bufp, "<script", 7)) {
 636                 parser->bufp += 7;
 637                 while ((p = strcasestr(parser->bufp, "</script>")) == NULL)
 638                         if (html_read_line(parser) == HTML_EOF) return;
 639                 parser->bufp = p + 9;
 640                 return;
 641         }
 642
 643         parser->bufp++;
 644         while ((p = strchr(parser->bufp, '>')) == NULL)
 645                 if (html_read_line(parser) == HTML_EOF) return;
 646
 647         strncpy2(buf, parser->bufp, MIN(p - parser->bufp + 1, len));
 648         g_strstrip(buf);
 649         parser->bufp = p + 1;
 650 }