src/html.c

   1 /*
   2  * Claws Mail -- a GTK+ based, lightweight, and fast e-mail client
   3  * Copyright (C) 1999-2016 Hiroyuki Yamamoto and the Claws Mail team
   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 3 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, see <http://www.gnu.org/licenses/>.
  17  */
  18
  19 #include <glib.h>
  20 #include <stdio.h>
  21 #include <string.h>
  22 #include <ctype.h>
  23
  24 #include "html.h"
  25 #include "codeconv.h"
  26 #include "utils.h"
  27 #include "entity.h"
  28
  29 #define SC_HTMLBUFSIZE  8192
  30 #define HR_STR          "────────────────────────────────────────────────"
  31 #define LI_STR          "• "
  32
  33 static SC_HTMLState sc_html_read_line   (SC_HTMLParser  *parser);
  34 static void sc_html_append_char                 (SC_HTMLParser  *parser,
  35                                          gchar           ch);
  36 static void sc_html_append_str                  (SC_HTMLParser  *parser,
  37                                          const gchar    *str,
  38                                          gint            len);
  39 static SC_HTMLState sc_html_parse_tag   (SC_HTMLParser  *parser);
  40 static void sc_html_parse_special               (SC_HTMLParser  *parser);
  41 static void sc_html_get_parenthesis             (SC_HTMLParser  *parser,
  42                                          gchar          *buf,
  43                                          gint            len);
  44
  45
  46 SC_HTMLParser *sc_html_parser_new(FILE *fp, CodeConverter *conv)
  47 {
  48         SC_HTMLParser *parser;
  49
  50         cm_return_val_if_fail(fp != NULL, NULL);
  51         cm_return_val_if_fail(conv != NULL, NULL);
  52
  53         parser = g_new0(SC_HTMLParser, 1);
  54         parser->fp = fp;
  55         parser->conv = conv;
  56         parser->str = g_string_new(NULL);
  57         parser->buf = g_string_new(NULL);
  58         parser->bufp = parser->buf->str;
  59         parser->state = SC_HTML_NORMAL;
  60         parser->href = NULL;
  61         parser->newline = TRUE;
  62         parser->empty_line = TRUE;
  63         parser->space = FALSE;
  64         parser->pre = FALSE;
  65         parser->indent = 0;
  66
  67         return parser;
  68 }
  69
  70 void sc_html_parser_destroy(SC_HTMLParser *parser)
  71 {
  72         g_string_free(parser->str, TRUE);
  73         g_string_free(parser->buf, TRUE);
  74         g_free(parser->href);
  75         g_free(parser);
  76 }
  77
  78 gchar *sc_html_parse(SC_HTMLParser *parser)
  79 {
  80         parser->state = SC_HTML_NORMAL;
  81         g_string_truncate(parser->str, 0);
  82
  83         if (*parser->bufp == '\0') {
  84                 g_string_truncate(parser->buf, 0);
  85                 parser->bufp = parser->buf->str;
  86                 if (sc_html_read_line(parser) == SC_HTML_EOF)
  87                         return NULL;
  88         }
  89
  90         while (*parser->bufp != '\0') {
  91                 switch (*parser->bufp) {
  92                 case '<': {
  93                         SC_HTMLState st;
  94                         st = sc_html_parse_tag(parser);
  95                         /* when we see an href, we need to flush the str
  96                          * buffer.  Then collect all the chars until we
  97                          * see the end anchor tag
  98                          */
  99                         if (SC_HTML_HREF_BEG == st || SC_HTML_HREF == st)
 100                                 return parser->str->str;
 101                         }
 102                         break;
 103                 case '&':
 104                         sc_html_parse_special(parser);
 105                         break;
 106                 case ' ':
 107                 case '\t':
 108                 case '\r':
 109                 case '\n':
 110                         if (parser->bufp[0] == '\r' && parser->bufp[1] == '\n')
 111                                 parser->bufp++;
 112
 113                         if (!parser->pre) {
 114                                 if (!parser->newline)
 115                                         parser->space = TRUE;
 116
 117                                 parser->bufp++;
 118                                 break;
 119                         }
 120                         /* fallthrough */
 121                 default:
 122                         sc_html_append_char(parser, *parser->bufp++);
 123                 }
 124         }
 125
 126         return parser->str->str;
 127 }
 128
 129 static SC_HTMLState sc_html_read_line(SC_HTMLParser *parser)
 130 {
 131         gchar buf[SC_HTMLBUFSIZE];
 132         gchar buf2[SC_HTMLBUFSIZE];
 133         gint index;
 134         gint n;
 135
 136         if (parser->fp == NULL)
 137                 return SC_HTML_EOF;
 138
 139         n = fread(buf, 1, sizeof(buf) - 1, parser->fp);
 140         if (n == 0) {
 141                 parser->state = SC_HTML_EOF;
 142                 return SC_HTML_EOF;
 143         } else
 144                 buf[n] = '\0';
 145
 146         if (conv_convert(parser->conv, buf2, sizeof(buf2), buf) < 0) {
 147                 index = parser->bufp - parser->buf->str;
 148
 149                 conv_utf8todisp(buf2, sizeof(buf2), buf);
 150                 g_string_append(parser->buf, buf2);
 151
 152                 parser->bufp = parser->buf->str + index;
 153
 154                 return SC_HTML_CONV_FAILED;
 155         }
 156
 157         index = parser->bufp - parser->buf->str;
 158
 159         g_string_append(parser->buf, buf2);
 160
 161         parser->bufp = parser->buf->str + index;
 162
 163         return SC_HTML_NORMAL;
 164 }
 165
 166 static void sc_html_append_char(SC_HTMLParser *parser, gchar ch)
 167 {
 168         GString *str = parser->str;
 169
 170         if (!parser->pre && parser->space) {
 171                 g_string_append_c(str, ' ');
 172                 parser->space = FALSE;
 173         }
 174
 175         g_string_append_c(str, ch);
 176
 177         parser->empty_line = FALSE;
 178         if (ch == '\n') {
 179                 parser->newline = TRUE;
 180                 if (str->len > 1 && str->str[str->len - 2] == '\n')
 181                         parser->empty_line = TRUE;
 182                 if (parser->indent > 0) {
 183                         gint i, n = parser->indent;
 184                         for (i = 0; i < n; i++)
 185                                 g_string_append_c(str, '>');
 186                         g_string_append_c(str, ' ');
 187                 }
 188         } else
 189                 parser->newline = FALSE;
 190 }
 191
 192 static void sc_html_append_str(SC_HTMLParser *parser, const gchar *str, gint len)
 193 {
 194         GString *string = parser->str;
 195
 196         if (!parser->pre && parser->space) {
 197                 g_string_append_c(string, ' ');
 198                 parser->space = FALSE;
 199         }
 200
 201         if (len == 0) return;
 202         if (len < 0)
 203                 g_string_append(string, str);
 204         else {
 205                 gchar *s;
 206                 Xstrndup_a(s, str, len, return);
 207                 g_string_append(string, s);
 208         }
 209
 210         parser->empty_line = FALSE;
 211         if (string->len > 0 && string->str[string->len - 1] == '\n') {
 212                 parser->newline = TRUE;
 213                 if (string->len > 1 && string->str[string->len - 2] == '\n')
 214                         parser->empty_line = TRUE;
 215         } else
 216                 parser->newline = FALSE;
 217 }
 218
 219 static SC_HTMLTag *sc_html_get_tag(const gchar *str)
 220 {
 221         SC_HTMLTag *tag;
 222         gchar *tmp;
 223         guchar *tmpp;
 224
 225         cm_return_val_if_fail(str != NULL, NULL);
 226
 227         if (*str == '\0' || *str == '!') return NULL;
 228
 229         Xstrdup_a(tmp, str, return NULL);
 230
 231         tag = g_new0(SC_HTMLTag, 1);
 232
 233         for (tmpp = tmp; *tmpp != '\0' && !g_ascii_isspace(*tmpp); tmpp++)
 234                 ;
 235
 236         if (*tmpp == '\0') {
 237                 tag->name = g_utf8_strdown(tmp, -1);
 238                 return tag;
 239         } else {
 240                 *tmpp++ = '\0';
 241                 tag->name = g_utf8_strdown(tmp, -1);
 242         }
 243
 244         while (*tmpp != '\0') {
 245                 SC_HTMLAttr *attr;
 246                 gchar *attr_name;
 247                 gchar *attr_value;
 248                 gchar *p;
 249                 gchar quote;
 250
 251                 while (g_ascii_isspace(*tmpp)) tmpp++;
 252                 attr_name = tmpp;
 253
 254                 while (*tmpp != '\0' && !g_ascii_isspace(*tmpp) &&
 255                        *tmpp != '=')
 256                         tmpp++;
 257                 if (*tmpp != '\0' && *tmpp != '=') {
 258                         *tmpp++ = '\0';
 259                         while (g_ascii_isspace(*tmpp)) tmpp++;
 260                 }
 261
 262                 if (*tmpp == '=') {
 263                         *tmpp++ = '\0';
 264                         while (g_ascii_isspace(*tmpp)) tmpp++;
 265
 266                         if (*tmpp == '"' || *tmpp == '\'') {
 267                                 /* name="value" */
 268                                 quote = *tmpp;
 269                                 tmpp++;
 270                                 attr_value = tmpp;
 271                                 if ((p = strchr(attr_value, quote)) == NULL) {
 272                                         if (debug_get_mode()) {
 273                                                 g_warning("sc_html_get_tag(): syntax error in tag: '%s'",
 274                                                                   str);
 275                                         } else {
 276                                                 gchar *cut = g_strndup(str, 100);
 277                                                 g_warning("sc_html_get_tag(): syntax error in tag: '%s%s'",
 278                                                                   cut, strlen(str)>100?"...":".");
 279                                                 g_free(cut);
 280                                         }
 281                                         return tag;
 282                                 }
 283                                 tmpp = p;
 284                                 *tmpp++ = '\0';
 285                                 while (g_ascii_isspace(*tmpp)) tmpp++;
 286                         } else {
 287                                 /* name=value */
 288                                 attr_value = tmpp;
 289                                 while (*tmpp != '\0' && !g_ascii_isspace(*tmpp)) tmpp++;
 290                                 if (*tmpp != '\0')
 291                                         *tmpp++ = '\0';
 292                         }
 293                 } else
 294                         attr_value = "";
 295
 296                 g_strchomp(attr_name);
 297                 attr = g_new(SC_HTMLAttr, 1);
 298                 attr->name = g_utf8_strdown(attr_name, -1);
 299                 attr->value = g_strdup(attr_value);
 300                 tag->attr = g_list_append(tag->attr, attr);
 301         }
 302
 303         return tag;
 304 }
 305
 306 static void sc_html_free_tag(SC_HTMLTag *tag)
 307 {
 308         if (!tag) return;
 309
 310         g_free(tag->name);
 311         while (tag->attr != NULL) {
 312                 SC_HTMLAttr *attr = (SC_HTMLAttr *)tag->attr->data;
 313                 g_free(attr->name);
 314                 g_free(attr->value);
 315                 g_free(attr);
 316                 tag->attr = g_list_remove(tag->attr, tag->attr->data);
 317         }
 318         g_free(tag);
 319 }
 320
 321 static void decode_href(SC_HTMLParser *parser)
 322 {
 323         gchar *tmp;
 324         SC_HTMLParser *tparser = g_new0(SC_HTMLParser, 1);
 325
 326         tparser->str = g_string_new(NULL);
 327         tparser->buf = g_string_new(parser->href);
 328         tparser->bufp = tparser->buf->str;
 329
 330         tmp = sc_html_parse(tparser);
 331
 332         g_free(parser->href);
 333         parser->href = g_strdup(tmp);
 334
 335         sc_html_parser_destroy(tparser);
 336 }
 337
 338 static SC_HTMLState sc_html_parse_tag(SC_HTMLParser *parser)
 339 {
 340         gchar buf[SC_HTMLBUFSIZE];
 341         SC_HTMLTag *tag;
 342
 343         sc_html_get_parenthesis(parser, buf, sizeof(buf));
 344
 345         tag = sc_html_get_tag(buf);
 346
 347         parser->state = SC_HTML_UNKNOWN;
 348         if (!tag) return SC_HTML_UNKNOWN;
 349
 350         if (!strcmp(tag->name, "br") || !strcmp(tag->name, "br/")) {
 351                 parser->space = FALSE;
 352                 sc_html_append_char(parser, '\n');
 353                 parser->state = SC_HTML_BR;
 354         } else if (!strcmp(tag->name, "a")) {
 355                 GList *cur;
 356                 parser->href = NULL;
 357                 for (cur = tag->attr; cur != NULL; cur = cur->next) {
 358                         if (cur->data && !strcmp(((SC_HTMLAttr *)cur->data)->name, "href")) {
 359                                 g_free(parser->href);
 360                                 parser->href = g_strdup(((SC_HTMLAttr *)cur->data)->value);
 361                                 decode_href(parser);
 362                                 parser->state = SC_HTML_HREF_BEG;
 363                                 break;
 364                         }
 365                 }
 366                 if (parser->href == NULL)
 367                         parser->href = g_strdup("");
 368                 parser->state = SC_HTML_HREF_BEG;
 369         } else if (!strcmp(tag->name, "/a")) {
 370                 parser->state = SC_HTML_HREF;
 371         } else if (!strcmp(tag->name, "p")) {
 372                 parser->space = FALSE;
 373                 if (!parser->empty_line) {
 374                         parser->space = FALSE;
 375                         if (!parser->newline) sc_html_append_char(parser, '\n');
 376                         sc_html_append_char(parser, '\n');
 377                 }
 378                 parser->state = SC_HTML_PAR;
 379         } else if (!strcmp(tag->name, "pre")) {
 380                 parser->pre = TRUE;
 381                 parser->state = SC_HTML_PRE;
 382         } else if (!strcmp(tag->name, "/pre")) {
 383                 parser->pre = FALSE;
 384                 parser->state = SC_HTML_NORMAL;
 385         } else if (!strcmp(tag->name, "hr")) {
 386                 if (!parser->newline) {
 387                         parser->space = FALSE;
 388                         sc_html_append_char(parser, '\n');
 389                 }
 390                 sc_html_append_str(parser, HR_STR, -1);
 391                 sc_html_append_char(parser, '\n');
 392                 parser->state = SC_HTML_HR;
 393         } else if (!strcmp(tag->name, "div")    ||
 394                    !strcmp(tag->name, "ul")     ||
 395                    !strcmp(tag->name, "li")     ||
 396                    !strcmp(tag->name, "table")  ||
 397                    !strcmp(tag->name, "dd")     ||
 398                    !strcmp(tag->name, "tr")     ||
 399                    (tag->name[0] == 'h' && g_ascii_isdigit(tag->name[1]))) {
 400                 if (!parser->newline) {
 401                         parser->space = FALSE;
 402                         sc_html_append_char(parser, '\n');
 403                 }
 404                 if (!strcmp(tag->name, "li")) {
 405                         sc_html_append_str(parser, LI_STR, -1);
 406                 }
 407                 parser->state = SC_HTML_NORMAL;
 408         } else if (!strcmp(tag->name, "blockquote")) {
 409                 parser->state = SC_HTML_NORMAL;
 410                 parser->indent++;
 411         } else if (!strcmp(tag->name, "/blockquote")) {
 412                 parser->state = SC_HTML_NORMAL;
 413                 parser->indent--;
 414         } else if (!strcmp(tag->name, "/table") ||
 415                    (tag->name[0] == '/' &&
 416                     tag->name[1] == 'h' &&
 417                     g_ascii_isdigit(tag->name[1]))) {
 418                 if (!parser->empty_line) {
 419                         parser->space = FALSE;
 420                         if (!parser->newline) sc_html_append_char(parser, '\n');
 421                         sc_html_append_char(parser, '\n');
 422                 }
 423                 parser->state = SC_HTML_NORMAL;
 424         } else if (!strcmp(tag->name, "/div")   ||
 425                    !strcmp(tag->name, "/ul")    ||
 426                    !strcmp(tag->name, "/li")) {
 427                 if (!parser->newline) {
 428                         parser->space = FALSE;
 429                         sc_html_append_char(parser, '\n');
 430                 }
 431                 parser->state = SC_HTML_NORMAL;
 432                         }
 433
 434         sc_html_free_tag(tag);
 435
 436         return parser->state;
 437 }
 438
 439 static void sc_html_parse_special(SC_HTMLParser *parser)
 440 {
 441         gchar *entity;
 442
 443         parser->state = SC_HTML_UNKNOWN;
 444         cm_return_if_fail(*parser->bufp == '&');
 445
 446         entity = entity_decode(parser->bufp);
 447         if (entity != NULL) {
 448                 sc_html_append_str(parser, entity, -1);
 449                 g_free(entity);
 450                 while (*parser->bufp++ != ';');
 451         } else {
 452                 /* output literal `&' */
 453                 sc_html_append_char(parser, *parser->bufp++);
 454         }
 455         parser->state = SC_HTML_NORMAL;
 456 }
 457
 458 static gchar *sc_html_find_tag(SC_HTMLParser *parser, const gchar *tag)
 459 {
 460         gchar *cur = parser->bufp;
 461         gint len = strlen(tag);
 462
 463         if (cur == NULL)
 464                 return NULL;
 465
 466         while ((cur = strstr(cur, "<")) != NULL) {
 467                 if (!g_ascii_strncasecmp(cur, tag, len))
 468                         return cur;
 469                 cur += 2;
 470         }
 471         return NULL;
 472 }
 473
 474 static void sc_html_get_parenthesis(SC_HTMLParser *parser, gchar *buf, gint len)
 475 {
 476         gchar *p;
 477
 478         buf[0] = '\0';
 479         cm_return_if_fail(*parser->bufp == '<');
 480
 481         /* ignore comment / CSS / script stuff */
 482         if (!strncmp(parser->bufp, "<!--", 4)) {
 483                 parser->bufp += 4;
 484                 while ((p = strstr(parser->bufp, "-->")) == NULL)
 485                         if (sc_html_read_line(parser) == SC_HTML_EOF) return;
 486                 parser->bufp = p + 3;
 487                 return;
 488         }
 489         if (!g_ascii_strncasecmp(parser->bufp, "<style", 6)) {
 490                 parser->bufp += 6;
 491                 while ((p = sc_html_find_tag(parser, "</style>")) == NULL)
 492                         if (sc_html_read_line(parser) == SC_HTML_EOF) return;
 493                 parser->bufp = p + 8;
 494                 return;
 495         }
 496         if (!g_ascii_strncasecmp(parser->bufp, "<script", 7)) {
 497                 parser->bufp += 7;
 498                 while ((p = sc_html_find_tag(parser, "</script>")) == NULL)
 499                         if (sc_html_read_line(parser) == SC_HTML_EOF) return;
 500                 parser->bufp = p + 9;
 501                 return;
 502         }
 503
 504         parser->bufp++;
 505         while ((p = strchr(parser->bufp, '>')) == NULL)
 506                 if (sc_html_read_line(parser) == SC_HTML_EOF) return;
 507
 508         strncpy2(buf, parser->bufp, MIN(p - parser->bufp + 1, len));
 509         g_strstrip(buf);
 510         parser->bufp = p + 1;
 511 }