modules/base/src/template_utils.c

   1 /*
   2  * LuCI Template - Utility functions
   3  *
   4  *   Copyright (C) 2010 Jo-Philipp Wich <xm@subsignal.org>
   5  *
   6  *  Licensed under the Apache License, Version 2.0 (the "License");
   7  *  you may not use this file except in compliance with the License.
   8  *  You may obtain a copy of the License at
   9  *
  10  *      http://www.apache.org/licenses/LICENSE-2.0
  11  *
  12  *  Unless required by applicable law or agreed to in writing, software
  13  *  distributed under the License is distributed on an "AS IS" BASIS,
  14  *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  15  *  See the License for the specific language governing permissions and
  16  *  limitations under the License.
  17  */
  18
  19 #include "template_utils.h"
  20 #include "template_lmo.h"
  21
  22 /* initialize a buffer object */
  23 struct template_buffer * buf_init(int size)
  24 {
  25         struct template_buffer *buf;
  26
  27         if (size <= 0)
  28                 size = 1024;
  29
  30         buf = (struct template_buffer *)malloc(sizeof(struct template_buffer));
  31
  32         if (buf != NULL)
  33         {
  34                 buf->fill = 0;
  35                 buf->size = size;
  36                 buf->data = malloc(buf->size);
  37
  38                 if (buf->data != NULL)
  39                 {
  40                         buf->dptr = buf->data;
  41                         buf->data[0] = 0;
  42
  43                         return buf;
  44                 }
  45
  46                 free(buf);
  47         }
  48
  49         return NULL;
  50 }
  51
  52 /* grow buffer */
  53 int buf_grow(struct template_buffer *buf, int size)
  54 {
  55         unsigned int off = (buf->dptr - buf->data);
  56         char *data;
  57
  58         if (size <= 0)
  59                 size = 1024;
  60
  61         data = realloc(buf->data, buf->size + size);
  62
  63         if (data != NULL)
  64         {
  65                 buf->data  = data;
  66                 buf->dptr  = data + off;
  67                 buf->size += size;
  68
  69                 return buf->size;
  70         }
  71
  72         return 0;
  73 }
  74
  75 /* put one char into buffer object */
  76 int buf_putchar(struct template_buffer *buf, char c)
  77 {
  78         if( ((buf->fill + 1) >= buf->size) && !buf_grow(buf, 0) )
  79                 return 0;
  80
  81         *(buf->dptr++) = c;
  82         *(buf->dptr) = 0;
  83
  84         buf->fill++;
  85         return 1;
  86 }
  87
  88 /* append data to buffer */
  89 int buf_append(struct template_buffer *buf, const char *s, int len)
  90 {
  91         if ((buf->fill + len + 1) >= buf->size)
  92         {
  93                 if (!buf_grow(buf, len + 1))
  94                         return 0;
  95         }
  96
  97         memcpy(buf->dptr, s, len);
  98         buf->fill += len;
  99         buf->dptr += len;
 100
 101         *(buf->dptr) = 0;
 102
 103         return len;
 104 }
 105
 106 /* read buffer length */
 107 int buf_length(struct template_buffer *buf)
 108 {
 109         return buf->fill;
 110 }
 111
 112 /* destroy buffer object and return pointer to data */
 113 char * buf_destroy(struct template_buffer *buf)
 114 {
 115         char *data = buf->data;
 116
 117         free(buf);
 118         return data;
 119 }
 120
 121
 122 /* calculate the number of expected continuation chars */
 123 static inline int mb_num_chars(unsigned char c)
 124 {
 125         if ((c & 0xE0) == 0xC0)
 126                 return 2;
 127         else if ((c & 0xF0) == 0xE0)
 128                 return 3;
 129         else if ((c & 0xF8) == 0xF0)
 130                 return 4;
 131         else if ((c & 0xFC) == 0xF8)
 132                 return 5;
 133         else if ((c & 0xFE) == 0xFC)
 134                 return 6;
 135
 136         return 1;
 137 }
 138
 139 /* test whether the given byte is a valid continuation char */
 140 static inline int mb_is_cont(unsigned char c)
 141 {
 142         return ((c >= 0x80) && (c <= 0xBF));
 143 }
 144
 145 /* test whether the byte sequence at the given pointer with the given
 146  * length is the shortest possible representation of the code point */
 147 static inline int mb_is_shortest(unsigned char *s, int n)
 148 {
 149         switch (n)
 150         {
 151                 case 2:
 152                         /* 1100000x (10xxxxxx) */
 153                         return !(((*s >> 1) == 0x60) &&
 154                                          ((*(s+1) >> 6) == 0x02));
 155
 156                 case 3:
 157                         /* 11100000 100xxxxx (10xxxxxx) */
 158                         return !((*s == 0xE0) &&
 159                                          ((*(s+1) >> 5) == 0x04) &&
 160                                          ((*(s+2) >> 6) == 0x02));
 161
 162                 case 4:
 163                         /* 11110000 1000xxxx (10xxxxxx 10xxxxxx) */
 164                         return !((*s == 0xF0) &&
 165                                          ((*(s+1) >> 4) == 0x08) &&
 166                                          ((*(s+2) >> 6) == 0x02) &&
 167                                          ((*(s+3) >> 6) == 0x02));
 168
 169                 case 5:
 170                         /* 11111000 10000xxx (10xxxxxx 10xxxxxx 10xxxxxx) */
 171                         return !((*s == 0xF8) &&
 172                                          ((*(s+1) >> 3) == 0x10) &&
 173                                          ((*(s+2) >> 6) == 0x02) &&
 174                                          ((*(s+3) >> 6) == 0x02) &&
 175                                          ((*(s+4) >> 6) == 0x02));
 176
 177                 case 6:
 178                         /* 11111100 100000xx (10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx) */
 179                         return !((*s == 0xF8) &&
 180                                          ((*(s+1) >> 2) == 0x20) &&
 181                                          ((*(s+2) >> 6) == 0x02) &&
 182                                          ((*(s+3) >> 6) == 0x02) &&
 183                                          ((*(s+4) >> 6) == 0x02) &&
 184                                          ((*(s+5) >> 6) == 0x02));
 185         }
 186
 187         return 1;
 188 }
 189
 190 /* test whether the byte sequence at the given pointer with the given
 191  * length is an UTF-16 surrogate */
 192 static inline int mb_is_surrogate(unsigned char *s, int n)
 193 {
 194         return ((n == 3) && (*s == 0xED) && (*(s+1) >= 0xA0) && (*(s+1) <= 0xBF));
 195 }
 196
 197 /* test whether the byte sequence at the given pointer with the given
 198  * length is an illegal UTF-8 code point */
 199 static inline int mb_is_illegal(unsigned char *s, int n)
 200 {
 201         return ((n == 3) && (*s == 0xEF) && (*(s+1) == 0xBF) &&
 202                         (*(s+2) >= 0xBE) && (*(s+2) <= 0xBF));
 203 }
 204
 205
 206 /* scan given source string, validate UTF-8 sequence and store result
 207  * in given buffer object */
 208 static int _validate_utf8(unsigned char **s, int l, struct template_buffer *buf)
 209 {
 210         unsigned char *ptr = *s;
 211         unsigned int o = 0, v, n;
 212
 213         /* ascii byte without null */
 214         if ((*(ptr+0) >= 0x01) && (*(ptr+0) <= 0x7F))
 215         {
 216                 if (!buf_putchar(buf, *ptr++))
 217                         return 0;
 218
 219                 o = 1;
 220         }
 221
 222         /* multi byte sequence */
 223         else if ((n = mb_num_chars(*ptr)) > 1)
 224         {
 225                 /* count valid chars */
 226                 for (v = 1; (v <= n) && ((o+v) < l) && mb_is_cont(*(ptr+v)); v++);
 227
 228                 switch (n)
 229                 {
 230                         case 6:
 231                         case 5:
 232                                 /* five and six byte sequences are always invalid */
 233                                 if (!buf_putchar(buf, '?'))
 234                                         return 0;
 235
 236                                 break;
 237
 238                         default:
 239                                 /* if the number of valid continuation bytes matches the
 240                                  * expected number and if the sequence is legal, copy
 241                                  * the bytes to the destination buffer */
 242                                 if ((v == n) && mb_is_shortest(ptr, n) &&
 243                                         !mb_is_surrogate(ptr, n) && !mb_is_illegal(ptr, n))
 244                                 {
 245                                         /* copy sequence */
 246                                         if (!buf_append(buf, (char *)ptr, n))
 247                                                 return 0;
 248                                 }
 249
 250                                 /* the found sequence is illegal, skip it */
 251                                 else
 252                                 {
 253                                         /* invalid sequence */
 254                                         if (!buf_putchar(buf, '?'))
 255                                                 return 0;
 256                                 }
 257
 258                                 break;
 259                 }
 260
 261                 /* advance beyound the last found valid continuation char */
 262                 o = v;
 263                 ptr += v;
 264         }
 265
 266         /* invalid byte (0x00) */
 267         else
 268         {
 269                 if (!buf_putchar(buf, '?')) /* or 0xEF, 0xBF, 0xBD */
 270                         return 0;
 271
 272                 o = 1;
 273                 ptr++;
 274         }
 275
 276         *s = ptr;
 277         return o;
 278 }
 279
 280 /* sanitize given string and replace all invalid UTF-8 sequences with "?" */
 281 char * utf8(const char *s, unsigned int l)
 282 {
 283         struct template_buffer *buf = buf_init(l);
 284         unsigned char *ptr = (unsigned char *)s;
 285         unsigned int v, o;
 286
 287         if (!buf)
 288                 return NULL;
 289
 290         for (o = 0; o < l; o++)
 291         {
 292                 /* ascii char */
 293                 if ((*ptr >= 0x01) && (*ptr <= 0x7F))
 294                 {
 295                         if (!buf_putchar(buf, (char)*ptr++))
 296                                 break;
 297                 }
 298
 299                 /* invalid byte or multi byte sequence */
 300                 else
 301                 {
 302                         if (!(v = _validate_utf8(&ptr, l - o, buf)))
 303                                 break;
 304
 305                         o += (v - 1);
 306                 }
 307         }
 308
 309         return buf_destroy(buf);
 310 }
 311
 312 /* Sanitize given string and strip all invalid XML bytes
 313  * Validate UTF-8 sequences
 314  * Escape XML control chars */
 315 char * pcdata(const char *s, unsigned int l)
 316 {
 317         struct template_buffer *buf = buf_init(l);
 318         unsigned char *ptr = (unsigned char *)s;
 319         unsigned int o, v;
 320         char esq[8];
 321         int esl;
 322
 323         if (!buf)
 324                 return NULL;
 325
 326         for (o = 0; o < l; o++)
 327         {
 328                 /* Invalid XML bytes */
 329                 if (((*ptr >= 0x00) && (*ptr <= 0x08)) ||
 330                     ((*ptr >= 0x0B) && (*ptr <= 0x0C)) ||
 331                     ((*ptr >= 0x0E) && (*ptr <= 0x1F)) ||
 332                     (*ptr == 0x7F))
 333                 {
 334                         ptr++;
 335                 }
 336
 337                 /* Escapes */
 338                 else if ((*ptr == 0x26) ||
 339                          (*ptr == 0x27) ||
 340                          (*ptr == 0x22) ||
 341                          (*ptr == 0x3C) ||
 342                          (*ptr == 0x3E))
 343                 {
 344                         esl = snprintf(esq, sizeof(esq), "&#%i;", *ptr);
 345
 346                         if (!buf_append(buf, esq, esl))
 347                                 break;
 348
 349                         ptr++;
 350                 }
 351
 352                 /* ascii char */
 353                 else if (*ptr <= 0x7F)
 354                 {
 355                         buf_putchar(buf, (char)*ptr++);
 356                 }
 357
 358                 /* multi byte sequence */
 359                 else
 360                 {
 361                         if (!(v = _validate_utf8(&ptr, l - o, buf)))
 362                                 break;
 363
 364                         o += (v - 1);
 365                 }
 366         }
 367
 368         return buf_destroy(buf);
 369 }
 370
 371 char * striptags(const char *s, unsigned int l)
 372 {
 373         struct template_buffer *buf = buf_init(l);
 374         unsigned char *ptr = (unsigned char *)s;
 375         unsigned char *end = ptr + l;
 376         unsigned char *tag;
 377         unsigned char prev;
 378         char esq[8];
 379         int esl;
 380
 381         for (prev = ' '; ptr < end; ptr++)
 382         {
 383                 if ((*ptr == '<') && ((ptr + 2) < end) &&
 384                         ((*(ptr + 1) == '/') || isalpha(*(ptr + 1))))
 385                 {
 386                         for (tag = ptr; tag < end; tag++)
 387                         {
 388                                 if (*tag == '>')
 389                                 {
 390                                         if (!isspace(prev))
 391                                                 buf_putchar(buf, ' ');
 392
 393                                         ptr = tag;
 394                                         prev = ' ';
 395                                         break;
 396                                 }
 397                         }
 398                 }
 399                 else if (isspace(*ptr))
 400                 {
 401                         if (!isspace(prev))
 402                                 buf_putchar(buf, *ptr);
 403
 404                         prev = *ptr;
 405                 }
 406                 else
 407                 {
 408                         switch(*ptr)
 409                         {
 410                                 case '"':
 411                                 case '\'':
 412                                 case '<':
 413                                 case '>':
 414                                 case '&':
 415                                         esl = snprintf(esq, sizeof(esq), "&#%i;", *ptr);
 416                                         buf_append(buf, esq, esl);
 417                                         break;
 418
 419                                 default:
 420                                         buf_putchar(buf, *ptr);
 421                                         break;
 422                         }
 423
 424                         prev = *ptr;
 425                 }
 426         }
 427
 428         return buf_destroy(buf);
 429 }
 430
 431 void luastr_escape(struct template_buffer *out, const char *s, unsigned int l,
 432                                    int escape_xml)
 433 {
 434         int esl;
 435         char esq[8];
 436         char *ptr;
 437
 438         for (ptr = (char *)s; ptr < (s + l); ptr++)
 439         {
 440                 switch (*ptr)
 441                 {
 442                 case '\\':
 443                         buf_append(out, "\\\\", 2);
 444                         break;
 445
 446                 case '"':
 447                         if (escape_xml)
 448                                 buf_append(out, "&#34;", 5);
 449                         else
 450                                 buf_append(out, "\\\"", 2);
 451                         break;
 452
 453                 case '\n':
 454                         buf_append(out, "\\n", 2);
 455                         break;
 456
 457                 case '\'':
 458                 case '&':
 459                 case '<':
 460                 case '>':
 461                         if (escape_xml)
 462                         {
 463                                 esl = snprintf(esq, sizeof(esq), "&#%i;", *ptr);
 464                                 buf_append(out, esq, esl);
 465                                 break;
 466                         }
 467
 468                 default:
 469                         buf_putchar(out, *ptr);
 470                 }
 471         }
 472 }
 473
 474 void luastr_translate(struct template_buffer *out, const char *s, unsigned int l,
 475                                           int escape_xml)
 476 {
 477         char *tr;
 478         int trlen;
 479
 480         switch (lmo_translate(s, l, &tr, &trlen))
 481         {
 482                 case 0:
 483                         luastr_escape(out, tr, trlen, escape_xml);
 484                         break;
 485
 486                 case -1:
 487                         luastr_escape(out, s, l, escape_xml);
 488                         break;
 489
 490                 default:
 491                         /* no catalog loaded */
 492                         break;
 493         }
 494 }