1 #include "first.h" 2 3 #include "buffer.h" 4 #include "settings.h" /* BUFFER_MAX_REUSE_SIZE */ 5 6 #include <stdlib.h> 7 #include <string.h> 8 #include <time.h> /* strftime() */ 9 10 static const char hex_chars_lc[] = "0123456789abcdef"; 11 static const char hex_chars_uc[] = "0123456789ABCDEF"; 12 13 /** 14 * init the buffer 15 * 16 */ 17 18 buffer* buffer_init(void) { 19 buffer *b; 20 21 b = malloc(sizeof(*b)); 22 force_assert(b); 23 24 b->ptr = NULL; 25 b->size = 0; 26 b->used = 0; 27 28 return b; 29 } 30 31 buffer *buffer_init_buffer(const buffer *src) { 32 buffer *b = buffer_init(); 33 buffer_copy_buffer(b, src); 34 return b; 35 } 36 37 buffer *buffer_init_string(const char *str) { 38 buffer *b = buffer_init(); 39 buffer_copy_string(b, str); 40 return b; 41 } 42 43 void buffer_free(buffer *b) { 44 if (NULL == b) return; 45 46 free(b->ptr); 47 free(b); 48 } 49 50 void buffer_reset(buffer *b) { 51 if (NULL == b) return; 52 53 /* limit don't reuse buffer larger than ... bytes */ 54 if (b->size > BUFFER_MAX_REUSE_SIZE) { 55 free(b->ptr); 56 b->ptr = NULL; 57 b->size = 0; 58 } else if (b->size > 0) { 59 b->ptr[0] = '\0'; 60 } 61 62 b->used = 0; 63 } 64 65 void buffer_move(buffer *b, buffer *src) { 66 buffer tmp; 67 68 if (NULL == b) { 69 buffer_reset(src); 70 return; 71 } 72 buffer_reset(b); 73 if (NULL == src) return; 74 75 tmp = *src; *src = *b; *b = tmp; 76 } 77 78 #define BUFFER_PIECE_SIZE 64 79 static size_t buffer_align_size(size_t size) { 80 size_t align = BUFFER_PIECE_SIZE - (size % BUFFER_PIECE_SIZE); 81 /* overflow on unsinged size_t is defined to wrap around */ 82 if (size + align < size) return size; 83 return size + align; 84 } 85 86 /* make sure buffer is at least "size" big. discard old data */ 87 static void buffer_alloc(buffer *b, size_t size) { 88 force_assert(NULL != b); 89 if (0 == size) size = 1; 90 91 if (size <= b->size) return; 92 93 if (NULL != b->ptr) free(b->ptr); 94 95 b->used = 0; 96 b->size = buffer_align_size(size); 97 b->ptr = malloc(b->size); 98 99 force_assert(NULL != b->ptr); 100 } 101 102 /* make sure buffer is at least "size" big. keep old data */ 103 static void buffer_realloc(buffer *b, size_t size) { 104 force_assert(NULL != b); 105 if (0 == size) size = 1; 106 107 if (size <= b->size) return; 108 109 b->size = buffer_align_size(size); 110 b->ptr = realloc(b->ptr, b->size); 111 112 force_assert(NULL != b->ptr); 113 } 114 115 116 char* buffer_string_prepare_copy(buffer *b, size_t size) { 117 force_assert(NULL != b); 118 force_assert(size + 1 > size); 119 120 buffer_alloc(b, size + 1); 121 122 b->used = 1; 123 b->ptr[0] = '\0'; 124 125 return b->ptr; 126 } 127 128 char* buffer_string_prepare_append(buffer *b, size_t size) { 129 force_assert(NULL != b); 130 131 if (buffer_string_is_empty(b)) { 132 return buffer_string_prepare_copy(b, size); 133 } else { 134 /* not empty, b->used already includes a terminating 0 */ 135 size_t req_size = b->used + size; 136 137 /* check for overflow: unsigned overflow is defined to wrap around */ 138 force_assert(req_size >= b->used); 139 140 buffer_realloc(b, req_size); 141 142 return b->ptr + b->used - 1; 143 } 144 } 145 146 void buffer_string_set_length(buffer *b, size_t len) { 147 force_assert(NULL != b); 148 force_assert(len + 1 > len); 149 150 buffer_realloc(b, len + 1); 151 152 b->used = len + 1; 153 b->ptr[len] = '\0'; 154 } 155 156 void buffer_commit(buffer *b, size_t size) 157 { 158 force_assert(NULL != b); 159 force_assert(b->size > 0); 160 161 if (0 == b->used) b->used = 1; 162 163 if (size > 0) { 164 /* check for overflow: unsigned overflow is defined to wrap around */ 165 force_assert(b->used + size > b->used); 166 167 force_assert(b->used + size <= b->size); 168 b->used += size; 169 } 170 171 b->ptr[b->used - 1] = '\0'; 172 } 173 174 void buffer_copy_string(buffer *b, const char *s) { 175 buffer_copy_string_len(b, s, NULL != s ? strlen(s) : 0); 176 } 177 178 void buffer_copy_string_len(buffer *b, const char *s, size_t s_len) { 179 force_assert(NULL != b); 180 force_assert(NULL != s || s_len == 0); 181 182 buffer_string_prepare_copy(b, s_len); 183 184 if (0 != s_len) memcpy(b->ptr, s, s_len); 185 186 buffer_commit(b, s_len); 187 } 188 189 void buffer_copy_buffer(buffer *b, const buffer *src) { 190 if (NULL == src || 0 == src->used) { 191 buffer_string_prepare_copy(b, 0); 192 b->used = 0; /* keep special empty state for now */ 193 } else { 194 buffer_copy_string_len(b, src->ptr, buffer_string_length(src)); 195 } 196 } 197 198 void buffer_append_string(buffer *b, const char *s) { 199 buffer_append_string_len(b, s, NULL != s ? strlen(s) : 0); 200 } 201 202 /** 203 * append a string to the end of the buffer 204 * 205 * the resulting buffer is terminated with a '\0' 206 * s is treated as a un-terminated string (a \0 is handled a normal character) 207 * 208 * @param b a buffer 209 * @param s the string 210 * @param s_len size of the string (without the terminating \0) 211 */ 212 213 void buffer_append_string_len(buffer *b, const char *s, size_t s_len) { 214 char *target_buf; 215 216 force_assert(NULL != b); 217 force_assert(NULL != s || s_len == 0); 218 219 target_buf = buffer_string_prepare_append(b, s_len); 220 221 if (0 == s_len) return; /* nothing to append */ 222 223 memcpy(target_buf, s, s_len); 224 225 buffer_commit(b, s_len); 226 } 227 228 void buffer_append_string_buffer(buffer *b, const buffer *src) { 229 if (NULL == src) { 230 buffer_append_string_len(b, NULL, 0); 231 } else { 232 buffer_append_string_len(b, src->ptr, buffer_string_length(src)); 233 } 234 } 235 236 void buffer_append_uint_hex_lc(buffer *b, uintmax_t value) { 237 char *buf; 238 unsigned int shift = 0; 239 240 { 241 uintmax_t copy = value; 242 do { 243 copy >>= 8; 244 shift += 8; /* counting bits */ 245 } while (0 != copy); 246 } 247 248 buf = buffer_string_prepare_append(b, shift >> 2); /*nibbles (4 bits)*/ 249 buffer_commit(b, shift >> 2); /* will fill below */ 250 251 while (shift > 0) { 252 shift -= 4; 253 *(buf++) = hex_chars_lc[(value >> shift) & 0x0F]; 254 } 255 } 256 257 static char* utostr(char * const buf_end, uintmax_t val) { 258 char *cur = buf_end; 259 do { 260 int mod = val % 10; 261 val /= 10; 262 /* prepend digit mod */ 263 *(--cur) = (char) ('0' + mod); 264 } while (0 != val); 265 return cur; 266 } 267 268 static char* itostr(char * const buf_end, intmax_t val) { 269 /* absolute value not defined for INTMAX_MIN, but can take absolute 270 * value of any negative number via twos complement cast to unsigned. 271 * negative sign is prepended after (now unsigned) value is converted 272 * to string */ 273 uintmax_t uval = val >= 0 ? (uintmax_t)val : ((uintmax_t)~val) + 1; 274 char *cur = utostr(buf_end, uval); 275 if (val < 0) *(--cur) = '-'; 276 277 return cur; 278 } 279 280 void buffer_append_int(buffer *b, intmax_t val) { 281 char buf[LI_ITOSTRING_LENGTH]; 282 char* const buf_end = buf + sizeof(buf); 283 char *str; 284 285 force_assert(NULL != b); 286 287 str = itostr(buf_end, val); 288 force_assert(buf_end > str && str >= buf); 289 290 buffer_append_string_len(b, str, buf_end - str); 291 } 292 293 void buffer_copy_int(buffer *b, intmax_t val) { 294 force_assert(NULL != b); 295 296 b->used = 0; 297 buffer_append_int(b, val); 298 } 299 300 void buffer_append_strftime(buffer *b, const char *format, const struct tm *tm) { 301 size_t r; 302 char* buf; 303 force_assert(NULL != b); 304 force_assert(NULL != tm); 305 306 if (NULL == format || '\0' == format[0]) { 307 /* empty format */ 308 buffer_string_prepare_append(b, 0); 309 return; 310 } 311 312 buf = buffer_string_prepare_append(b, 255); 313 r = strftime(buf, buffer_string_space(b), format, tm); 314 315 /* 0 (in some apis buffer_string_space(b)) signals the string may have 316 * been too small; but the format could also just have lead to an empty 317 * string 318 */ 319 if (0 == r || r >= buffer_string_space(b)) { 320 /* give it a second try with a larger string */ 321 buf = buffer_string_prepare_append(b, 4095); 322 r = strftime(buf, buffer_string_space(b), format, tm); 323 } 324 325 if (r >= buffer_string_space(b)) r = 0; 326 327 buffer_commit(b, r); 328 } 329 330 331 void li_itostrn(char *buf, size_t buf_len, intmax_t val) { 332 char p_buf[LI_ITOSTRING_LENGTH]; 333 char* const p_buf_end = p_buf + sizeof(p_buf); 334 char* str = p_buf_end - 1; 335 *str = '\0'; 336 337 str = itostr(str, val); 338 force_assert(p_buf_end > str && str >= p_buf); 339 340 force_assert(buf_len >= (size_t) (p_buf_end - str)); 341 memcpy(buf, str, p_buf_end - str); 342 } 343 344 void li_utostrn(char *buf, size_t buf_len, uintmax_t val) { 345 char p_buf[LI_ITOSTRING_LENGTH]; 346 char* const p_buf_end = p_buf + sizeof(p_buf); 347 char* str = p_buf_end - 1; 348 *str = '\0'; 349 350 str = utostr(str, val); 351 force_assert(p_buf_end > str && str >= p_buf); 352 353 force_assert(buf_len >= (size_t) (p_buf_end - str)); 354 memcpy(buf, str, p_buf_end - str); 355 } 356 357 #define li_ntox_lc(n) ((n) <= 9 ? (n) + '0' : (n) + 'a' - 10) 358 359 char int2hex(char c) { 360 /*return li_ntox_lc(c & 0xF);*/ 361 return hex_chars_lc[(c & 0x0F)]; 362 } 363 364 /* c (char) and n (nibble) MUST be unsigned integer types */ 365 #define li_cton(c,n) \ 366 (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) 367 368 /* converts hex char (0-9, A-Z, a-z) to decimal. 369 * returns 0xFF on invalid input. 370 */ 371 char hex2int(unsigned char hex) { 372 unsigned char n; 373 return li_cton(hex,n) ? (char)n : 0xFF; 374 } 375 376 /** 377 * check if two buffer contain the same data 378 * 379 * HISTORY: this function was pretty much optimized, but didn't handled 380 * alignment properly. 381 */ 382 383 int buffer_is_equal(const buffer *a, const buffer *b) { 384 force_assert(NULL != a && NULL != b); 385 386 if (a->used != b->used) return 0; 387 if (a->used == 0) return 1; 388 389 return (0 == memcmp(a->ptr, b->ptr, a->used)); 390 } 391 392 int buffer_is_equal_string(const buffer *a, const char *s, size_t b_len) { 393 force_assert(NULL != a && NULL != s); 394 force_assert(b_len + 1 > b_len); 395 396 if (a->used != b_len + 1) return 0; 397 if (0 != memcmp(a->ptr, s, b_len)) return 0; 398 if ('\0' != a->ptr[a->used-1]) return 0; 399 400 return 1; 401 } 402 403 /* buffer_is_equal_caseless_string(b, CONST_STR_LEN("value")) */ 404 int buffer_is_equal_caseless_string(const buffer *a, const char *s, size_t b_len) { 405 force_assert(NULL != a); 406 if (a->used != b_len + 1) return 0; 407 force_assert('\0' == a->ptr[a->used - 1]); 408 409 return (0 == strcasecmp(a->ptr, s)); 410 } 411 412 int buffer_caseless_compare(const char *a, size_t a_len, const char *b, size_t b_len) { 413 size_t const len = (a_len < b_len) ? a_len : b_len; 414 size_t i; 415 416 for (i = 0; i < len; ++i) { 417 unsigned char ca = a[i], cb = b[i]; 418 if (ca == cb) continue; 419 420 /* always lowercase for transitive results */ 421 if (ca >= 'A' && ca <= 'Z') ca |= 32; 422 if (cb >= 'A' && cb <= 'Z') cb |= 32; 423 424 if (ca == cb) continue; 425 return ((int)ca) - ((int)cb); 426 } 427 if (a_len == b_len) return 0; 428 return a_len < b_len ? -1 : 1; 429 } 430 431 int buffer_is_equal_right_len(const buffer *b1, const buffer *b2, size_t len) { 432 /* no len -> equal */ 433 if (len == 0) return 1; 434 435 /* len > 0, but empty buffers -> not equal */ 436 if (b1->used == 0 || b2->used == 0) return 0; 437 438 /* buffers too small -> not equal */ 439 if (b1->used - 1 < len || b2->used - 1 < len) return 0; 440 441 return 0 == memcmp(b1->ptr + b1->used - 1 - len, b2->ptr + b2->used - 1 - len, len); 442 } 443 444 445 void li_tohex_lc(char *buf, size_t buf_len, const char *s, size_t s_len) { 446 force_assert(2 * s_len > s_len); 447 force_assert(2 * s_len < buf_len); 448 449 for (size_t i = 0; i < s_len; ++i) { 450 buf[2*i] = hex_chars_lc[(s[i] >> 4) & 0x0F]; 451 buf[2*i+1] = hex_chars_lc[s[i] & 0x0F]; 452 } 453 buf[2*s_len] = '\0'; 454 } 455 456 void li_tohex_uc(char *buf, size_t buf_len, const char *s, size_t s_len) { 457 force_assert(2 * s_len > s_len); 458 force_assert(2 * s_len < buf_len); 459 460 for (size_t i = 0; i < s_len; ++i) { 461 buf[2*i] = hex_chars_uc[(s[i] >> 4) & 0x0F]; 462 buf[2*i+1] = hex_chars_uc[s[i] & 0x0F]; 463 } 464 buf[2*s_len] = '\0'; 465 } 466 467 468 void buffer_substr_replace (buffer * const b, const size_t offset, 469 const size_t len, const buffer * const replace) 470 { 471 const size_t blen = buffer_string_length(b); 472 const size_t rlen = buffer_string_length(replace); 473 474 if (rlen > len) { 475 buffer_string_set_length(b, blen-len+rlen); 476 memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len); 477 } 478 479 memcpy(b->ptr+offset, replace->ptr, rlen); 480 481 if (rlen < len) { 482 memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len); 483 buffer_string_set_length(b, blen-len+rlen); 484 } 485 } 486 487 488 void buffer_append_string_encoded_hex_lc(buffer *b, const char *s, size_t len) { 489 unsigned char * const p = 490 (unsigned char*) buffer_string_prepare_append(b, len*2); 491 buffer_commit(b, len*2); /* fill below */ 492 for (size_t i = 0; i < len; ++i) { 493 p[(i<<1)] = hex_chars_lc[(s[i] >> 4) & 0x0F]; 494 p[(i<<1)+1] = hex_chars_lc[(s[i]) & 0x0F]; 495 } 496 } 497 498 void buffer_append_string_encoded_hex_uc(buffer *b, const char *s, size_t len) { 499 unsigned char * const p = 500 (unsigned char*) buffer_string_prepare_append(b, len*2); 501 buffer_commit(b, len*2); /* fill below */ 502 for (size_t i = 0; i < len; ++i) { 503 p[(i<<1)] = hex_chars_uc[(s[i] >> 4) & 0x0F]; 504 p[(i<<1)+1] = hex_chars_uc[(s[i]) & 0x0F]; 505 } 506 } 507 508 509 /* everything except: ! ( ) * - . 0-9 A-Z _ a-z */ 510 static const char encoded_chars_rel_uri_part[] = { 511 /* 512 0 1 2 3 4 5 6 7 8 9 A B C D E F 513 */ 514 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 515 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 516 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, /* 20 - 2F space " # $ % & ' + , / */ 517 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, /* 30 - 3F : ; < = > ? */ 518 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F @ */ 519 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, /* 50 - 5F [ \ ] ^ */ 520 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 521 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, /* 70 - 7F { | } DEL */ 522 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 523 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 524 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 525 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 526 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 527 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 528 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 529 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 530 }; 531 532 /* everything except: ! ( ) * - . / 0-9 A-Z _ a-z */ 533 static const char encoded_chars_rel_uri[] = { 534 /* 535 0 1 2 3 4 5 6 7 8 9 A B C D E F 536 */ 537 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 538 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 539 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, /* 20 - 2F space " # $ % & ' + , */ 540 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, /* 30 - 3F : ; < = > ? */ 541 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F @ */ 542 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, /* 50 - 5F [ \ ] ^ */ 543 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 544 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, /* 70 - 7F { | } DEL */ 545 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 546 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 547 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 548 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 549 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 550 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 551 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 552 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 553 }; 554 555 static const char encoded_chars_html[] = { 556 /* 557 0 1 2 3 4 5 6 7 8 9 A B C D E F 558 */ 559 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 560 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 561 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, /* 20 - 2F " & ' */ 562 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, /* 30 - 3F < > */ 563 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F */ 564 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 50 - 5F */ 565 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 566 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, /* 70 - 7F DEL */ 567 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 568 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 569 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 570 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 571 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 572 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 573 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 574 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 575 }; 576 577 static const char encoded_chars_minimal_xml[] = { 578 /* 579 0 1 2 3 4 5 6 7 8 9 A B C D E F 580 */ 581 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 582 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 583 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, /* 20 - 2F " & ' */ 584 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, /* 30 - 3F < > */ 585 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F */ 586 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 50 - 5F */ 587 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 588 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, /* 70 - 7F DEL */ 589 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */ 590 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */ 591 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */ 592 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */ 593 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */ 594 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */ 595 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */ 596 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */ 597 }; 598 599 600 601 void buffer_append_string_encoded(buffer *b, const char *s, size_t s_len, buffer_encoding_t encoding) { 602 unsigned char *ds, *d; 603 size_t d_len, ndx; 604 const char *map = NULL; 605 606 force_assert(NULL != b); 607 force_assert(NULL != s || 0 == s_len); 608 609 if (0 == s_len) return; 610 611 switch(encoding) { 612 case ENCODING_REL_URI: 613 map = encoded_chars_rel_uri; 614 break; 615 case ENCODING_REL_URI_PART: 616 map = encoded_chars_rel_uri_part; 617 break; 618 case ENCODING_HTML: 619 map = encoded_chars_html; 620 break; 621 case ENCODING_MINIMAL_XML: 622 map = encoded_chars_minimal_xml; 623 break; 624 } 625 626 force_assert(NULL != map); 627 628 /* count to-be-encoded-characters */ 629 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 630 if (map[*ds]) { 631 switch(encoding) { 632 case ENCODING_REL_URI: 633 case ENCODING_REL_URI_PART: 634 d_len += 3; 635 break; 636 case ENCODING_HTML: 637 case ENCODING_MINIMAL_XML: 638 d_len += 6; 639 break; 640 } 641 } else { 642 d_len++; 643 } 644 } 645 646 d = (unsigned char*) buffer_string_prepare_append(b, d_len); 647 buffer_commit(b, d_len); /* fill below */ 648 force_assert('\0' == *d); 649 650 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 651 if (map[*ds]) { 652 switch(encoding) { 653 case ENCODING_REL_URI: 654 case ENCODING_REL_URI_PART: 655 d[d_len++] = '%'; 656 d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F]; 657 d[d_len++] = hex_chars_uc[(*ds) & 0x0F]; 658 break; 659 case ENCODING_HTML: 660 case ENCODING_MINIMAL_XML: 661 d[d_len++] = '&'; 662 d[d_len++] = '#'; 663 d[d_len++] = 'x'; 664 d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F]; 665 d[d_len++] = hex_chars_uc[(*ds) & 0x0F]; 666 d[d_len++] = ';'; 667 break; 668 } 669 } else { 670 d[d_len++] = *ds; 671 } 672 } 673 } 674 675 void buffer_append_string_c_escaped(buffer *b, const char *s, size_t s_len) { 676 unsigned char *ds, *d; 677 size_t d_len, ndx; 678 679 force_assert(NULL != b); 680 force_assert(NULL != s || 0 == s_len); 681 682 if (0 == s_len) return; 683 684 /* count to-be-encoded-characters */ 685 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 686 if ((*ds < 0x20) /* control character */ 687 || (*ds >= 0x7f)) { /* DEL + non-ASCII characters */ 688 switch (*ds) { 689 case '\t': 690 case '\r': 691 case '\n': 692 d_len += 2; 693 break; 694 default: 695 d_len += 4; /* \xCC */ 696 break; 697 } 698 } else { 699 d_len++; 700 } 701 } 702 703 d = (unsigned char*) buffer_string_prepare_append(b, d_len); 704 buffer_commit(b, d_len); /* fill below */ 705 force_assert('\0' == *d); 706 707 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 708 if ((*ds < 0x20) /* control character */ 709 || (*ds >= 0x7f)) { /* DEL + non-ASCII characters */ 710 d[d_len++] = '\\'; 711 switch (*ds) { 712 case '\t': 713 d[d_len++] = 't'; 714 break; 715 case '\r': 716 d[d_len++] = 'r'; 717 break; 718 case '\n': 719 d[d_len++] = 'n'; 720 break; 721 default: 722 d[d_len++] = 'x'; 723 d[d_len++] = hex_chars_lc[((*ds) >> 4) & 0x0F]; 724 d[d_len++] = hex_chars_lc[(*ds) & 0x0F]; 725 break; 726 } 727 } else { 728 d[d_len++] = *ds; 729 } 730 } 731 } 732 733 734 void buffer_copy_string_encoded_cgi_varnames(buffer *b, const char *s, size_t s_len, int is_http_header) { 735 size_t i, j; 736 737 force_assert(NULL != b); 738 force_assert(NULL != s || 0 == s_len); 739 740 buffer_reset(b); 741 742 if (is_http_header && NULL != s && 0 != strcasecmp(s, "CONTENT-TYPE")) { 743 buffer_string_prepare_append(b, s_len + 5); 744 buffer_copy_string_len(b, CONST_STR_LEN("HTTP_")); 745 } else { 746 buffer_string_prepare_append(b, s_len); 747 } 748 749 j = buffer_string_length(b); 750 for (i = 0; i < s_len; ++i) { 751 unsigned char cr = s[i]; 752 if (light_isalpha(cr)) { 753 /* upper-case */ 754 cr &= ~32; 755 } else if (!light_isdigit(cr)) { 756 cr = '_'; 757 } 758 b->ptr[j++] = cr; 759 } 760 b->used = j; 761 b->ptr[b->used++] = '\0'; 762 } 763 764 /* decodes url-special-chars inplace. 765 * replaces non-printable characters with '_' 766 */ 767 768 static void buffer_urldecode_internal(buffer *url, int is_query) { 769 unsigned char high, low; 770 char *src; 771 char *dst; 772 773 force_assert(NULL != url); 774 if (buffer_string_is_empty(url)) return; 775 776 force_assert('\0' == url->ptr[url->used-1]); 777 778 src = (char*) url->ptr; 779 780 while ('\0' != *src) { 781 if ('%' == *src) break; 782 if (is_query && '+' == *src) *src = ' '; 783 src++; 784 } 785 dst = src; 786 787 while ('\0' != *src) { 788 if (is_query && *src == '+') { 789 *dst = ' '; 790 } else if (*src == '%') { 791 *dst = '%'; 792 793 high = hex2int(*(src + 1)); 794 if (0xFF != high) { 795 low = hex2int(*(src + 2)); 796 if (0xFF != low) { 797 high = (high << 4) | low; 798 799 /* map control-characters out */ 800 if (high < 32 || high == 127) high = '_'; 801 802 *dst = high; 803 src += 2; 804 } 805 } 806 } else { 807 *dst = *src; 808 } 809 810 dst++; 811 src++; 812 } 813 814 *dst = '\0'; 815 url->used = (dst - url->ptr) + 1; 816 } 817 818 void buffer_urldecode_path(buffer *url) { 819 buffer_urldecode_internal(url, 0); 820 } 821 822 void buffer_urldecode_query(buffer *url) { 823 buffer_urldecode_internal(url, 1); 824 } 825 826 /* - special case: empty string returns empty string 827 * - on windows or cygwin: replace \ with / 828 * - strip leading spaces 829 * - prepends "/" if not present already 830 * - resolve "/../", "//" and "/./" the usual way: 831 * the first one removes a preceding component, the other two 832 * get compressed to "/". 833 * - "/." and "/.." at the end are similar, but always leave a trailing 834 * "/" 835 * 836 * /blah/.. gets / 837 * /blah/../foo gets /foo 838 * /abc/./xyz gets /abc/xyz 839 * /abc//xyz gets /abc/xyz 840 * 841 * NOTE: src and dest can point to the same buffer, in which case, 842 * the operation is performed in-place. 843 */ 844 845 void buffer_path_simplify(buffer *dest, buffer *src) 846 { 847 /* current character, the one before, and the one before that from input */ 848 char c, pre1, pre2; 849 char *start, *slash, *walk, *out; 850 851 force_assert(NULL != dest && NULL != src); 852 853 if (buffer_string_is_empty(src)) { 854 buffer_string_prepare_copy(dest, 0); 855 return; 856 } 857 858 force_assert('\0' == src->ptr[src->used-1]); 859 860 /* might need one character more for the '/' prefix */ 861 if (src == dest) { 862 buffer_string_prepare_append(dest, 1); 863 } else { 864 buffer_string_prepare_copy(dest, buffer_string_length(src) + 1); 865 } 866 867 #if defined(__WIN32) || defined(__CYGWIN__) 868 /* cygwin is treating \ and / the same, so we have to that too */ 869 { 870 char *p; 871 for (p = src->ptr; *p; p++) { 872 if (*p == '\\') *p = '/'; 873 } 874 } 875 #endif 876 877 walk = src->ptr; 878 start = dest->ptr; 879 out = dest->ptr; 880 slash = dest->ptr; 881 882 /* skip leading spaces */ 883 while (*walk == ' ') { 884 walk++; 885 } 886 887 pre1 = 0; 888 c = *(walk++); 889 /* prefix with '/' if not already present */ 890 if (c != '/') { 891 pre1 = '/'; 892 *(out++) = '/'; 893 } 894 895 while (c != '\0') { 896 /* assert((src != dest || out <= walk) && slash <= out); */ 897 /* the following comments about out and walk are only interesting if 898 * src == dest; otherwise the memory areas don't overlap anyway. 899 */ 900 pre2 = pre1; 901 pre1 = c; 902 903 /* possibly: out == walk - need to read first */ 904 c = *walk; 905 *out = pre1; 906 907 out++; 908 walk++; 909 /* (out <= walk) still true; also now (slash < out) */ 910 911 if (c == '/' || c == '\0') { 912 const size_t toklen = out - slash; 913 if (toklen == 3 && pre2 == '.' && pre1 == '.') { 914 /* "/../" or ("/.." at end of string) */ 915 out = slash; 916 /* if there is something before "/..", there is at least one 917 * component, which needs to be removed */ 918 if (out > start) { 919 out--; 920 while (out > start && *out != '/') out--; 921 } 922 923 /* don't kill trailing '/' at end of path */ 924 if (c == '\0') out++; 925 /* slash < out before, so out_new <= slash + 1 <= out_before <= walk */ 926 } else if (toklen == 1 || (pre2 == '/' && pre1 == '.')) { 927 /* "//" or "/./" or (("/" or "/.") at end of string) */ 928 out = slash; 929 /* don't kill trailing '/' at end of path */ 930 if (c == '\0') out++; 931 /* slash < out before, so out_new <= slash + 1 <= out_before <= walk */ 932 } 933 934 slash = out; 935 } 936 } 937 938 buffer_string_set_length(dest, out - start); 939 } 940 941 void buffer_to_lower(buffer *b) { 942 size_t i; 943 944 for (i = 0; i < b->used; ++i) { 945 char c = b->ptr[i]; 946 if (c >= 'A' && c <= 'Z') b->ptr[i] |= 0x20; 947 } 948 } 949 950 951 void buffer_to_upper(buffer *b) { 952 size_t i; 953 954 for (i = 0; i < b->used; ++i) { 955 char c = b->ptr[i]; 956 if (c >= 'a' && c <= 'z') b->ptr[i] &= ~0x20; 957 } 958 } 959 960 961 #include <stdio.h> 962 963 #ifdef HAVE_LIBUNWIND 964 # define UNW_LOCAL_ONLY 965 # include <libunwind.h> 966 967 static void print_backtrace(FILE *file) { 968 unw_cursor_t cursor; 969 unw_context_t context; 970 int ret; 971 unsigned int frame = 0; 972 973 if (0 != (ret = unw_getcontext(&context))) goto error; 974 if (0 != (ret = unw_init_local(&cursor, &context))) goto error; 975 976 fprintf(file, "Backtrace:\n"); 977 978 while (0 < (ret = unw_step(&cursor))) { 979 unw_word_t proc_ip = 0; 980 unw_proc_info_t procinfo; 981 char procname[256]; 982 unw_word_t proc_offset = 0; 983 984 if (0 != (ret = unw_get_reg(&cursor, UNW_REG_IP, &proc_ip))) goto error; 985 986 if (0 == proc_ip) { 987 /* without an IP the other functions are useless; unw_get_proc_name would return UNW_EUNSPEC */ 988 ++frame; 989 fprintf(file, "%u: (nil)\n", frame); 990 continue; 991 } 992 993 if (0 != (ret = unw_get_proc_info(&cursor, &procinfo))) goto error; 994 995 if (0 != (ret = unw_get_proc_name(&cursor, procname, sizeof(procname), &proc_offset))) { 996 switch (-ret) { 997 case UNW_ENOMEM: 998 memset(procname + sizeof(procname) - 4, '.', 3); 999 procname[sizeof(procname) - 1] = '\0'; 1000 break; 1001 case UNW_ENOINFO: 1002 procname[0] = '?'; 1003 procname[1] = '\0'; 1004 proc_offset = 0; 1005 break; 1006 default: 1007 snprintf(procname, sizeof(procname), "?? (unw_get_proc_name error %d)", -ret); 1008 break; 1009 } 1010 } 1011 1012 ++frame; 1013 fprintf(file, "%u: %s (+0x%x) [%p]\n", 1014 frame, 1015 procname, 1016 (unsigned int) proc_offset, 1017 (void*)(uintptr_t)proc_ip); 1018 } 1019 1020 if (0 != ret) goto error; 1021 1022 return; 1023 1024 error: 1025 fprintf(file, "Error while generating backtrace: unwind error %i\n", (int) -ret); 1026 } 1027 #else 1028 static void print_backtrace(FILE *file) { 1029 UNUSED(file); 1030 } 1031 #endif 1032 1033 void log_failed_assert(const char *filename, unsigned int line, const char *msg) { 1034 /* can't use buffer here; could lead to recursive assertions */ 1035 fprintf(stderr, "%s.%u: %s\n", filename, line, msg); 1036 print_backtrace(stderr); 1037 fflush(stderr); 1038 abort(); 1039 } 1040