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 __attribute_cold__ 51 static void buffer_free_ptr(buffer *b) { 52 free(b->ptr); 53 b->ptr = NULL; 54 b->used = 0; 55 b->size = 0; 56 } 57 58 void buffer_reset(buffer *b) { 59 force_assert(NULL != b); 60 b->used = 0; 61 /* release buffer larger than ... bytes */ 62 if (b->size > BUFFER_MAX_REUSE_SIZE) buffer_free_ptr(b); 63 } 64 65 void buffer_move(buffer *b, buffer *src) { 66 buffer tmp; 67 force_assert(NULL != b); 68 force_assert(NULL != src); 69 70 buffer_clear(b); 71 tmp = *src; *src = *b; *b = tmp; 72 } 73 74 /* make sure buffer is at least "size" big + 1 for '\0'. keep old data */ 75 __attribute_cold__ 76 static void buffer_realloc(buffer *b, size_t len) { 77 #define BUFFER_PIECE_SIZE 64uL /*(must be power-of-2)*/ 78 const size_t sz = (len + 1 + BUFFER_PIECE_SIZE-1) & ~(BUFFER_PIECE_SIZE-1); 79 force_assert(sz > len); 80 81 b->size = sz; 82 b->ptr = realloc(b->ptr, sz); 83 84 force_assert(NULL != b->ptr); 85 } 86 87 __attribute_cold__ 88 static void buffer_alloc_replace(buffer *b, size_t size) { 89 /*(discard old data so realloc() does not copy)*/ 90 if (NULL != b->ptr) { 91 free(b->ptr); 92 b->ptr = NULL; 93 } 94 buffer_realloc(b, size); 95 } 96 97 char* buffer_string_prepare_copy(buffer *b, size_t size) { 98 force_assert(NULL != b); 99 100 if (size >= b->size) buffer_alloc_replace(b, size); 101 102 b->used = 0; 103 return b->ptr; 104 } 105 106 char* buffer_string_prepare_append(buffer *b, size_t size) { 107 force_assert(NULL != b); 108 109 if (b->used && size < b->size - b->used) 110 return b->ptr + b->used - 1; 111 112 if (buffer_string_is_empty(b)) { 113 return buffer_string_prepare_copy(b, size); 114 } else { 115 /* not empty, b->used already includes a terminating 0 */ 116 size_t req_size = b->used + size; 117 118 /* check for overflow: unsigned overflow is defined to wrap around */ 119 force_assert(req_size >= b->used); 120 121 buffer_realloc(b, req_size); 122 123 return b->ptr + b->used - 1; 124 } 125 } 126 127 void buffer_string_set_length(buffer *b, size_t len) { 128 force_assert(NULL != b); 129 130 if (len >= b->size) buffer_realloc(b, len); 131 132 b->used = len + 1; 133 b->ptr[len] = '\0'; 134 } 135 136 void buffer_commit(buffer *b, size_t size) 137 { 138 force_assert(NULL != b); 139 force_assert(b->size > 0); 140 141 if (0 == b->used) b->used = 1; 142 143 if (size > 0) { 144 /* check for overflow: unsigned overflow is defined to wrap around */ 145 size_t sz = b->used + size; 146 force_assert(sz > b->used); 147 force_assert(sz <= b->size); 148 b->used = sz; 149 } 150 151 b->ptr[b->used - 1] = '\0'; 152 } 153 154 void buffer_copy_string(buffer *b, const char *s) { 155 buffer_copy_string_len(b, s, NULL != s ? strlen(s) : 0); 156 } 157 158 void buffer_copy_string_len(buffer *b, const char *s, size_t s_len) { 159 force_assert(NULL != b); 160 force_assert(NULL != s || s_len == 0); 161 162 if (s_len >= b->size) buffer_string_prepare_copy(b, s_len); 163 164 if (0 != s_len) memcpy(b->ptr, s, s_len); /*(s might be NULL)*/ 165 b->ptr[s_len] = '\0'; 166 b->used = s_len + 1; 167 } 168 169 void buffer_append_string(buffer *b, const char *s) { 170 buffer_append_string_len(b, s, NULL != s ? strlen(s) : 0); 171 } 172 173 /** 174 * append a string to the end of the buffer 175 * 176 * the resulting buffer is terminated with a '\0' 177 * s is treated as a un-terminated string (a \0 is handled a normal character) 178 * 179 * @param b a buffer 180 * @param s the string 181 * @param s_len size of the string (without the terminating \0) 182 */ 183 184 void buffer_append_string_len(buffer *b, const char *s, size_t s_len) { 185 char *target_buf; 186 187 force_assert(NULL != b); 188 force_assert(NULL != s || s_len == 0); 189 190 target_buf = buffer_string_prepare_append(b, s_len); 191 if (0 == b->used) ++b->used; /*(must include '\0' for append below)*/ 192 193 /*(s might be NULL if 0 == s_len)*/ 194 if (s_len) memcpy(target_buf, s, s_len); 195 target_buf[s_len] = '\0'; 196 b->used += s_len; 197 } 198 199 void buffer_append_path_len(buffer *b, const char *a, size_t alen) { 200 size_t blen = buffer_string_length(b); 201 int aslash = (alen && a[0] == '/'); 202 buffer_string_prepare_append(b, alen+2); /*(+ '/' and + '\0' if 0 == blen)*/ 203 if (blen && b->ptr[blen-1] == '/') { 204 if (aslash) --b->used; 205 } 206 else { 207 if (!b->used) ++b->used; 208 if (!aslash) b->ptr[++b->used - 2] = '/'; 209 } 210 memcpy(b->ptr+b->used-1, a, alen); 211 b->ptr[(b->used += alen)-1] = '\0'; 212 } 213 214 void buffer_append_uint_hex_lc(buffer *b, uintmax_t value) { 215 char *buf; 216 unsigned int shift = 0; 217 218 { 219 uintmax_t copy = value; 220 do { 221 copy >>= 8; 222 shift += 8; /* counting bits */ 223 } while (0 != copy); 224 } 225 226 buf = buffer_string_prepare_append(b, shift >> 2); /*nibbles (4 bits)*/ 227 buffer_commit(b, shift >> 2); /* will fill below */ 228 229 while (shift > 0) { 230 shift -= 4; 231 *(buf++) = hex_chars_lc[(value >> shift) & 0x0F]; 232 } 233 } 234 235 static char* utostr(char * const buf_end, uintmax_t val) { 236 char *cur = buf_end; 237 do { 238 int mod = val % 10; 239 val /= 10; 240 /* prepend digit mod */ 241 *(--cur) = (char) ('0' + mod); 242 } while (0 != val); 243 return cur; 244 } 245 246 static char* itostr(char * const buf_end, intmax_t val) { 247 /* absolute value not defined for INTMAX_MIN, but can take absolute 248 * value of any negative number via twos complement cast to unsigned. 249 * negative sign is prepended after (now unsigned) value is converted 250 * to string */ 251 uintmax_t uval = val >= 0 ? (uintmax_t)val : ((uintmax_t)~val) + 1; 252 char *cur = utostr(buf_end, uval); 253 if (val < 0) *(--cur) = '-'; 254 255 return cur; 256 } 257 258 void buffer_append_int(buffer *b, intmax_t val) { 259 char buf[LI_ITOSTRING_LENGTH]; 260 char* const buf_end = buf + sizeof(buf); 261 char *str; 262 263 force_assert(NULL != b); 264 265 str = itostr(buf_end, val); 266 force_assert(buf_end > str && str >= buf); 267 268 buffer_append_string_len(b, str, buf_end - str); 269 } 270 271 void buffer_copy_int(buffer *b, intmax_t val) { 272 force_assert(NULL != b); 273 274 b->used = 0; 275 buffer_append_int(b, val); 276 } 277 278 void buffer_append_strftime(buffer *b, const char *format, const struct tm *tm) { 279 size_t r; 280 char* buf; 281 force_assert(NULL != b); 282 force_assert(NULL != format); 283 force_assert(NULL != tm); 284 285 buf = buffer_string_prepare_append(b, 255); 286 r = strftime(buf, buffer_string_space(b), format, tm); 287 288 /* 0 (in some apis buffer_string_space(b)) signals the string may have 289 * been too small; but the format could also just have lead to an empty 290 * string 291 */ 292 if (0 == r || r >= buffer_string_space(b)) { 293 /* give it a second try with a larger string */ 294 buf = buffer_string_prepare_append(b, 4095); 295 r = strftime(buf, buffer_string_space(b), format, tm); 296 } 297 298 if (r >= buffer_string_space(b)) r = 0; 299 300 buffer_commit(b, r); 301 } 302 303 304 void li_itostrn(char *buf, size_t buf_len, intmax_t val) { 305 char p_buf[LI_ITOSTRING_LENGTH]; 306 char* const p_buf_end = p_buf + sizeof(p_buf); 307 char* str = p_buf_end - 1; 308 *str = '\0'; 309 310 str = itostr(str, val); 311 force_assert(p_buf_end > str && str >= p_buf); 312 313 force_assert(buf_len >= (size_t) (p_buf_end - str)); 314 memcpy(buf, str, p_buf_end - str); 315 } 316 317 void li_utostrn(char *buf, size_t buf_len, uintmax_t val) { 318 char p_buf[LI_ITOSTRING_LENGTH]; 319 char* const p_buf_end = p_buf + sizeof(p_buf); 320 char* str = p_buf_end - 1; 321 *str = '\0'; 322 323 str = utostr(str, val); 324 force_assert(p_buf_end > str && str >= p_buf); 325 326 force_assert(buf_len >= (size_t) (p_buf_end - str)); 327 memcpy(buf, str, p_buf_end - str); 328 } 329 330 #define li_ntox_lc(n) ((n) <= 9 ? (n) + '0' : (n) + 'a' - 10) 331 332 char int2hex(char c) { 333 /*return li_ntox_lc(c & 0xF);*/ 334 return hex_chars_lc[(c & 0x0F)]; 335 } 336 337 /* c (char) and n (nibble) MUST be unsigned integer types */ 338 #define li_cton(c,n) \ 339 (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) 340 341 /* converts hex char (0-9, A-Z, a-z) to decimal. 342 * returns 0xFF on invalid input. 343 */ 344 char hex2int(unsigned char hex) { 345 unsigned char n; 346 return li_cton(hex,n) ? (char)n : 0xFF; 347 } 348 349 350 int buffer_eq_icase_ssn(const char * const a, const char * const b, const size_t len) { 351 for (size_t i = 0; i < len; ++i) { 352 unsigned int ca = ((unsigned char *)a)[i]; 353 unsigned int cb = ((unsigned char *)b)[i]; 354 if (ca != cb) { 355 ca |= 0x20; 356 cb |= 0x20; 357 if (ca != cb) return 0; 358 if (ca < 'a' || 'z' < ca) return 0; 359 if (cb < 'a' || 'z' < cb) return 0; 360 } 361 } 362 return 1; 363 } 364 365 int buffer_eq_icase_ss(const char * const a, const size_t alen, const char * const b, const size_t blen) { 366 /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */ 367 return (alen == blen && buffer_eq_icase_ssn(a, b, blen)); 368 } 369 370 int buffer_eq_icase_slen(const buffer * const b, const char * const s, const size_t slen) { 371 /* Note: b must be initialized, i.e. 0 != b->used; uninitialized is not eq*/ 372 /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */ 373 return (b->used == slen + 1 && buffer_eq_icase_ssn(b->ptr, s, slen)); 374 } 375 376 int buffer_eq_slen(const buffer * const b, const char * const s, const size_t slen) { 377 /* Note: b must be initialized, i.e. 0 != b->used; uninitialized is not eq*/ 378 /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */ 379 return (b->used == slen + 1 && 0 == memcmp(b->ptr, s, slen)); 380 } 381 382 383 /** 384 * check if two buffer contain the same data 385 */ 386 387 int buffer_is_equal(const buffer *a, const buffer *b) { 388 force_assert(NULL != a && NULL != b); 389 390 /* 1 = equal; 0 = not equal */ 391 return (a->used == b->used && 0 == memcmp(a->ptr, b->ptr, a->used)); 392 } 393 394 int buffer_is_equal_string(const buffer *a, const char *s, size_t b_len) { 395 force_assert(NULL != a && NULL != s); 396 force_assert(b_len + 1 > b_len); 397 398 /* 1 = equal; 0 = not equal */ 399 return (a->used == b_len + 1 && 0 == memcmp(a->ptr, s, b_len)); 400 } 401 402 /* buffer_is_equal_caseless_string(b, CONST_STR_LEN("value")) */ 403 int buffer_is_equal_caseless_string(const buffer *a, const char *s, size_t b_len) { 404 force_assert(NULL != a && NULL != s); 405 force_assert(b_len + 1 > b_len); 406 /* 1 = equal; 0 = not equal */ 407 return buffer_eq_icase_slen(a, s, b_len); 408 } 409 410 int buffer_is_equal_right_len(const buffer *b1, const buffer *b2, size_t len) { 411 /* no len -> equal */ 412 if (len == 0) return 1; 413 414 /* len > 0, but empty buffers -> not equal */ 415 if (b1->used == 0 || b2->used == 0) return 0; 416 417 /* buffers too small -> not equal */ 418 if (b1->used - 1 < len || b2->used - 1 < len) return 0; 419 420 return 0 == memcmp(b1->ptr + b1->used - 1 - len, b2->ptr + b2->used - 1 - len, len); 421 } 422 423 424 void li_tohex_lc(char *buf, size_t buf_len, const char *s, size_t s_len) { 425 force_assert(2 * s_len > s_len); 426 force_assert(2 * s_len < buf_len); 427 428 for (size_t i = 0; i < s_len; ++i) { 429 buf[2*i] = hex_chars_lc[(s[i] >> 4) & 0x0F]; 430 buf[2*i+1] = hex_chars_lc[s[i] & 0x0F]; 431 } 432 buf[2*s_len] = '\0'; 433 } 434 435 void li_tohex_uc(char *buf, size_t buf_len, const char *s, size_t s_len) { 436 force_assert(2 * s_len > s_len); 437 force_assert(2 * s_len < buf_len); 438 439 for (size_t i = 0; i < s_len; ++i) { 440 buf[2*i] = hex_chars_uc[(s[i] >> 4) & 0x0F]; 441 buf[2*i+1] = hex_chars_uc[s[i] & 0x0F]; 442 } 443 buf[2*s_len] = '\0'; 444 } 445 446 447 void buffer_substr_replace (buffer * const b, const size_t offset, 448 const size_t len, const buffer * const replace) 449 { 450 const size_t blen = buffer_string_length(b); 451 const size_t rlen = buffer_string_length(replace); 452 453 if (rlen > len) { 454 buffer_string_set_length(b, blen-len+rlen); 455 memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len); 456 } 457 458 memcpy(b->ptr+offset, replace->ptr, rlen); 459 460 if (rlen < len) { 461 memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len); 462 buffer_string_set_length(b, blen-len+rlen); 463 } 464 } 465 466 467 void buffer_append_string_encoded_hex_lc(buffer *b, const char *s, size_t len) { 468 unsigned char * const p = 469 (unsigned char*) buffer_string_prepare_append(b, len*2); 470 buffer_commit(b, len*2); /* fill below */ 471 for (size_t i = 0; i < len; ++i) { 472 p[(i<<1)] = hex_chars_lc[(s[i] >> 4) & 0x0F]; 473 p[(i<<1)+1] = hex_chars_lc[(s[i]) & 0x0F]; 474 } 475 } 476 477 void buffer_append_string_encoded_hex_uc(buffer *b, const char *s, size_t len) { 478 unsigned char * const p = 479 (unsigned char*) buffer_string_prepare_append(b, len*2); 480 buffer_commit(b, len*2); /* fill below */ 481 for (size_t i = 0; i < len; ++i) { 482 p[(i<<1)] = hex_chars_uc[(s[i] >> 4) & 0x0F]; 483 p[(i<<1)+1] = hex_chars_uc[(s[i]) & 0x0F]; 484 } 485 } 486 487 488 /* everything except: ! ( ) * - . 0-9 A-Z _ a-z */ 489 static const char encoded_chars_rel_uri_part[] = { 490 /* 491 0 1 2 3 4 5 6 7 8 9 A B C D E F 492 */ 493 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 494 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 495 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, /* 20 - 2F space " # $ % & ' + , / */ 496 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, /* 30 - 3F : ; < = > ? */ 497 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F @ */ 498 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, /* 50 - 5F [ \ ] ^ */ 499 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 500 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, /* 70 - 7F { | } DEL */ 501 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 502 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 503 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 504 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 505 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 506 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 507 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 508 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 509 }; 510 511 /* everything except: ! ( ) * - . / 0-9 A-Z _ a-z */ 512 static const char encoded_chars_rel_uri[] = { 513 /* 514 0 1 2 3 4 5 6 7 8 9 A B C D E F 515 */ 516 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 517 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 518 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, /* 20 - 2F space " # $ % & ' + , */ 519 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, /* 30 - 3F : ; < = > ? */ 520 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F @ */ 521 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, /* 50 - 5F [ \ ] ^ */ 522 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 523 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, /* 70 - 7F { | } DEL */ 524 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 525 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 526 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 527 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 528 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 529 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 530 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 531 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 532 }; 533 534 static const char encoded_chars_html[] = { 535 /* 536 0 1 2 3 4 5 6 7 8 9 A B C D E F 537 */ 538 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 539 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 540 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, /* 20 - 2F " & ' */ 541 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, /* 30 - 3F < > */ 542 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F */ 543 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 50 - 5F */ 544 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 545 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, /* 70 - 7F DEL */ 546 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 80 - 8F */ 547 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 90 - 9F */ 548 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* A0 - AF */ 549 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* B0 - BF */ 550 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* C0 - CF */ 551 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* D0 - DF */ 552 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* E0 - EF */ 553 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* F0 - FF */ 554 }; 555 556 static const char encoded_chars_minimal_xml[] = { 557 /* 558 0 1 2 3 4 5 6 7 8 9 A B C D E F 559 */ 560 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 00 - 0F control chars */ 561 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 10 - 1F */ 562 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, /* 20 - 2F " & ' */ 563 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, /* 30 - 3F < > */ 564 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 40 - 4F */ 565 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 50 - 5F */ 566 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F ` */ 567 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, /* 70 - 7F DEL */ 568 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */ 569 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */ 570 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */ 571 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */ 572 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */ 573 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */ 574 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */ 575 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */ 576 }; 577 578 579 580 void buffer_append_string_encoded(buffer *b, const char *s, size_t s_len, buffer_encoding_t encoding) { 581 unsigned char *ds, *d; 582 size_t d_len, ndx; 583 const char *map = NULL; 584 585 force_assert(NULL != b); 586 force_assert(NULL != s || 0 == s_len); 587 588 if (0 == s_len) return; 589 590 switch(encoding) { 591 case ENCODING_REL_URI: 592 map = encoded_chars_rel_uri; 593 break; 594 case ENCODING_REL_URI_PART: 595 map = encoded_chars_rel_uri_part; 596 break; 597 case ENCODING_HTML: 598 map = encoded_chars_html; 599 break; 600 case ENCODING_MINIMAL_XML: 601 map = encoded_chars_minimal_xml; 602 break; 603 } 604 605 force_assert(NULL != map); 606 607 /* count to-be-encoded-characters */ 608 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 609 if (map[*ds]) { 610 switch(encoding) { 611 case ENCODING_REL_URI: 612 case ENCODING_REL_URI_PART: 613 d_len += 3; 614 break; 615 case ENCODING_HTML: 616 case ENCODING_MINIMAL_XML: 617 d_len += 6; 618 break; 619 } 620 } else { 621 d_len++; 622 } 623 } 624 625 d = (unsigned char*) buffer_string_prepare_append(b, d_len); 626 buffer_commit(b, d_len); /* fill below */ 627 628 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 629 if (map[*ds]) { 630 switch(encoding) { 631 case ENCODING_REL_URI: 632 case ENCODING_REL_URI_PART: 633 d[d_len++] = '%'; 634 d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F]; 635 d[d_len++] = hex_chars_uc[(*ds) & 0x0F]; 636 break; 637 case ENCODING_HTML: 638 case ENCODING_MINIMAL_XML: 639 d[d_len++] = '&'; 640 d[d_len++] = '#'; 641 d[d_len++] = 'x'; 642 d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F]; 643 d[d_len++] = hex_chars_uc[(*ds) & 0x0F]; 644 d[d_len++] = ';'; 645 break; 646 } 647 } else { 648 d[d_len++] = *ds; 649 } 650 } 651 } 652 653 void buffer_append_string_c_escaped(buffer *b, const char *s, size_t s_len) { 654 unsigned char *ds, *d; 655 size_t d_len, ndx; 656 657 force_assert(NULL != b); 658 force_assert(NULL != s || 0 == s_len); 659 660 if (0 == s_len) return; 661 662 /* count to-be-encoded-characters */ 663 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 664 if ((*ds < 0x20) /* control character */ 665 || (*ds >= 0x7f)) { /* DEL + non-ASCII characters */ 666 switch (*ds) { 667 case '\t': 668 case '\r': 669 case '\n': 670 d_len += 2; 671 break; 672 default: 673 d_len += 4; /* \xCC */ 674 break; 675 } 676 } else { 677 d_len++; 678 } 679 } 680 681 d = (unsigned char*) buffer_string_prepare_append(b, d_len); 682 buffer_commit(b, d_len); /* fill below */ 683 684 for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) { 685 if ((*ds < 0x20) /* control character */ 686 || (*ds >= 0x7f)) { /* DEL + non-ASCII characters */ 687 d[d_len++] = '\\'; 688 switch (*ds) { 689 case '\t': 690 d[d_len++] = 't'; 691 break; 692 case '\r': 693 d[d_len++] = 'r'; 694 break; 695 case '\n': 696 d[d_len++] = 'n'; 697 break; 698 default: 699 d[d_len++] = 'x'; 700 d[d_len++] = hex_chars_lc[((*ds) >> 4) & 0x0F]; 701 d[d_len++] = hex_chars_lc[(*ds) & 0x0F]; 702 break; 703 } 704 } else { 705 d[d_len++] = *ds; 706 } 707 } 708 } 709 710 711 void buffer_copy_string_encoded_cgi_varnames(buffer *b, const char *s, size_t s_len, int is_http_header) { 712 size_t i, j = 0; 713 714 force_assert(NULL != b); 715 force_assert(NULL != s || 0 == s_len); 716 717 buffer_string_prepare_copy(b, s_len + 5); 718 719 if (is_http_header) { 720 if (s_len == 12 && buffer_eq_icase_ssn(s, "Content-Type", 12)) { 721 buffer_copy_string_len(b, CONST_STR_LEN("CONTENT_TYPE")); 722 return; 723 } 724 buffer_copy_string_len(b, CONST_STR_LEN("HTTP_")); 725 j = 5; /* "HTTP_" */ 726 } 727 728 for (i = 0; i < s_len; ++i) { 729 unsigned char cr = s[i]; 730 if (light_isalpha(cr)) { 731 /* upper-case */ 732 cr &= ~32; 733 } else if (!light_isdigit(cr)) { 734 cr = '_'; 735 } 736 b->ptr[j++] = cr; 737 } 738 b->used = j; 739 b->ptr[b->used++] = '\0'; 740 } 741 742 /* decodes url-special-chars inplace. 743 * replaces non-printable characters with '_' 744 */ 745 746 static void buffer_urldecode_internal(buffer *url, int is_query) { 747 unsigned char high, low; 748 char *src; 749 char *dst; 750 751 force_assert(NULL != url); 752 if (buffer_string_is_empty(url)) return; 753 754 force_assert('\0' == url->ptr[url->used-1]); 755 756 src = (char*) url->ptr; 757 758 while ('\0' != *src) { 759 if ('%' == *src) break; 760 if (is_query && '+' == *src) *src = ' '; 761 src++; 762 } 763 dst = src; 764 765 while ('\0' != *src) { 766 if (is_query && *src == '+') { 767 *dst = ' '; 768 } else if (*src == '%') { 769 *dst = '%'; 770 771 high = hex2int(*(src + 1)); 772 if (0xFF != high) { 773 low = hex2int(*(src + 2)); 774 if (0xFF != low) { 775 high = (high << 4) | low; 776 777 /* map control-characters out */ 778 if (high < 32 || high == 127) high = '_'; 779 780 *dst = high; 781 src += 2; 782 } 783 } 784 } else { 785 *dst = *src; 786 } 787 788 dst++; 789 src++; 790 } 791 792 *dst = '\0'; 793 url->used = (dst - url->ptr) + 1; 794 } 795 796 void buffer_urldecode_path(buffer *url) { 797 buffer_urldecode_internal(url, 0); 798 } 799 800 void buffer_urldecode_query(buffer *url) { 801 buffer_urldecode_internal(url, 1); 802 } 803 804 int buffer_is_valid_UTF8(const buffer *b) { 805 /* https://www.w3.org/International/questions/qa-forms-utf-8 */ 806 const unsigned char *c = (unsigned char *)b->ptr; 807 while (*c) { 808 809 /*(note: includes ctrls)*/ 810 if ( c[0] < 0x80 ) { ++c; continue; } 811 812 if ( 0xc2 <= c[0] && c[0] <= 0xdf 813 && 0x80 <= c[1] && c[1] <= 0xbf ) { c+=2; continue; } 814 815 if ( ( ( 0xe0 == c[0] 816 && 0xa0 <= c[1] && c[1] <= 0xbf) 817 || ( 0xe1 <= c[0] && c[0] <= 0xef && c[0] != 0xed 818 && 0x80 <= c[1] && c[1] <= 0xbf) 819 || ( 0xed == c[0] 820 && 0x80 <= c[1] && c[1] <= 0x9f) ) 821 && 0x80 <= c[2] && c[2] <= 0xbf ) { c+=3; continue; } 822 823 if ( ( ( 0xf0 == c[0] 824 && 0x90 <= c[1] && c[1] <= 0xbf) 825 || ( 0xf1 <= c[0] && c[0] <= 0xf3 826 && 0x80 <= c[1] && c[1] <= 0xbf) 827 || ( 0xf4 == c[0] 828 && 0x80 <= c[1] && c[1] <= 0x8f) ) 829 && 0x80 <= c[2] && c[2] <= 0xbf 830 && 0x80 <= c[3] && c[3] <= 0xbf ) { c+=4; continue; } 831 832 return 0; /* invalid */ 833 } 834 return 1; /* valid */ 835 } 836 837 /* - special case: empty string returns empty string 838 * - on windows or cygwin: replace \ with / 839 * - strip leading spaces 840 * - prepends "/" if not present already 841 * - resolve "/../", "//" and "/./" the usual way: 842 * the first one removes a preceding component, the other two 843 * get compressed to "/". 844 * - "/." and "/.." at the end are similar, but always leave a trailing 845 * "/" 846 * 847 * /blah/.. gets / 848 * /blah/../foo gets /foo 849 * /abc/./xyz gets /abc/xyz 850 * /abc//xyz gets /abc/xyz 851 * 852 * NOTE: src and dest can point to the same buffer, in which case, 853 * the operation is performed in-place. 854 */ 855 856 void buffer_path_simplify(buffer *dest, buffer *src) 857 { 858 /* current character, the one before, and the one before that from input */ 859 char c, pre1, pre2; 860 char *start, *slash, *walk, *out; 861 862 force_assert(NULL != dest && NULL != src); 863 864 if (buffer_string_is_empty(src)) { 865 buffer_copy_string_len(dest, CONST_STR_LEN("")); 866 return; 867 } 868 869 force_assert('\0' == src->ptr[src->used-1]); 870 871 #if defined(__WIN32) || defined(__CYGWIN__) 872 /* cygwin is treating \ and / the same, so we have to that too */ 873 { 874 char *p; 875 for (p = src->ptr; *p; p++) { 876 if (*p == '\\') *p = '/'; 877 } 878 } 879 #endif 880 881 walk = src->ptr; 882 start = dest->ptr; 883 out = dest->ptr; 884 slash = dest->ptr; 885 886 /* skip leading spaces */ 887 while (*walk == ' ') { 888 walk++; 889 } 890 if (*walk == '.') { 891 if (walk[1] == '/' || walk[1] == '\0') 892 ++walk; 893 else if (walk[1] == '.' && (walk[2] == '/' || walk[2] == '\0')) 894 walk+=2; 895 } 896 897 pre1 = 0; 898 c = *(walk++); 899 900 while (c != '\0') { 901 /* assert((src != dest || out <= walk) && slash <= out); */ 902 /* the following comments about out and walk are only interesting if 903 * src == dest; otherwise the memory areas don't overlap anyway. 904 */ 905 pre2 = pre1; 906 pre1 = c; 907 908 /* possibly: out == walk - need to read first */ 909 c = *walk; 910 *out = pre1; 911 912 out++; 913 walk++; 914 /* (out <= walk) still true; also now (slash < out) */ 915 916 if (c == '/' || c == '\0') { 917 const size_t toklen = out - slash; 918 if (toklen == 3 && pre2 == '.' && pre1 == '.' && *slash == '/') { 919 /* "/../" or ("/.." at end of string) */ 920 out = slash; 921 /* if there is something before "/..", there is at least one 922 * component, which needs to be removed */ 923 if (out > start) { 924 out--; 925 while (out > start && *out != '/') out--; 926 } 927 928 /* don't kill trailing '/' at end of path */ 929 if (c == '\0') out++; 930 /* slash < out before, so out_new <= slash + 1 <= out_before <= walk */ 931 } else if (toklen == 1 || (pre2 == '/' && pre1 == '.')) { 932 /* "//" or "/./" or (("/" or "/.") at end of string) */ 933 out = slash; 934 /* don't kill trailing '/' at end of path */ 935 if (c == '\0') out++; 936 /* slash < out before, so out_new <= slash + 1 <= out_before <= walk */ 937 } 938 939 slash = out; 940 } 941 } 942 943 buffer_string_set_length(dest, out - start); 944 } 945 946 void buffer_to_lower(buffer *b) { 947 size_t i; 948 949 for (i = 0; i < b->used; ++i) { 950 char c = b->ptr[i]; 951 if (c >= 'A' && c <= 'Z') b->ptr[i] |= 0x20; 952 } 953 } 954 955 956 void buffer_to_upper(buffer *b) { 957 size_t i; 958 959 for (i = 0; i < b->used; ++i) { 960 char c = b->ptr[i]; 961 if (c >= 'a' && c <= 'z') b->ptr[i] &= ~0x20; 962 } 963 } 964 965 966 #include <stdio.h> 967 968 #ifdef HAVE_LIBUNWIND 969 # define UNW_LOCAL_ONLY 970 # include <libunwind.h> 971 972 static void print_backtrace(FILE *file) { 973 unw_cursor_t cursor; 974 unw_context_t context; 975 int ret; 976 unsigned int frame = 0; 977 978 if (0 != (ret = unw_getcontext(&context))) goto error; 979 if (0 != (ret = unw_init_local(&cursor, &context))) goto error; 980 981 fprintf(file, "Backtrace:\n"); 982 983 while (0 < (ret = unw_step(&cursor))) { 984 unw_word_t proc_ip = 0; 985 unw_proc_info_t procinfo; 986 char procname[256]; 987 unw_word_t proc_offset = 0; 988 989 if (0 != (ret = unw_get_reg(&cursor, UNW_REG_IP, &proc_ip))) goto error; 990 991 if (0 == proc_ip) { 992 /* without an IP the other functions are useless; unw_get_proc_name would return UNW_EUNSPEC */ 993 ++frame; 994 fprintf(file, "%u: (nil)\n", frame); 995 continue; 996 } 997 998 if (0 != (ret = unw_get_proc_info(&cursor, &procinfo))) goto error; 999 1000 if (0 != (ret = unw_get_proc_name(&cursor, procname, sizeof(procname), &proc_offset))) { 1001 switch (-ret) { 1002 case UNW_ENOMEM: 1003 memset(procname + sizeof(procname) - 4, '.', 3); 1004 procname[sizeof(procname) - 1] = '\0'; 1005 break; 1006 case UNW_ENOINFO: 1007 procname[0] = '?'; 1008 procname[1] = '\0'; 1009 proc_offset = 0; 1010 break; 1011 default: 1012 snprintf(procname, sizeof(procname), "?? (unw_get_proc_name error %d)", -ret); 1013 break; 1014 } 1015 } 1016 1017 ++frame; 1018 fprintf(file, "%u: %s (+0x%x) [%p]\n", 1019 frame, 1020 procname, 1021 (unsigned int) proc_offset, 1022 (void*)(uintptr_t)proc_ip); 1023 } 1024 1025 if (0 != ret) goto error; 1026 1027 return; 1028 1029 error: 1030 fprintf(file, "Error while generating backtrace: unwind error %i\n", (int) -ret); 1031 } 1032 #else 1033 static void print_backtrace(FILE *file) { 1034 UNUSED(file); 1035 } 1036 #endif 1037 1038 void log_failed_assert(const char *filename, unsigned int line, const char *msg) { 1039 /* can't use buffer here; could lead to recursive assertions */ 1040 fprintf(stderr, "%s.%u: %s\n", filename, line, msg); 1041 print_backtrace(stderr); 1042 fflush(stderr); 1043 abort(); 1044 } 1045