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