run_burl_normalize: 26| 1.64k| size_t in_len) { 27| 1.64k| int qs; 28| 1.64k| buffer_copy_string_len(psrc, in, in_len); 29| 1.64k| qs = burl_normalize(psrc, ptmp, flags); 30| 1.64k|} LLVMFuzzerTestOneInput: 32| 1.64k|int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { 33| 1.64k| if (size <= 4) { ------------------ | Branch (33:9): [True: 4, False: 1.64k] ------------------ 34| 4| return 0; 35| 4| } 36| 1.64k| int flags = ((int*)data)[0]; 37| 1.64k| data += 4; 38| 1.64k| size -= 4; 39| 1.64k| char *new_str = (char *)malloc(size+1); 40| 1.64k| if (new_str == NULL){ ------------------ | Branch (40:9): [True: 0, False: 1.64k] ------------------ 41| 0| return 0; 42| 0| } 43| 1.64k| memcpy(new_str, data, size); 44| 1.64k| new_str[size] = '\0'; 45| | 46| | /* main fuzzer entrypoint for library */ 47| 1.64k| buffer *psrc = buffer_init(); 48| 1.64k| buffer *ptmp = buffer_init(); 49| 1.64k| run_burl_normalize(psrc, ptmp, flags, __LINE__, new_str, size); 50| 1.64k| buffer_urldecode_path(psrc); 51| | 52| 1.64k| buffer_free(psrc); 53| 1.64k| buffer_free(ptmp); 54| 1.64k| free(new_str); 55| 1.64k| return 0; 56| 1.64k|} buffer_init: 14| 3.28k|buffer* buffer_init(void) { 15| | #if 0 /* buffer_init() and chunk_init() can be hot, 16| | * so avoid the additional hop of indirection */ 17| | return ck_calloc(1, sizeof(buffer)); 18| | #else 19| 3.28k| buffer * const b = calloc(1, sizeof(*b)); 20| 3.28k| force_assert(b); ------------------ | | 448| 3.28k|#define force_assert(x) ck_assert(x) | | ------------------ | | | | 115| 3.28k| do { if (!(x)) ck_assert_failed(__FILE__, __LINE__, #x); } while (0) | | | | ------------------ | | | | | Branch (115:18): [True: 0, False: 3.28k] | | | | | Branch (115:75): [Folded, False: 3.28k] | | | | ------------------ | | ------------------ ------------------ 21| 3.28k| return b; 22| 3.28k| #endif 23| 3.28k|} buffer_free: 25| 3.28k|void buffer_free(buffer *b) { 26| 3.28k| if (NULL == b) return; ------------------ | Branch (26:6): [True: 0, False: 3.28k] ------------------ 27| 3.28k| free(b->ptr); 28| 3.28k| free(b); 29| 3.28k|} buffer_string_prepare_copy: 82| 1.27k|char* buffer_string_prepare_copy(buffer * const b, const size_t size) { 83| 1.27k| b->used = 0; 84| | #ifdef __COVERITY__ /*(b->ptr is not NULL if b->size is not 0)*/ 85| | force_assert(size >= b->size || b->ptr); 86| | #endif 87| 1.27k| return (size < b->size) ------------------ | Branch (87:12): [True: 0, False: 1.27k] ------------------ 88| 1.27k| ? b->ptr 89| 1.27k| : buffer_alloc_replace(b, size); 90| 1.27k|} buffer_extend: 127| 150|{ 128| | /* extend buffer to append x (reallocate by power-2 (or larger), if needed) 129| | * (combine buffer_string_prepare_append() and buffer_commit()) 130| | * (future: might make buffer.h static inline func for HTTP/1.1 performance) 131| | * pre-sets '\0' byte and b->used (unlike buffer_string_prepare_append())*/ 132| | #if 0 133| | char * const s = buffer_string_prepare_append(b, x); 134| | b->used += x + (0 == b->used); 135| | #else 136| 150| const uint32_t len = b->used ? b->used-1 : 0; ------------------ | Branch (136:26): [True: 150, False: 0] ------------------ 137| 150| char * const s = (b->size - len >= x + 1) ------------------ | Branch (137:22): [True: 150, False: 0] ------------------ 138| 150| ? b->ptr + len 139| 150| : buffer_string_prepare_append_resize(b, x); 140| 150| b->used = len+x+1; 141| 150| #endif 142| 150| s[x] = '\0'; 143| 150| return s; 144| 150|} buffer_copy_string_len: 172| 3.06k|void buffer_copy_string_len(buffer * const restrict b, const char * const restrict s, const size_t len) { 173| 3.06k| b->used = len + 1; 174| 3.06k| char * const restrict d = (len < b->size) ------------------ | Branch (174:31): [True: 944, False: 2.12k] ------------------ 175| 3.06k| ? b->ptr 176| 3.06k| : buffer_alloc_replace(b, len); 177| 3.06k| d[len] = '\0'; 178| 3.06k| memcpy(d, s, len); 179| 3.06k|} buffer_append_string_len: 198| 150|void buffer_append_string_len(buffer * const restrict b, const char * const restrict s, const size_t len) { 199| 150| memcpy(buffer_extend(b, len), s, len); 200| 150|} hex2int: 386| 99.4M|char hex2int(unsigned char hex) { 387| 99.4M| unsigned char n; 388| 99.4M| return li_cton(hex,n) ? (char)n : 0xFF; ------------------ | | 381| 99.4M| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (381:4): [True: 89.8M, False: 9.63M] | | | Branch (381:30): [True: 9.63M, False: 0] | | | Branch (381:31): [True: 9.63M, False: 0] | | ------------------ ------------------ 389| 99.4M|} buffer_urldecode_path: 792| 1.64k|void buffer_urldecode_path(buffer * const b) { 793| 1.64k| const size_t len = buffer_clen(b); 794| 1.64k| char *src = len ? memchr(b->ptr, '%', len) : NULL; ------------------ | Branch (794:17): [True: 1.63k, False: 6] ------------------ 795| 1.64k| if (NULL == src) return; ------------------ | Branch (795:9): [True: 350, False: 1.29k] ------------------ 796| | 797| 1.29k| char *dst = src; 798| 49.7M| do { 799| | /* *src == '%' */ 800| 49.7M| unsigned char high = ((unsigned char *)src)[1]; 801| 49.7M| unsigned char low = high ? hex2int(((unsigned char *)src)[2]) : 0xFF; ------------------ | Branch (801:29): [True: 49.7M, False: 0] ------------------ 802| 49.7M| if (0xFF != (high = hex2int(high)) && 0xFF != low) { ------------------ | Branch (802:13): [True: 49.7M, False: 0] | Branch (802:47): [True: 49.7M, False: 0] ------------------ 803| 49.7M| high = (high << 4) | low; /* map ctrls to '_' */ 804| 49.7M| *dst = (high >= 32 && high != 127) ? high : '_'; ------------------ | Branch (804:21): [True: 6.95M, False: 42.7M] | Branch (804:35): [True: 5.93M, False: 1.02M] ------------------ 805| 49.7M| src += 2; 806| 49.7M| } /* else ignore this '%'; leave as-is and move on */ 807| | 808| 53.3M| while ((*++dst = *++src) != '%' && *src) ; ------------------ | Branch (808:16): [True: 3.60M, False: 49.7M] | Branch (808:44): [True: 3.60M, False: 1.29k] ------------------ 809| 49.7M| } while (*src); ------------------ | Branch (809:14): [True: 49.7M, False: 1.29k] ------------------ 810| 1.29k| b->used = (dst - b->ptr) + 1; 811| 1.29k|} buffer_path_simplify: 864| 360|{ 865| 360| char *out = b->ptr; 866| 360| char * const end = b->ptr + b->used - 1; 867| | 868| 360| if (__builtin_expect( (buffer_is_blank(b)), 0)) { ------------------ | Branch (868:9): [True: 0, False: 360] ------------------ 869| 0| buffer_blank(b); 870| 0| return; 871| 0| } 872| | 873| | #if defined(_WIN32) || defined(__CYGWIN__) 874| | /* cygwin is treating \ and / the same, so we have to that too */ 875| | for (char *p = b->ptr; *p; p++) { 876| | if (*p == '\\') *p = '/'; 877| | } 878| | #endif 879| | 880| 360| *end = '/'; /*(end of path modified to avoid need to check '\0')*/ 881| | 882| 360| char *walk = out; 883| 360| if (__builtin_expect( (*walk == '/'), 1)) { ------------------ | Branch (883:9): [True: 135, False: 225] ------------------ 884| | /* scan to detect (potential) need for path simplification 885| | * (repeated '/' or "/.") */ 886| 508| do { 887| 508| if (*++walk == '.' || *walk == '/') ------------------ | Branch (887:17): [True: 72, False: 436] | Branch (887:35): [True: 63, False: 373] ------------------ 888| 135| break; 889| 1.22M| do { ++walk; } while (*walk != '/'); ------------------ | Branch (889:35): [True: 1.22M, False: 373] ------------------ 890| 373| } while (walk != end); ------------------ | Branch (890:18): [True: 373, False: 0] ------------------ 891| 135| if (__builtin_expect( (walk == end), 1)) { ------------------ | Branch (891:13): [True: 0, False: 135] ------------------ 892| | /* common case: no repeated '/' or "/." */ 893| 0| *end = '\0'; /* overwrite extra '/' added to end of path */ 894| 0| return; 895| 0| } 896| 135| out = walk-1; 897| 135| } 898| 225| else { 899| 225| if (walk[0] == '.' && walk[1] == '/') ------------------ | Branch (899:13): [True: 121, False: 104] | Branch (899:31): [True: 71, False: 50] ------------------ 900| 71| *out = *++walk; 901| 154| else if (walk[0] == '.' && walk[1] == '.' && walk[2] == '/') ------------------ | Branch (901:18): [True: 50, False: 104] | Branch (901:36): [True: 8, False: 42] | Branch (901:54): [True: 2, False: 6] ------------------ 902| 2| *out = *(walk += 2); 903| 152| else { 904| 10.5M| while (*++walk != '/') ; ------------------ | Branch (904:20): [True: 10.5M, False: 152] ------------------ 905| 152| out = walk; 906| 152| } 907| 225| ++walk; 908| 225| } 909| | 910| 9.05k| while (walk <= end) { ------------------ | Branch (910:12): [True: 8.84k, False: 215] ------------------ 911| | /* previous char is '/' at this point (or start of string w/o '/') */ 912| 8.84k| if (__builtin_expect( (walk[0] == '/'), 0)) { ------------------ | Branch (912:13): [True: 2.20k, False: 6.64k] ------------------ 913| | /* skip repeated '/' (e.g. "///" -> "/") */ 914| 2.20k| if (++walk < end) ------------------ | Branch (914:17): [True: 2.14k, False: 54] ------------------ 915| 2.14k| continue; 916| 54| else { 917| 54| ++out; 918| 54| break; 919| 54| } 920| 2.20k| } 921| 6.64k| else if (__builtin_expect( (walk[0] == '.'), 0)) { ------------------ | Branch (921:18): [True: 4.81k, False: 1.82k] ------------------ 922| | /* handle "./" and "../" */ 923| 4.81k| if (walk[1] == '.' && walk[2] == '/') { ------------------ | Branch (923:17): [True: 2.27k, False: 2.54k] | Branch (923:35): [True: 884, False: 1.38k] ------------------ 924| | /* handle "../" */ 925| 5.62M| while (out > b->ptr && *--out != '/') ; ------------------ | Branch (925:24): [True: 5.62M, False: 493] | Branch (925:40): [True: 5.62M, False: 391] ------------------ 926| 884| *out = '/'; /*(in case path had not started with '/')*/ 927| 884| if ((walk += 3) >= end) { ------------------ | Branch (927:21): [True: 48, False: 836] ------------------ 928| 48| ++out; 929| 48| break; 930| 48| } 931| 836| else 932| 836| continue; 933| 884| } 934| 3.93k| else if (walk[1] == '/') { ------------------ | Branch (934:22): [True: 978, False: 2.95k] ------------------ 935| | /* handle "./" */ 936| 978| if ((walk += 2) >= end) { ------------------ | Branch (936:21): [True: 43, False: 935] ------------------ 937| 43| ++out; 938| 43| break; 939| 43| } 940| 935| continue; 941| 978| } 942| 2.95k| else { 943| | /* accept "." if not part of "../" or "./" */ 944| 2.95k| *++out = '.'; 945| 2.95k| ++walk; 946| 2.95k| } 947| 4.81k| } 948| | 949| 28.6M| while ((*++out = *walk++) != '/') ; ------------------ | Branch (949:16): [True: 28.6M, False: 4.77k] ------------------ 950| 4.77k| } 951| 360| *out = *end = '\0'; /* overwrite extra '/' added to end of path */ 952| 360| b->used = (out - b->ptr) + 1; 953| | /*buffer_truncate(b, out - b->ptr);*/ 954| 360|} buffer.c:buffer_alloc_replace: 71| 3.39k|static char* buffer_alloc_replace(buffer * const restrict b, const size_t size) { 72| | /*(discard old data so realloc() does not copy)*/ 73| 3.39k| if (NULL != b->ptr) { ------------------ | Branch (73:9): [True: 468, False: 2.92k] ------------------ 74| 468| free(b->ptr); 75| 468| b->ptr = NULL; 76| 468| } 77| | /*(note: if size larger than one lshift, use size instead of power-2)*/ 78| 3.39k| const size_t bsize2x = (b->size & ~1uL) << 1; 79| 3.39k| return buffer_realloc(b, bsize2x > size ? bsize2x-1 : size); ------------------ | Branch (79:30): [True: 234, False: 3.15k] ------------------ 80| 3.39k|} buffer.c:buffer_realloc: 49| 3.39k|static char* buffer_realloc(buffer * const restrict b, const size_t len) { 50| 3.39k| #define BUFFER_PIECE_SIZE 64uL /*(must be power-of-2)*/ 51| 3.39k| size_t sz = (len + 1 + BUFFER_PIECE_SIZE-1) & ~(BUFFER_PIECE_SIZE-1); ------------------ | | 50| 3.39k| #define BUFFER_PIECE_SIZE 64uL /*(must be power-of-2)*/ ------------------ size_t sz = (len + 1 + BUFFER_PIECE_SIZE-1) & ~(BUFFER_PIECE_SIZE-1); ------------------ | | 50| 3.39k| #define BUFFER_PIECE_SIZE 64uL /*(must be power-of-2)*/ ------------------ 52| 3.39k| force_assert(sz > len); ------------------ | | 448| 3.39k|#define force_assert(x) ck_assert(x) | | ------------------ | | | | 115| 3.39k| do { if (!(x)) ck_assert_failed(__FILE__, __LINE__, #x); } while (0) | | | | ------------------ | | | | | Branch (115:18): [True: 0, False: 3.39k] | | | | | Branch (115:75): [Folded, False: 3.39k] | | | | ------------------ | | ------------------ ------------------ 53| 3.39k| if ((sz & (sz-1)) && sz < INT_MAX) {/* not power-2; huge val not expected */ ------------------ | Branch (53:9): [True: 1.17k, False: 2.22k] | Branch (53:26): [True: 1.17k, False: 0] ------------------ 54| | /*(optimizer should recognize this and use ffs or clz or equivalent)*/ 55| 1.17k| const size_t psz = sz; 56| 8.06k| for (sz = 256; sz < psz; sz <<= 1) ; ------------------ | Branch (56:24): [True: 6.89k, False: 1.17k] ------------------ 57| 1.17k| } 58| 3.39k| sz |= 1; /*(extra +1 for '\0' when needed buffer size is exact power-2)*/ 59| | 60| 3.39k| b->size = sz; 61| 3.39k| b->ptr = realloc(b->ptr, sz); 62| | 63| 3.39k| force_assert(NULL != b->ptr); ------------------ | | 448| 3.39k|#define force_assert(x) ck_assert(x) | | ------------------ | | | | 115| 3.39k| do { if (!(x)) ck_assert_failed(__FILE__, __LINE__, #x); } while (0) | | | | ------------------ | | | | | Branch (115:18): [True: 0, False: 3.39k] | | | | | Branch (115:75): [Folded, False: 3.39k] | | | | ------------------ | | ------------------ ------------------ 64| 3.39k| return b->ptr; 65| 3.39k|} burl.c:buffer_truncate: 349| 429|static inline void buffer_truncate(buffer *b, uint32_t len) { 350| 429| b->ptr[len] = '\0'; /* b->ptr must exist; use buffer_blank() for trunc 0 */ 351| 429| b->used = len + 1; 352| 429|} burl.c:light_isalnum: 232| 11.0k|static inline int light_isalnum(int c) { 233| 11.0k| return light_isdigit(c) || light_isalpha(c); ------------------ | Branch (233:9): [True: 241, False: 10.7k] | Branch (233:29): [True: 215, False: 10.5k] ------------------ 234| 11.0k|} burl.c:light_isdigit: 214| 11.0k|static inline int light_isdigit(int c) { 215| 11.0k| return ((uint32_t)c-'0' <= '9'-'0'); 216| 11.0k|} burl.c:light_isalpha: 226| 10.7k|static inline int light_isalpha(int c) { 227| 10.7k| return (((uint32_t)c | 0x20)-'a' <= 'z'-'a'); 228| 10.7k|} burl.c:buffer_clen: 323| 6.09k|static inline uint32_t buffer_clen (const buffer *b) { 324| 6.09k| return b->used - (0 != b->used); 325| 6.09k|} buffer.c:buffer_clen: 323| 1.64k|static inline uint32_t buffer_clen (const buffer *b) { 324| 1.64k| return b->used - (0 != b->used); 325| 1.64k|} buffer.c:buffer_is_blank: 315| 360|static inline int buffer_is_blank(const buffer *b) { 316| 360| return b->used < 2; /* buffer_is_blank() || buffer_is_unset() */ 317| 360|} burl_normalize: 338| 1.64k|{ 339| 1.64k| int qs; 340| | 341| | #if defined(_WIN32) || defined(__CYGWIN__) 342| | /* Windows and Cygwin treat '\\' as '/' if '\\' is present in path; 343| | * convert to '/' for consistency before percent-encoding 344| | * normalization which will convert '\\' to "%5C" in the URL. 345| | * (Clients still should not be sending '\\' unencoded in requests.) */ 346| | if (flags & HTTP_PARSEOPT_URL_NORMALIZE_PATH_BACKSLASH_TRANS) { 347| | for (char *p = b->ptr; *p != '?' && *p != '\0'; ++p) { 348| | if (*p == '\\') *p = '/'; 349| | } 350| | } 351| | #endif 352| | 353| 1.64k| qs = (flags & HTTP_PARSEOPT_URL_NORMALIZE_REQUIRED) ------------------ | Branch (353:10): [True: 895, False: 749] ------------------ 354| 1.64k| ? burl_normalize_basic_required(b, t) 355| 1.64k| : burl_normalize_basic_unreserved(b, t); 356| 1.64k| if (-2 == qs) { ------------------ | Branch (356:9): [True: 489, False: 1.15k] ------------------ 357| 489| if (flags & HTTP_PARSEOPT_URL_NORMALIZE_INVALID_UTF8_REJECT) return -2; ------------------ | Branch (357:13): [True: 102, False: 387] ------------------ 358| 387| qs = burl_scan_qmark(b); 359| 387| } 360| | 361| 1.54k| if (flags & HTTP_PARSEOPT_URL_NORMALIZE_CTRLS_REJECT) { ------------------ | Branch (361:9): [True: 325, False: 1.21k] ------------------ 362| 325| if (burl_contains_ctrls(b)) return -2; ------------------ | Branch (362:13): [True: 90, False: 235] ------------------ 363| 325| } 364| | 365| 1.45k| if (flags & (HTTP_PARSEOPT_URL_NORMALIZE_PATH_2F_DECODE ------------------ | Branch (365:9): [True: 1.18k, False: 268] ------------------ 366| 1.45k| |HTTP_PARSEOPT_URL_NORMALIZE_PATH_2F_REJECT)) { 367| 1.18k| qs = burl_normalize_2F_to_slash(b, qs, flags); 368| 1.18k| if (-2 == qs) return -2; ------------------ | Branch (368:13): [True: 12, False: 1.17k] ------------------ 369| 1.18k| } 370| | 371| 1.44k| if (flags & (HTTP_PARSEOPT_URL_NORMALIZE_PATH_DOTSEG_REMOVE ------------------ | Branch (371:9): [True: 1.14k, False: 292] ------------------ 372| 1.44k| |HTTP_PARSEOPT_URL_NORMALIZE_PATH_DOTSEG_REJECT)) { 373| 1.14k| qs = burl_normalize_path(b, t, qs, flags); 374| 1.14k| if (-2 == qs) return -2; ------------------ | Branch (374:13): [True: 24, False: 1.12k] ------------------ 375| 1.14k| } 376| | 377| 1.41k| if (flags & HTTP_PARSEOPT_URL_NORMALIZE_QUERY_20_PLUS) { ------------------ | Branch (377:9): [True: 831, False: 585] ------------------ 378| 831| if (qs >= 0) burl_normalize_qs20_to_plus(b, qs); ------------------ | Branch (378:13): [True: 407, False: 424] ------------------ 379| 831| } 380| | 381| 1.41k| return qs; 382| 1.44k|} burl.c:burl_normalize_basic_required: 175| 895|{ 176| 895| const unsigned char * const s = (unsigned char *)b->ptr; 177| 895| const int used = (int)buffer_clen(b); 178| 895| unsigned int n1, n2, x; 179| 895| int qs = -1; 180| 895| int invalid_utf8 = 0; 181| | 182| 1.16M| for (int i = 0; i < used; ++i) { ------------------ | Branch (182:21): [True: 1.16M, False: 188] ------------------ 183| 1.16M| if (!encoded_chars_http_uri_reqd[s[i]]) { ------------------ | Branch (183:13): [True: 1.09M, False: 75.3k] ------------------ 184| 1.09M| if (s[i] == '?') qs = i; ------------------ | Branch (184:17): [True: 59.9k, False: 1.03M] ------------------ 185| 1.09M| } 186| 75.3k| else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2) ------------------ | | 43| 150k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 70.2k, False: 4.60k] | | | Branch (43:30): [True: 4.52k, False: 82] | | | Branch (43:31): [True: 4.52k, False: 82] | | ------------------ ------------------ else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2) ------------------ | | 43| 150k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 44.4k, False: 30.3k] | | | Branch (43:30): [True: 30.2k, False: 57] | | | Branch (43:31): [True: 30.2k, False: 57] | | ------------------ ------------------ | Branch (186:18): [True: 74.8k, False: 498] ------------------ 187| 74.6k| && (encoded_chars_http_uri_reqd[(x = (n1 << 4) | n2)] ------------------ | Branch (187:22): [True: 13.2k, False: 61.4k] ------------------ 188| 61.4k| || (qs < 0 ------------------ | Branch (188:25): [True: 61.3k, False: 70] | Branch (188:26): [True: 710, False: 60.6k] ------------------ 189| 61.4k| ? (x == '/' || x == '?') ------------------ | Branch (189:29): [True: 431, False: 279] | Branch (189:41): [True: 248, False: 31] ------------------ 190| 74.6k| : (x == '&' || x == '=' || x == ';' || x == '+')))) { ------------------ | Branch (190:29): [True: 33.3k, False: 27.2k] | Branch (190:41): [True: 16.3k, False: 10.9k] | Branch (190:53): [True: 3.04k, False: 7.89k] | Branch (190:65): [True: 7.85k, False: 39] ------------------ 191| 74.6k| invalid_utf8 |= li_utf8_invalid_byte(x); ------------------ | | 49| 74.6k|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 74.6k| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 600, False: 74.0k] | | | | ------------------ | | | | 261| 74.6k| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 194, False: 73.8k] | | | | ------------------ | | ------------------ ------------------ 192| 74.6k| if (s[i+1] >= 'a') b->ptr[i+1] &= 0xdf; /* uppercase hex */ ------------------ | Branch (192:17): [True: 825, False: 73.7k] ------------------ 193| 74.6k| if (s[i+2] >= 'a') b->ptr[i+2] &= 0xdf; /* uppercase hex */ ------------------ | Branch (193:17): [True: 11.8k, False: 62.7k] ------------------ 194| 74.6k| i+=2; 195| 74.6k| } 196| 707| else if (s[i] == '#') { /* ignore fragment */ ------------------ | Branch (196:18): [True: 1, False: 706] ------------------ 197| 1| buffer_truncate(b, (size_t)i); 198| 1| break; 199| 1| } 200| 706| else { 201| 706| qs = burl_normalize_basic_required_fix(b, t, i, qs); 202| 706| break; 203| 706| } 204| 1.16M| } 205| | 206| 895| return !invalid_utf8 ? qs : -2; ------------------ | Branch (206:12): [True: 866, False: 29] ------------------ 207| 895|} burl.c:burl_normalize_basic_required_fix: 131| 706|{ 132| 706| int j = i; 133| 706| const int used = (int)buffer_clen(b); 134| 706| const unsigned char * const s = (unsigned char *)b->ptr; 135| 706| unsigned char * const p = 136| 706| (unsigned char *)buffer_string_prepare_copy(t,i+(used-i)*3+1); 137| 706| unsigned int n1, n2; 138| 706| int invalid_utf8 = 0; 139| 706| memcpy(p, s, (size_t)i); 140| 29.5M| for (; i < used; ++i, ++j) { ------------------ | Branch (140:12): [True: 29.5M, False: 693] ------------------ 141| 29.5M| if (!encoded_chars_http_uri_reqd[s[i]]) { ------------------ | Branch (141:13): [True: 746k, False: 28.7M] ------------------ 142| 746k| p[j] = s[i]; 143| 746k| if (__builtin_expect( (s[i] == '?'), 0)) qs = j; ------------------ | Branch (143:17): [True: 13.5k, False: 733k] ------------------ 144| 746k| } 145| 28.7M| else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2)) { ------------------ | | 43| 28.8M| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 26.7k, False: 37.7k] | | | Branch (43:30): [True: 8.35k, False: 29.4k] | | | Branch (43:31): [True: 8.35k, False: 29.4k] | | ------------------ ------------------ else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2)) { ------------------ | | 43| 35.0k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 16.4k, False: 18.6k] | | | Branch (43:30): [True: 15.0k, False: 3.55k] | | | Branch (43:31): [True: 15.0k, False: 3.55k] | | ------------------ ------------------ | Branch (145:18): [True: 64.4k, False: 28.6M] ------------------ 146| 31.5k| const unsigned int x = (n1 << 4) | n2; 147| 31.5k| if (!encoded_chars_http_uri_reqd[x] ------------------ | Branch (147:17): [True: 20.6k, False: 10.8k] ------------------ 148| 20.6k| && (qs < 0 ------------------ | Branch (148:20): [True: 5.61k, False: 15.0k] | Branch (148:21): [True: 1.97k, False: 18.7k] ------------------ 149| 20.6k| ? (x != '/' && x != '?') ------------------ | Branch (149:24): [True: 1.39k, False: 577] | Branch (149:36): [True: 1.03k, False: 364] ------------------ 150| 20.6k| : (x != '&' && x != '=' && x != ';' && x != '+'))) { ------------------ | Branch (150:24): [True: 12.1k, False: 6.61k] | Branch (150:36): [True: 8.66k, False: 3.43k] | Branch (150:48): [True: 6.99k, False: 1.67k] | Branch (150:60): [True: 4.58k, False: 2.41k] ------------------ 151| 5.61k| p[j] = x; 152| 5.61k| } 153| 25.9k| else { 154| 25.9k| p[j] = '%'; 155| 25.9k| p[++j] = hex_chars_uc[n1]; /*(s[i+1] & 0xdf)*/ 156| 25.9k| p[++j] = hex_chars_uc[n2]; /*(s[i+2] & 0xdf)*/ 157| 25.9k| invalid_utf8 |= li_utf8_invalid_byte(x); ------------------ | | 49| 25.9k|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 25.9k| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 1.92k, False: 23.9k] | | | | ------------------ | | | | 261| 25.9k| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 78, False: 23.8k] | | | | ------------------ | | ------------------ ------------------ 158| 25.9k| } 159| 31.5k| i+=2; 160| 31.5k| } 161| 28.7M| else if (s[i] == '#') break; /* ignore fragment */ ------------------ | Branch (161:18): [True: 13, False: 28.7M] ------------------ 162| 28.7M| else { 163| 28.7M| p[j] = '%'; 164| 28.7M| p[++j] = hex_chars_uc[(s[i] >> 4) & 0xF]; 165| 28.7M| p[++j] = hex_chars_uc[s[i] & 0xF]; 166| 28.7M| invalid_utf8 |= li_utf8_invalid_byte(s[i]); ------------------ | | 49| 28.7M|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 28.7M| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 1.88M, False: 26.8M] | | | | ------------------ | | | | 261| 28.7M| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 726, False: 26.8M] | | | | ------------------ | | ------------------ ------------------ 167| 28.7M| } 168| 29.5M| } 169| 706| buffer_copy_string_len(b, (char *)p, (size_t)j); 170| 706| return !invalid_utf8 ? qs : -2; ------------------ | Branch (170:12): [True: 447, False: 259] ------------------ 171| 706|} burl.c:burl_normalize_basic_unreserved: 99| 749|{ 100| 749| const unsigned char * const s = (unsigned char *)b->ptr; 101| 749| const int used = (int)buffer_clen(b); 102| 749| unsigned int n1, n2, x; 103| 749| int qs = -1; 104| | 105| 765k| for (int i = 0; i < used; ++i) { ------------------ | Branch (105:21): [True: 765k, False: 184] ------------------ 106| 765k| if (!encoded_chars_http_uri_reqd[s[i]]) { ------------------ | Branch (106:13): [True: 761k, False: 4.06k] ------------------ 107| 761k| if (__builtin_expect( (s[i] == '?'), 0) && -1 == qs) qs = i; ------------------ | Branch (107:17): [True: 1.81k, False: 759k] | Branch (107:56): [True: 66, False: 1.74k] ------------------ 108| 761k| } 109| 4.06k| else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2) ------------------ | | 43| 7.74k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 2.54k, False: 1.12k] | | | Branch (43:30): [True: 1.06k, False: 59] | | | Branch (43:31): [True: 1.06k, False: 59] | | ------------------ ------------------ else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2) ------------------ | | 43| 7.68k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 2.00k, False: 1.61k] | | | Branch (43:30): [True: 1.55k, False: 58] | | | Branch (43:31): [True: 1.55k, False: 58] | | ------------------ ------------------ | Branch (109:18): [True: 3.67k, False: 395] ------------------ 110| 3.55k| && !burl_is_unreserved((x = (n1 << 4) | n2))) { ------------------ | Branch (110:21): [True: 3.50k, False: 53] ------------------ 111| 3.50k| if (li_utf8_invalid_byte(x)) qs = -2; ------------------ | | 49| 3.50k|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 3.50k| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 438, False: 3.06k] | | | | ------------------ | | | | 261| 3.50k| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 376, False: 2.68k] | | | | ------------------ | | ------------------ ------------------ 112| 3.50k| if (s[i+1] >= 'a') b->ptr[i+1] &= 0xdf; /* uppercase hex */ ------------------ | Branch (112:17): [True: 928, False: 2.57k] ------------------ 113| 3.50k| if (s[i+2] >= 'a') b->ptr[i+2] &= 0xdf; /* uppercase hex */ ------------------ | Branch (113:17): [True: 764, False: 2.73k] ------------------ 114| 3.50k| i+=2; 115| 3.50k| } 116| 565| else if (s[i] == '#') { /* ignore fragment */ ------------------ | Branch (116:18): [True: 1, False: 564] ------------------ 117| 1| buffer_truncate(b, (size_t)i); 118| 1| break; 119| 1| } 120| 564| else { 121| 564| qs = burl_normalize_basic_unreserved_fix(b, t, i, qs); 122| 564| break; 123| 564| } 124| 765k| } 125| | 126| 749| return qs; 127| 749|} burl.c:burl_is_unreserved: 53| 11.0k|{ 54| 11.0k| return (light_isalnum(c) || c == '-' || c == '.' || c == '_' || c == '~'); ------------------ | Branch (54:13): [True: 456, False: 10.5k] | Branch (54:33): [True: 310, False: 10.2k] | Branch (54:45): [True: 204, False: 10.0k] | Branch (54:57): [True: 207, False: 9.82k] | Branch (54:69): [True: 56, False: 9.76k] ------------------ 55| 11.0k|} burl.c:burl_normalize_basic_unreserved_fix: 59| 564|{ 60| 564| int j = i; 61| 564| const int used = (int)buffer_clen(b); 62| 564| const unsigned char * const s = (unsigned char *)b->ptr; 63| 564| unsigned char * const p = 64| 564| (unsigned char *)buffer_string_prepare_copy(t,i+(used-i)*3+1); 65| 564| unsigned int n1, n2; 66| 564| memcpy(p, s, (size_t)i); 67| 24.0M| for (; i < used; ++i, ++j) { ------------------ | Branch (67:12): [True: 24.0M, False: 559] ------------------ 68| 24.0M| if (!encoded_chars_http_uri_reqd[s[i]]) { ------------------ | Branch (68:13): [True: 1.23M, False: 22.7M] ------------------ 69| 1.23M| p[j] = s[i]; 70| 1.23M| if (__builtin_expect( (s[i] == '?'), 0) && -1 == qs) qs = j; ------------------ | Branch (70:17): [True: 1.47k, False: 1.23M] | Branch (70:56): [True: 87, False: 1.38k] ------------------ 71| 1.23M| } 72| 22.7M| else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2)) { ------------------ | | 43| 22.8M| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 3.83k, False: 6.76k] | | | Branch (43:30): [True: 4.84k, False: 1.92k] | | | Branch (43:31): [True: 4.84k, False: 1.92k] | | ------------------ ------------------ else if (s[i]=='%' && li_cton(s[i+1], n1) && li_cton(s[i+2], n2)) { ------------------ | | 43| 8.67k| (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0)) | | ------------------ | | | Branch (43:4): [True: 3.12k, False: 5.55k] | | | Branch (43:30): [True: 4.32k, False: 1.23k] | | | Branch (43:31): [True: 4.32k, False: 1.23k] | | ------------------ ------------------ | Branch (72:18): [True: 10.5k, False: 22.7M] ------------------ 73| 7.44k| const unsigned int x = (n1 << 4) | n2; 74| 7.44k| if (burl_is_unreserved(x)) { ------------------ | Branch (74:17): [True: 1.18k, False: 6.26k] ------------------ 75| 1.18k| p[j] = x; 76| 1.18k| } 77| 6.26k| else { 78| 6.26k| p[j] = '%'; 79| 6.26k| p[++j] = hex_chars_uc[n1]; /*(s[i+1] & 0xdf)*/ 80| 6.26k| p[++j] = hex_chars_uc[n2]; /*(s[i+2] & 0xdf)*/ 81| 6.26k| if (li_utf8_invalid_byte(x)) qs = -2; ------------------ | | 49| 6.26k|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 6.26k| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 2.47k, False: 3.78k] | | | | ------------------ | | | | 261| 6.26k| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 1.03k, False: 2.75k] | | | | ------------------ | | ------------------ ------------------ 82| 6.26k| } 83| 7.44k| i+=2; 84| 7.44k| } 85| 22.7M| else if (s[i] == '#') break; /* ignore fragment */ ------------------ | Branch (85:18): [True: 5, False: 22.7M] ------------------ 86| 22.7M| else { 87| 22.7M| p[j] = '%'; 88| 22.7M| p[++j] = hex_chars_uc[(s[i] >> 4) & 0xF]; 89| 22.7M| p[++j] = hex_chars_uc[s[i] & 0xF]; 90| 22.7M| if (li_utf8_invalid_byte(s[i])) qs = -2; ------------------ | | 49| 22.7M|#define li_utf8_invalid_byte(b) light_utf8_invalid_byte(b) | | ------------------ | | | | 260| 22.7M| ( __builtin_expect( ((c) >= 0xF5), 0) \ | | | | ------------------ | | | | | Branch (260:7): [True: 22.6k, False: 22.7M] | | | | ------------------ | | | | 261| 22.7M| || __builtin_expect( (((c)|0x1) == 0xC1), 0) ) | | | | ------------------ | | | | | Branch (261:7): [True: 2.61k, False: 22.7M] | | | | ------------------ | | ------------------ ------------------ 91| 22.7M| } 92| 24.0M| } 93| 564| buffer_copy_string_len(b, (char *)p, (size_t)j); 94| 564| return qs; 95| 564|} burl.c:burl_scan_qmark: 331| 387|static int burl_scan_qmark (const buffer * const b) { 332| 387| const char * const qmark = strchr(b->ptr, '?'); 333| 387| return qmark ? (int)(qmark - b->ptr) : -1; ------------------ | Branch (333:12): [True: 185, False: 202] ------------------ 334| 387|} burl.c:burl_contains_ctrls: 211| 325|{ 212| 325| const char * const s = b->ptr; 213| 325| const int used = (int)buffer_clen(b); 214| 7.76M| for (int i = 0; i < used; ++i) { ------------------ | Branch (214:21): [True: 7.76M, False: 235] ------------------ 215| 7.76M| if (s[i] == '%' && (s[i+1] < '2' || (s[i+1] == '7' && s[i+2] == 'F'))) ------------------ | Branch (215:13): [True: 2.16M, False: 5.59M] | Branch (215:29): [True: 82, False: 2.16M] | Branch (215:46): [True: 347, False: 2.16M] | Branch (215:63): [True: 8, False: 339] ------------------ 216| 90| return 1; 217| 7.76M| } 218| 235| return 0; 219| 325|} burl.c:burl_normalize_2F_to_slash: 276| 1.18k|{ 277| | /*("%2F" must already have been uppercased during normalization)*/ 278| 1.18k| const char * const s = b->ptr; 279| 1.18k| const int used = qs < 0 ? (int)buffer_clen(b) : qs; ------------------ | Branch (279:22): [True: 728, False: 456] ------------------ 280| 78.1M| for (int i = 0; i < used; ++i) { ------------------ | Branch (280:21): [True: 78.1M, False: 1.00k] ------------------ 281| 78.1M| if (s[i] == '%' && s[i+1] == '2' && s[i+2] == 'F') { ------------------ | Branch (281:13): [True: 25.6M, False: 52.4M] | Branch (281:28): [True: 48.2k, False: 25.6M] | Branch (281:45): [True: 184, False: 48.0k] ------------------ 282| 184| return (flags & HTTP_PARSEOPT_URL_NORMALIZE_PATH_2F_DECODE) ------------------ | Branch (282:20): [True: 172, False: 12] ------------------ 283| 184| ? burl_normalize_2F_to_slash_fix(b, qs, i) 284| 184| : -2; /*(flags & HTTP_PARSEOPT_URL_NORMALIZE_PATH_2F_REJECT)*/ 285| 184| } 286| 78.1M| } 287| 1.00k| return qs; 288| 1.18k|} burl.c:burl_normalize_2F_to_slash_fix: 252| 172|{ 253| 172| char * const s = b->ptr; 254| 172| const int blen = (int)buffer_clen(b); 255| 172| const int used = qs < 0 ? blen : qs; ------------------ | Branch (255:22): [True: 107, False: 65] ------------------ 256| 172| int j = i; 257| 13.3M| for (; i < used; ++i, ++j) { ------------------ | Branch (257:12): [True: 13.3M, False: 172] ------------------ 258| 13.3M| s[j] = s[i]; 259| 13.3M| if (s[i] == '%' && s[i+1] == '2' && s[i+2] == 'F') { ------------------ | Branch (259:13): [True: 4.45M, False: 8.90M] | Branch (259:28): [True: 20.6k, False: 4.43M] | Branch (259:45): [True: 895, False: 19.7k] ------------------ 260| 895| s[j] = '/'; 261| 895| i+=2; 262| 895| } 263| 13.3M| } 264| 172| if (qs >= 0) { ------------------ | Branch (264:9): [True: 65, False: 107] ------------------ 265| 65| const int qslen = blen - qs; 266| 65| memmove(s+j, s+qs, (size_t)qslen); 267| 65| qs = j; 268| 65| j += qslen; 269| 65| } 270| 172| buffer_truncate(b, j); 271| 172| return qs; 272| 172|} burl.c:burl_normalize_path: 292| 1.14k|{ 293| 1.14k| const unsigned char * const s = (unsigned char *)b->ptr; 294| 1.14k| const int used = (int)buffer_clen(b); 295| 1.14k| int path_simplify = 0; 296| 4.09k| for (int i = 0, len = qs < 0 ? used : qs; i < len; ++i) { ------------------ | Branch (296:27): [True: 715, False: 433] | Branch (296:47): [True: 3.33k, False: 764] ------------------ 297| 3.33k| if (s[i] == '.' && (s[i+1] != '.' || ++i) ------------------ | Branch (297:13): [True: 1.45k, False: 1.87k] | Branch (297:29): [True: 870, False: 583] | Branch (297:46): [True: 583, False: 0] ------------------ 298| 1.45k| && (s[i+1] == '/' || s[i+1] == '?' || s[i+1] == '\0')) { ------------------ | Branch (298:17): [True: 89, False: 1.36k] | Branch (298:34): [True: 83, False: 1.28k] | Branch (298:51): [True: 65, False: 1.21k] ------------------ 299| 237| path_simplify = 1; 300| 237| break; 301| 237| } 302| 68.8M| while (i < len && s[i] != '/') ++i; ------------------ | Branch (302:16): [True: 68.8M, False: 635] | Branch (302:27): [True: 68.8M, False: 2.45k] ------------------ 303| 3.09k| if (s[i] == '/' && s[i+1] == '/') { /*(s[len] != '/')*/ ------------------ | Branch (303:13): [True: 2.45k, False: 635] | Branch (303:28): [True: 147, False: 2.31k] ------------------ 304| 147| path_simplify = 1; 305| 147| break; 306| 147| } 307| 3.09k| } 308| | 309| 1.14k| if (path_simplify) { ------------------ | Branch (309:9): [True: 384, False: 764] ------------------ 310| 384| if (flags & HTTP_PARSEOPT_URL_NORMALIZE_PATH_DOTSEG_REJECT) return -2; ------------------ | Branch (310:13): [True: 24, False: 360] ------------------ 311| 360| if (qs >= 0) { ------------------ | Branch (311:13): [True: 150, False: 210] ------------------ 312| 150| buffer_copy_string_len(t, b->ptr+qs, used - qs); 313| 150| buffer_truncate(b, qs); 314| 150| } 315| | 316| 360| buffer_path_simplify(b); 317| | 318| 360| if (qs >= 0) { ------------------ | Branch (318:13): [True: 150, False: 210] ------------------ 319| 150| qs = (int)buffer_clen(b); 320| 150| buffer_append_string_len(b, BUF_PTR_LEN(t)); ------------------ | | 297| 150|#define BUF_PTR_LEN(x) (x)->ptr, buffer_clen(x) ------------------ 321| 150| } 322| 360| } 323| | 324| 1.12k| return qs; 325| 1.14k|} burl.c:burl_normalize_qs20_to_plus: 239| 407|{ 240| 407| const char * const s = b->ptr; 241| 407| const int used = qs < 0 ? 0 : (int)buffer_clen(b); ------------------ | Branch (241:22): [True: 0, False: 407] ------------------ 242| 407| int i; 243| 407| if (qs < 0) return; ------------------ | Branch (243:9): [True: 0, False: 407] ------------------ 244| 7.68M| for (i = qs+1; i < used; ++i) { ------------------ | Branch (244:20): [True: 7.68M, False: 302] ------------------ 245| 7.68M| if (s[i] == '%' && s[i+1] == '2' && s[i+2] == '0') break; ------------------ | Branch (245:13): [True: 2.51M, False: 5.17M] | Branch (245:28): [True: 2.35k, False: 2.51M] | Branch (245:45): [True: 105, False: 2.24k] ------------------ 246| 7.68M| } 247| 407| if (i != used) burl_normalize_qs20_to_plus_fix(b, i); ------------------ | Branch (247:9): [True: 105, False: 302] ------------------ 248| 407|} burl.c:burl_normalize_qs20_to_plus_fix: 223| 105|{ 224| 105| char * const s = b->ptr; 225| 105| const int used = (int)buffer_clen(b); 226| 105| int j = i; 227| 6.83M| for (; i < used; ++i, ++j) { ------------------ | Branch (227:12): [True: 6.83M, False: 105] ------------------ 228| 6.83M| s[j] = s[i]; 229| 6.83M| if (s[i] == '%' && s[i+1] == '2' && s[i+2] == '0') { ------------------ | Branch (229:13): [True: 2.26M, False: 4.57M] | Branch (229:28): [True: 12.3k, False: 2.24M] | Branch (229:45): [True: 3.10k, False: 9.28k] ------------------ 230| 3.10k| s[j] = '+'; 231| 3.10k| i+=2; 232| 3.10k| } 233| 6.83M| } 234| 105| buffer_truncate(b, j); 235| 105|}