| File: | builds/wireshark/wireshark/wsutil/str_util.c |
| Warning: | line 1472, column 9 Value stored to 'printable_bytes' is never read |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
| 1 | /* str_util.c |
| 2 | * String utility routines |
| 3 | * |
| 4 | * Wireshark - Network traffic analyzer |
| 5 | * By Gerald Combs <gerald@wireshark.org> |
| 6 | * Copyright 1998 Gerald Combs |
| 7 | * |
| 8 | * SPDX-License-Identifier: GPL-2.0-or-later |
| 9 | */ |
| 10 | |
| 11 | #define _GNU_SOURCE |
| 12 | #include "config.h" |
| 13 | #include "str_util.h" |
| 14 | |
| 15 | #include <string.h> |
| 16 | #include <locale.h> |
| 17 | #include <math.h> |
| 18 | |
| 19 | #include <ws_codepoints.h> |
| 20 | |
| 21 | #include <wsutil/to_str.h> |
| 22 | #include <wsutil/strtoi.h> |
| 23 | #include <wsutil/unicode-utils.h> |
| 24 | |
| 25 | struct prefix_parameters { |
| 26 | const char * const *prefix; /**< array of prefixes to represent unit multiplication factors. */ |
| 27 | int prefix_count; /**< number of elements in the prefix array. */ |
| 28 | int power; /**< multiplication factor between prefixes. */ |
| 29 | int prefix_offset; /**< index of element within the prefix array for "no prefix". */ |
| 30 | }; |
| 31 | |
| 32 | static const char hex[16] = { '0', '1', '2', '3', '4', '5', '6', '7', |
| 33 | '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' }; |
| 34 | |
| 35 | /* Given a "flags" value passed into a formatting function, determine which |
| 36 | * formatting parameters should apply. |
| 37 | */ |
| 38 | static const struct prefix_parameters * |
| 39 | prefix_parameters_for_flags(uint16_t flags) { |
| 40 | static const char * const si_prefixes[] = {" a", " f", " p", " n", " μ", " m", " ", " k", " M", " G", " T", " P", " E"}; |
| 41 | static const struct prefix_parameters si_parameters = {si_prefixes, G_N_ELEMENTS(si_prefixes)(sizeof (si_prefixes) / sizeof ((si_prefixes)[0])), 1000, 6}; |
| 42 | static const char * const iec_prefixes[] = {" ", " Ki", " Mi", " Gi", " Ti", " Pi", " Ei"}; |
| 43 | static const struct prefix_parameters iec_parameters = {iec_prefixes, G_N_ELEMENTS(iec_prefixes)(sizeof (iec_prefixes) / sizeof ((iec_prefixes)[0])), 1024, 0}; |
| 44 | |
| 45 | return (flags & FORMAT_SIZE_PREFIX_IEC(1 << 1)) != 0 ? &iec_parameters : &si_parameters; |
| 46 | } |
| 47 | |
| 48 | char * |
| 49 | wmem_strconcat(wmem_allocator_t *allocator, const char *first, ...) |
| 50 | { |
| 51 | size_t len; |
| 52 | va_list args; |
| 53 | char *s; |
| 54 | char *concat; |
| 55 | char *ptr; |
| 56 | |
| 57 | if (!first) |
| 58 | return NULL((void*)0); |
| 59 | |
| 60 | len = 1 + strlen(first); |
| 61 | va_start(args, first)__builtin_va_start(args, first); |
| 62 | while ((s = va_arg(args, char*)__builtin_va_arg(args, char*))) { |
| 63 | len += strlen(s); |
| 64 | } |
| 65 | va_end(args)__builtin_va_end(args); |
| 66 | |
| 67 | ptr = concat = (char *)wmem_alloc(allocator, len); |
| 68 | |
| 69 | ptr = g_stpcpy(ptr, first); |
| 70 | va_start(args, first)__builtin_va_start(args, first); |
| 71 | while ((s = va_arg(args, char*)__builtin_va_arg(args, char*))) { |
| 72 | ptr = g_stpcpy(ptr, s); |
| 73 | } |
| 74 | va_end(args)__builtin_va_end(args); |
| 75 | |
| 76 | return concat; |
| 77 | } |
| 78 | |
| 79 | char * |
| 80 | wmem_strjoin(wmem_allocator_t *allocator, |
| 81 | const char *separator, const char *first, ...) |
| 82 | { |
| 83 | size_t len; |
| 84 | va_list args; |
| 85 | size_t separator_len; |
| 86 | char *s; |
| 87 | char *concat; |
| 88 | char *ptr; |
| 89 | |
| 90 | if (!first) |
| 91 | return NULL((void*)0); |
| 92 | |
| 93 | if (separator == NULL((void*)0)) { |
| 94 | separator = ""; |
| 95 | } |
| 96 | |
| 97 | separator_len = strlen (separator); |
| 98 | |
| 99 | len = 1 + strlen(first); /* + 1 for null byte */ |
| 100 | va_start(args, first)__builtin_va_start(args, first); |
| 101 | while ((s = va_arg(args, char*)__builtin_va_arg(args, char*))) { |
| 102 | len += (separator_len + strlen(s)); |
| 103 | } |
| 104 | va_end(args)__builtin_va_end(args); |
| 105 | |
| 106 | ptr = concat = (char *)wmem_alloc(allocator, len); |
| 107 | ptr = g_stpcpy(ptr, first); |
| 108 | va_start(args, first)__builtin_va_start(args, first); |
| 109 | while ((s = va_arg(args, char*)__builtin_va_arg(args, char*))) { |
| 110 | ptr = g_stpcpy(ptr, separator); |
| 111 | ptr = g_stpcpy(ptr, s); |
| 112 | } |
| 113 | va_end(args)__builtin_va_end(args); |
| 114 | |
| 115 | return concat; |
| 116 | |
| 117 | } |
| 118 | |
| 119 | char * |
| 120 | wmem_strjoinv(wmem_allocator_t *allocator, |
| 121 | const char *separator, char **str_array) |
| 122 | { |
| 123 | char *string = NULL((void*)0); |
| 124 | |
| 125 | ws_return_val_if(!str_array, NULL)do { if (1 && (!str_array)) { ws_log_full("InvalidArg" , LOG_LEVEL_WARNING, "wsutil/str_util.c", 125, __func__, "invalid argument: %s" , "!str_array"); return (((void*)0)); } } while (0); |
| 126 | |
| 127 | if (separator == NULL((void*)0)) { |
| 128 | separator = ""; |
| 129 | } |
| 130 | |
| 131 | if (str_array[0]) { |
| 132 | int i; |
| 133 | char *ptr; |
| 134 | size_t len, separator_len; |
| 135 | |
| 136 | separator_len = strlen(separator); |
| 137 | |
| 138 | /* Get first part of length. Plus one for null byte. */ |
| 139 | len = 1 + strlen(str_array[0]); |
| 140 | /* Get the full length, including the separators. */ |
| 141 | for (i = 1; str_array[i] != NULL((void*)0); i++) { |
| 142 | len += separator_len; |
| 143 | len += strlen(str_array[i]); |
| 144 | } |
| 145 | |
| 146 | /* Allocate and build the string. */ |
| 147 | string = (char *)wmem_alloc(allocator, len); |
| 148 | ptr = g_stpcpy(string, str_array[0]); |
| 149 | for (i = 1; str_array[i] != NULL((void*)0); i++) { |
| 150 | ptr = g_stpcpy(ptr, separator); |
| 151 | ptr = g_stpcpy(ptr, str_array[i]); |
| 152 | } |
| 153 | } else { |
| 154 | string = wmem_strdup(allocator, ""); |
| 155 | } |
| 156 | |
| 157 | return string; |
| 158 | |
| 159 | } |
| 160 | |
| 161 | char ** |
| 162 | wmem_strsplit(wmem_allocator_t *allocator, const char *src, |
| 163 | const char *delimiter, int max_tokens) |
| 164 | { |
| 165 | char *splitted; |
| 166 | char *s; |
| 167 | unsigned tokens; |
| 168 | unsigned sep_len; |
| 169 | unsigned i; |
| 170 | char **vec; |
| 171 | |
| 172 | if (!src || !delimiter || !delimiter[0]) |
| 173 | return NULL((void*)0); |
| 174 | |
| 175 | /* An empty string results in an empty vector. */ |
| 176 | if (!src[0]) { |
| 177 | vec = wmem_new0(allocator, char *)((char **)wmem_alloc0((allocator), sizeof(char *))); |
| 178 | return vec; |
| 179 | } |
| 180 | |
| 181 | splitted = wmem_strdup(allocator, src); |
| 182 | sep_len = (unsigned)strlen(delimiter); |
| 183 | |
| 184 | if (max_tokens < 1) |
| 185 | max_tokens = INT_MAX2147483647; |
| 186 | |
| 187 | /* Calculate the number of fields. */ |
| 188 | s = splitted; |
| 189 | tokens = 1; |
| 190 | while (tokens < (unsigned)max_tokens && (s = strstr(s, delimiter)_Generic (0 ? (s) : (void *) 1, const void *: (const char *) ( strstr (s, delimiter)), default: strstr (s, delimiter)))) { |
| 191 | s += sep_len; |
| 192 | tokens++; |
| 193 | } |
| 194 | |
| 195 | vec = wmem_alloc_array(allocator, char *, tokens + 1)((char **)wmem_alloc((allocator), (((((tokens + 1)) <= 0) || ((size_t)sizeof(char *) > (9223372036854775807L / (size_t )((tokens + 1))))) ? 0 : (sizeof(char *) * ((tokens + 1)))))); |
| 196 | |
| 197 | /* Populate the array of string tokens. */ |
| 198 | s = splitted; |
| 199 | vec[0] = s; |
| 200 | tokens = 1; |
| 201 | while (tokens < (unsigned)max_tokens && (s = strstr(s, delimiter)_Generic (0 ? (s) : (void *) 1, const void *: (const char *) ( strstr (s, delimiter)), default: strstr (s, delimiter)))) { |
| 202 | for (i = 0; i < sep_len; i++) |
| 203 | s[i] = '\0'; |
| 204 | s += sep_len; |
| 205 | vec[tokens] = s; |
| 206 | tokens++; |
| 207 | |
| 208 | } |
| 209 | |
| 210 | vec[tokens] = NULL((void*)0); |
| 211 | |
| 212 | return vec; |
| 213 | } |
| 214 | |
| 215 | /* |
| 216 | * wmem_ascii_strdown: |
| 217 | * based on g_ascii_strdown. |
| 218 | */ |
| 219 | char* |
| 220 | wmem_ascii_strdown(wmem_allocator_t *allocator, const char *str, ssize_t len) |
| 221 | { |
| 222 | char *result, *s; |
| 223 | size_t abs_len; |
| 224 | |
| 225 | g_return_val_if_fail (str != NULL, NULL)do { if ((str != ((void*)0))) { } else { g_return_if_fail_warning (((gchar*) 0), ((const char*) (__func__)), "str != NULL"); return (((void*)0)); } } while (0); |
| 226 | |
| 227 | abs_len = (len < 0) ? strlen(str) : (size_t)len; |
| 228 | |
| 229 | result = wmem_strndup(allocator, str, abs_len); |
| 230 | for (s = result; *s; s++) |
| 231 | *s = g_ascii_tolower (*s); |
| 232 | |
| 233 | return result; |
| 234 | } |
| 235 | |
| 236 | int |
| 237 | ws_xton(char ch) |
| 238 | { |
| 239 | switch (ch) { |
| 240 | case '0': return 0; |
| 241 | case '1': return 1; |
| 242 | case '2': return 2; |
| 243 | case '3': return 3; |
| 244 | case '4': return 4; |
| 245 | case '5': return 5; |
| 246 | case '6': return 6; |
| 247 | case '7': return 7; |
| 248 | case '8': return 8; |
| 249 | case '9': return 9; |
| 250 | case 'a': case 'A': return 10; |
| 251 | case 'b': case 'B': return 11; |
| 252 | case 'c': case 'C': return 12; |
| 253 | case 'd': case 'D': return 13; |
| 254 | case 'e': case 'E': return 14; |
| 255 | case 'f': case 'F': return 15; |
| 256 | default: return -1; |
| 257 | } |
| 258 | } |
| 259 | |
| 260 | /* Convert all ASCII letters to lower case, in place. */ |
| 261 | char * |
| 262 | ascii_strdown_inplace(char *str) |
| 263 | { |
| 264 | char *s; |
| 265 | |
| 266 | for (s = str; *s; s++) |
| 267 | /* What 'g_ascii_tolower (char c)' does, this should be slightly more efficient */ |
| 268 | *s = g_ascii_isupper (*s)((g_ascii_table[(guchar) (*s)] & G_ASCII_UPPER) != 0) ? *s - 'A' + 'a' : *s; |
| 269 | |
| 270 | return (str); |
| 271 | } |
| 272 | |
| 273 | /* Convert all ASCII letters to upper case, in place. */ |
| 274 | char * |
| 275 | ascii_strup_inplace(char *str) |
| 276 | { |
| 277 | char *s; |
| 278 | |
| 279 | for (s = str; *s; s++) |
| 280 | /* What 'g_ascii_toupper (char c)' does, this should be slightly more efficient */ |
| 281 | *s = g_ascii_islower (*s)((g_ascii_table[(guchar) (*s)] & G_ASCII_LOWER) != 0) ? *s - 'a' + 'A' : *s; |
| 282 | |
| 283 | return (str); |
| 284 | } |
| 285 | |
| 286 | /* Check if an entire string is printable. */ |
| 287 | bool_Bool |
| 288 | isprint_string(const char *str) |
| 289 | { |
| 290 | unsigned pos; |
| 291 | |
| 292 | /* Loop until we reach the end of the string (a null) */ |
| 293 | for(pos = 0; str[pos] != '\0'; pos++){ |
| 294 | if(!g_ascii_isprint(str[pos])((g_ascii_table[(guchar) (str[pos])] & G_ASCII_PRINT) != 0 )){ |
| 295 | /* The string contains a non-printable character */ |
| 296 | return false0; |
| 297 | } |
| 298 | } |
| 299 | |
| 300 | /* The string contains only printable characters */ |
| 301 | return true1; |
| 302 | } |
| 303 | |
| 304 | /* Check if an entire UTF-8 string is printable. */ |
| 305 | bool_Bool |
| 306 | isprint_utf8_string(const char *str, const unsigned length) |
| 307 | { |
| 308 | const char *strend = str + length; |
| 309 | |
| 310 | if (!g_utf8_validate(str, length, NULL((void*)0))) { |
| 311 | return false0; |
| 312 | } |
| 313 | |
| 314 | while (str < strend) { |
| 315 | /* This returns false for G_UNICODE_CONTROL | G_UNICODE_FORMAT | |
| 316 | * G_UNICODE_UNASSIGNED | G_UNICODE_SURROGATE |
| 317 | * XXX: Could it be ok to have certain format characters, e.g. |
| 318 | * U+00AD SOFT HYPHEN? If so, format_text() should be changed too. |
| 319 | */ |
| 320 | if (!g_unichar_isprint(g_utf8_get_char(str))) { |
| 321 | return false0; |
| 322 | } |
| 323 | str = g_utf8_next_char(str)((str) + g_utf8_skip[*(const guchar *)(str)]); |
| 324 | } |
| 325 | |
| 326 | return true1; |
| 327 | } |
| 328 | |
| 329 | /* Check if an entire string is digits. */ |
| 330 | bool_Bool |
| 331 | isdigit_string(const char *str) |
| 332 | { |
| 333 | unsigned pos; |
| 334 | |
| 335 | /* Loop until we reach the end of the string (a null) */ |
| 336 | for(pos = 0; str[pos] != '\0'; pos++){ |
| 337 | if(!g_ascii_isdigit(str[pos])((g_ascii_table[(guchar) (str[pos])] & G_ASCII_DIGIT) != 0 )){ |
| 338 | /* The string contains a non-digit character */ |
| 339 | return false0; |
| 340 | } |
| 341 | } |
| 342 | |
| 343 | /* The string contains only digits */ |
| 344 | return true1; |
| 345 | } |
| 346 | |
| 347 | const char * |
| 348 | ws_ascii_strcasestr(const char *haystack, const char *needle) |
| 349 | { |
| 350 | /* Do not use strcasestr() here, even if a system has it, as it is |
| 351 | * locale-dependent (and has different results for e.g. Turkic languages.) |
| 352 | * FreeBSD, NetBSD, macOS have a strcasestr_l() that could be used. |
| 353 | */ |
| 354 | size_t hlen = strlen(haystack); |
| 355 | size_t nlen = strlen(needle); |
| 356 | |
| 357 | while (hlen-- >= nlen) { |
| 358 | if (!g_ascii_strncasecmp(haystack, needle, nlen)) |
| 359 | return haystack; |
| 360 | haystack++; |
| 361 | } |
| 362 | return NULL((void*)0); |
| 363 | } |
| 364 | |
| 365 | /* Return the last occurrence of ch in the n bytes of haystack. |
| 366 | * If not found or n is 0, return NULL. */ |
| 367 | const uint8_t * |
| 368 | ws_memrchr(const void *_haystack, int ch, size_t n) |
| 369 | { |
| 370 | #ifdef HAVE_MEMRCHR1 |
| 371 | return memrchr(_haystack, ch, n); |
| 372 | #else |
| 373 | /* A generic implementation. This could be optimized considerably, |
| 374 | * e.g. by fetching a word at a time. |
| 375 | */ |
| 376 | if (n == 0) { |
| 377 | return NULL((void*)0); |
| 378 | } |
| 379 | const uint8_t *haystack = _haystack; |
| 380 | const uint8_t *p; |
| 381 | uint8_t c = (uint8_t)ch; |
| 382 | |
| 383 | const uint8_t *const end = haystack + n - 1; |
| 384 | |
| 385 | for (p = end; p >= haystack; --p) { |
| 386 | if (*p == c) { |
| 387 | return p; |
| 388 | } |
| 389 | } |
| 390 | |
| 391 | return NULL((void*)0); |
| 392 | #endif /* HAVE_MEMRCHR */ |
| 393 | } |
| 394 | |
| 395 | /* Return the first occurrence of ch in the null-terminated string str. |
| 396 | * If not found, return a pointer to the null-terminator. */ |
| 397 | char * |
| 398 | ws_strchrnul(const char *str, int ch) |
| 399 | { |
| 400 | #ifdef HAVE_STRCHRNUL1 |
| 401 | #ifdef __APPLE__ |
| 402 | /* strchrnul was introduced in macOS 15.4, runtime check if building |
| 403 | * with a newer SDK than that but an older deployment target. */ |
| 404 | if (__builtin_available(macOS 15.4, *)) { |
| 405 | return (char*)strchrnul(str, ch); |
| 406 | } else { |
| 407 | /* Minimal generic implementation. */ |
| 408 | while (*str != '\0' && *str != (char)ch) { |
| 409 | str++; |
| 410 | } |
| 411 | return (char *)str; |
| 412 | } |
| 413 | #else |
| 414 | /* Cast in case someone has an implementation that works like one of those |
| 415 | * fancy C23 qualifier-preserving versions. */ |
| 416 | return (char*)strchrnul(str, ch); |
| 417 | #endif |
| 418 | #else |
| 419 | /* Minimal generic implementation. */ |
| 420 | while (*str != '\0' && *str != (char)ch) { |
| 421 | str++; |
| 422 | } |
| 423 | return (char *)str; |
| 424 | #endif |
| 425 | } |
| 426 | |
| 427 | static const char *thousands_grouping_fmt; |
| 428 | static const char *thousands_grouping_fmt_flt; |
| 429 | |
| 430 | DIAG_OFF(format)clang diagnostic push clang diagnostic ignored "-Wformat" |
| 431 | static void test_printf_thousands_grouping(void) { |
| 432 | /* test whether wmem_strbuf works with "'" flag character */ |
| 433 | wmem_strbuf_t *buf = wmem_strbuf_new(NULL((void*)0), NULL((void*)0)); |
| 434 | wmem_strbuf_append_printf(buf, "%'d", 22); |
| 435 | if (g_strcmp0(wmem_strbuf_get_str(buf), "22") == 0) { |
| 436 | thousands_grouping_fmt = "%'"PRId64"l" "d"; |
| 437 | thousands_grouping_fmt_flt = "%'.*f"; |
| 438 | } else { |
| 439 | /* Don't use */ |
| 440 | thousands_grouping_fmt = "%"PRId64"l" "d"; |
| 441 | thousands_grouping_fmt_flt = "%.*f"; |
| 442 | } |
| 443 | wmem_strbuf_destroy(buf); |
| 444 | } |
| 445 | DIAG_ON(format)clang diagnostic pop |
| 446 | |
| 447 | static const char* decimal_point = NULL((void*)0); |
| 448 | |
| 449 | static void truncate_numeric_strbuf(wmem_strbuf_t *strbuf, int n) { |
| 450 | |
| 451 | const char *s = wmem_strbuf_get_str(strbuf); |
| 452 | const char *p; |
| 453 | int count; |
| 454 | |
| 455 | if (decimal_point == NULL((void*)0)) { |
| 456 | decimal_point = localeconv()->decimal_point; |
| 457 | } |
| 458 | |
| 459 | p = strchr(s, decimal_point[0])_Generic (0 ? (s) : (void *) 1, const void *: (const char *) ( strchr (s, decimal_point[0])), default: strchr (s, decimal_point [0])); |
| 460 | if (p != NULL((void*)0)) { |
| 461 | count = n; |
| 462 | while (count >= 0) { |
| 463 | count--; |
| 464 | if (*p == '\0') |
| 465 | break; |
| 466 | p++; |
| 467 | } |
| 468 | |
| 469 | p--; |
| 470 | while (*p == '0') { |
| 471 | p--; |
| 472 | } |
| 473 | |
| 474 | if (*p != decimal_point[0]) { |
| 475 | p++; |
| 476 | } |
| 477 | wmem_strbuf_truncate(strbuf, (size_t)(p - s)); |
| 478 | } |
| 479 | } |
| 480 | |
| 481 | /* Given a floating point value, return it in a human-readable format, |
| 482 | * using units with metric prefixes (falling back to scientific notation |
| 483 | * with the base units if outside the range.) |
| 484 | */ |
| 485 | char * |
| 486 | format_units(wmem_allocator_t *allocator, double size, |
| 487 | format_size_units_e unit, uint16_t flags, |
| 488 | int precision) |
| 489 | { |
| 490 | wmem_strbuf_t *human_str = wmem_strbuf_new(allocator, NULL((void*)0)); |
| 491 | bool_Bool is_small = false0; |
| 492 | /* is_small is when to use the longer, spelled out unit. |
| 493 | * We use it for inf, NaN, 0, and unprefixed small values, |
| 494 | * but not for unprefixed values using scientific notation |
| 495 | * the value is outside the supported prefix range. |
| 496 | */ |
| 497 | bool_Bool scientific = false0; |
| 498 | double abs_size = fabs(size); |
| 499 | const struct prefix_parameters * const pp = prefix_parameters_for_flags(flags); |
| 500 | int prefix_index = pp->prefix_offset; |
| 501 | char *ret_val; |
| 502 | |
| 503 | if (thousands_grouping_fmt == NULL((void*)0)) |
| 504 | test_printf_thousands_grouping(); |
| 505 | |
| 506 | if (isfinite(size)__builtin_isfinite (size) && size != 0.0) { |
| 507 | |
| 508 | double comp = precision == 0 ? 10.0 : 1.0; |
| 509 | |
| 510 | /* For precision 0, use the range [10, 10*power) because only |
| 511 | * one significant digit is not as useful. This is what format_size |
| 512 | * does for integers. ("ls -h" uses one digit after the decimal |
| 513 | * point only for the [1, 10) range, g_format_size() always displays |
| 514 | * tenths.) Prefer non-prefixed units for the range [1,10), though. |
| 515 | * |
| 516 | * We have a limited number of units to check, so this (which |
| 517 | * can be unrolled) is presumably faster than log + floor + pow/exp |
| 518 | */ |
| 519 | if (abs_size < 1.0) { |
| 520 | while (abs_size < comp) { |
| 521 | abs_size *= pp->power; |
| 522 | if (prefix_index == 0) { |
| 523 | scientific = true1; |
| 524 | break; |
| 525 | } |
| 526 | prefix_index--; |
| 527 | } |
| 528 | } else { |
| 529 | while (abs_size >= comp * pp->power) { |
| 530 | abs_size /= pp->power; |
| 531 | if (prefix_index == pp->prefix_count - 1) { |
| 532 | scientific = true1; |
| 533 | break; |
| 534 | } |
| 535 | prefix_index++; |
| 536 | } |
| 537 | } |
| 538 | } |
| 539 | |
| 540 | if (scientific) { |
| 541 | wmem_strbuf_append_printf(human_str, "%.*g", precision + 1, size); |
| 542 | prefix_index = pp->prefix_offset; |
| 543 | } else { |
| 544 | if (prefix_index == pp->prefix_offset) { |
| 545 | is_small = true1; |
| 546 | } |
| 547 | size = copysign(abs_size, size); |
| 548 | // Truncate trailing zeros, but do it this way because we know |
| 549 | // we don't want scientific notation, and we don't want %g to |
| 550 | // switch to that if precision is small. (We could always use |
| 551 | // %g when precision is large.) |
| 552 | wmem_strbuf_append_printf(human_str, thousands_grouping_fmt_flt, precision, size); |
| 553 | truncate_numeric_strbuf(human_str, precision); |
| 554 | // XXX - when rounding to a certain precision, printf might |
| 555 | // round up to "power" from something like 999.99999995, which |
| 556 | // looks a little odd on a graph when transitioning from 1,000 bytes |
| 557 | // (for values just under 1 kB) to 1 kB (for values 1 kB and larger.) |
| 558 | // Due to edge cases in binary fp representation and how printf might |
| 559 | // round things, the right way to handle it is taking the printf output |
| 560 | // and comparing it to "1000" and "1024" and adjusting the exponent |
| 561 | // if so - though we need to compare to the version with the thousands |
| 562 | // separator if we have that (which makes it harder to use strnatcmp |
| 563 | // as is.) |
| 564 | } |
| 565 | |
| 566 | wmem_strbuf_append(human_str, pp->prefix[prefix_index]); |
| 567 | |
| 568 | switch (unit) { |
| 569 | case FORMAT_SIZE_UNIT_NONE: |
| 570 | break; |
| 571 | case FORMAT_SIZE_UNIT_BYTES: |
| 572 | wmem_strbuf_append(human_str, is_small ? "bytes" : "B"); |
| 573 | break; |
| 574 | case FORMAT_SIZE_UNIT_BITS: |
| 575 | wmem_strbuf_append(human_str, is_small ? "bits" : "b"); |
| 576 | break; |
| 577 | case FORMAT_SIZE_UNIT_BITS_S: |
| 578 | wmem_strbuf_append(human_str, is_small ? "bits/s" : "bps"); |
| 579 | break; |
| 580 | case FORMAT_SIZE_UNIT_BYTES_S: |
| 581 | wmem_strbuf_append(human_str, is_small ? "bytes/s" : "Bps"); |
| 582 | break; |
| 583 | case FORMAT_SIZE_UNIT_PACKETS: |
| 584 | wmem_strbuf_append(human_str, is_small ? "packets" : "pkts"); |
| 585 | break; |
| 586 | case FORMAT_SIZE_UNIT_PACKETS_S: |
| 587 | wmem_strbuf_append(human_str, is_small ? "packets/s" : "pkts/s"); |
| 588 | break; |
| 589 | case FORMAT_SIZE_UNIT_EVENTS: |
| 590 | wmem_strbuf_append(human_str, is_small ? "events" : "evts"); |
| 591 | break; |
| 592 | case FORMAT_SIZE_UNIT_EVENTS_S: |
| 593 | wmem_strbuf_append(human_str, is_small ? "events/s" : "evts/s"); |
| 594 | break; |
| 595 | case FORMAT_SIZE_UNIT_FIELDS: |
| 596 | wmem_strbuf_append(human_str, is_small ? "fields" : "flds"); |
| 597 | break; |
| 598 | case FORMAT_SIZE_UNIT_SECONDS: |
| 599 | wmem_strbuf_append(human_str, is_small ? "seconds" : "s"); |
| 600 | break; |
| 601 | case FORMAT_SIZE_UNIT_ERLANGS: |
| 602 | wmem_strbuf_append(human_str, is_small ? "erlangs" : "E"); |
| 603 | break; |
| 604 | default: |
| 605 | ws_assert_not_reached()ws_log_fatal_full("", LOG_LEVEL_ERROR, "wsutil/str_util.c", 605 , __func__, "assertion \"not reached\" failed"); |
| 606 | } |
| 607 | |
| 608 | ret_val = wmem_strbuf_finalize(human_str); |
| 609 | /* Convention is a space between the value and the units. If we have |
| 610 | * a prefix, the space is before the prefix. There are two possible |
| 611 | * uses of FORMAT_SIZE_UNIT_NONE: |
| 612 | * 1. Add a unit immediately after the string returned. In this case, |
| 613 | * we would want the string to end with a space if there's no prefix. |
| 614 | * 2. The unit appears somewhere else, e.g. in a legend, header, or |
| 615 | * different column. In this case, we don't want the string to end |
| 616 | * with a space if there's no prefix. |
| 617 | * chomping the string here, as we've traditionally done, optimizes for |
| 618 | * the latter case but makes the former case harder. |
| 619 | * Perhaps the right approach is to distinguish the cases with a new |
| 620 | * enum value. |
| 621 | */ |
| 622 | return g_strchomp(ret_val); |
| 623 | } |
| 624 | |
| 625 | /* Given a size, return its value in a human-readable format */ |
| 626 | /* This doesn't handle fractional values. We might want to just |
| 627 | * call the version with the double and precision 0 (possibly |
| 628 | * slower due to the use of floating point math, but do we care?) |
| 629 | */ |
| 630 | char * |
| 631 | format_size_wmem(wmem_allocator_t *allocator, int64_t size, |
| 632 | format_size_units_e unit, uint16_t flags) |
| 633 | { |
| 634 | wmem_strbuf_t *human_str = wmem_strbuf_new(allocator, NULL((void*)0)); |
| 635 | bool_Bool is_small = false0; |
| 636 | const struct prefix_parameters * const pp = prefix_parameters_for_flags(flags); |
| 637 | char *ret_val; |
| 638 | |
| 639 | if (thousands_grouping_fmt == NULL((void*)0)) |
| 640 | test_printf_thousands_grouping(); |
| 641 | |
| 642 | int prefix_index = pp->prefix_offset; |
| 643 | int64_t scale = 1; |
| 644 | while (prefix_index + 1 < pp->prefix_count && scale < INT64_MAX(9223372036854775807L) / (10 * pp->power) && size >= scale * pp->power * 10) { |
| 645 | prefix_index++; |
| 646 | scale *= pp->power; |
| 647 | } |
| 648 | |
| 649 | wmem_strbuf_append_printf(human_str, thousands_grouping_fmt, size / scale); |
| 650 | wmem_strbuf_append(human_str, pp->prefix[prefix_index]); |
| 651 | is_small = prefix_index == pp->prefix_offset; |
| 652 | |
| 653 | switch (unit) { |
| 654 | case FORMAT_SIZE_UNIT_NONE: |
| 655 | break; |
| 656 | case FORMAT_SIZE_UNIT_BYTES: |
| 657 | wmem_strbuf_append(human_str, is_small ? "bytes" : "B"); |
| 658 | break; |
| 659 | case FORMAT_SIZE_UNIT_BITS: |
| 660 | wmem_strbuf_append(human_str, is_small ? "bits" : "b"); |
| 661 | break; |
| 662 | case FORMAT_SIZE_UNIT_BITS_S: |
| 663 | wmem_strbuf_append(human_str, is_small ? "bits/s" : "bps"); |
| 664 | break; |
| 665 | case FORMAT_SIZE_UNIT_BYTES_S: |
| 666 | wmem_strbuf_append(human_str, is_small ? "bytes/s" : "Bps"); |
| 667 | break; |
| 668 | case FORMAT_SIZE_UNIT_PACKETS: |
| 669 | wmem_strbuf_append(human_str, is_small ? "packets" : "pkts"); |
| 670 | break; |
| 671 | case FORMAT_SIZE_UNIT_PACKETS_S: |
| 672 | wmem_strbuf_append(human_str, is_small ? "packets/s" : "pkts/s"); |
| 673 | break; |
| 674 | case FORMAT_SIZE_UNIT_EVENTS: |
| 675 | wmem_strbuf_append(human_str, is_small ? "events" : "evts"); |
| 676 | break; |
| 677 | case FORMAT_SIZE_UNIT_EVENTS_S: |
| 678 | wmem_strbuf_append(human_str, is_small ? "events/s" : "evts/s"); |
| 679 | break; |
| 680 | case FORMAT_SIZE_UNIT_FIELDS: |
| 681 | wmem_strbuf_append(human_str, is_small ? "fields" : "flds"); |
| 682 | break; |
| 683 | case FORMAT_SIZE_UNIT_SECONDS: |
| 684 | wmem_strbuf_append(human_str, is_small ? "seconds" : "s"); |
| 685 | break; |
| 686 | case FORMAT_SIZE_UNIT_ERLANGS: |
| 687 | wmem_strbuf_append(human_str, is_small ? "erlangs" : "E"); |
| 688 | break; |
| 689 | default: |
| 690 | ws_assert_not_reached()ws_log_fatal_full("", LOG_LEVEL_ERROR, "wsutil/str_util.c", 690 , __func__, "assertion \"not reached\" failed"); |
| 691 | } |
| 692 | |
| 693 | ret_val = wmem_strbuf_finalize(human_str); |
| 694 | return g_strchomp(ret_val); |
| 695 | } |
| 696 | |
| 697 | char |
| 698 | printable_char_or_period(char c) |
| 699 | { |
| 700 | return g_ascii_isprint(c)((g_ascii_table[(guchar) (c)] & G_ASCII_PRINT) != 0) ? c : '.'; |
| 701 | } |
| 702 | |
| 703 | /* |
| 704 | * This is used by the display filter engine and must be compatible |
| 705 | * with display filter syntax. |
| 706 | */ |
| 707 | static inline bool_Bool |
| 708 | escape_char(char c, char *p) |
| 709 | { |
| 710 | int r = -1; |
| 711 | ws_assert(p)do { if ((1) && !(p)) ws_log_fatal_full("", LOG_LEVEL_ERROR , "wsutil/str_util.c", 711, __func__, "assertion failed: %s", "p"); } while (0); |
| 712 | |
| 713 | /* |
| 714 | * backslashes and double-quotes must be escaped (double-quotes |
| 715 | * are escaped by passing '"' as quote_char in escape_string_len) |
| 716 | * whitespace is also escaped. |
| 717 | */ |
| 718 | switch (c) { |
| 719 | case '\a': r = 'a'; break; |
| 720 | case '\b': r = 'b'; break; |
| 721 | case '\f': r = 'f'; break; |
| 722 | case '\n': r = 'n'; break; |
| 723 | case '\r': r = 'r'; break; |
| 724 | case '\t': r = 't'; break; |
| 725 | case '\v': r = 'v'; break; |
| 726 | case '\\': r = '\\'; break; |
| 727 | case '\0': r = '0'; break; |
| 728 | } |
| 729 | |
| 730 | if (r != -1) { |
| 731 | *p = r; |
| 732 | return true1; |
| 733 | } |
| 734 | return false0; |
| 735 | } |
| 736 | |
| 737 | static inline bool_Bool |
| 738 | escape_null(char c, char *p) |
| 739 | { |
| 740 | ws_assert(p)do { if ((1) && !(p)) ws_log_fatal_full("", LOG_LEVEL_ERROR , "wsutil/str_util.c", 740, __func__, "assertion failed: %s", "p"); } while (0); |
| 741 | if (c == '\0') { |
| 742 | *p = '0'; |
| 743 | return true1; |
| 744 | } |
| 745 | return false0; |
| 746 | } |
| 747 | |
| 748 | static char * |
| 749 | escape_string_len(wmem_allocator_t *alloc, const char *string, ssize_t len, |
| 750 | bool_Bool (*escape_func)(char c, char *p), bool_Bool add_quotes, |
| 751 | char quote_char, bool_Bool double_quote) |
| 752 | { |
| 753 | char c, r; |
| 754 | wmem_strbuf_t *buf; |
| 755 | size_t abs_len, alloc_size, i; |
| 756 | |
| 757 | abs_len = (len < 0) ? strlen(string) : (size_t)len; |
| 758 | |
| 759 | alloc_size = abs_len; |
| 760 | if (add_quotes) |
| 761 | alloc_size += 2; |
| 762 | |
| 763 | buf = wmem_strbuf_new_sized(alloc, alloc_size); |
| 764 | |
| 765 | if (add_quotes && quote_char != '\0') |
| 766 | wmem_strbuf_append_c(buf, quote_char); |
| 767 | |
| 768 | for (i = 0; i < abs_len; i++) { |
| 769 | c = string[i]; |
| 770 | if ((escape_func(c, &r))) { |
| 771 | wmem_strbuf_append_c(buf, '\\'); |
| 772 | wmem_strbuf_append_c(buf, r); |
| 773 | } |
| 774 | else if (c == quote_char && quote_char != '\0') { |
| 775 | /* If quoting, we must escape the quote_char somehow. */ |
| 776 | if (double_quote) { |
| 777 | wmem_strbuf_append_c(buf, c); |
| 778 | wmem_strbuf_append_c(buf, c); |
| 779 | } else { |
| 780 | wmem_strbuf_append_c(buf, '\\'); |
| 781 | wmem_strbuf_append_c(buf, c); |
| 782 | } |
| 783 | } |
| 784 | else if (c == '\\' && quote_char != '\0' && !double_quote) { |
| 785 | /* If quoting, and escaping the quote_char with a backslash, |
| 786 | * then backslash must be escaped, even if escape_func doesn't. */ |
| 787 | wmem_strbuf_append_c(buf, '\\'); |
| 788 | wmem_strbuf_append_c(buf, '\\'); |
| 789 | } |
| 790 | else { |
| 791 | /* Other UTF-8 bytes are passed through. */ |
| 792 | wmem_strbuf_append_c(buf, c); |
| 793 | } |
| 794 | } |
| 795 | |
| 796 | if (add_quotes && quote_char != '\0') |
| 797 | wmem_strbuf_append_c(buf, quote_char); |
| 798 | |
| 799 | return wmem_strbuf_finalize(buf); |
| 800 | } |
| 801 | |
| 802 | char * |
| 803 | ws_escape_string_len(wmem_allocator_t *alloc, const char *string, ssize_t len, bool_Bool add_quotes) |
| 804 | { |
| 805 | return escape_string_len(alloc, string, len, escape_char, add_quotes, '"', false0); |
| 806 | } |
| 807 | |
| 808 | char * |
| 809 | ws_escape_string(wmem_allocator_t *alloc, const char *string, bool_Bool add_quotes) |
| 810 | { |
| 811 | return escape_string_len(alloc, string, -1, escape_char, add_quotes, '"', false0); |
| 812 | } |
| 813 | |
| 814 | char *ws_escape_null(wmem_allocator_t *alloc, const char *string, size_t len, bool_Bool add_quotes) |
| 815 | { |
| 816 | /* XXX: The existing behavior (maintained) here is not to escape |
| 817 | * backslashes even though NUL is escaped. |
| 818 | */ |
| 819 | return escape_string_len(alloc, string, len, escape_null, add_quotes, add_quotes ? '"' : '\0', false0); |
| 820 | } |
| 821 | |
| 822 | char *ws_escape_csv(wmem_allocator_t *alloc, const char *string, bool_Bool add_quotes, char quote_char, bool_Bool double_quote, bool_Bool escape_whitespace) |
| 823 | { |
| 824 | if (escape_whitespace) |
| 825 | return escape_string_len(alloc, string, -1, escape_char, add_quotes, quote_char, double_quote); |
| 826 | else |
| 827 | return escape_string_len(alloc, string, -1, escape_null, add_quotes, quote_char, double_quote); |
| 828 | } |
| 829 | |
| 830 | bool_Bool ws_csv_value_is_formula(const char *string) |
| 831 | { |
| 832 | if (string == NULL((void*)0)) |
| 833 | return false0; |
| 834 | |
| 835 | /* Some spreadsheet applications strip leading spaces on import and then |
| 836 | * evaluate what follows, so look past them. */ |
| 837 | while (*string == ' ') |
| 838 | string++; |
| 839 | |
| 840 | switch (*string) { |
| 841 | case '=': /* Formula */ |
| 842 | case '+': /* Formula */ |
| 843 | case '-': /* Formula */ |
| 844 | case '@': /* Lotus-style function, still honored for compatibility */ |
| 845 | case '\t': /* Stripped on import, so it can precede any of the above */ |
| 846 | case '\r': |
| 847 | return true1; |
| 848 | default: |
| 849 | return false0; |
| 850 | } |
| 851 | } |
| 852 | |
| 853 | typedef enum { |
| 854 | UNESCAPE_NONE, |
| 855 | UNESCAPE_ESCAPE, |
| 856 | } unescape_state_e; |
| 857 | |
| 858 | wmem_strbuf_t *ws_unescape_string_len(wmem_allocator_t *alloc, const uint8_t *string, ssize_t len, GError **err) |
| 859 | { |
| 860 | wmem_strbuf_t *buf; |
| 861 | size_t abs_len, i; |
| 862 | unsigned char c; |
| 863 | unescape_state_e state = UNESCAPE_NONE; |
| 864 | bool_Bool possibly_invalid = false0; |
| 865 | int value; |
| 866 | gunichar cp; |
| 867 | |
| 868 | abs_len = (len < 0) ? strlen((const char*)string) : (size_t)len; |
| 869 | |
| 870 | buf = wmem_strbuf_new_sized(alloc, abs_len); |
| 871 | |
| 872 | /* With \u and \U escapes, we can only produce a result by knowing the |
| 873 | * intended final encoding. (We assume UTF-8 here.) \x escapes do not |
| 874 | * assume the final encoding, but neither do they ensure that the result |
| 875 | * will be valid in whatever encoding is used. The other escapes still |
| 876 | * assume an ASCII-like encoding for the C0 control characters. |
| 877 | * |
| 878 | * This attempts to be permissive and allow C, C++, Python, and JSON |
| 879 | * escapes, though C and C++ \x escapes longer than two characters are |
| 880 | * not allowed, since the target encoding is UTF-8. */ |
| 881 | for (i = 0; i < abs_len; i++) { |
| 882 | c = string[i]; |
| 883 | switch (state) { |
| 884 | case UNESCAPE_NONE: |
| 885 | switch (c) { |
| 886 | case '\\': |
| 887 | state = UNESCAPE_ESCAPE; |
| 888 | break; |
| 889 | default: |
| 890 | if (c & 0x80) { |
| 891 | /* XXX - Use ws_utf8_char_len and validate now? |
| 892 | * Trickier because there might, somehow, be a single |
| 893 | * UTF-8 character which is a mix of escaped and |
| 894 | * unescaped bytes. */ |
| 895 | possibly_invalid = true1; |
| 896 | } |
| 897 | /* Pass through. */ |
| 898 | wmem_strbuf_append_c(buf, c); |
| 899 | } |
| 900 | break; |
| 901 | case UNESCAPE_ESCAPE: |
| 902 | switch (c) { |
| 903 | case 'a': |
| 904 | wmem_strbuf_append_c(buf, '\a'); |
| 905 | break; |
| 906 | case 'b': |
| 907 | wmem_strbuf_append_c(buf, '\b'); |
| 908 | break; |
| 909 | case 'f': |
| 910 | wmem_strbuf_append_c(buf, '\f'); |
| 911 | break; |
| 912 | case 'n': |
| 913 | wmem_strbuf_append_c(buf, '\n'); |
| 914 | break; |
| 915 | case 'r': |
| 916 | wmem_strbuf_append_c(buf, '\r'); |
| 917 | break; |
| 918 | case 't': |
| 919 | wmem_strbuf_append_c(buf, '\t'); |
| 920 | break; |
| 921 | case 'v': |
| 922 | wmem_strbuf_append_c(buf, '\v'); |
| 923 | break; |
| 924 | case '\\': |
| 925 | wmem_strbuf_append_c(buf, '\\'); |
| 926 | break; |
| 927 | case '\"': |
| 928 | wmem_strbuf_append_c(buf, '\"'); |
| 929 | break; |
| 930 | case '\'': |
| 931 | // C, C++, Python (not JSON) |
| 932 | wmem_strbuf_append_c(buf, '\''); |
| 933 | break; |
| 934 | case '?': |
| 935 | // C and C++ \? for supporting the late unlamented trigraphs |
| 936 | wmem_strbuf_append_c(buf, '?'); |
| 937 | break; |
| 938 | case 'x': |
| 939 | // C and C++ allow an unlimited number of hex characters, |
| 940 | // but the end value must fit in a code unit of the string |
| 941 | // type, e.g. 8-bit for UTF-8. (Not a Unicode code point.) |
| 942 | // Python requires two characters exactly (which also means |
| 943 | // that if the third is valid hex digit that's fine.) |
| 944 | // JSON doesn't support. |
| 945 | ++i; |
| 946 | if (i <= abs_len) { |
| 947 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_PARTIAL_INPUT, |
| 948 | "Hexadecimal character missing after \\x"); |
| 949 | goto out; |
| 950 | } |
| 951 | value = ws_xton(string[i]); |
| 952 | if (value == -1) { |
| 953 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 954 | "Non-hexadecimal character after \\x"); |
| 955 | goto out; |
| 956 | } |
| 957 | if (((abs_len - i) > 1) && ws_xton(string[i+1])) { |
| 958 | ++i; |
| 959 | value <<= 4; |
| 960 | value |= ws_xton(string[i]); |
| 961 | } |
| 962 | c = (char)value; |
| 963 | if (c & 0x80) { |
| 964 | possibly_invalid = true1; |
| 965 | } |
| 966 | wmem_strbuf_append_c(buf, c); |
| 967 | break; |
| 968 | case 'u': |
| 969 | ++i; |
| 970 | if ((abs_len - i) < 4) { |
| 971 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_PARTIAL_INPUT, |
| 972 | "\\u must be followed by four characters"); |
| 973 | goto out; |
| 974 | } |
| 975 | if (!ws_hexbuftou32(&string[i], 4, NULL((void*)0), &cp)) { |
| 976 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 977 | "\\u must be followed by four hexadecimal characters"); |
| 978 | goto out; |
| 979 | } |
| 980 | i += 4; |
| 981 | /* JSON (but not C, C++, Python) can represent code points outside |
| 982 | * the BMP with UTF-16 surrogate pairs. */ |
| 983 | if (IS_LEAD_SURROGATE(cp)((cp) >= 0xd800 && (cp) < 0xdc00)) { |
| 984 | /* high surrogate */ |
| 985 | if ((abs_len - i) < 6) { |
| 986 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_PARTIAL_INPUT, |
| 987 | "\\u high surrogate must be followed by escaped low surrogate"); |
| 988 | goto out; |
| 989 | } |
| 990 | uint16_t second_code; |
| 991 | if (string[i] != '\\' || string[i + 1] != 'u' || |
| 992 | !ws_hexbuftou16(&string[i+2], 4, NULL((void*)0), &second_code)) { |
| 993 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 994 | "\\u high surrogate must be followed by escaped low surrogate"); |
| 995 | goto out; |
| 996 | } |
| 997 | i += 6; |
| 998 | if (!IS_TRAIL_SURROGATE(second_code)((second_code) >= 0xdc00 && (second_code) < 0xe000 )) { |
| 999 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 1000 | "\\u followed by unpaired UTF-16 surrogate"); |
| 1001 | goto out; |
| 1002 | } |
| 1003 | cp = SURROGATE_VALUE(cp, second_code)(((((cp) - 0xd800) << 10) | ((second_code) - 0xdc00)) + 0x10000); |
| 1004 | } else if (IS_TRAIL_SURROGATE(cp)((cp) >= 0xdc00 && (cp) < 0xe000)) { |
| 1005 | /* isolated low surrogate, not allowed. */ |
| 1006 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 1007 | "\\u followed by isolated low surrogate"); |
| 1008 | goto out; |
| 1009 | } |
| 1010 | wmem_strbuf_append_unichar_validated(buf, cp); |
| 1011 | break; |
| 1012 | case 'U': |
| 1013 | ++i; |
| 1014 | if ((abs_len - i) < 8) { |
| 1015 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_PARTIAL_INPUT, |
| 1016 | "\\U must be followed by eight characters"); |
| 1017 | goto out; |
| 1018 | } |
| 1019 | if (!ws_hexbuftou32(&string[i], 8, NULL((void*)0), &cp)) { |
| 1020 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 1021 | "\\U must be followed by eight hexadecimal characters"); |
| 1022 | goto out; |
| 1023 | } |
| 1024 | wmem_strbuf_append_unichar_validated(buf, cp); |
| 1025 | break; |
| 1026 | default: |
| 1027 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_ILLEGAL_SEQUENCE, |
| 1028 | "Unknown escape sequence"); |
| 1029 | goto out; |
| 1030 | } |
| 1031 | state = UNESCAPE_NONE; |
| 1032 | break; |
| 1033 | default: |
| 1034 | /* Pass through. */ |
| 1035 | wmem_strbuf_append_c(buf, c); |
| 1036 | } |
| 1037 | } |
| 1038 | switch (state) { |
| 1039 | case UNESCAPE_ESCAPE: |
| 1040 | g_set_error(err, G_CONVERT_ERRORg_convert_error_quark(), G_CONVERT_ERROR_PARTIAL_INPUT, |
| 1041 | "\\ at the end of input"); |
| 1042 | goto out; |
| 1043 | default: |
| 1044 | break; |
| 1045 | } |
| 1046 | |
| 1047 | out: |
| 1048 | /* Right now on error this passes through what was successfully decoded. |
| 1049 | * We could also replace with replacement characters and continue if |
| 1050 | * possible, or return a strbuf with an empty string. */ |
| 1051 | if (possibly_invalid) { |
| 1052 | wmem_strbuf_utf8_make_valid(buf); |
| 1053 | } |
| 1054 | return buf; |
| 1055 | } |
| 1056 | |
| 1057 | const char * |
| 1058 | ws_strerrorname_r(int errnum, char *buf, size_t buf_size) |
| 1059 | { |
| 1060 | #ifdef HAVE_STRERRORNAME_NP1 |
| 1061 | const char *errstr = strerrorname_np(errnum); |
| 1062 | if (errstr != NULL((void*)0)) { |
| 1063 | (void)g_strlcpy(buf, errstr, buf_size); |
| 1064 | return buf; |
| 1065 | } |
| 1066 | #endif |
| 1067 | snprintf(buf, buf_size, "Errno(%d)", errnum); |
| 1068 | return buf; |
| 1069 | } |
| 1070 | |
| 1071 | char * |
| 1072 | ws_strdup_underline(wmem_allocator_t *allocator, long offset, size_t len) |
| 1073 | { |
| 1074 | if (offset < 0) |
| 1075 | return NULL((void*)0); |
| 1076 | |
| 1077 | wmem_strbuf_t *buf = wmem_strbuf_new_sized(allocator, offset + len); |
| 1078 | |
| 1079 | for (int i = 0; i < offset; i++) { |
| 1080 | wmem_strbuf_append_c(buf, ' '); |
| 1081 | } |
| 1082 | wmem_strbuf_append_c(buf, '^'); |
| 1083 | |
| 1084 | for (size_t l = len; l > 1; l--) { |
| 1085 | wmem_strbuf_append_c(buf, '~'); |
| 1086 | } |
| 1087 | |
| 1088 | return wmem_strbuf_finalize(buf); |
| 1089 | } |
| 1090 | |
| 1091 | #define INITIAL_FMTBUF_SIZE128 128 |
| 1092 | |
| 1093 | /* |
| 1094 | * Declare, and initialize, the variables used for an output buffer. |
| 1095 | */ |
| 1096 | #define FMTBUF_VARSchar *fmtbuf = (char*)wmem_alloc(allocator, 128); unsigned fmtbuf_len = 128; unsigned column = 0 \ |
| 1097 | char *fmtbuf = (char*)wmem_alloc(allocator, INITIAL_FMTBUF_SIZE128); \ |
| 1098 | unsigned fmtbuf_len = INITIAL_FMTBUF_SIZE128; \ |
| 1099 | unsigned column = 0 |
| 1100 | |
| 1101 | /* |
| 1102 | * Expand the buffer to be large enough to add nbytes bytes, plus a |
| 1103 | * terminating '\0'. |
| 1104 | */ |
| 1105 | #define FMTBUF_EXPAND(nbytes)if (column+(nbytes+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1105, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(nbytes+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + nbytes + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1105, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); } \ |
| 1106 | /* \ |
| 1107 | * Is there enough room for those bytes and also enough room for \ |
| 1108 | * a terminating '\0'? \ |
| 1109 | */ \ |
| 1110 | if (column+(nbytes+1) >= fmtbuf_len) { \ |
| 1111 | /* \ |
| 1112 | * Double the buffer's size if it's not big enough. \ |
| 1113 | * The size of the buffer starts at 128, so doubling its size \ |
| 1114 | * adds at least another 128 bytes, which is more than enough \ |
| 1115 | * for one more character plus a terminating '\0'. \ |
| 1116 | */ \ |
| 1117 | if (ckd_mul(&fmtbuf_len, fmtbuf_len, 2)__builtin_mul_overflow((fmtbuf_len), (2), (&fmtbuf_len))) { \ |
| 1118 | ws_debug("overflow.")do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1118, __func__, "overflow."); } } while (0); \ |
| 1119 | FMTBUF_ENDSTRfmtbuf[column] = '\0'; \ |
| 1120 | return fmtbuf; \ |
| 1121 | } \ |
| 1122 | if (column+(nbytes+1) >= fmtbuf_len) { \ |
| 1123 | if (ckd_add(&fmtbuf_len, fmtbuf_len, (column + nbytes + 2) - fmtbuf_len)__builtin_add_overflow((fmtbuf_len), ((column + nbytes + 2) - fmtbuf_len), (&fmtbuf_len))) { \ |
| 1124 | ws_debug("overflow.")do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1124, __func__, "overflow."); } } while (0); \ |
| 1125 | FMTBUF_ENDSTRfmtbuf[column] = '\0'; \ |
| 1126 | return fmtbuf; \ |
| 1127 | } \ |
| 1128 | } \ |
| 1129 | fmtbuf = (char *)wmem_realloc(allocator, fmtbuf, fmtbuf_len); \ |
| 1130 | } |
| 1131 | |
| 1132 | /* |
| 1133 | * Put a byte into the buffer; space must have been ensured for it. |
| 1134 | */ |
| 1135 | #define FMTBUF_PUTCHAR(b)fmtbuf[column] = (b); column++ \ |
| 1136 | fmtbuf[column] = (b); \ |
| 1137 | column++ |
| 1138 | |
| 1139 | /* |
| 1140 | * Add the one-byte argument, as an octal escape sequence, to the end |
| 1141 | * of the buffer. |
| 1142 | */ |
| 1143 | #define FMTBUF_PUTBYTE_OCTAL(b)fmtbuf[column] = ((((b)>>6)&03) + '0'); column++; fmtbuf [column] = ((((b)>>3)&07) + '0'); column++; fmtbuf[ column] = ((((b)>>0)&07) + '0'); column++ \ |
| 1144 | FMTBUF_PUTCHAR((((b)>>6)&03) + '0')fmtbuf[column] = ((((b)>>6)&03) + '0'); column++; \ |
| 1145 | FMTBUF_PUTCHAR((((b)>>3)&07) + '0')fmtbuf[column] = ((((b)>>3)&07) + '0'); column++; \ |
| 1146 | FMTBUF_PUTCHAR((((b)>>0)&07) + '0')fmtbuf[column] = ((((b)>>0)&07) + '0'); column++ |
| 1147 | |
| 1148 | /* |
| 1149 | * Add the one-byte argument, as a hex escape sequence, to the end |
| 1150 | * of the buffer. |
| 1151 | */ |
| 1152 | #define FMTBUF_PUTBYTE_HEX(b)fmtbuf[column] = ('\\'); column++; fmtbuf[column] = ('x'); column ++; fmtbuf[column] = (hex[((b) >> 4) & 0xF]); column ++; fmtbuf[column] = (hex[((b) >> 0) & 0xF]); column ++ \ |
| 1153 | FMTBUF_PUTCHAR('\\')fmtbuf[column] = ('\\'); column++; \ |
| 1154 | FMTBUF_PUTCHAR('x')fmtbuf[column] = ('x'); column++; \ |
| 1155 | FMTBUF_PUTCHAR(hex[((b) >> 4) & 0xF])fmtbuf[column] = (hex[((b) >> 4) & 0xF]); column++; \ |
| 1156 | FMTBUF_PUTCHAR(hex[((b) >> 0) & 0xF])fmtbuf[column] = (hex[((b) >> 0) & 0xF]); column++ |
| 1157 | |
| 1158 | #define FMTBUF_PUTBYTES(bytes, len)if (column+(len+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1158, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(len+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + len + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1158, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); } memcpy(&fmtbuf[column], bytes, len ); column += (unsigned)len; \ |
| 1159 | FMTBUF_EXPAND(len)if (column+(len+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1159, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(len+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + len + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1159, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); } \ |
| 1160 | memcpy(&fmtbuf[column], bytes, len); \ |
| 1161 | column += (unsigned)len; // FMTBUF_EXPAND checks for overflow |
| 1162 | |
| 1163 | /* |
| 1164 | * Put the trailing '\0' at the end of the buffer. |
| 1165 | */ |
| 1166 | #define FMTBUF_ENDSTRfmtbuf[column] = '\0' \ |
| 1167 | fmtbuf[column] = '\0' |
| 1168 | |
| 1169 | static char * |
| 1170 | format_text_internal(wmem_allocator_t *allocator, |
| 1171 | const unsigned char *string, size_t len, |
| 1172 | bool_Bool replace_space) |
| 1173 | { |
| 1174 | FMTBUF_VARSchar *fmtbuf = (char*)wmem_alloc(allocator, 128); unsigned fmtbuf_len = 128; unsigned column = 0; |
| 1175 | const unsigned char *prev = string; |
| 1176 | const unsigned char *stringend = string + len; |
| 1177 | unsigned char c; |
| 1178 | size_t printable_bytes = 0; |
| 1179 | |
| 1180 | while (string < stringend) { |
| 1181 | /* |
| 1182 | * Get the first byte of this character. |
| 1183 | */ |
| 1184 | c = *string++; |
| 1185 | if ((0x20 <= c) && (c < 0x7F)) { |
| 1186 | /* |
| 1187 | * Printable ASCII, so not part of a multi-byte UTF-8 sequence. |
| 1188 | * Make sure there's enough room for one more byte, and add |
| 1189 | * the character. |
| 1190 | */ |
| 1191 | printable_bytes++; |
| 1192 | } else { |
| 1193 | if (printable_bytes) { |
| 1194 | FMTBUF_PUTBYTES(prev, printable_bytes)if (column+(printable_bytes+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1194, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(printable_bytes+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + printable_bytes + 2) - fmtbuf_len), (&fmtbuf_len))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG , "wsutil/str_util.c", 1194, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator, fmtbuf, fmtbuf_len); } memcpy(& fmtbuf[column], prev, printable_bytes); column += (unsigned)printable_bytes ;; |
| 1195 | printable_bytes = 0; |
| 1196 | } |
| 1197 | if (replace_space && g_ascii_isspace(c)((g_ascii_table[(guchar) (c)] & G_ASCII_SPACE) != 0)) { |
| 1198 | /* |
| 1199 | * ASCII, so not part of a multi-byte UTF-8 sequence, but |
| 1200 | * not printable, but is a space character; show it as a |
| 1201 | * blank. |
| 1202 | * |
| 1203 | * Make sure there's enough room for one more byte, and add |
| 1204 | * the blank. |
| 1205 | */ |
| 1206 | FMTBUF_EXPAND(1)if (column+(1 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1206, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(1 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 1 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1206, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1207 | FMTBUF_PUTCHAR(' ')fmtbuf[column] = (' '); column++; |
| 1208 | } else if (c < 128) { |
| 1209 | /* |
| 1210 | * ASCII, so not part of a multi-byte UTF-8 sequence, but not |
| 1211 | * printable. |
| 1212 | * |
| 1213 | * That requires a minimum of 2 bytes, one for the backslash |
| 1214 | * and one for a letter, so make sure we have enough room |
| 1215 | * for that, plus a trailing '\0'. |
| 1216 | */ |
| 1217 | FMTBUF_EXPAND(2)if (column+(2 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1217, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(2 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 2 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1217, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1218 | FMTBUF_PUTCHAR('\\')fmtbuf[column] = ('\\'); column++; |
| 1219 | switch (c) { |
| 1220 | |
| 1221 | case '\a': |
| 1222 | FMTBUF_PUTCHAR('a')fmtbuf[column] = ('a'); column++; |
| 1223 | break; |
| 1224 | |
| 1225 | case '\b': |
| 1226 | FMTBUF_PUTCHAR('b')fmtbuf[column] = ('b'); column++; /* BS */ |
| 1227 | break; |
| 1228 | |
| 1229 | case '\f': |
| 1230 | FMTBUF_PUTCHAR('f')fmtbuf[column] = ('f'); column++; /* FF */ |
| 1231 | break; |
| 1232 | |
| 1233 | case '\n': |
| 1234 | FMTBUF_PUTCHAR('n')fmtbuf[column] = ('n'); column++; /* NL */ |
| 1235 | break; |
| 1236 | |
| 1237 | case '\r': |
| 1238 | FMTBUF_PUTCHAR('r')fmtbuf[column] = ('r'); column++; /* CR */ |
| 1239 | break; |
| 1240 | |
| 1241 | case '\t': |
| 1242 | FMTBUF_PUTCHAR('t')fmtbuf[column] = ('t'); column++; /* tab */ |
| 1243 | break; |
| 1244 | |
| 1245 | case '\v': |
| 1246 | FMTBUF_PUTCHAR('v')fmtbuf[column] = ('v'); column++; |
| 1247 | break; |
| 1248 | |
| 1249 | default: |
| 1250 | /* |
| 1251 | * We've already put the backslash, but this |
| 1252 | * will put 3 more characters for the octal |
| 1253 | * number; make sure we have enough room for |
| 1254 | * that, plus the trailing '\0'. |
| 1255 | */ |
| 1256 | FMTBUF_EXPAND(3)if (column+(3 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1256, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(3 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 3 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1256, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1257 | FMTBUF_PUTBYTE_OCTAL(c)fmtbuf[column] = ((((c)>>6)&03) + '0'); column++; fmtbuf [column] = ((((c)>>3)&07) + '0'); column++; fmtbuf[ column] = ((((c)>>0)&07) + '0'); column++; |
| 1258 | break; |
| 1259 | } |
| 1260 | } else { |
| 1261 | /* |
| 1262 | * We've fetched the first byte of a multi-byte UTF-8 |
| 1263 | * sequence into c. |
| 1264 | */ |
| 1265 | int utf8_len; |
| 1266 | unsigned char mask; |
| 1267 | gunichar uc; |
| 1268 | unsigned char first; |
| 1269 | |
| 1270 | if ((c & 0xe0) == 0xc0) { |
| 1271 | /* Starts a 2-byte UTF-8 sequence; 1 byte left */ |
| 1272 | utf8_len = 1; |
| 1273 | mask = 0x1f; |
| 1274 | } else if ((c & 0xf0) == 0xe0) { |
| 1275 | /* Starts a 3-byte UTF-8 sequence; 2 bytes left */ |
| 1276 | utf8_len = 2; |
| 1277 | mask = 0x0f; |
| 1278 | } else if ((c & 0xf8) == 0xf0) { |
| 1279 | /* Starts a 4-byte UTF-8 sequence; 3 bytes left */ |
| 1280 | utf8_len = 3; |
| 1281 | mask = 0x07; |
| 1282 | } else if ((c & 0xfc) == 0xf8) { |
| 1283 | /* Starts an old-style 5-byte UTF-8 sequence; 4 bytes left */ |
| 1284 | utf8_len = 4; |
| 1285 | mask = 0x03; |
| 1286 | } else if ((c & 0xfe) == 0xfc) { |
| 1287 | /* Starts an old-style 6-byte UTF-8 sequence; 5 bytes left */ |
| 1288 | utf8_len = 5; |
| 1289 | mask = 0x01; |
| 1290 | } else { |
| 1291 | /* 0xfe or 0xff or a continuation byte - not valid */ |
| 1292 | utf8_len = -1; |
| 1293 | } |
| 1294 | if (utf8_len > 0) { |
| 1295 | /* Try to construct the Unicode character */ |
| 1296 | uc = c & mask; |
| 1297 | for (int i = 0; i < utf8_len; i++) { |
| 1298 | if (string >= stringend) { |
| 1299 | /* |
| 1300 | * Ran out of octets, so the character is |
| 1301 | * incomplete. Put in a REPLACEMENT CHARACTER |
| 1302 | * instead, and then continue the loop, which |
| 1303 | * will terminate. |
| 1304 | */ |
| 1305 | uc = UNICODE_REPLACEMENT_CHARACTER0x00FFFD; |
| 1306 | break; |
| 1307 | } |
| 1308 | c = *string; |
| 1309 | if ((c & 0xc0) != 0x80) { |
| 1310 | /* |
| 1311 | * Not valid UTF-8 continuation character; put in |
| 1312 | * a replacement character, and then re-process |
| 1313 | * this octet as the beginning of a new character. |
| 1314 | */ |
| 1315 | uc = UNICODE_REPLACEMENT_CHARACTER0x00FFFD; |
| 1316 | break; |
| 1317 | } |
| 1318 | string++; |
| 1319 | uc = (uc << 6) | (c & 0x3f); |
| 1320 | } |
| 1321 | |
| 1322 | /* |
| 1323 | * If this isn't a valid Unicode character, put in |
| 1324 | * a REPLACEMENT CHARACTER. |
| 1325 | */ |
| 1326 | if (!g_unichar_validate(uc)) |
| 1327 | uc = UNICODE_REPLACEMENT_CHARACTER0x00FFFD; |
| 1328 | } else { |
| 1329 | /* 0xfe or 0xff; put it a REPLACEMENT CHARACTER */ |
| 1330 | uc = UNICODE_REPLACEMENT_CHARACTER0x00FFFD; |
| 1331 | } |
| 1332 | |
| 1333 | /* |
| 1334 | * OK, is it a printable Unicode character? |
| 1335 | */ |
| 1336 | if (g_unichar_isprint(uc)) { |
| 1337 | /* |
| 1338 | * Yes - put it into the string as UTF-8. |
| 1339 | * This means that if it was an overlong |
| 1340 | * encoding, this will put out the right |
| 1341 | * sized encoding. |
| 1342 | */ |
| 1343 | if (uc < 0x80) { |
| 1344 | first = 0; |
| 1345 | utf8_len = 1; |
| 1346 | } else if (uc < 0x800) { |
| 1347 | first = 0xc0; |
| 1348 | utf8_len = 2; |
| 1349 | } else if (uc < 0x10000) { |
| 1350 | first = 0xe0; |
| 1351 | utf8_len = 3; |
| 1352 | } else if (uc < 0x200000) { |
| 1353 | first = 0xf0; |
| 1354 | utf8_len = 4; |
| 1355 | } else if (uc < 0x4000000) { |
| 1356 | /* |
| 1357 | * This should never happen, as Unicode doesn't |
| 1358 | * go that high. |
| 1359 | */ |
| 1360 | first = 0xf8; |
| 1361 | utf8_len = 5; |
| 1362 | } else { |
| 1363 | /* |
| 1364 | * This should never happen, as Unicode doesn't |
| 1365 | * go that high. |
| 1366 | */ |
| 1367 | first = 0xfc; |
| 1368 | utf8_len = 6; |
| 1369 | } |
| 1370 | FMTBUF_EXPAND(utf8_len)if (column+(utf8_len+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1370, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(utf8_len+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + utf8_len + 2) - fmtbuf_len), (& fmtbuf_len))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG , "wsutil/str_util.c", 1370, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator, fmtbuf, fmtbuf_len); }; |
| 1371 | for (int i = utf8_len - 1; i > 0; i--) { |
| 1372 | fmtbuf[column + i] = (uc & 0x3f) | 0x80; |
| 1373 | uc >>= 6; |
| 1374 | } |
| 1375 | fmtbuf[column] = uc | first; |
| 1376 | column += utf8_len; |
| 1377 | } else if (replace_space && g_unichar_isspace(uc)) { |
| 1378 | /* |
| 1379 | * Not printable, but is a space character; show it |
| 1380 | * as a blank. |
| 1381 | * |
| 1382 | * Make sure there's enough room for one more byte, |
| 1383 | * and add the blank. |
| 1384 | */ |
| 1385 | FMTBUF_EXPAND(1)if (column+(1 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1385, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(1 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 1 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1385, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1386 | FMTBUF_PUTCHAR(' ')fmtbuf[column] = (' '); column++; |
| 1387 | } else if (c < 128) { |
| 1388 | /* |
| 1389 | * ASCII, but not printable. |
| 1390 | * Yes, this could happen with an overlong encoding. |
| 1391 | * |
| 1392 | * That requires a minimum of 2 bytes, one for the |
| 1393 | * backslash and one for a letter, so make sure we |
| 1394 | * have enough room for that, plus a trailing '\0'. |
| 1395 | */ |
| 1396 | FMTBUF_EXPAND(2)if (column+(2 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1396, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(2 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 2 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1396, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1397 | FMTBUF_PUTCHAR('\\')fmtbuf[column] = ('\\'); column++; |
| 1398 | switch (c) { |
| 1399 | |
| 1400 | case '\a': |
| 1401 | FMTBUF_PUTCHAR('a')fmtbuf[column] = ('a'); column++; |
| 1402 | break; |
| 1403 | |
| 1404 | case '\b': |
| 1405 | FMTBUF_PUTCHAR('b')fmtbuf[column] = ('b'); column++; /* BS */ |
| 1406 | break; |
| 1407 | |
| 1408 | case '\f': |
| 1409 | FMTBUF_PUTCHAR('f')fmtbuf[column] = ('f'); column++; /* FF */ |
| 1410 | break; |
| 1411 | |
| 1412 | case '\n': |
| 1413 | FMTBUF_PUTCHAR('n')fmtbuf[column] = ('n'); column++; /* NL */ |
| 1414 | break; |
| 1415 | |
| 1416 | case '\r': |
| 1417 | FMTBUF_PUTCHAR('r')fmtbuf[column] = ('r'); column++; /* CR */ |
| 1418 | break; |
| 1419 | |
| 1420 | case '\t': |
| 1421 | FMTBUF_PUTCHAR('t')fmtbuf[column] = ('t'); column++; /* tab */ |
| 1422 | break; |
| 1423 | |
| 1424 | case '\v': |
| 1425 | FMTBUF_PUTCHAR('v')fmtbuf[column] = ('v'); column++; |
| 1426 | break; |
| 1427 | |
| 1428 | default: |
| 1429 | /* |
| 1430 | * We've already put the backslash, but this |
| 1431 | * will put 3 more characters for the octal |
| 1432 | * number; make sure we have enough room for |
| 1433 | * that, plus the trailing '\0'. |
| 1434 | */ |
| 1435 | FMTBUF_EXPAND(3)if (column+(3 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1435, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(3 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 3 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1435, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1436 | FMTBUF_PUTBYTE_OCTAL(c)fmtbuf[column] = ((((c)>>6)&03) + '0'); column++; fmtbuf [column] = ((((c)>>3)&07) + '0'); column++; fmtbuf[ column] = ((((c)>>0)&07) + '0'); column++; |
| 1437 | break; |
| 1438 | } |
| 1439 | } else { |
| 1440 | /* |
| 1441 | * Unicode, but not printable, and not ASCII; |
| 1442 | * put it out as \uxxxx or \Uxxxxxxxx. |
| 1443 | */ |
| 1444 | if (uc <= 0xFFFF) { |
| 1445 | FMTBUF_EXPAND(6)if (column+(6 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1445, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(6 +1) >= fmtbuf_len) { if (__builtin_add_overflow( (fmtbuf_len), ((column + 6 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1445, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1446 | FMTBUF_PUTCHAR('\\')fmtbuf[column] = ('\\'); column++; |
| 1447 | FMTBUF_PUTCHAR('u')fmtbuf[column] = ('u'); column++; |
| 1448 | FMTBUF_PUTCHAR(hex[(uc >> 12) & 0xF])fmtbuf[column] = (hex[(uc >> 12) & 0xF]); column++; |
| 1449 | FMTBUF_PUTCHAR(hex[(uc >> 8) & 0xF])fmtbuf[column] = (hex[(uc >> 8) & 0xF]); column++; |
| 1450 | FMTBUF_PUTCHAR(hex[(uc >> 4) & 0xF])fmtbuf[column] = (hex[(uc >> 4) & 0xF]); column++; |
| 1451 | FMTBUF_PUTCHAR(hex[(uc >> 0) & 0xF])fmtbuf[column] = (hex[(uc >> 0) & 0xF]); column++; |
| 1452 | } else { |
| 1453 | FMTBUF_EXPAND(10)if (column+(10 +1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1453, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(10 +1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + 10 + 2) - fmtbuf_len), (&fmtbuf_len ))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG, "wsutil/str_util.c" , 1453, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator , fmtbuf, fmtbuf_len); }; |
| 1454 | FMTBUF_PUTCHAR('\\')fmtbuf[column] = ('\\'); column++; |
| 1455 | FMTBUF_PUTCHAR('U')fmtbuf[column] = ('U'); column++; |
| 1456 | FMTBUF_PUTCHAR(hex[(uc >> 28) & 0xF])fmtbuf[column] = (hex[(uc >> 28) & 0xF]); column++; |
| 1457 | FMTBUF_PUTCHAR(hex[(uc >> 24) & 0xF])fmtbuf[column] = (hex[(uc >> 24) & 0xF]); column++; |
| 1458 | FMTBUF_PUTCHAR(hex[(uc >> 20) & 0xF])fmtbuf[column] = (hex[(uc >> 20) & 0xF]); column++; |
| 1459 | FMTBUF_PUTCHAR(hex[(uc >> 16) & 0xF])fmtbuf[column] = (hex[(uc >> 16) & 0xF]); column++; |
| 1460 | FMTBUF_PUTCHAR(hex[(uc >> 12) & 0xF])fmtbuf[column] = (hex[(uc >> 12) & 0xF]); column++; |
| 1461 | FMTBUF_PUTCHAR(hex[(uc >> 8) & 0xF])fmtbuf[column] = (hex[(uc >> 8) & 0xF]); column++; |
| 1462 | FMTBUF_PUTCHAR(hex[(uc >> 4) & 0xF])fmtbuf[column] = (hex[(uc >> 4) & 0xF]); column++; |
| 1463 | FMTBUF_PUTCHAR(hex[(uc >> 0) & 0xF])fmtbuf[column] = (hex[(uc >> 0) & 0xF]); column++; |
| 1464 | } |
| 1465 | } |
| 1466 | } |
| 1467 | prev = string; |
| 1468 | } |
| 1469 | } |
| 1470 | if (printable_bytes) { |
| 1471 | FMTBUF_PUTBYTES(prev, printable_bytes)if (column+(printable_bytes+1) >= fmtbuf_len) { if (__builtin_mul_overflow ((fmtbuf_len), (2), (&fmtbuf_len))) { do { if (1) { ws_log_full ("", LOG_LEVEL_DEBUG, "wsutil/str_util.c", 1471, __func__, "overflow." ); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } if ( column+(printable_bytes+1) >= fmtbuf_len) { if (__builtin_add_overflow ((fmtbuf_len), ((column + printable_bytes + 2) - fmtbuf_len), (&fmtbuf_len))) { do { if (1) { ws_log_full("", LOG_LEVEL_DEBUG , "wsutil/str_util.c", 1471, __func__, "overflow."); } } while (0); fmtbuf[column] = '\0'; return fmtbuf; } } fmtbuf = (char *)wmem_realloc(allocator, fmtbuf, fmtbuf_len); } memcpy(& fmtbuf[column], prev, printable_bytes); column += (unsigned)printable_bytes ;; |
| 1472 | printable_bytes = 0; |
Value stored to 'printable_bytes' is never read | |
| 1473 | } |
| 1474 | |
| 1475 | FMTBUF_ENDSTRfmtbuf[column] = '\0'; |
| 1476 | |
| 1477 | return fmtbuf; |
| 1478 | } |
| 1479 | |
| 1480 | /* |
| 1481 | * Given a wmem scope, a not-necessarily-null-terminated string, |
| 1482 | * expected to be in UTF-8 but possibly containing invalid sequences |
| 1483 | * (as it may have come from packet data), and the length of the string, |
| 1484 | * generate a valid UTF-8 string from it, allocated in the specified |
| 1485 | * wmem scope, that: |
| 1486 | * |
| 1487 | * shows printable Unicode characters as themselves; |
| 1488 | * |
| 1489 | * shows non-printable ASCII characters as C-style escapes (octal |
| 1490 | * if not one of the standard ones such as LF -> '\n'); |
| 1491 | * |
| 1492 | * shows non-printable Unicode-but-not-ASCII characters as |
| 1493 | * their universal character names; |
| 1494 | * |
| 1495 | * shows illegal UTF-8 sequences as a sequence of bytes represented |
| 1496 | * as C-style hex escapes (XXX: Does not actually do this. Some illegal |
| 1497 | * sequences, such as overlong encodings, the sequences reserved for |
| 1498 | * UTF-16 surrogate halves (paired or unpaired), and values outside |
| 1499 | * Unicode (i.e., the old sequences for code points above U+10FFFF) |
| 1500 | * will be decoded in a permissive way. Other illegal sequences, |
| 1501 | * such 0xFE and 0xFF and the presence of a continuation byte where |
| 1502 | * not expected (or vice versa its absence), are replaced with |
| 1503 | * REPLACEMENT CHARACTER.) |
| 1504 | * |
| 1505 | * and return a pointer to it. |
| 1506 | */ |
| 1507 | char * |
| 1508 | format_text(wmem_allocator_t *allocator, |
| 1509 | const char *string, size_t len) |
| 1510 | { |
| 1511 | return format_text_internal(allocator, (const uint8_t*)string, len, false0); |
| 1512 | } |
| 1513 | |
| 1514 | /** Given a wmem scope and a null-terminated string, expected to be in |
| 1515 | * UTF-8 but possibly containing invalid sequences (as it may have come |
| 1516 | * from packet data), and the length of the string, generate a valid |
| 1517 | * UTF-8 string from it, allocated in the specified wmem scope, that: |
| 1518 | * |
| 1519 | * shows printable Unicode characters as themselves; |
| 1520 | * |
| 1521 | * shows non-printable ASCII characters as C-style escapes (octal |
| 1522 | * if not one of the standard ones such as LF -> '\n'); |
| 1523 | * |
| 1524 | * shows non-printable Unicode-but-not-ASCII characters as |
| 1525 | * their universal character names; |
| 1526 | * |
| 1527 | * shows illegal UTF-8 sequences as a sequence of bytes represented |
| 1528 | * as C-style hex escapes; |
| 1529 | * |
| 1530 | * and return a pointer to it. |
| 1531 | */ |
| 1532 | char * |
| 1533 | format_text_string(wmem_allocator_t* allocator, const char *string) |
| 1534 | { |
| 1535 | return format_text_internal(allocator, (const uint8_t*)string, strlen(string), false0); |
| 1536 | } |
| 1537 | |
| 1538 | /* |
| 1539 | * Given a string, generate a string from it that shows non-printable |
| 1540 | * characters as C-style escapes except a whitespace character |
| 1541 | * (space, tab, carriage return, new line, vertical tab, or formfeed) |
| 1542 | * which will be replaced by a space, and return a pointer to it. |
| 1543 | */ |
| 1544 | char * |
| 1545 | format_text_wsp(wmem_allocator_t* allocator, const char *string, size_t len) |
| 1546 | { |
| 1547 | return format_text_internal(allocator, (const uint8_t*)string, len, true1); |
| 1548 | } |
| 1549 | |
| 1550 | /* |
| 1551 | * Given a string, generate a string from it that shows non-printable |
| 1552 | * characters as the chr parameter passed, except a whitespace character |
| 1553 | * (space, tab, carriage return, new line, vertical tab, or formfeed) |
| 1554 | * which will be replaced by a space, and return a pointer to it. |
| 1555 | * |
| 1556 | * This does *not* treat the input string as UTF-8. |
| 1557 | * |
| 1558 | * This is useful for displaying binary data that frequently but not always |
| 1559 | * contains text; otherwise the number of C escape codes makes it unreadable. |
| 1560 | */ |
| 1561 | char * |
| 1562 | format_text_chr(wmem_allocator_t *allocator, const char *string, size_t len, char chr) |
| 1563 | { |
| 1564 | wmem_strbuf_t *buf; |
| 1565 | |
| 1566 | buf = wmem_strbuf_new_sized(allocator, len + 1); |
| 1567 | for (const char *p = string; p < string + len; p++) { |
| 1568 | if (g_ascii_isprint(*p)((g_ascii_table[(guchar) (*p)] & G_ASCII_PRINT) != 0)) { |
| 1569 | wmem_strbuf_append_c(buf, *p); |
| 1570 | } |
| 1571 | else if (g_ascii_isspace(*p)((g_ascii_table[(guchar) (*p)] & G_ASCII_SPACE) != 0)) { |
| 1572 | wmem_strbuf_append_c(buf, ' '); |
| 1573 | } |
| 1574 | else { |
| 1575 | wmem_strbuf_append_c(buf, chr); |
| 1576 | } |
| 1577 | } |
| 1578 | return wmem_strbuf_finalize(buf); |
| 1579 | } |
| 1580 | |
| 1581 | char * |
| 1582 | format_char(wmem_allocator_t *allocator, char c) |
| 1583 | { |
| 1584 | char *buf; |
| 1585 | char r; |
| 1586 | |
| 1587 | if (g_ascii_isprint(c)((g_ascii_table[(guchar) (c)] & G_ASCII_PRINT) != 0)) { |
| 1588 | buf = wmem_alloc_array(allocator, char, 2)((char*)wmem_alloc((allocator), (((((2)) <= 0) || ((size_t )sizeof(char) > (9223372036854775807L / (size_t)((2))))) ? 0 : (sizeof(char) * ((2)))))); |
| 1589 | buf[0] = c; |
| 1590 | buf[1] = '\0'; |
| 1591 | return buf; |
| 1592 | } |
| 1593 | if (escape_char(c, &r)) { |
| 1594 | buf = wmem_alloc_array(allocator, char, 3)((char*)wmem_alloc((allocator), (((((3)) <= 0) || ((size_t )sizeof(char) > (9223372036854775807L / (size_t)((3))))) ? 0 : (sizeof(char) * ((3)))))); |
| 1595 | buf[0] = '\\'; |
| 1596 | buf[1] = r; |
| 1597 | buf[2] = '\0'; |
| 1598 | return buf; |
| 1599 | } |
| 1600 | buf = wmem_alloc_array(allocator, char, 5)((char*)wmem_alloc((allocator), (((((5)) <= 0) || ((size_t )sizeof(char) > (9223372036854775807L / (size_t)((5))))) ? 0 : (sizeof(char) * ((5)))))); |
| 1601 | buf[0] = '\\'; |
| 1602 | buf[1] = 'x'; |
| 1603 | buf[2] = hex[((uint8_t)c >> 4) & 0xF]; |
| 1604 | buf[3] = hex[((uint8_t)c >> 0) & 0xF]; |
| 1605 | buf[4] = '\0'; |
| 1606 | return buf; |
| 1607 | } |
| 1608 | |
| 1609 | char* |
| 1610 | ws_utf8_truncate(char *string, size_t len) |
| 1611 | { |
| 1612 | char* last_char; |
| 1613 | |
| 1614 | /* Ensure that it is null terminated */ |
| 1615 | string[len] = '\0'; |
| 1616 | last_char = g_utf8_find_prev_char(string, string + len); |
| 1617 | if (last_char != NULL((void*)0) && g_utf8_get_char_validated(last_char, -1) == (gunichar)-2) { |
| 1618 | /* The last UTF-8 character was truncated into a partial sequence. */ |
| 1619 | *last_char = '\0'; |
| 1620 | } |
| 1621 | return string; |
| 1622 | } |
| 1623 | |
| 1624 | /* ASCII/EBCDIC conversion tables from |
| 1625 | * https://web.archive.org/web/20060813174742/http://www.room42.com/store/computer_center/code_tables.shtml |
| 1626 | */ |
| 1627 | #if 0 |
| 1628 | static const uint8_t ASCII_translate_EBCDIC [ 256 ] = { |
| 1629 | 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, |
| 1630 | 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, |
| 1631 | 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, |
| 1632 | 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, |
| 1633 | 0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D, 0x4D, |
| 1634 | 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61, |
| 1635 | 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, |
| 1636 | 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F, |
| 1637 | 0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, |
| 1638 | 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, |
| 1639 | 0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, |
| 1640 | 0xE8, 0xE9, 0xAD, 0xE0, 0xBD, 0x5F, 0x6D, |
| 1641 | 0x7D, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, |
| 1642 | 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, |
| 1643 | 0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, |
| 1644 | 0xA8, 0xA9, 0xC0, 0x6A, 0xD0, 0xA1, 0x4B, |
| 1645 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1646 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1647 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1648 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1649 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1650 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1651 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1652 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1653 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1654 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1655 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1656 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1657 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1658 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1659 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, |
| 1660 | 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B, 0x4B |
| 1661 | }; |
| 1662 | |
| 1663 | void |
| 1664 | ASCII_to_EBCDIC(uint8_t *buf, unsigned bytes) |
| 1665 | { |
| 1666 | unsigned i; |
| 1667 | uint8_t *bufptr; |
| 1668 | |
| 1669 | bufptr = buf; |
| 1670 | |
| 1671 | for (i = 0; i < bytes; i++, bufptr++) { |
| 1672 | *bufptr = ASCII_translate_EBCDIC[*bufptr]; |
| 1673 | } |
| 1674 | } |
| 1675 | |
| 1676 | uint8_t |
| 1677 | ASCII_to_EBCDIC1(uint8_t c) |
| 1678 | { |
| 1679 | return ASCII_translate_EBCDIC[c]; |
| 1680 | } |
| 1681 | #endif |
| 1682 | |
| 1683 | static const uint8_t EBCDIC_translate_ASCII [ 256 ] = { |
| 1684 | 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, |
| 1685 | 0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, |
| 1686 | 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, |
| 1687 | 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, |
| 1688 | 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, |
| 1689 | 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F, |
| 1690 | 0x2E, 0x2E, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, |
| 1691 | 0x38, 0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x2E, 0x3F, |
| 1692 | 0x20, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1693 | 0x2E, 0x2E, 0x2E, 0x2E, 0x3C, 0x28, 0x2B, 0x7C, |
| 1694 | 0x26, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1695 | 0x2E, 0x2E, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0x5E, |
| 1696 | 0x2D, 0x2F, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1697 | 0x2E, 0x2E, 0x7C, 0x2C, 0x25, 0x5F, 0x3E, 0x3F, |
| 1698 | 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1699 | 0x2E, 0x2E, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22, |
| 1700 | 0x2E, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, |
| 1701 | 0x68, 0x69, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1702 | 0x2E, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, |
| 1703 | 0x71, 0x72, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1704 | 0x2E, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, |
| 1705 | 0x79, 0x7A, 0x2E, 0x2E, 0x2E, 0x5B, 0x2E, 0x2E, |
| 1706 | 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1707 | 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x5D, 0x2E, 0x2E, |
| 1708 | 0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, |
| 1709 | 0x48, 0x49, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1710 | 0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, |
| 1711 | 0x51, 0x52, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1712 | 0x5C, 0x2E, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, |
| 1713 | 0x59, 0x5A, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, |
| 1714 | 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, |
| 1715 | 0x38, 0x39, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E, 0x2E |
| 1716 | }; |
| 1717 | |
| 1718 | void |
| 1719 | EBCDIC_to_ASCII(uint8_t *buf, unsigned bytes) |
| 1720 | { |
| 1721 | unsigned i; |
| 1722 | uint8_t *bufptr; |
| 1723 | |
| 1724 | bufptr = buf; |
| 1725 | |
| 1726 | for (i = 0; i < bytes; i++, bufptr++) { |
| 1727 | *bufptr = EBCDIC_translate_ASCII[*bufptr]; |
| 1728 | } |
| 1729 | } |
| 1730 | |
| 1731 | uint8_t |
| 1732 | EBCDIC_to_ASCII1(uint8_t c) |
| 1733 | { |
| 1734 | return EBCDIC_translate_ASCII[c]; |
| 1735 | } |
| 1736 | |
| 1737 | /* |
| 1738 | * This routine is based on a routine created by Dan Lasley |
| 1739 | * <DLASLEY@PROMUS.com>. |
| 1740 | * |
| 1741 | * It was modified for Wireshark by Gilbert Ramirez and others. |
| 1742 | */ |
| 1743 | |
| 1744 | #define MAX_OFFSET_LEN8 8 /* max length of hex offset of bytes */ |
| 1745 | #define BYTES_PER_LINE16 16 /* max byte values printed on a line */ |
| 1746 | #define HEX_DUMP_LEN(16*3) (BYTES_PER_LINE16*3) |
| 1747 | /* max number of characters hex dump takes - |
| 1748 | 2 digits plus trailing blank */ |
| 1749 | #define DATA_DUMP_LEN((16*3) + 2 + 2 + 16) (HEX_DUMP_LEN(16*3) + 2 + 2 + BYTES_PER_LINE16) |
| 1750 | /* number of characters those bytes take; |
| 1751 | 3 characters per byte of hex dump, |
| 1752 | 2 blanks separating hex from ASCII, |
| 1753 | 2 optional ASCII dump delimiters, |
| 1754 | 1 character per byte of ASCII dump */ |
| 1755 | #define MAX_LINE_LEN(8 + 2 + ((16*3) + 2 + 2 + 16)) (MAX_OFFSET_LEN8 + 2 + DATA_DUMP_LEN((16*3) + 2 + 2 + 16)) |
| 1756 | /* number of characters per line; |
| 1757 | offset, 2 blanks separating offset |
| 1758 | from data dump, data dump */ |
| 1759 | |
| 1760 | bool_Bool |
| 1761 | hex_dump_buffer(bool_Bool (*print_line)(void *, const char *), void *fp, |
| 1762 | const unsigned char *cp, unsigned length, |
| 1763 | hex_dump_enc encoding, |
| 1764 | unsigned ascii_option) |
| 1765 | { |
| 1766 | register unsigned int ad, i, j, k, l; |
| 1767 | unsigned char c; |
| 1768 | char line[MAX_LINE_LEN(8 + 2 + ((16*3) + 2 + 2 + 16)) + 1]; |
| 1769 | unsigned int use_digits; |
| 1770 | |
| 1771 | static const char binhex[16] = { |
| 1772 | '0', '1', '2', '3', '4', '5', '6', '7', |
| 1773 | '8', '9', 'a', 'b', 'c', 'd', 'e', 'f'}; |
| 1774 | |
| 1775 | /* |
| 1776 | * How many of the leading digits of the offset will we supply? |
| 1777 | * We always supply at least 4 digits, but if the maximum offset |
| 1778 | * won't fit in 4 digits, we use as many digits as will be needed. |
| 1779 | */ |
| 1780 | if (((length - 1) & 0xF0000000) != 0) |
| 1781 | use_digits = 8; /* need all 8 digits */ |
| 1782 | else if (((length - 1) & 0x0F000000) != 0) |
| 1783 | use_digits = 7; /* need 7 digits */ |
| 1784 | else if (((length - 1) & 0x00F00000) != 0) |
| 1785 | use_digits = 6; /* need 6 digits */ |
| 1786 | else if (((length - 1) & 0x000F0000) != 0) |
| 1787 | use_digits = 5; /* need 5 digits */ |
| 1788 | else |
| 1789 | use_digits = 4; /* we'll supply 4 digits */ |
| 1790 | |
| 1791 | ad = 0; |
| 1792 | i = 0; |
| 1793 | j = 0; |
| 1794 | k = 0; |
| 1795 | while (i < length) { |
| 1796 | if ((i & 15) == 0) { |
| 1797 | /* |
| 1798 | * Start of a new line. |
| 1799 | */ |
| 1800 | j = 0; |
| 1801 | l = use_digits; |
| 1802 | do { |
| 1803 | l--; |
| 1804 | c = (ad >> (l*4)) & 0xF; |
| 1805 | line[j++] = binhex[c]; |
| 1806 | } while (l != 0); |
| 1807 | line[j++] = ' '; |
| 1808 | line[j++] = ' '; |
| 1809 | memset(line+j, ' ', DATA_DUMP_LEN((16*3) + 2 + 2 + 16)); |
| 1810 | |
| 1811 | /* |
| 1812 | * Offset in line of ASCII dump. |
| 1813 | */ |
| 1814 | k = j + HEX_DUMP_LEN(16*3) + 2; |
| 1815 | if (ascii_option == HEXDUMP_ASCII_DELIMIT(0x0001U)) |
| 1816 | line[k++] = '|'; |
| 1817 | } |
| 1818 | c = *cp++; |
| 1819 | line[j++] = binhex[c>>4]; |
| 1820 | line[j++] = binhex[c&0xf]; |
| 1821 | j++; |
| 1822 | if (ascii_option != HEXDUMP_ASCII_EXCLUDE(0x0002U) ) { |
| 1823 | if (encoding == HEXDUMP_ENC_EBCDIC) { |
| 1824 | c = EBCDIC_to_ASCII1(c); |
| 1825 | } |
| 1826 | line[k++] = ((c >= ' ') && (c < 0x7f)) ? c : '.'; |
| 1827 | } |
| 1828 | i++; |
| 1829 | if (((i & 15) == 0) || (i == length)) { |
| 1830 | /* |
| 1831 | * We'll be starting a new line, or |
| 1832 | * we're finished printing this buffer; |
| 1833 | * dump out the line we've constructed, |
| 1834 | * and advance the offset. |
| 1835 | */ |
| 1836 | if (ascii_option == HEXDUMP_ASCII_DELIMIT(0x0001U)) |
| 1837 | line[k++] = '|'; |
| 1838 | line[k] = '\0'; |
| 1839 | if (!print_line(fp, line)) |
| 1840 | return false0; |
| 1841 | ad += 16; |
| 1842 | } |
| 1843 | } |
| 1844 | return true1; |
| 1845 | } |
| 1846 | |
| 1847 | /* |
| 1848 | * Editor modelines - https://www.wireshark.org/tools/modelines.html |
| 1849 | * |
| 1850 | * Local variables: |
| 1851 | * c-basic-offset: 4 |
| 1852 | * tab-width: 8 |
| 1853 | * indent-tabs-mode: nil |
| 1854 | * End: |
| 1855 | * |
| 1856 | * vi: set shiftwidth=4 tabstop=8 expandtab: |
| 1857 | * :indentSize=4:tabSize=8:noTabs=true: |
| 1858 | */ |