src/basic/utf8.c

   1 /* SPDX-License-Identifier: LGPL-2.1+ */
   2 /***
   3   This file is part of systemd.
   4
   5   Copyright 2008-2011 Kay Sievers
   6   Copyright 2012 Lennart Poettering
   7 ***/
   8
   9 /* Parts of this file are based on the GLIB utf8 validation functions. The
  10  * original license text follows. */
  11
  12 /* gutf8.c - Operations on UTF-8 strings.
  13  *
  14  * Copyright (C) 1999 Tom Tromey
  15  * Copyright (C) 2000 Red Hat, Inc.
  16  *
  17  * This library is free software; you can redistribute it and/or
  18  * modify it under the terms of the GNU Library General Public
  19  * License as published by the Free Software Foundation; either
  20  * version 2 of the License, or (at your option) any later version.
  21  *
  22  * This library is distributed in the hope that it will be useful,
  23  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  24  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  25  * Library General Public License for more details.
  26  *
  27  * You should have received a copy of the GNU Library General Public
  28  * License along with this library; if not, write to the Free Software
  29  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  30  */
  31
  32 #include <errno.h>
  33 #include <stdbool.h>
  34 #include <stdlib.h>
  35 #include <string.h>
  36
  37 #include "alloc-util.h"
  38 #include "gunicode.h"
  39 #include "hexdecoct.h"
  40 #include "macro.h"
  41 #include "utf8.h"
  42
  43 bool unichar_is_valid(char32_t ch) {
  44
  45         if (ch >= 0x110000) /* End of unicode space */
  46                 return false;
  47         if ((ch & 0xFFFFF800) == 0xD800) /* Reserved area for UTF-16 */
  48                 return false;
  49         if ((ch >= 0xFDD0) && (ch <= 0xFDEF)) /* Reserved */
  50                 return false;
  51         if ((ch & 0xFFFE) == 0xFFFE) /* BOM (Byte Order Mark) */
  52                 return false;
  53
  54         return true;
  55 }
  56
  57 static bool unichar_is_control(char32_t ch) {
  58
  59         /*
  60           0 to ' '-1 is the C0 range.
  61           DEL=0x7F, and DEL+1 to 0x9F is C1 range.
  62           '\t' is in C0 range, but more or less harmless and commonly used.
  63         */
  64
  65         return (ch < ' ' && !IN_SET(ch, '\t', '\n')) ||
  66                 (0x7F <= ch && ch <= 0x9F);
  67 }
  68
  69 /* count of characters used to encode one unicode char */
  70 static int utf8_encoded_expected_len(const char *str) {
  71         unsigned char c;
  72
  73         assert(str);
  74
  75         c = (unsigned char) str[0];
  76         if (c < 0x80)
  77                 return 1;
  78         if ((c & 0xe0) == 0xc0)
  79                 return 2;
  80         if ((c & 0xf0) == 0xe0)
  81                 return 3;
  82         if ((c & 0xf8) == 0xf0)
  83                 return 4;
  84         if ((c & 0xfc) == 0xf8)
  85                 return 5;
  86         if ((c & 0xfe) == 0xfc)
  87                 return 6;
  88
  89         return 0;
  90 }
  91
  92 /* decode one unicode char */
  93 int utf8_encoded_to_unichar(const char *str, char32_t *ret_unichar) {
  94         char32_t unichar;
  95         int len, i;
  96
  97         assert(str);
  98
  99         len = utf8_encoded_expected_len(str);
 100
 101         switch (len) {
 102         case 1:
 103                 *ret_unichar = (char32_t)str[0];
 104                 return 0;
 105         case 2:
 106                 unichar = str[0] & 0x1f;
 107                 break;
 108         case 3:
 109                 unichar = (char32_t)str[0] & 0x0f;
 110                 break;
 111         case 4:
 112                 unichar = (char32_t)str[0] & 0x07;
 113                 break;
 114         case 5:
 115                 unichar = (char32_t)str[0] & 0x03;
 116                 break;
 117         case 6:
 118                 unichar = (char32_t)str[0] & 0x01;
 119                 break;
 120         default:
 121                 return -EINVAL;
 122         }
 123
 124         for (i = 1; i < len; i++) {
 125                 if (((char32_t)str[i] & 0xc0) != 0x80)
 126                         return -EINVAL;
 127                 unichar <<= 6;
 128                 unichar |= (char32_t)str[i] & 0x3f;
 129         }
 130
 131         *ret_unichar = unichar;
 132
 133         return 0;
 134 }
 135
 136 bool utf8_is_printable_newline(const char* str, size_t length, bool newline) {
 137         const char *p;
 138
 139         assert(str);
 140
 141         for (p = str; length;) {
 142                 int encoded_len, r;
 143                 char32_t val;
 144
 145                 encoded_len = utf8_encoded_valid_unichar(p);
 146                 if (encoded_len < 0 ||
 147                     (size_t) encoded_len > length)
 148                         return false;
 149
 150                 r = utf8_encoded_to_unichar(p, &val);
 151                 if (r < 0 ||
 152                     unichar_is_control(val) ||
 153                     (!newline && val == '\n'))
 154                         return false;
 155
 156                 length -= encoded_len;
 157                 p += encoded_len;
 158         }
 159
 160         return true;
 161 }
 162
 163 const char *utf8_is_valid(const char *str) {
 164         const uint8_t *p;
 165
 166         assert(str);
 167
 168         for (p = (const uint8_t*) str; *p; ) {
 169                 int len;
 170
 171                 len = utf8_encoded_valid_unichar((const char *)p);
 172                 if (len < 0)
 173                         return NULL;
 174
 175                 p += len;
 176         }
 177
 178         return str;
 179 }
 180
 181 char *utf8_escape_invalid(const char *str) {
 182         char *p, *s;
 183
 184         assert(str);
 185
 186         p = s = malloc(strlen(str) * 4 + 1);
 187         if (!p)
 188                 return NULL;
 189
 190         while (*str) {
 191                 int len;
 192
 193                 len = utf8_encoded_valid_unichar(str);
 194                 if (len > 0) {
 195                         s = mempcpy(s, str, len);
 196                         str += len;
 197                 } else {
 198                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 199                         str += 1;
 200                 }
 201         }
 202
 203         *s = '\0';
 204
 205         return p;
 206 }
 207
 208 char *utf8_escape_non_printable(const char *str) {
 209         char *p, *s;
 210
 211         assert(str);
 212
 213         p = s = malloc(strlen(str) * 4 + 1);
 214         if (!p)
 215                 return NULL;
 216
 217         while (*str) {
 218                 int len;
 219
 220                 len = utf8_encoded_valid_unichar(str);
 221                 if (len > 0) {
 222                         if (utf8_is_printable(str, len)) {
 223                                 s = mempcpy(s, str, len);
 224                                 str += len;
 225                         } else {
 226                                 while (len > 0) {
 227                                         *(s++) = '\\';
 228                                         *(s++) = 'x';
 229                                         *(s++) = hexchar((int) *str >> 4);
 230                                         *(s++) = hexchar((int) *str);
 231
 232                                         str += 1;
 233                                         len--;
 234                                 }
 235                         }
 236                 } else {
 237                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 238                         str += 1;
 239                 }
 240         }
 241
 242         *s = '\0';
 243
 244         return p;
 245 }
 246
 247 char *ascii_is_valid(const char *str) {
 248         const char *p;
 249
 250         /* Check whether the string consists of valid ASCII bytes,
 251          * i.e values between 0 and 127, inclusive. */
 252
 253         assert(str);
 254
 255         for (p = str; *p; p++)
 256                 if ((unsigned char) *p >= 128)
 257                         return NULL;
 258
 259         return (char*) str;
 260 }
 261
 262 char *ascii_is_valid_n(const char *str, size_t len) {
 263         size_t i;
 264
 265         /* Very similar to ascii_is_valid(), but checks exactly len
 266          * bytes and rejects any NULs in that range. */
 267
 268         assert(str);
 269
 270         for (i = 0; i < len; i++)
 271                 if ((unsigned char) str[i] >= 128 || str[i] == 0)
 272                         return NULL;
 273
 274         return (char*) str;
 275 }
 276
 277 /**
 278  * utf8_encode_unichar() - Encode single UCS-4 character as UTF-8
 279  * @out_utf8: output buffer of at least 4 bytes or NULL
 280  * @g: UCS-4 character to encode
 281  *
 282  * This encodes a single UCS-4 character as UTF-8 and writes it into @out_utf8.
 283  * The length of the character is returned. It is not zero-terminated! If the
 284  * output buffer is NULL, only the length is returned.
 285  *
 286  * Returns: The length in bytes that the UTF-8 representation does or would
 287  *          occupy.
 288  */
 289 size_t utf8_encode_unichar(char *out_utf8, char32_t g) {
 290
 291         if (g < (1 << 7)) {
 292                 if (out_utf8)
 293                         out_utf8[0] = g & 0x7f;
 294                 return 1;
 295         } else if (g < (1 << 11)) {
 296                 if (out_utf8) {
 297                         out_utf8[0] = 0xc0 | ((g >> 6) & 0x1f);
 298                         out_utf8[1] = 0x80 | (g & 0x3f);
 299                 }
 300                 return 2;
 301         } else if (g < (1 << 16)) {
 302                 if (out_utf8) {
 303                         out_utf8[0] = 0xe0 | ((g >> 12) & 0x0f);
 304                         out_utf8[1] = 0x80 | ((g >> 6) & 0x3f);
 305                         out_utf8[2] = 0x80 | (g & 0x3f);
 306                 }
 307                 return 3;
 308         } else if (g < (1 << 21)) {
 309                 if (out_utf8) {
 310                         out_utf8[0] = 0xf0 | ((g >> 18) & 0x07);
 311                         out_utf8[1] = 0x80 | ((g >> 12) & 0x3f);
 312                         out_utf8[2] = 0x80 | ((g >> 6) & 0x3f);
 313                         out_utf8[3] = 0x80 | (g & 0x3f);
 314                 }
 315                 return 4;
 316         }
 317
 318         return 0;
 319 }
 320
 321 char *utf16_to_utf8(const void *s, size_t length) {
 322         const uint8_t *f;
 323         char *r, *t;
 324
 325         r = new(char, (length * 4 + 1) / 2 + 1);
 326         if (!r)
 327                 return NULL;
 328
 329         f = s;
 330         t = r;
 331
 332         while (f < (const uint8_t*) s + length) {
 333                 char16_t w1, w2;
 334
 335                 /* see RFC 2781 section 2.2 */
 336
 337                 w1 = f[1] << 8 | f[0];
 338                 f += 2;
 339
 340                 if (!utf16_is_surrogate(w1)) {
 341                         t += utf8_encode_unichar(t, w1);
 342
 343                         continue;
 344                 }
 345
 346                 if (utf16_is_trailing_surrogate(w1))
 347                         continue;
 348                 else if (f >= (const uint8_t*) s + length)
 349                         break;
 350
 351                 w2 = f[1] << 8 | f[0];
 352                 f += 2;
 353
 354                 if (!utf16_is_trailing_surrogate(w2)) {
 355                         f -= 2;
 356                         continue;
 357                 }
 358
 359                 t += utf8_encode_unichar(t, utf16_surrogate_pair_to_unichar(w1, w2));
 360         }
 361
 362         *t = 0;
 363         return r;
 364 }
 365
 366 /* expected size used to encode one unicode char */
 367 static int utf8_unichar_to_encoded_len(char32_t unichar) {
 368
 369         if (unichar < 0x80)
 370                 return 1;
 371         if (unichar < 0x800)
 372                 return 2;
 373         if (unichar < 0x10000)
 374                 return 3;
 375         if (unichar < 0x200000)
 376                 return 4;
 377         if (unichar < 0x4000000)
 378                 return 5;
 379
 380         return 6;
 381 }
 382
 383 /* validate one encoded unicode char and return its length */
 384 int utf8_encoded_valid_unichar(const char *str) {
 385         int len, i, r;
 386         char32_t unichar;
 387
 388         assert(str);
 389
 390         len = utf8_encoded_expected_len(str);
 391         if (len == 0)
 392                 return -EINVAL;
 393
 394         /* ascii is valid */
 395         if (len == 1)
 396                 return 1;
 397
 398         /* check if expected encoded chars are available */
 399         for (i = 0; i < len; i++)
 400                 if ((str[i] & 0x80) != 0x80)
 401                         return -EINVAL;
 402
 403         r = utf8_encoded_to_unichar(str, &unichar);
 404         if (r < 0)
 405                 return r;
 406
 407         /* check if encoded length matches encoded value */
 408         if (utf8_unichar_to_encoded_len(unichar) != len)
 409                 return -EINVAL;
 410
 411         /* check if value has valid range */
 412         if (!unichar_is_valid(unichar))
 413                 return -EINVAL;
 414
 415         return len;
 416 }
 417
 418 size_t utf8_n_codepoints(const char *str) {
 419         size_t n = 0;
 420
 421         /* Returns the number of UTF-8 codepoints in this string, or (size_t) -1 if the string is not valid UTF-8. */
 422
 423         while (*str != 0) {
 424                 int k;
 425
 426                 k = utf8_encoded_valid_unichar(str);
 427                 if (k < 0)
 428                         return (size_t) -1;
 429
 430                 str += k;
 431                 n++;
 432         }
 433
 434         return n;
 435 }
 436
 437 size_t utf8_console_width(const char *str) {
 438         size_t n = 0;
 439
 440         /* Returns the approximate width a string will take on screen when printed on a character cell
 441          * terminal/console. */
 442
 443         while (*str != 0) {
 444                 char32_t c;
 445
 446                 if (utf8_encoded_to_unichar(str, &c) < 0)
 447                         return (size_t) -1;
 448
 449                 str = utf8_next_char(str);
 450
 451                 n += unichar_iswide(c) ? 2 : 1;
 452         }
 453
 454         return n;
 455 }