src/basic/utf8.c

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