src/basic/utf8.c

   1 /* SPDX-License-Identifier: LGPL-2.1+ */
   2
   3 /* Parts of this file are based on the GLIB utf8 validation functions. The
   4  * original license text follows. */
   5
   6 /* gutf8.c - Operations on UTF-8 strings.
   7  *
   8  * Copyright (C) 1999 Tom Tromey
   9  * Copyright (C) 2000 Red Hat, Inc.
  10  *
  11  * This library is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Library General Public
  13  * License as published by the Free Software Foundation; either
  14  * version 2 of the License, or (at your option) any later version.
  15  *
  16  * This library is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Library General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Library General Public
  22  * License along with this library; if not, write to the Free Software
  23  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  24  */
  25
  26 #include <errno.h>
  27 #include <stdbool.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30
  31 #include "alloc-util.h"
  32 #include "gunicode.h"
  33 #include "hexdecoct.h"
  34 #include "macro.h"
  35 #include "utf8.h"
  36
  37 bool unichar_is_valid(char32_t ch) {
  38
  39         if (ch >= 0x110000) /* End of unicode space */
  40                 return false;
  41         if ((ch & 0xFFFFF800) == 0xD800) /* Reserved area for UTF-16 */
  42                 return false;
  43         if ((ch >= 0xFDD0) && (ch <= 0xFDEF)) /* Reserved */
  44                 return false;
  45         if ((ch & 0xFFFE) == 0xFFFE) /* BOM (Byte Order Mark) */
  46                 return false;
  47
  48         return true;
  49 }
  50
  51 static bool unichar_is_control(char32_t ch) {
  52
  53         /*
  54           0 to ' '-1 is the C0 range.
  55           DEL=0x7F, and DEL+1 to 0x9F is C1 range.
  56           '\t' is in C0 range, but more or less harmless and commonly used.
  57         */
  58
  59         return (ch < ' ' && !IN_SET(ch, '\t', '\n')) ||
  60                 (0x7F <= ch && ch <= 0x9F);
  61 }
  62
  63 /* count of characters used to encode one unicode char */
  64 static size_t utf8_encoded_expected_len(const char *str) {
  65         uint8_t c;
  66
  67         assert(str);
  68
  69         c = (uint8_t) str[0];
  70         if (c < 0x80)
  71                 return 1;
  72         if ((c & 0xe0) == 0xc0)
  73                 return 2;
  74         if ((c & 0xf0) == 0xe0)
  75                 return 3;
  76         if ((c & 0xf8) == 0xf0)
  77                 return 4;
  78         if ((c & 0xfc) == 0xf8)
  79                 return 5;
  80         if ((c & 0xfe) == 0xfc)
  81                 return 6;
  82
  83         return 0;
  84 }
  85
  86 /* decode one unicode char */
  87 int utf8_encoded_to_unichar(const char *str, char32_t *ret_unichar) {
  88         char32_t unichar;
  89         size_t len, i;
  90
  91         assert(str);
  92
  93         len = utf8_encoded_expected_len(str);
  94
  95         switch (len) {
  96         case 1:
  97                 *ret_unichar = (char32_t)str[0];
  98                 return 0;
  99         case 2:
 100                 unichar = str[0] & 0x1f;
 101                 break;
 102         case 3:
 103                 unichar = (char32_t)str[0] & 0x0f;
 104                 break;
 105         case 4:
 106                 unichar = (char32_t)str[0] & 0x07;
 107                 break;
 108         case 5:
 109                 unichar = (char32_t)str[0] & 0x03;
 110                 break;
 111         case 6:
 112                 unichar = (char32_t)str[0] & 0x01;
 113                 break;
 114         default:
 115                 return -EINVAL;
 116         }
 117
 118         for (i = 1; i < len; i++) {
 119                 if (((char32_t)str[i] & 0xc0) != 0x80)
 120                         return -EINVAL;
 121
 122                 unichar <<= 6;
 123                 unichar |= (char32_t)str[i] & 0x3f;
 124         }
 125
 126         *ret_unichar = unichar;
 127
 128         return 0;
 129 }
 130
 131 bool utf8_is_printable_newline(const char* str, size_t length, bool newline) {
 132         const char *p;
 133
 134         assert(str);
 135
 136         for (p = str; length;) {
 137                 int encoded_len, r;
 138                 char32_t val;
 139
 140                 encoded_len = utf8_encoded_valid_unichar(p);
 141                 if (encoded_len < 0 ||
 142                     (size_t) encoded_len > length)
 143                         return false;
 144
 145                 r = utf8_encoded_to_unichar(p, &val);
 146                 if (r < 0 ||
 147                     unichar_is_control(val) ||
 148                     (!newline && val == '\n'))
 149                         return false;
 150
 151                 length -= encoded_len;
 152                 p += encoded_len;
 153         }
 154
 155         return true;
 156 }
 157
 158 char *utf8_is_valid(const char *str) {
 159         const char *p;
 160
 161         assert(str);
 162
 163         p = str;
 164         while (*p) {
 165                 int len;
 166
 167                 len = utf8_encoded_valid_unichar(p);
 168                 if (len < 0)
 169                         return NULL;
 170
 171                 p += len;
 172         }
 173
 174         return (char*) str;
 175 }
 176
 177 char *utf8_escape_invalid(const char *str) {
 178         char *p, *s;
 179
 180         assert(str);
 181
 182         p = s = malloc(strlen(str) * 4 + 1);
 183         if (!p)
 184                 return NULL;
 185
 186         while (*str) {
 187                 int len;
 188
 189                 len = utf8_encoded_valid_unichar(str);
 190                 if (len > 0) {
 191                         s = mempcpy(s, str, len);
 192                         str += len;
 193                 } else {
 194                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 195                         str += 1;
 196                 }
 197         }
 198
 199         *s = '\0';
 200
 201         return p;
 202 }
 203
 204 char *utf8_escape_non_printable(const char *str) {
 205         char *p, *s;
 206
 207         assert(str);
 208
 209         p = s = malloc(strlen(str) * 4 + 1);
 210         if (!p)
 211                 return NULL;
 212
 213         while (*str) {
 214                 int len;
 215
 216                 len = utf8_encoded_valid_unichar(str);
 217                 if (len > 0) {
 218                         if (utf8_is_printable(str, len)) {
 219                                 s = mempcpy(s, str, len);
 220                                 str += len;
 221                         } else {
 222                                 while (len > 0) {
 223                                         *(s++) = '\\';
 224                                         *(s++) = 'x';
 225                                         *(s++) = hexchar((int) *str >> 4);
 226                                         *(s++) = hexchar((int) *str);
 227
 228                                         str += 1;
 229                                         len--;
 230                                 }
 231                         }
 232                 } else {
 233                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 234                         str += 1;
 235                 }
 236         }
 237
 238         *s = '\0';
 239
 240         return p;
 241 }
 242
 243 char *ascii_is_valid(const char *str) {
 244         const char *p;
 245
 246         /* Check whether the string consists of valid ASCII bytes,
 247          * i.e values between 0 and 127, inclusive. */
 248
 249         assert(str);
 250
 251         for (p = str; *p; p++)
 252                 if ((unsigned char) *p >= 128)
 253                         return NULL;
 254
 255         return (char*) str;
 256 }
 257
 258 char *ascii_is_valid_n(const char *str, size_t len) {
 259         size_t i;
 260
 261         /* Very similar to ascii_is_valid(), but checks exactly len
 262          * bytes and rejects any NULs in that range. */
 263
 264         assert(str);
 265
 266         for (i = 0; i < len; i++)
 267                 if ((unsigned char) str[i] >= 128 || str[i] == 0)
 268                         return NULL;
 269
 270         return (char*) str;
 271 }
 272
 273 /**
 274  * utf8_encode_unichar() - Encode single UCS-4 character as UTF-8
 275  * @out_utf8: output buffer of at least 4 bytes or NULL
 276  * @g: UCS-4 character to encode
 277  *
 278  * This encodes a single UCS-4 character as UTF-8 and writes it into @out_utf8.
 279  * The length of the character is returned. It is not zero-terminated! If the
 280  * output buffer is NULL, only the length is returned.
 281  *
 282  * Returns: The length in bytes that the UTF-8 representation does or would
 283  *          occupy.
 284  */
 285 size_t utf8_encode_unichar(char *out_utf8, char32_t g) {
 286
 287         if (g < (1 << 7)) {
 288                 if (out_utf8)
 289                         out_utf8[0] = g & 0x7f;
 290                 return 1;
 291         } else if (g < (1 << 11)) {
 292                 if (out_utf8) {
 293                         out_utf8[0] = 0xc0 | ((g >> 6) & 0x1f);
 294                         out_utf8[1] = 0x80 | (g & 0x3f);
 295                 }
 296                 return 2;
 297         } else if (g < (1 << 16)) {
 298                 if (out_utf8) {
 299                         out_utf8[0] = 0xe0 | ((g >> 12) & 0x0f);
 300                         out_utf8[1] = 0x80 | ((g >> 6) & 0x3f);
 301                         out_utf8[2] = 0x80 | (g & 0x3f);
 302                 }
 303                 return 3;
 304         } else if (g < (1 << 21)) {
 305                 if (out_utf8) {
 306                         out_utf8[0] = 0xf0 | ((g >> 18) & 0x07);
 307                         out_utf8[1] = 0x80 | ((g >> 12) & 0x3f);
 308                         out_utf8[2] = 0x80 | ((g >> 6) & 0x3f);
 309                         out_utf8[3] = 0x80 | (g & 0x3f);
 310                 }
 311                 return 4;
 312         }
 313
 314         return 0;
 315 }
 316
 317 char *utf16_to_utf8(const char16_t *s, size_t length /* bytes! */) {
 318         const uint8_t *f;
 319         char *r, *t;
 320
 321         assert(s);
 322
 323         /* Input length is in bytes, i.e. the shortest possible character takes 2 bytes. Each unicode character may
 324          * take up to 4 bytes in UTF-8. Let's also account for a trailing NUL byte. */
 325         if (length * 2 < length)
 326                 return NULL; /* overflow */
 327
 328         r = new(char, length * 2 + 1);
 329         if (!r)
 330                 return NULL;
 331
 332         f = (const uint8_t*) s;
 333         t = r;
 334
 335         while (f + 1 < (const uint8_t*) s + length) {
 336                 char16_t w1, w2;
 337
 338                 /* see RFC 2781 section 2.2 */
 339
 340                 w1 = f[1] << 8 | f[0];
 341                 f += 2;
 342
 343                 if (!utf16_is_surrogate(w1)) {
 344                         t += utf8_encode_unichar(t, w1);
 345                         continue;
 346                 }
 347
 348                 if (utf16_is_trailing_surrogate(w1))
 349                         continue; /* spurious trailing surrogate, ignore */
 350
 351                 if (f + 1 >= (const uint8_t*) s + length)
 352                         break;
 353
 354                 w2 = f[1] << 8 | f[0];
 355                 f += 2;
 356
 357                 if (!utf16_is_trailing_surrogate(w2)) {
 358                         f -= 2;
 359                         continue; /* surrogate missing its trailing surrogate, ignore */
 360                 }
 361
 362                 t += utf8_encode_unichar(t, utf16_surrogate_pair_to_unichar(w1, w2));
 363         }
 364
 365         *t = 0;
 366         return r;
 367 }
 368
 369 /* expected size used to encode one unicode char */
 370 static int utf8_unichar_to_encoded_len(char32_t unichar) {
 371
 372         if (unichar < 0x80)
 373                 return 1;
 374         if (unichar < 0x800)
 375                 return 2;
 376         if (unichar < 0x10000)
 377                 return 3;
 378         if (unichar < 0x200000)
 379                 return 4;
 380         if (unichar < 0x4000000)
 381                 return 5;
 382
 383         return 6;
 384 }
 385
 386 /* validate one encoded unicode char and return its length */
 387 int utf8_encoded_valid_unichar(const char *str) {
 388         char32_t unichar;
 389         size_t len, i;
 390         int r;
 391
 392         assert(str);
 393
 394         len = utf8_encoded_expected_len(str);
 395         if (len == 0)
 396                 return -EINVAL;
 397
 398         /* ascii is valid */
 399         if (len == 1)
 400                 return 1;
 401
 402         /* check if expected encoded chars are available */
 403         for (i = 0; i < len; i++)
 404                 if ((str[i] & 0x80) != 0x80)
 405                         return -EINVAL;
 406
 407         r = utf8_encoded_to_unichar(str, &unichar);
 408         if (r < 0)
 409                 return r;
 410
 411         /* check if encoded length matches encoded value */
 412         if (utf8_unichar_to_encoded_len(unichar) != (int) len)
 413                 return -EINVAL;
 414
 415         /* check if value has valid range */
 416         if (!unichar_is_valid(unichar))
 417                 return -EINVAL;
 418
 419         return (int) len;
 420 }
 421
 422 size_t utf8_n_codepoints(const char *str) {
 423         size_t n = 0;
 424
 425         /* Returns the number of UTF-8 codepoints in this string, or (size_t) -1 if the string is not valid UTF-8. */
 426
 427         while (*str != 0) {
 428                 int k;
 429
 430                 k = utf8_encoded_valid_unichar(str);
 431                 if (k < 0)
 432                         return (size_t) -1;
 433
 434                 str += k;
 435                 n++;
 436         }
 437
 438         return n;
 439 }
 440
 441 size_t utf8_console_width(const char *str) {
 442         size_t n = 0;
 443
 444         /* Returns the approximate width a string will take on screen when printed on a character cell
 445          * terminal/console. */
 446
 447         while (*str != 0) {
 448                 char32_t c;
 449
 450                 if (utf8_encoded_to_unichar(str, &c) < 0)
 451                         return (size_t) -1;
 452
 453                 str = utf8_next_char(str);
 454
 455                 n += unichar_iswide(c) ? 2 : 1;
 456         }
 457
 458         return n;
 459 }