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 "hexdecoct.h"
  39 #include "macro.h"
  40 #include "utf8.h"
  41
  42 bool unichar_is_valid(char32_t ch) {
  43
  44         if (ch >= 0x110000) /* End of unicode space */
  45                 return false;
  46         if ((ch & 0xFFFFF800) == 0xD800) /* Reserved area for UTF-16 */
  47                 return false;
  48         if ((ch >= 0xFDD0) && (ch <= 0xFDEF)) /* Reserved */
  49                 return false;
  50         if ((ch & 0xFFFE) == 0xFFFE) /* BOM (Byte Order Mark) */
  51                 return false;
  52
  53         return true;
  54 }
  55
  56 static bool unichar_is_control(char32_t ch) {
  57
  58         /*
  59           0 to ' '-1 is the C0 range.
  60           DEL=0x7F, and DEL+1 to 0x9F is C1 range.
  61           '\t' is in C0 range, but more or less harmless and commonly used.
  62         */
  63
  64         return (ch < ' ' && !IN_SET(ch, '\t', '\n')) ||
  65                 (0x7F <= ch && ch <= 0x9F);
  66 }
  67
  68 /* count of characters used to encode one unicode char */
  69 static int utf8_encoded_expected_len(const char *str) {
  70         unsigned char c;
  71
  72         assert(str);
  73
  74         c = (unsigned char) str[0];
  75         if (c < 0x80)
  76                 return 1;
  77         if ((c & 0xe0) == 0xc0)
  78                 return 2;
  79         if ((c & 0xf0) == 0xe0)
  80                 return 3;
  81         if ((c & 0xf8) == 0xf0)
  82                 return 4;
  83         if ((c & 0xfc) == 0xf8)
  84                 return 5;
  85         if ((c & 0xfe) == 0xfc)
  86                 return 6;
  87
  88         return 0;
  89 }
  90
  91 /* decode one unicode char */
  92 int utf8_encoded_to_unichar(const char *str, char32_t *ret_unichar) {
  93         char32_t unichar;
  94         int len, i;
  95
  96         assert(str);
  97
  98         len = utf8_encoded_expected_len(str);
  99
 100         switch (len) {
 101         case 1:
 102                 *ret_unichar = (char32_t)str[0];
 103                 return 0;
 104         case 2:
 105                 unichar = str[0] & 0x1f;
 106                 break;
 107         case 3:
 108                 unichar = (char32_t)str[0] & 0x0f;
 109                 break;
 110         case 4:
 111                 unichar = (char32_t)str[0] & 0x07;
 112                 break;
 113         case 5:
 114                 unichar = (char32_t)str[0] & 0x03;
 115                 break;
 116         case 6:
 117                 unichar = (char32_t)str[0] & 0x01;
 118                 break;
 119         default:
 120                 return -EINVAL;
 121         }
 122
 123         for (i = 1; i < len; i++) {
 124                 if (((char32_t)str[i] & 0xc0) != 0x80)
 125                         return -EINVAL;
 126                 unichar <<= 6;
 127                 unichar |= (char32_t)str[i] & 0x3f;
 128         }
 129
 130         *ret_unichar = unichar;
 131
 132         return 0;
 133 }
 134
 135 bool utf8_is_printable_newline(const char* str, size_t length, bool newline) {
 136         const char *p;
 137
 138         assert(str);
 139
 140         for (p = str; length;) {
 141                 int encoded_len, r;
 142                 char32_t val;
 143
 144                 encoded_len = utf8_encoded_valid_unichar(p);
 145                 if (encoded_len < 0 ||
 146                     (size_t) encoded_len > length)
 147                         return false;
 148
 149                 r = utf8_encoded_to_unichar(p, &val);
 150                 if (r < 0 ||
 151                     unichar_is_control(val) ||
 152                     (!newline && val == '\n'))
 153                         return false;
 154
 155                 length -= encoded_len;
 156                 p += encoded_len;
 157         }
 158
 159         return true;
 160 }
 161
 162 const char *utf8_is_valid(const char *str) {
 163         const uint8_t *p;
 164
 165         assert(str);
 166
 167         for (p = (const uint8_t*) str; *p; ) {
 168                 int len;
 169
 170                 len = utf8_encoded_valid_unichar((const char *)p);
 171                 if (len < 0)
 172                         return NULL;
 173
 174                 p += len;
 175         }
 176
 177         return str;
 178 }
 179
 180 char *utf8_escape_invalid(const char *str) {
 181         char *p, *s;
 182
 183         assert(str);
 184
 185         p = s = malloc(strlen(str) * 4 + 1);
 186         if (!p)
 187                 return NULL;
 188
 189         while (*str) {
 190                 int len;
 191
 192                 len = utf8_encoded_valid_unichar(str);
 193                 if (len > 0) {
 194                         s = mempcpy(s, str, len);
 195                         str += len;
 196                 } else {
 197                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 198                         str += 1;
 199                 }
 200         }
 201
 202         *s = '\0';
 203
 204         return p;
 205 }
 206
 207 char *utf8_escape_non_printable(const char *str) {
 208         char *p, *s;
 209
 210         assert(str);
 211
 212         p = s = malloc(strlen(str) * 4 + 1);
 213         if (!p)
 214                 return NULL;
 215
 216         while (*str) {
 217                 int len;
 218
 219                 len = utf8_encoded_valid_unichar(str);
 220                 if (len > 0) {
 221                         if (utf8_is_printable(str, len)) {
 222                                 s = mempcpy(s, str, len);
 223                                 str += len;
 224                         } else {
 225                                 while (len > 0) {
 226                                         *(s++) = '\\';
 227                                         *(s++) = 'x';
 228                                         *(s++) = hexchar((int) *str >> 4);
 229                                         *(s++) = hexchar((int) *str);
 230
 231                                         str += 1;
 232                                         len--;
 233                                 }
 234                         }
 235                 } else {
 236                         s = stpcpy(s, UTF8_REPLACEMENT_CHARACTER);
 237                         str += 1;
 238                 }
 239         }
 240
 241         *s = '\0';
 242
 243         return p;
 244 }
 245
 246 char *ascii_is_valid(const char *str) {
 247         const char *p;
 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 /**
 259  * utf8_encode_unichar() - Encode single UCS-4 character as UTF-8
 260  * @out_utf8: output buffer of at least 4 bytes or NULL
 261  * @g: UCS-4 character to encode
 262  *
 263  * This encodes a single UCS-4 character as UTF-8 and writes it into @out_utf8.
 264  * The length of the character is returned. It is not zero-terminated! If the
 265  * output buffer is NULL, only the length is returned.
 266  *
 267  * Returns: The length in bytes that the UTF-8 representation does or would
 268  *          occupy.
 269  */
 270 size_t utf8_encode_unichar(char *out_utf8, char32_t g) {
 271
 272         if (g < (1 << 7)) {
 273                 if (out_utf8)
 274                         out_utf8[0] = g & 0x7f;
 275                 return 1;
 276         } else if (g < (1 << 11)) {
 277                 if (out_utf8) {
 278                         out_utf8[0] = 0xc0 | ((g >> 6) & 0x1f);
 279                         out_utf8[1] = 0x80 | (g & 0x3f);
 280                 }
 281                 return 2;
 282         } else if (g < (1 << 16)) {
 283                 if (out_utf8) {
 284                         out_utf8[0] = 0xe0 | ((g >> 12) & 0x0f);
 285                         out_utf8[1] = 0x80 | ((g >> 6) & 0x3f);
 286                         out_utf8[2] = 0x80 | (g & 0x3f);
 287                 }
 288                 return 3;
 289         } else if (g < (1 << 21)) {
 290                 if (out_utf8) {
 291                         out_utf8[0] = 0xf0 | ((g >> 18) & 0x07);
 292                         out_utf8[1] = 0x80 | ((g >> 12) & 0x3f);
 293                         out_utf8[2] = 0x80 | ((g >> 6) & 0x3f);
 294                         out_utf8[3] = 0x80 | (g & 0x3f);
 295                 }
 296                 return 4;
 297         }
 298
 299         return 0;
 300 }
 301
 302 char *utf16_to_utf8(const void *s, size_t length) {
 303         const uint8_t *f;
 304         char *r, *t;
 305
 306         r = new(char, (length * 4 + 1) / 2 + 1);
 307         if (!r)
 308                 return NULL;
 309
 310         f = s;
 311         t = r;
 312
 313         while (f < (const uint8_t*) s + length) {
 314                 char16_t w1, w2;
 315
 316                 /* see RFC 2781 section 2.2 */
 317
 318                 w1 = f[1] << 8 | f[0];
 319                 f += 2;
 320
 321                 if (!utf16_is_surrogate(w1)) {
 322                         t += utf8_encode_unichar(t, w1);
 323
 324                         continue;
 325                 }
 326
 327                 if (utf16_is_trailing_surrogate(w1))
 328                         continue;
 329                 else if (f >= (const uint8_t*) s + length)
 330                         break;
 331
 332                 w2 = f[1] << 8 | f[0];
 333                 f += 2;
 334
 335                 if (!utf16_is_trailing_surrogate(w2)) {
 336                         f -= 2;
 337                         continue;
 338                 }
 339
 340                 t += utf8_encode_unichar(t, utf16_surrogate_pair_to_unichar(w1, w2));
 341         }
 342
 343         *t = 0;
 344         return r;
 345 }
 346
 347 /* expected size used to encode one unicode char */
 348 static int utf8_unichar_to_encoded_len(char32_t unichar) {
 349
 350         if (unichar < 0x80)
 351                 return 1;
 352         if (unichar < 0x800)
 353                 return 2;
 354         if (unichar < 0x10000)
 355                 return 3;
 356         if (unichar < 0x200000)
 357                 return 4;
 358         if (unichar < 0x4000000)
 359                 return 5;
 360
 361         return 6;
 362 }
 363
 364 /* validate one encoded unicode char and return its length */
 365 int utf8_encoded_valid_unichar(const char *str) {
 366         int len, i, r;
 367         char32_t unichar;
 368
 369         assert(str);
 370
 371         len = utf8_encoded_expected_len(str);
 372         if (len == 0)
 373                 return -EINVAL;
 374
 375         /* ascii is valid */
 376         if (len == 1)
 377                 return 1;
 378
 379         /* check if expected encoded chars are available */
 380         for (i = 0; i < len; i++)
 381                 if ((str[i] & 0x80) != 0x80)
 382                         return -EINVAL;
 383
 384         r = utf8_encoded_to_unichar(str, &unichar);
 385         if (r < 0)
 386                 return r;
 387
 388         /* check if encoded length matches encoded value */
 389         if (utf8_unichar_to_encoded_len(unichar) != len)
 390                 return -EINVAL;
 391
 392         /* check if value has valid range */
 393         if (!unichar_is_valid(unichar))
 394                 return -EINVAL;
 395
 396         return len;
 397 }
 398
 399 size_t utf8_n_codepoints(const char *str) {
 400         size_t n = 0;
 401
 402         /* Returns the number of UTF-8 codepoints in this string, or (size_t) -1 if the string is not valid UTF-8. */
 403
 404         while (*str != 0) {
 405                 int k;
 406
 407                 k = utf8_encoded_valid_unichar(str);
 408                 if (k < 0)
 409                         return (size_t) -1;
 410
 411                 str += k;
 412                 n++;
 413         }
 414
 415         return n;
 416 }