libidn/punycode.c

   1 /* punycode.c   Implementation of punycode used to ASCII encode IDN's.
   2  * Copyright (C) 2002, 2003  Simon Josefsson
   3  *
   4  * This file is part of GNU Libidn.
   5  *
   6  * GNU Libidn is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * GNU Libidn is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with GNU Libidn; if not, write to the Free Software
  18  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  19  *
  20  */
  21
  22 /*
  23  * This file is derived from RFC 3492bis written by Adam M. Costello.
  24  *
  25  * Disclaimer and license: Regarding this entire document or any
  26  * portion of it (including the pseudocode and C code), the author
  27  * makes no guarantees and is not responsible for any damage resulting
  28  * from its use.  The author grants irrevocable permission to anyone
  29  * to use, modify, and distribute it in any way that does not diminish
  30  * the rights of anyone else to use, modify, and distribute it,
  31  * provided that redistributed derivative works do not contain
  32  * misleading author or version information.  Derivative works need
  33  * not be licensed under similar terms.
  34  *
  35  * Copyright (C) The Internet Society (2003).  All Rights Reserved.
  36  *
  37  * This document and translations of it may be copied and furnished to
  38  * others, and derivative works that comment on or otherwise explain it
  39  * or assist in its implementation may be prepared, copied, published
  40  * and distributed, in whole or in part, without restriction of any
  41  * kind, provided that the above copyright notice and this paragraph are
  42  * included on all such copies and derivative works.  However, this
  43  * document itself may not be modified in any way, such as by removing
  44  * the copyright notice or references to the Internet Society or other
  45  * Internet organizations, except as needed for the purpose of
  46  * developing Internet standards in which case the procedures for
  47  * copyrights defined in the Internet Standards process must be
  48  * followed, or as required to translate it into languages other than
  49  * English.
  50  *
  51  * The limited permissions granted above are perpetual and will not be
  52  * revoked by the Internet Society or its successors or assigns.
  53  *
  54  * This document and the information contained herein is provided on an
  55  * "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
  56  * TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
  57  * BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
  58  * HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
  59  * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
  60  */
  61
  62 #include <string.h>
  63
  64 #include "punycode.h"
  65
  66 /*** Bootstring parameters for Punycode ***/
  67
  68 enum
  69 { base = 36, tmin = 1, tmax = 26, skew = 38, damp = 700,
  70   initial_bias = 72, initial_n = 0x80, delimiter = 0x2D
  71 };
  72
  73 /* basic(cp) tests whether cp is a basic code point: */
  74 #define basic(cp) ((punycode_uint)(cp) < 0x80)
  75
  76 /* delim(cp) tests whether cp is a delimiter: */
  77 #define delim(cp) ((cp) == delimiter)
  78
  79 /* decode_digit(cp) returns the numeric value of a basic code */
  80 /* point (for use in representing integers) in the range 0 to */
  81 /* base-1, or base if cp does not represent a value.          */
  82
  83 static punycode_uint
  84 decode_digit (punycode_uint cp)
  85 {
  86   return cp - 48 < 10 ? cp - 22 : cp - 65 < 26 ? cp - 65 :
  87     cp - 97 < 26 ? cp - 97 : base;
  88 }
  89
  90 /* encode_digit(d,flag) returns the basic code point whose value      */
  91 /* (when used for representing integers) is d, which needs to be in   */
  92 /* the range 0 to base-1.  The lowercase form is used unless flag is  */
  93 /* nonzero, in which case the uppercase form is used.  The behavior   */
  94 /* is undefined if flag is nonzero and digit d has no uppercase form. */
  95
  96 static char
  97 encode_digit (punycode_uint d, int flag)
  98 {
  99   return d + 22 + 75 * (d < 26) - ((flag != 0) << 5);
 100   /*  0..25 map to ASCII a..z or A..Z */
 101   /* 26..35 map to ASCII 0..9         */
 102 }
 103
 104 /* flagged(bcp) tests whether a basic code point is flagged */
 105 /* (uppercase).  The behavior is undefined if bcp is not a  */
 106 /* basic code point.                                        */
 107
 108 #define flagged(bcp) ((punycode_uint)(bcp) - 65 < 26)
 109
 110 /* encode_basic(bcp,flag) forces a basic code point to lowercase */
 111 /* if flag is zero, uppercase if flag is nonzero, and returns    */
 112 /* the resulting code point.  The code point is unchanged if it  */
 113 /* is caseless.  The behavior is undefined if bcp is not a basic */
 114 /* code point.                                                   */
 115
 116 static char
 117 encode_basic (punycode_uint bcp, int flag)
 118 {
 119   bcp -= (bcp - 97 < 26) << 5;
 120   return bcp + ((!flag && (bcp - 65 < 26)) << 5);
 121 }
 122
 123 /*** Platform-specific constants ***/
 124
 125 /* maxint is the maximum value of a punycode_uint variable: */
 126 static const punycode_uint maxint = -1;
 127 /* Because maxint is unsigned, -1 becomes the maximum value. */
 128
 129 /*** Bias adaptation function ***/
 130
 131 static punycode_uint
 132 adapt (punycode_uint delta, punycode_uint numpoints, int firsttime)
 133 {
 134   punycode_uint k;
 135
 136   delta = firsttime ? delta / damp : delta >> 1;
 137   /* delta >> 1 is a faster way of doing delta / 2 */
 138   delta += delta / numpoints;
 139
 140   for (k = 0; delta > ((base - tmin) * tmax) / 2; k += base)
 141     {
 142       delta /= base - tmin;
 143     }
 144
 145   return k + (base - tmin + 1) * delta / (delta + skew);
 146 }
 147
 148 /*** Main encode function ***/
 149
 150 /**
 151  * punycode_encode:
 152  * @input_length: The number of code points in the @input array and
 153  *   the number of flags in the @case_flags array.
 154  * @input: An array of code points.  They are presumed to be Unicode
 155  *   code points, but that is not strictly REQUIRED.  The array
 156  *   contains code points, not code units.  UTF-16 uses code units
 157  *   D800 through DFFF to refer to code points 10000..10FFFF.  The
 158  *   code points D800..DFFF do not occur in any valid Unicode string.
 159  *   The code points that can occur in Unicode strings (0..D7FF and
 160  *   E000..10FFFF) are also called Unicode scalar values.
 161  * @case_flags: A %NULL pointer or an array of boolean values parallel
 162  *   to the @input array.  Nonzero (true, flagged) suggests that the
 163  *   corresponding Unicode character be forced to uppercase after
 164  *   being decoded (if possible), and zero (false, unflagged) suggests
 165  *   that it be forced to lowercase (if possible).  ASCII code points
 166  *   (0..7F) are encoded literally, except that ASCII letters are
 167  *   forced to uppercase or lowercase according to the corresponding
 168  *   case flags.  If @case_flags is a %NULL pointer then ASCII letters
 169  *   are left as they are, and other code points are treated as
 170  *   unflagged.
 171  * @output_length: The caller passes in the maximum number of ASCII
 172  *   code points that it can receive.  On successful return it will
 173  *   contain the number of ASCII code points actually output.
 174  * @output: An array of ASCII code points.  It is *not*
 175  *   null-terminated; it will contain zeros if and only if the @input
 176  *   contains zeros.  (Of course the caller can leave room for a
 177  *   terminator and add one if needed.)
 178  *
 179  * Converts a sequence of code points (presumed to be Unicode code
 180  * points) to Punycode.
 181  *
 182  * Return value: The return value can be any of the punycode_status
 183  *   values defined above except %punycode_bad_input.  If not
 184  *   %punycode_success, then @output_size and @output might contain
 185  *   garbage.
 186  **/
 187 int
 188 punycode_encode (size_t input_length,
 189                  const punycode_uint input[],
 190                  const unsigned char case_flags[],
 191                  size_t * output_length, char output[])
 192 {
 193   punycode_uint input_len, n, delta, h, b, bias, j, m, q, k, t;
 194   size_t out, max_out;
 195
 196   /* The Punycode spec assumes that the input length is the same type */
 197   /* of integer as a code point, so we need to convert the size_t to  */
 198   /* a punycode_uint, which could overflow.                           */
 199
 200   if (input_length > maxint)
 201     return punycode_overflow;
 202   input_len = (punycode_uint) input_length;
 203
 204   /* Initialize the state: */
 205
 206   n = initial_n;
 207   delta = 0;
 208   out = 0;
 209   max_out = *output_length;
 210   bias = initial_bias;
 211
 212   /* Handle the basic code points: */
 213
 214   for (j = 0; j < input_len; ++j)
 215     {
 216       if (basic (input[j]))
 217         {
 218           if (max_out - out < 2)
 219             return punycode_big_output;
 220           output[out++] = case_flags ?
 221             encode_basic (input[j], case_flags[j]) : (char) input[j];
 222         }
 223       /* else if (input[j] < n) return punycode_bad_input; */
 224       /* (not needed for Punycode with unsigned code points) */
 225     }
 226
 227   h = b = (punycode_uint) out;
 228   /* cannot overflow because out <= input_len <= maxint */
 229
 230   /* h is the number of code points that have been handled, b is the  */
 231   /* number of basic code points, and out is the number of ASCII code */
 232   /* points that have been output.                                    */
 233
 234   if (b > 0)
 235     output[out++] = delimiter;
 236
 237   /* Main encoding loop: */
 238
 239   while (h < input_len)
 240     {
 241       /* All non-basic code points < n have been     */
 242       /* handled already.  Find the next larger one: */
 243
 244       for (m = maxint, j = 0; j < input_len; ++j)
 245         {
 246           /* if (basic(input[j])) continue; */
 247           /* (not needed for Punycode) */
 248           if (input[j] >= n && input[j] < m)
 249             m = input[j];
 250         }
 251
 252       /* Increase delta enough to advance the decoder's    */
 253       /* <n,i> state to <m,0>, but guard against overflow: */
 254
 255       if (m - n > (maxint - delta) / (h + 1))
 256         return punycode_overflow;
 257       delta += (m - n) * (h + 1);
 258       n = m;
 259
 260       for (j = 0; j < input_len; ++j)
 261         {
 262           /* Punycode does not need to check whether input[j] is basic: */
 263           if (input[j] < n /* || basic(input[j]) */ )
 264             {
 265               if (++delta == 0)
 266                 return punycode_overflow;
 267             }
 268
 269           if (input[j] == n)
 270             {
 271               /* Represent delta as a generalized variable-length integer: */
 272
 273               for (q = delta, k = base;; k += base)
 274                 {
 275                   if (out >= max_out)
 276                     return punycode_big_output;
 277                   t = k <= bias /* + tmin */ ? tmin :   /* +tmin not needed */
 278                     k >= bias + tmax ? tmax : k - bias;
 279                   if (q < t)
 280                     break;
 281                   output[out++] = encode_digit (t + (q - t) % (base - t), 0);
 282                   q = (q - t) / (base - t);
 283                 }
 284
 285               output[out++] = encode_digit (q, case_flags && case_flags[j]);
 286               bias = adapt (delta, h + 1, h == b);
 287               delta = 0;
 288               ++h;
 289             }
 290         }
 291
 292       ++delta, ++n;
 293     }
 294
 295   *output_length = out;
 296   return punycode_success;
 297 }
 298
 299 /*** Main decode function ***/
 300
 301 /**
 302  * punycode_decode:
 303  * @input_length: The number of ASCII code points in the @input array.
 304  * @input: An array of ASCII code points (0..7F).
 305  * @output_length: The caller passes in the maximum number of code
 306  *   points that it can receive into the @output array (which is also
 307  *   the maximum number of flags that it can receive into the
 308  *   @case_flags array, if @case_flags is not a %NULL pointer).  On
 309  *   successful return it will contain the number of code points
 310  *   actually output (which is also the number of flags actually
 311  *   output, if case_flags is not a null pointer).  The decoder will
 312  *   never need to output more code points than the number of ASCII
 313  *   code points in the input, because of the way the encoding is
 314  *   defined.  The number of code points output cannot exceed the
 315  *   maximum possible value of a punycode_uint, even if the supplied
 316  *   @output_length is greater than that.
 317  * @output: An array of code points like the input argument of
 318  *   punycode_encode() (see above).
 319  * @case_flags: A %NULL pointer (if the flags are not needed by the
 320  *   caller) or an array of boolean values parallel to the @output
 321  *   array.  Nonzero (true, flagged) suggests that the corresponding
 322  *   Unicode character be forced to uppercase by the caller (if
 323  *   possible), and zero (false, unflagged) suggests that it be forced
 324  *   to lowercase (if possible).  ASCII code points (0..7F) are output
 325  *   already in the proper case, but their flags will be set
 326  *   appropriately so that applying the flags would be harmless.
 327  *
 328  * Converts Punycode to a sequence of code points (presumed to be
 329  * Unicode code points).
 330  *
 331  * Return value: The return value can be any of the punycode_status
 332  *   values defined above.  If not %punycode_success, then
 333  *   @output_length, @output, and @case_flags might contain garbage.
 334  *
 335  **/
 336 int
 337 punycode_decode (size_t input_length,
 338                  const char input[],
 339                  size_t * output_length,
 340                  punycode_uint output[], unsigned char case_flags[])
 341 {
 342   punycode_uint n, out, i, max_out, bias, oldi, w, k, digit, t;
 343   size_t b, j, in;
 344
 345   /* Initialize the state: */
 346
 347   n = initial_n;
 348   out = i = 0;
 349   max_out = *output_length > maxint ? maxint
 350     : (punycode_uint) * output_length;
 351   bias = initial_bias;
 352
 353   /* Handle the basic code points:  Let b be the number of input code */
 354   /* points before the last delimiter, or 0 if there is none, then    */
 355   /* copy the first b code points to the output.                      */
 356
 357   for (b = j = 0; j < input_length; ++j)
 358     if (delim (input[j]))
 359       b = j;
 360   if (b > max_out)
 361     return punycode_big_output;
 362
 363   for (j = 0; j < b; ++j)
 364     {
 365       if (case_flags)
 366         case_flags[out] = flagged (input[j]);
 367       if (!basic (input[j]))
 368         return punycode_bad_input;
 369       output[out++] = input[j];
 370     }
 371
 372   /* Main decoding loop:  Start just after the last delimiter if any  */
 373   /* basic code points were copied; start at the beginning otherwise. */
 374
 375   for (in = b > 0 ? b + 1 : 0; in < input_length; ++out)
 376     {
 377
 378       /* in is the index of the next ASCII code point to be consumed, */
 379       /* and out is the number of code points in the output array.    */
 380
 381       /* Decode a generalized variable-length integer into delta,  */
 382       /* which gets added to i.  The overflow checking is easier   */
 383       /* if we increase i as we go, then subtract off its starting */
 384       /* value at the end to obtain delta.                         */
 385
 386       for (oldi = i, w = 1, k = base;; k += base)
 387         {
 388           if (in >= input_length)
 389             return punycode_bad_input;
 390           digit = decode_digit (input[in++]);
 391           if (digit >= base)
 392             return punycode_bad_input;
 393           if (digit > (maxint - i) / w)
 394             return punycode_overflow;
 395           i += digit * w;
 396           t = k <= bias /* + tmin */ ? tmin :   /* +tmin not needed */
 397             k >= bias + tmax ? tmax : k - bias;
 398           if (digit < t)
 399             break;
 400           if (w > maxint / (base - t))
 401             return punycode_overflow;
 402           w *= (base - t);
 403         }
 404
 405       bias = adapt (i - oldi, out + 1, oldi == 0);
 406
 407       /* i was supposed to wrap around from out+1 to 0,   */
 408       /* incrementing n each time, so we'll fix that now: */
 409
 410       if (i / (out + 1) > maxint - n)
 411         return punycode_overflow;
 412       n += i / (out + 1);
 413       i %= (out + 1);
 414
 415       /* Insert n at position i of the output: */
 416
 417       /* not needed for Punycode: */
 418       /* if (basic(n)) return punycode_invalid_input; */
 419       if (out >= max_out)
 420         return punycode_big_output;
 421
 422       if (case_flags)
 423         {
 424           memmove (case_flags + i + 1, case_flags + i, out - i);
 425           /* Case of last ASCII code point determines case flag: */
 426           case_flags[i] = flagged (input[in - 1]);
 427         }
 428
 429       memmove (output + i + 1, output + i, (out - i) * sizeof *output);
 430       output[i++] = n;
 431     }
 432
 433   *output_length = (size_t) out;
 434   /* cannot overflow because out <= old value of *output_length */
 435   return punycode_success;
 436 }
 437
 438 /**
 439  * punycode_uint
 440  *
 441  * Unicode code point data type, this is always a 32 bit unsigned
 442  * integer.
 443  */
 444
 445 /**
 446  * Punycode_status
 447  * @PUNYCODE_SUCCESS: Successful operation.  This value is guaranteed
 448  *   to always be zero, the remaining ones are only guaranteed to hold
 449  *   non-zero values, for logical comparison purposes.
 450  * @PUNYCODE_BAD_INPUT: Input is invalid.
 451  * @PUNYCODE_BIG_OUTPUT: Output would exceed the space provided.
 452  * @PUNYCODE_OVERFLOW: Input needs wider integers to process.
 453  *
 454  * Enumerated return codes of punycode_encode() and punycode_decode().
 455  * The value 0 is guaranteed to always correspond to success.
 456  */