gcc/rust/util/rust-punycode.cc

   1 // Copyright (C) 2020-2024 Free Software Foundation, Inc.
   2
   3 // This file is part of GCC.
   4
   5 // GCC is free software; you can redistribute it and/or modify it under
   6 // the terms of the GNU General Public License as published by the Free
   7 // Software Foundation; either version 3, or (at your option) any later
   8 // version.
   9
  10 // GCC is distributed in the hope that it will be useful, but WITHOUT ANY
  11 // WARRANTY; without even the implied warranty of MERCHANTABILITY or
  12 // FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  13 // for more details.
  14
  15 // You should have received a copy of the GNU General Public License
  16 // along with GCC; see the file COPYING3.  If not see
  17 // <http://www.gnu.org/licenses/>.
  18
  19 // This file provides functions for punycode conversion
  20 // See https://datatracker.ietf.org/doc/html/rfc3492
  21
  22 #include "rust-system.h"
  23 #include "rust-unicode.h"
  24 #include "optional.h"
  25 #include "selftest.h"
  26
  27 namespace Rust {
  28
  29 // https://tools.ietf.org/html/rfc3492#section-4.
  30 constexpr uint32_t BASE = 36;
  31 constexpr uint32_t TMIN = 1;
  32 constexpr uint32_t TMAX = 26;
  33 constexpr uint32_t SKEW = 38;
  34 constexpr uint32_t DAMP = 700;
  35 constexpr uint32_t INITIAL_BIAS = 72;
  36 constexpr uint32_t INITIAL_N = 128;
  37 constexpr char DELIMITER = '-';
  38
  39 std::string
  40 extract_basic_string (const std::vector<Codepoint> &src)
  41 {
  42   std::string basic_string;
  43   for (auto c : src)
  44     {
  45       if (c.is_ascii ())
  46         basic_string += c.as_string ();
  47     }
  48   return basic_string;
  49 }
  50
  51 uint32_t
  52 adapt_bias (uint32_t delta, const uint32_t n_points, const bool is_first)
  53 {
  54   delta /= is_first ? DAMP : 2;
  55   delta += delta / n_points;
  56   uint32_t k = 0;
  57
  58   while (delta > (BASE - TMIN) * TMAX / 2)
  59     {
  60       delta /= BASE - TMIN;
  61       k += BASE;
  62     }
  63   return k + (BASE - TMIN + 1) * delta / (delta + SKEW);
  64 }
  65
  66 uint32_t
  67 clamped_sub (const uint32_t min, const uint32_t lhs, const uint32_t rhs,
  68              const uint32_t max)
  69 {
  70   if (min + rhs >= lhs)
  71     return min;
  72   else if (max + rhs <= lhs)
  73     return max;
  74   else
  75     return lhs - rhs;
  76 }
  77
  78 uint32_t
  79 min_gt_or_eq (const std::vector<Codepoint> &l, const uint32_t threshold)
  80 {
  81   uint32_t min = UINT32_MAX;
  82   for (auto c : l)
  83     if (c.value >= threshold && c.value < min)
  84       min = c.value;
  85   return min;
  86 }
  87
  88 char
  89 encode_digit (const uint32_t d)
  90 {
  91   return d + 22 + (d < 26 ? 75 : 0);
  92 }
  93
  94 tl::optional<std::string>
  95 encode_punycode (const Utf8String &input)
  96 {
  97   std::vector<Codepoint> input_chars = input.get_chars ();
  98
  99   uint32_t n = INITIAL_N;
 100   uint32_t delta = 0;
 101   uint32_t bias = INITIAL_BIAS;
 102
 103   std::string output = extract_basic_string (input_chars);
 104   uint32_t h = output.size ();
 105   const uint32_t b = h;
 106   if (b > 0)
 107     output += DELIMITER;
 108
 109   while (h < input_chars.size ())
 110     {
 111       const uint32_t m = min_gt_or_eq (input_chars, n);
 112
 113       if (m - n > ((UINT32_MAX - delta) / (h + 1)))
 114         return tl::nullopt;
 115
 116       delta += (m - n) * (h + 1);
 117       n = m;
 118
 119       for (const auto c : input_chars)
 120         {
 121           if (c.value < n)
 122             delta++;
 123           else if (c.value == n)
 124             {
 125               uint32_t q = delta;
 126               // encode as a variable length integer
 127               for (uint32_t k = 1;; k++)
 128                 {
 129                   const uint32_t kb = k * BASE;
 130                   const uint32_t t = clamped_sub (TMIN, kb, bias, TMAX);
 131                   if (q < t)
 132                     break;
 133
 134                   output += encode_digit (t + (q - t) % (BASE - t));
 135                   q = (q - t) / (BASE - t);
 136                 }
 137               output += encode_digit (q);
 138
 139               bias = adapt_bias (delta, h + 1, h == b);
 140               delta = 0;
 141               h++;
 142             }
 143         }
 144       delta++;
 145       n++;
 146     }
 147
 148   return {output};
 149 }
 150
 151 } // namespace Rust
 152
 153 #if CHECKING_P
 154
 155 namespace selftest {
 156
 157 void
 158 encode_assert (const std::string &input, const std::string &expected)
 159 {
 160   Rust::Utf8String input_utf8
 161     = Rust::Utf8String::make_utf8_string (input).value ();
 162   std::string actual = Rust::encode_punycode (input_utf8).value ();
 163   ASSERT_EQ (actual, expected);
 164 }
 165
 166 void
 167 rust_punycode_encode_test ()
 168 {
 169   encode_assert ("abc", "abc-");
 170   encode_assert ("12345", "12345-");
 171   encode_assert ("香港", "j6w193g");
 172
 173   // Examples from https://datatracker.ietf.org/doc/html/rfc3492#section-7.1
 174   encode_assert ("ليهمابتكلموشعربي؟", "egbpdaj6bu4bxfgehfvwxn");
 175   encode_assert ("他们为什么不说中文", "ihqwcrb4cv8a8dqg056pqjye");
 176   encode_assert ("他們爲什麽不說中文", "ihqwctvzc91f659drss3x8bo0yb");
 177   encode_assert ("Pročprostěnemluvíčesky", "Proprostnemluvesky-uyb24dma41a");
 178 }
 179
 180 } // namespace selftest
 181
 182 #endif // CHECKING_P