xf.li | bdd93d5 | 2023-05-12 07:10:14 -0700 | [diff] [blame] | 1 | /* Internal header for proving correct grouping in strings of numbers. |
| 2 | Copyright (C) 1995-2016 Free Software Foundation, Inc. |
| 3 | This file is part of the GNU C Library. |
| 4 | Contributed by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995. |
| 5 | |
| 6 | The GNU C Library 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 | The GNU C Library 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 the GNU C Library; if not, see |
| 18 | <http://www.gnu.org/licenses/>. */ |
| 19 | |
| 20 | #include <limits.h> |
| 21 | #include <stddef.h> |
| 22 | #include <string.h> |
| 23 | |
| 24 | #ifndef MAX |
| 25 | #define MAX(a,b) ({ typeof(a) _a = (a); typeof(b) _b = (b); \ |
| 26 | _a > _b ? _a : _b; }) |
| 27 | #endif |
| 28 | |
| 29 | #ifdef USE_WIDE_CHAR |
| 30 | # include <wctype.h> |
| 31 | # define L_(Ch) L##Ch |
| 32 | # define UCHAR_TYPE wint_t |
| 33 | # define STRING_TYPE wchar_t |
| 34 | #else |
| 35 | # define L_(Ch) Ch |
| 36 | # define UCHAR_TYPE unsigned char |
| 37 | # define STRING_TYPE char |
| 38 | #endif |
| 39 | |
| 40 | #include "grouping.h" |
| 41 | |
| 42 | /* Find the maximum prefix of the string between BEGIN and END which |
| 43 | satisfies the grouping rules. It is assumed that at least one digit |
| 44 | follows BEGIN directly. */ |
| 45 | |
| 46 | const STRING_TYPE * |
| 47 | #ifdef USE_WIDE_CHAR |
| 48 | __correctly_grouped_prefixwc (const STRING_TYPE *begin, const STRING_TYPE *end, |
| 49 | wchar_t thousands, |
| 50 | #else |
| 51 | __correctly_grouped_prefixmb (const STRING_TYPE *begin, const STRING_TYPE *end, |
| 52 | const char *thousands, |
| 53 | #endif |
| 54 | const char *grouping) |
| 55 | { |
| 56 | #ifndef USE_WIDE_CHAR |
| 57 | size_t thousands_len; |
| 58 | int cnt; |
| 59 | #endif |
| 60 | |
| 61 | if (grouping == NULL) |
| 62 | return end; |
| 63 | |
| 64 | #ifndef USE_WIDE_CHAR |
| 65 | thousands_len = strlen (thousands); |
| 66 | #endif |
| 67 | |
| 68 | while (end > begin) |
| 69 | { |
| 70 | const STRING_TYPE *cp = end - 1; |
| 71 | const char *gp = grouping; |
| 72 | |
| 73 | /* Check first group. */ |
| 74 | while (cp >= begin) |
| 75 | { |
| 76 | #ifdef USE_WIDE_CHAR |
| 77 | if (*cp == thousands) |
| 78 | break; |
| 79 | #else |
| 80 | if (cp[thousands_len - 1] == *thousands) |
| 81 | { |
| 82 | for (cnt = 1; thousands[cnt] != '\0'; ++cnt) |
| 83 | if (thousands[cnt] != cp[thousands_len - 1 - cnt]) |
| 84 | break; |
| 85 | if (thousands[cnt] == '\0') |
| 86 | break; |
| 87 | } |
| 88 | #endif |
| 89 | --cp; |
| 90 | } |
| 91 | |
| 92 | /* We allow the representation to contain no grouping at all even if |
| 93 | the locale specifies we can have grouping. */ |
| 94 | if (cp < begin) |
| 95 | return end; |
| 96 | |
| 97 | if (end - cp == (int) *gp + 1) |
| 98 | { |
| 99 | /* This group matches the specification. */ |
| 100 | |
| 101 | const STRING_TYPE *new_end; |
| 102 | |
| 103 | if (cp < begin) |
| 104 | /* There is just one complete group. We are done. */ |
| 105 | return end; |
| 106 | |
| 107 | /* CP points to a thousands separator character. The preceding |
| 108 | remainder of the string from BEGIN to NEW_END is the part we |
| 109 | will consider if there is a grouping error in this trailing |
| 110 | portion from CP to END. */ |
| 111 | new_end = cp - 1; |
| 112 | |
| 113 | /* Loop while the grouping is correct. */ |
| 114 | while (1) |
| 115 | { |
| 116 | /* Get the next grouping rule. */ |
| 117 | ++gp; |
| 118 | if (*gp == 0) |
| 119 | /* If end is reached use last rule. */ |
| 120 | --gp; |
| 121 | |
| 122 | /* Skip the thousands separator. */ |
| 123 | --cp; |
| 124 | |
| 125 | if (*gp == CHAR_MAX |
| 126 | #if CHAR_MIN < 0 |
| 127 | || *gp < 0 |
| 128 | #endif |
| 129 | ) |
| 130 | { |
| 131 | /* No more thousands separators are allowed to follow. */ |
| 132 | while (cp >= begin) |
| 133 | { |
| 134 | #ifdef USE_WIDE_CHAR |
| 135 | if (*cp == thousands) |
| 136 | break; |
| 137 | #else |
| 138 | for (cnt = 0; thousands[cnt] != '\0'; ++cnt) |
| 139 | if (thousands[cnt] != cp[thousands_len - cnt - 1]) |
| 140 | break; |
| 141 | if (thousands[cnt] == '\0') |
| 142 | break; |
| 143 | #endif |
| 144 | --cp; |
| 145 | } |
| 146 | |
| 147 | if (cp < begin) |
| 148 | /* OK, only digits followed. */ |
| 149 | return end; |
| 150 | } |
| 151 | else |
| 152 | { |
| 153 | /* Check the next group. */ |
| 154 | const STRING_TYPE *group_end = cp; |
| 155 | |
| 156 | while (cp >= begin) |
| 157 | { |
| 158 | #ifdef USE_WIDE_CHAR |
| 159 | if (*cp == thousands) |
| 160 | break; |
| 161 | #else |
| 162 | for (cnt = 0; thousands[cnt] != '\0'; ++cnt) |
| 163 | if (thousands[cnt] != cp[thousands_len - cnt - 1]) |
| 164 | break; |
| 165 | if (thousands[cnt] == '\0') |
| 166 | break; |
| 167 | #endif |
| 168 | --cp; |
| 169 | } |
| 170 | |
| 171 | if (cp < begin && group_end - cp <= (int) *gp) |
| 172 | /* Final group is correct. */ |
| 173 | return end; |
| 174 | |
| 175 | if (cp < begin || group_end - cp != (int) *gp) |
| 176 | /* Incorrect group. Punt. */ |
| 177 | break; |
| 178 | } |
| 179 | } |
| 180 | |
| 181 | /* The trailing portion of the string starting at NEW_END |
| 182 | contains a grouping error. So we will look for a correctly |
| 183 | grouped number in the preceding portion instead. */ |
| 184 | end = new_end; |
| 185 | } |
| 186 | else |
| 187 | { |
| 188 | /* Even the first group was wrong; determine maximum shift. */ |
| 189 | if (end - cp > (int) *gp + 1) |
| 190 | end = cp + (int) *gp + 1; |
| 191 | else if (cp < begin) |
| 192 | /* This number does not fill the first group, but is correct. */ |
| 193 | return end; |
| 194 | else |
| 195 | /* CP points to a thousands separator character. */ |
| 196 | end = cp; |
| 197 | } |
| 198 | } |
| 199 | |
| 200 | return MAX (begin, end); |
| 201 | } |