1 | // Licensed to the .NET Foundation under one or more agreements. |
2 | // The .NET Foundation licenses this file to you under the MIT license. |
3 | // See the LICENSE file in the project root for more information. |
4 | // |
5 | |
6 | #include <assert.h> |
7 | #include <stdint.h> |
8 | #include <string.h> |
9 | #include <stdlib.h> |
10 | #include <locale.h> |
11 | |
12 | #include "icushim.h" |
13 | #include "locale.hpp" |
14 | |
15 | int32_t UErrorCodeToBool(UErrorCode status) |
16 | { |
17 | if (U_SUCCESS(status)) |
18 | { |
19 | return 1; |
20 | } |
21 | |
22 | // assert errors that should never occur |
23 | assert(status != U_BUFFER_OVERFLOW_ERROR); |
24 | assert(status != U_INTERNAL_PROGRAM_ERROR); |
25 | |
26 | // add possible SetLastError support here |
27 | |
28 | return 0; |
29 | } |
30 | |
31 | int32_t GetLocale( |
32 | const UChar* localeName, char* localeNameResult, int32_t localeNameResultLength, bool canonicalize, UErrorCode* err) |
33 | { |
34 | char localeNameTemp[ULOC_FULLNAME_CAPACITY] = {0}; |
35 | int32_t localeLength; |
36 | |
37 | // Convert ourselves instead of doing u_UCharsToChars as that function considers '@' a variant and stops. |
38 | for (int i = 0; i < ULOC_FULLNAME_CAPACITY - 1; i++) |
39 | { |
40 | UChar c = localeName[i]; |
41 | |
42 | if (c > (UChar)0x7F) |
43 | { |
44 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
45 | return ULOC_FULLNAME_CAPACITY; |
46 | } |
47 | |
48 | localeNameTemp[i] = (char)c; |
49 | |
50 | if (c == (UChar)0x0) |
51 | { |
52 | break; |
53 | } |
54 | } |
55 | |
56 | if (canonicalize) |
57 | { |
58 | localeLength = uloc_canonicalize(localeNameTemp, localeNameResult, localeNameResultLength, err); |
59 | } |
60 | else |
61 | { |
62 | localeLength = uloc_getName(localeNameTemp, localeNameResult, localeNameResultLength, err); |
63 | } |
64 | |
65 | if (U_SUCCESS(*err)) |
66 | { |
67 | // Make sure the "language" part of the locale is reasonable (i.e. we can fetch it and it is within range). |
68 | // This mimics how the C++ ICU API determines if a locale is "bogus" or not. |
69 | |
70 | char language[ULOC_LANG_CAPACITY]; |
71 | uloc_getLanguage(localeNameTemp, language, ULOC_LANG_CAPACITY, err); |
72 | |
73 | if (*err == U_BUFFER_OVERFLOW_ERROR || *err == U_STRING_NOT_TERMINATED_WARNING) |
74 | { |
75 | // ULOC_LANG_CAPACITY includes the null terminator, so if we couldn't extract the language with the null |
76 | // terminator, the language must be invalid. |
77 | |
78 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
79 | } |
80 | } |
81 | |
82 | return localeLength; |
83 | } |
84 | |
85 | UErrorCode u_charsToUChars_safe(const char* str, UChar* value, int32_t valueLength) |
86 | { |
87 | int len = strlen(str); |
88 | |
89 | if (len >= valueLength) |
90 | { |
91 | return U_BUFFER_OVERFLOW_ERROR; |
92 | } |
93 | |
94 | u_charsToUChars(str, value, len + 1); |
95 | return U_ZERO_ERROR; |
96 | } |
97 | |
98 | int32_t FixupLocaleName(UChar* value, int32_t valueLength) |
99 | { |
100 | int32_t i = 0; |
101 | for (; i < valueLength; i++) |
102 | { |
103 | if (value[i] == (UChar)'\0') |
104 | { |
105 | break; |
106 | } |
107 | else if (value[i] == (UChar)'_') |
108 | { |
109 | value[i] = (UChar)'-'; |
110 | } |
111 | } |
112 | |
113 | return i; |
114 | } |
115 | |
116 | bool IsEnvVarSet(const char* name) |
117 | { |
118 | const char* value = getenv(name); |
119 | |
120 | return (value != nullptr) && (strcmp("" , value) != 0); |
121 | } |
122 | |
123 | // The behavior of uloc_getDefault() on POSIX systems is to query |
124 | // setlocale(LC_MESSAGES) and use that value, unless it is C or |
125 | // POSIX. In that case it tries to read LC_ALL, LC_MESSAGES and LANG |
126 | // and then falls back to en_US_POSIX if none of them are set. |
127 | // |
128 | // en_US_POSIX is a weird locale since the collation rules treat 'a' |
129 | // and 'A' as different letters even when ignoring case. Furthermore |
130 | // it's common for LC_ALL, LC_MESSAGES and LANG to be unset when |
131 | // running under Docker. |
132 | // |
133 | // We'd rather default to invariant in this case. If any of these |
134 | // are set, we'll just call into ICU and let it do whatever |
135 | // normalization it would do. |
136 | const char* DetectDefaultLocaleName() |
137 | { |
138 | char* loc = setlocale(LC_MESSAGES, nullptr); |
139 | |
140 | if (loc != nullptr && (strcmp("C" , loc) == 0 || strcmp("POSIX" , loc) == 0)) |
141 | { |
142 | if (!IsEnvVarSet("LC_ALL" ) && !IsEnvVarSet("LC_MESSAGES" ) && !IsEnvVarSet("LANG" )) |
143 | { |
144 | return "" ; |
145 | } |
146 | } |
147 | |
148 | return uloc_getDefault(); |
149 | } |
150 | |
151 | // GlobalizationNative_GetLocales gets all locale names and store it in the value buffer |
152 | // in case of success, it returns the count of the characters stored in value buffer |
153 | // in case of failure, it returns negative number. |
154 | // if the input value buffer is null, it returns the length needed to store the |
155 | // locale names list. |
156 | // if the value is not null, it fills the value with locale names separated by the length |
157 | // of each name. |
158 | extern "C" int32_t GlobalizationNative_GetLocales(UChar *value, int32_t valueLength) |
159 | { |
160 | int32_t totalLength = 0; |
161 | int32_t index = 0; |
162 | int32_t localeCount = uloc_countAvailable(); |
163 | |
164 | if (localeCount <= 0) |
165 | return -1; // failed |
166 | |
167 | for (int32_t i = 0; i < localeCount; i++) |
168 | { |
169 | const char *pLocaleName = uloc_getAvailable(i); |
170 | if (pLocaleName[0] == 0) // unexpected empty name |
171 | return -2; |
172 | |
173 | int32_t localeNameLength = strlen(pLocaleName); |
174 | |
175 | totalLength += localeNameLength + 1; // add 1 for the name length |
176 | |
177 | if (value != nullptr) |
178 | { |
179 | if (totalLength > valueLength) |
180 | return -3; |
181 | |
182 | value[index++] = (UChar) localeNameLength; |
183 | |
184 | for (int j=0; j<localeNameLength; j++) |
185 | { |
186 | if (pLocaleName[j] == '_') // fix the locale name |
187 | { |
188 | value[index++] = (UChar) '-'; |
189 | } |
190 | else |
191 | { |
192 | value[index++] = (UChar) pLocaleName[j]; |
193 | } |
194 | } |
195 | } |
196 | } |
197 | |
198 | return totalLength; |
199 | } |
200 | |
201 | extern "C" int32_t GlobalizationNative_GetLocaleName(const UChar* localeName, UChar* value, int32_t valueLength) |
202 | { |
203 | UErrorCode status = U_ZERO_ERROR; |
204 | |
205 | char localeNameBuffer[ULOC_FULLNAME_CAPACITY]; |
206 | GetLocale(localeName, localeNameBuffer, ULOC_FULLNAME_CAPACITY, true, &status); |
207 | |
208 | if (U_SUCCESS(status)) |
209 | { |
210 | status = u_charsToUChars_safe(localeNameBuffer, value, valueLength); |
211 | |
212 | if (U_SUCCESS(status)) |
213 | { |
214 | FixupLocaleName(value, valueLength); |
215 | } |
216 | } |
217 | |
218 | return UErrorCodeToBool(status); |
219 | } |
220 | |
221 | extern "C" int32_t GlobalizationNative_GetDefaultLocaleName(UChar* value, int32_t valueLength) |
222 | { |
223 | char localeNameBuffer[ULOC_FULLNAME_CAPACITY]; |
224 | UErrorCode status = U_ZERO_ERROR; |
225 | |
226 | const char* defaultLocale = DetectDefaultLocaleName(); |
227 | |
228 | uloc_getBaseName(defaultLocale, localeNameBuffer, ULOC_FULLNAME_CAPACITY, &status); |
229 | |
230 | if (U_SUCCESS(status)) |
231 | { |
232 | status = u_charsToUChars_safe(localeNameBuffer, value, valueLength); |
233 | |
234 | if (U_SUCCESS(status)) |
235 | { |
236 | int localeNameLen = FixupLocaleName(value, valueLength); |
237 | |
238 | char collationValueTemp[ULOC_KEYWORDS_CAPACITY]; |
239 | int32_t collationLen = |
240 | uloc_getKeywordValue(defaultLocale, "collation" , collationValueTemp, ULOC_KEYWORDS_CAPACITY, &status); |
241 | |
242 | if (U_SUCCESS(status) && collationLen > 0) |
243 | { |
244 | // copy the collation; managed uses a "_" to represent collation (not |
245 | // "@collation=") |
246 | status = u_charsToUChars_safe("_" , &value[localeNameLen], valueLength - localeNameLen); |
247 | if (U_SUCCESS(status)) |
248 | { |
249 | status = u_charsToUChars_safe( |
250 | collationValueTemp, &value[localeNameLen + 1], valueLength - localeNameLen - 1); |
251 | } |
252 | } |
253 | } |
254 | } |
255 | |
256 | return UErrorCodeToBool(status); |
257 | } |
258 | |