DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Header

Mercurial (b6d82b1a6b02)

VCS Links

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
 *******************************************************************************
 * Copyright (C) 2002-2014, International Business Machines Corporation and
 * others. All Rights Reserved.
 *******************************************************************************
 */
#include "unicode/utypes.h"

#if !UCONFIG_NO_SERVICE || !UCONFIG_NO_TRANSLITERATION

#include "unicode/resbund.h"
#include "unicode/uenum.h"
#include "cmemory.h"
#include "ustrfmt.h"
#include "locutil.h"
#include "charstr.h"
#include "ucln_cmn.h"
#include "uassert.h"
#include "umutex.h"

// see LocaleUtility::getAvailableLocaleNames
static icu::UInitOnce   LocaleUtilityInitOnce = U_INITONCE_INITIALIZER;
static icu::Hashtable * LocaleUtility_cache = NULL;

#define UNDERSCORE_CHAR ((UChar)0x005f)
#define AT_SIGN_CHAR    ((UChar)64)
#define PERIOD_CHAR     ((UChar)46)

/*
 ******************************************************************
 */

/**
 * Release all static memory held by Locale Utility.  
 */
U_CDECL_BEGIN
static UBool U_CALLCONV service_cleanup(void) {
    if (LocaleUtility_cache) {
        delete LocaleUtility_cache;
        LocaleUtility_cache = NULL;
    }
    return TRUE;
}


static void U_CALLCONV locale_utility_init(UErrorCode &status) {
    using namespace icu;
    U_ASSERT(LocaleUtility_cache == NULL);
    ucln_common_registerCleanup(UCLN_COMMON_SERVICE, service_cleanup);
    LocaleUtility_cache = new Hashtable(status);
    if (U_FAILURE(status)) {
        delete LocaleUtility_cache;
        LocaleUtility_cache = NULL;
        return;
    }
    if (LocaleUtility_cache == NULL) {
        status = U_MEMORY_ALLOCATION_ERROR;
        return;
    }
    LocaleUtility_cache->setValueDeleter(uhash_deleteHashtable);
}

U_CDECL_END

U_NAMESPACE_BEGIN

UnicodeString&
LocaleUtility::canonicalLocaleString(const UnicodeString* id, UnicodeString& result)
{
  if (id == NULL) {
    result.setToBogus();
  } else {
    // Fix case only (no other changes) up to the first '@' or '.' or
    // end of string, whichever comes first.  In 3.0 I changed this to
    // stop at first '@' or '.'.  It used to run out to the end of
    // string.  My fix makes the tests pass but is probably
    // structurally incorrect.  See below.  [alan 3.0]

    // TODO: Doug, you might want to revise this...
    result = *id;
    int32_t i = 0;
    int32_t end = result.indexOf(AT_SIGN_CHAR);
    int32_t n = result.indexOf(PERIOD_CHAR);
    if (n >= 0 && n < end) {
        end = n;
    }
    if (end < 0) {
        end = result.length();
    }
    n = result.indexOf(UNDERSCORE_CHAR);
    if (n < 0) {
      n = end;
    }
    for (; i < n; ++i) {
      UChar c = result.charAt(i);
      if (c >= 0x0041 && c <= 0x005a) {
        c += 0x20;
        result.setCharAt(i, c);
      }
    }
    for (n = end; i < n; ++i) {
      UChar c = result.charAt(i);
      if (c >= 0x0061 && c <= 0x007a) {
        c -= 0x20;
        result.setCharAt(i, c);
      }
    }
  }
  return result;

#if 0
    // This code does a proper full level 2 canonicalization of id.
    // It's nasty to go from UChar to char to char to UChar -- but
    // that's what you have to do to use the uloc_canonicalize
    // function on UnicodeStrings.

    // I ended up doing the alternate fix (see above) not for
    // performance reasons, although performance will certainly be
    // better, but because doing a full level 2 canonicalization
    // causes some tests to fail.  [alan 3.0]

    // TODO: Doug, you might want to revisit this...
    result.setToBogus();
    if (id != 0) {
        int32_t buflen = id->length() + 8; // space for NUL
        char* buf = (char*) uprv_malloc(buflen);
        char* canon = (buf == 0) ? 0 : (char*) uprv_malloc(buflen);
        if (buf != 0 && canon != 0) {
            U_ASSERT(id->extract(0, INT32_MAX, buf, buflen) < buflen);
            UErrorCode ec = U_ZERO_ERROR;
            uloc_canonicalize(buf, canon, buflen, &ec);
            if (U_SUCCESS(ec)) {
                result = UnicodeString(canon);
            }
        }
        uprv_free(buf);
        uprv_free(canon);
    }
    return result;
#endif
}

Locale&
LocaleUtility::initLocaleFromName(const UnicodeString& id, Locale& result)
{
    enum { BUFLEN = 128 }; // larger than ever needed

    if (id.isBogus() || id.length() >= BUFLEN) {
        result.setToBogus();
    } else {
        /*
         * We need to convert from a UnicodeString to char * in order to
         * create a Locale.
         *
         * Problem: Locale ID strings may contain '@' which is a variant
         * character and cannot be handled by invariant-character conversion.
         *
         * Hack: Since ICU code can handle locale IDs with multiple encodings
         * of '@' (at least for EBCDIC; it's not known to be a problem for
         * ASCII-based systems),
         * we use regular invariant-character conversion for everything else
         * and manually convert U+0040 into a compiler-char-constant '@'.
         * While this compilation-time constant may not match the runtime
         * encoding of '@', it should be one of the encodings which ICU
         * recognizes.
         *
         * There should be only at most one '@' in a locale ID.
         */
        char buffer[BUFLEN];
        int32_t prev, i;
        prev = 0;
        for(;;) {
            i = id.indexOf((UChar)0x40, prev);
            if(i < 0) {
                // no @ between prev and the rest of the string
                id.extract(prev, INT32_MAX, buffer + prev, BUFLEN - prev, US_INV);
                break; // done
            } else {
                // normal invariant-character conversion for text between @s
                id.extract(prev, i - prev, buffer + prev, BUFLEN - prev, US_INV);
                // manually "convert" U+0040 at id[i] into '@' at buffer[i]
                buffer[i] = '@';
                prev = i + 1;
            }
        }
        result = Locale::createFromName(buffer);
    }
    return result;
}

UnicodeString&
LocaleUtility::initNameFromLocale(const Locale& locale, UnicodeString& result)
{
    if (locale.isBogus()) {
        result.setToBogus();
    } else {
        result.append(UnicodeString(locale.getName(), -1, US_INV));
    }
    return result;
}

const Hashtable*
LocaleUtility::getAvailableLocaleNames(const UnicodeString& bundleID)
{
    // LocaleUtility_cache is a hash-of-hashes.  The top-level keys
    // are path strings ('bundleID') passed to
    // ures_openAvailableLocales.  The top-level values are
    // second-level hashes.  The second-level keys are result strings
    // from ures_openAvailableLocales.  The second-level values are
    // garbage ((void*)1 or other random pointer).

    UErrorCode status = U_ZERO_ERROR;
    umtx_initOnce(LocaleUtilityInitOnce, locale_utility_init, status);
    Hashtable *cache = LocaleUtility_cache;
    if (cache == NULL) {
        // Catastrophic failure.
        return NULL;
    }

    Hashtable* htp;
    umtx_lock(NULL);
    htp = (Hashtable*) cache->get(bundleID);
    umtx_unlock(NULL);

    if (htp == NULL) {
        htp = new Hashtable(status);
        if (htp && U_SUCCESS(status)) {
            CharString cbundleID;
            cbundleID.appendInvariantChars(bundleID, status);
            const char* path = cbundleID.isEmpty() ? NULL : cbundleID.data();
            icu::LocalUEnumerationPointer uenum(ures_openAvailableLocales(path, &status));
            for (;;) {
                const UChar* id = uenum_unext(uenum.getAlias(), NULL, &status);
                if (id == NULL) {
                    break;
                }
                htp->put(UnicodeString(id), (void*)htp, status);
            }
            if (U_FAILURE(status)) {
                delete htp;
                return NULL;
            }
            umtx_lock(NULL);
            Hashtable *t = static_cast<Hashtable *>(cache->get(bundleID));
            if (t != NULL) {
                // Another thread raced through this code, creating the cache entry first.
                // Discard ours and return theirs.
                umtx_unlock(NULL);
                delete htp;
                htp = t;
            } else {
                cache->put(bundleID, (void*)htp, status);
                umtx_unlock(NULL);
            }
        }
    }
    return htp;
}

UBool
LocaleUtility::isFallbackOf(const UnicodeString& root, const UnicodeString& child)
{
    return child.indexOf(root) == 0 &&
      (child.length() == root.length() ||
       child.charAt(root.length()) == UNDERSCORE_CHAR);
}

U_NAMESPACE_END

/* !UCONFIG_NO_SERVICE */
#endif