1/*
2 * Copyright (C) 2015 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 *      http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17#define LOG_TAG "Minikin"
18
19#include "FontLanguageListCache.h"
20
21#include <unicode/uloc.h>
22#include <unordered_set>
23
24#include <log/log.h>
25
26#include "FontLanguage.h"
27#include "MinikinInternal.h"
28
29namespace minikin {
30
31const uint32_t FontLanguageListCache::kEmptyListId;
32
33// Returns the text length of output.
34static size_t toLanguageTag(char* output, size_t outSize, const std::string& locale) {
35    output[0] = '\0';
36    if (locale.empty()) {
37        return 0;
38    }
39
40    size_t outLength = 0;
41    UErrorCode uErr = U_ZERO_ERROR;
42    outLength = uloc_canonicalize(locale.c_str(), output, outSize, &uErr);
43    if (U_FAILURE(uErr)) {
44        // unable to build a proper language identifier
45        ALOGD("uloc_canonicalize(\"%s\") failed: %s", locale.c_str(), u_errorName(uErr));
46        output[0] = '\0';
47        return 0;
48    }
49
50    // Preserve "und" and "und-****" since uloc_addLikelySubtags changes "und" to "en-Latn-US".
51    if (strncmp(output, "und", 3) == 0 &&
52        (outLength == 3 || (outLength == 8 && output[3]  == '_'))) {
53        return outLength;
54    }
55
56    char likelyChars[ULOC_FULLNAME_CAPACITY];
57    uErr = U_ZERO_ERROR;
58    uloc_addLikelySubtags(output, likelyChars, ULOC_FULLNAME_CAPACITY, &uErr);
59    if (U_FAILURE(uErr)) {
60        // unable to build a proper language identifier
61        ALOGD("uloc_addLikelySubtags(\"%s\") failed: %s", output, u_errorName(uErr));
62        output[0] = '\0';
63        return 0;
64    }
65
66    uErr = U_ZERO_ERROR;
67    outLength = uloc_toLanguageTag(likelyChars, output, outSize, FALSE, &uErr);
68    if (U_FAILURE(uErr)) {
69        // unable to build a proper language identifier
70        ALOGD("uloc_toLanguageTag(\"%s\") failed: %s", likelyChars, u_errorName(uErr));
71        output[0] = '\0';
72        return 0;
73    }
74#ifdef VERBOSE_DEBUG
75    ALOGD("ICU normalized '%s' to '%s'", locale.c_str(), output);
76#endif
77    return outLength;
78}
79
80static std::vector<FontLanguage> parseLanguageList(const std::string& input) {
81    std::vector<FontLanguage> result;
82    size_t currentIdx = 0;
83    size_t commaLoc = 0;
84    char langTag[ULOC_FULLNAME_CAPACITY];
85    std::unordered_set<uint64_t> seen;
86    std::string locale(input.size(), 0);
87
88    while ((commaLoc = input.find_first_of(',', currentIdx)) != std::string::npos) {
89        locale.assign(input, currentIdx, commaLoc - currentIdx);
90        currentIdx = commaLoc + 1;
91        size_t length = toLanguageTag(langTag, ULOC_FULLNAME_CAPACITY, locale);
92        FontLanguage lang(langTag, length);
93        uint64_t identifier = lang.getIdentifier();
94        if (!lang.isUnsupported() && seen.count(identifier) == 0) {
95            result.push_back(lang);
96            if (result.size() == FONT_LANGUAGES_LIMIT) {
97              break;
98            }
99            seen.insert(identifier);
100        }
101    }
102    if (result.size() < FONT_LANGUAGES_LIMIT) {
103      locale.assign(input, currentIdx, input.size() - currentIdx);
104      size_t length = toLanguageTag(langTag, ULOC_FULLNAME_CAPACITY, locale);
105      FontLanguage lang(langTag, length);
106      uint64_t identifier = lang.getIdentifier();
107      if (!lang.isUnsupported() && seen.count(identifier) == 0) {
108          result.push_back(lang);
109      }
110    }
111    return result;
112}
113
114// static
115uint32_t FontLanguageListCache::getId(const std::string& languages) {
116    FontLanguageListCache* inst = FontLanguageListCache::getInstance();
117    std::unordered_map<std::string, uint32_t>::const_iterator it =
118            inst->mLanguageListLookupTable.find(languages);
119    if (it != inst->mLanguageListLookupTable.end()) {
120        return it->second;
121    }
122
123    // Given language list is not in cache. Insert it and return newly assigned ID.
124    const uint32_t nextId = inst->mLanguageLists.size();
125    FontLanguages fontLanguages(parseLanguageList(languages));
126    if (fontLanguages.empty()) {
127        return kEmptyListId;
128    }
129    inst->mLanguageLists.push_back(std::move(fontLanguages));
130    inst->mLanguageListLookupTable.insert(std::make_pair(languages, nextId));
131    return nextId;
132}
133
134// static
135const FontLanguages& FontLanguageListCache::getById(uint32_t id) {
136    FontLanguageListCache* inst = FontLanguageListCache::getInstance();
137    LOG_ALWAYS_FATAL_IF(id >= inst->mLanguageLists.size(), "Lookup by unknown language list ID.");
138    return inst->mLanguageLists[id];
139}
140
141// static
142FontLanguageListCache* FontLanguageListCache::getInstance() {
143    assertMinikinLocked();
144    static FontLanguageListCache* instance = nullptr;
145    if (instance == nullptr) {
146        instance = new FontLanguageListCache();
147
148        // Insert an empty language list for mapping default language list to kEmptyListId.
149        // The default language list has only one FontLanguage and it is the unsupported language.
150        instance->mLanguageLists.push_back(FontLanguages());
151        instance->mLanguageListLookupTable.insert(std::make_pair("", kEmptyListId));
152    }
153    return instance;
154}
155
156}  // namespace minikin
157