std::string utf16ToUtf8(const StringPiece16& utf16) {
ssize_t utf8Length = utf16_to_utf8_length(utf16.data(), utf16.length());
if (utf8Length <= 0) {
return {};
}
std::string utf8;
utf8.resize(utf8Length);
utf16_to_utf8(utf16.data(), utf16.length(), &*utf8.begin());
return utf8;
}
/*
* Style parent's are a bit different. We accept the following formats:
*
* @[[*]package:][style/]<entry>
* ?[[*]package:]style/<entry>
* <[*]package>:[style/]<entry>
* [[*]package:style/]<entry>
*/
Maybe<Reference> parseStyleParentReference(const StringPiece16& str, std::string* outError) {
if (str.empty()) {
return {};
}
StringPiece16 name = str;
bool hasLeadingIdentifiers = false;
bool privateRef = false;
// Skip over these identifiers. A style's parent is a normal reference.
if (name.data()[0] == u'@' || name.data()[0] == u'?') {
hasLeadingIdentifiers = true;
name = name.substr(1, name.size() - 1);
}
if (name.data()[0] == u'*') {
privateRef = true;
name = name.substr(1, name.size() - 1);
}
ResourceNameRef ref;
ref.type = ResourceType::kStyle;
StringPiece16 typeStr;
extractResourceName(name, &ref.package, &typeStr, &ref.entry);
if (!typeStr.empty()) {
// If we have a type, make sure it is a Style.
const ResourceType* parsedType = parseResourceType(typeStr);
if (!parsedType || *parsedType != ResourceType::kStyle) {
std::stringstream err;
err << "invalid resource type '" << typeStr << "' for parent of style";
*outError = err.str();
return {};
}
}
if (!hasLeadingIdentifiers && ref.package.empty() && !typeStr.empty()) {
std::stringstream err;
err << "invalid parent reference '" << str << "'";
*outError = err.str();
return {};
}
Reference result(ref);
result.privateReference = privateRef;
return result;
}
std::unique_ptr<BinaryPrimitive> tryParseFloat(const StringPiece16& str) {
android::Res_value value;
if (!android::ResTable::stringToFloat(str.data(), str.size(), &value)) {
return {};
}
return util::make_unique<BinaryPrimitive>(value);
}
StringPiece16 trimWhitespace(const StringPiece16& str) {
if (str.size() == 0 || str.data() == nullptr) {
return str;
}
const char16_t* start = str.data();
const char16_t* end = str.data() + str.length();
while (start != end && util::isspace16(*start)) {
start++;
}
while (end != start && util::isspace16(*(end - 1))) {
end--;
}
return StringPiece16(start, end - start);
}
std::u16string Pseudolocalizer::text(const StringPiece16& text) {
std::u16string out;
size_t depth = mLastDepth;
size_t lastpos, pos;
const size_t length = text.size();
const char16_t* str = text.data();
bool escaped = false;
for (lastpos = pos = 0; pos < length; pos++) {
char16_t c = str[pos];
if (escaped) {
escaped = false;
continue;
}
if (c == '\'') {
escaped = true;
continue;
}
if (c == k_arg_start) {
depth++;
} else if (c == k_arg_end && depth) {
depth--;
}
if (mLastDepth != depth || pos == length - 1) {
bool pseudo = ((mLastDepth % 2) == 0);
size_t nextpos = pos;
if (!pseudo || depth == mLastDepth) {
nextpos++;
}
size_t size = nextpos - lastpos;
if (size) {
std::u16string chunk = text.substr(lastpos, size).toString();
if (pseudo) {
chunk = mImpl->text(chunk);
} else if (str[lastpos] == k_arg_start && str[nextpos - 1] == k_arg_end) {
chunk = mImpl->placeholder(chunk);
}
out.append(chunk);
}
if (pseudo && depth < mLastDepth) { // End of message
out.append(mImpl->end());
} else if (!pseudo && depth > mLastDepth) { // Start of message
out.append(mImpl->start());
}
lastpos = nextpos;
mLastDepth = depth;
}
}
return out;
}
/*
* The device ResXMLParser in libandroidfw differentiates between empty namespace and null
* namespace.
*/
TEST_F(XmlFlattenerTest, NoNamespaceIsNotTheSameAsEmptyNamespace) {
std::unique_ptr<xml::XmlResource> doc = test::buildXmlDom("<View package=\"android\"/>");
android::ResXMLTree tree;
ASSERT_TRUE(flatten(doc.get(), &tree));
while (tree.next() != android::ResXMLTree::START_TAG) {
ASSERT_NE(tree.getEventType(), android::ResXMLTree::BAD_DOCUMENT);
ASSERT_NE(tree.getEventType(), android::ResXMLTree::END_DOCUMENT);
}
const StringPiece16 kPackage = u"package";
EXPECT_GE(tree.indexOfAttribute(nullptr, 0, kPackage.data(), kPackage.size()), 0);
}
Maybe<std::u16string> getFullyQualifiedClassName(const StringPiece16& package,
const StringPiece16& className) {
if (className.empty()) {
return {};
}
if (util::isJavaClassName(className)) {
return className.toString();
}
if (package.empty()) {
return {};
}
std::u16string result(package.data(), package.size());
if (className.data()[0] != u'.') {
result += u'.';
}
result.append(className.data(), className.size());
if (!isJavaClassName(result)) {
return {};
}
return result;
}
bool parseResourceName(const StringPiece16& str, ResourceNameRef* outRef, bool* outPrivate) {
if (str.empty()) {
return false;
}
size_t offset = 0;
bool priv = false;
if (str.data()[0] == u'*') {
priv = true;
offset = 1;
}
StringPiece16 package;
StringPiece16 type;
StringPiece16 entry;
if (!extractResourceName(str.substr(offset, str.size() - offset), &package, &type, &entry)) {
return false;
}
const ResourceType* parsedType = parseResourceType(type);
if (!parsedType) {
return false;
}
if (entry.empty()) {
return false;
}
if (outRef) {
outRef->package = package;
outRef->type = *parsedType;
outRef->entry = entry;
}
if (outPrivate) {
*outPrivate = priv;
}
return true;
}
bool extractResourceName(const StringPiece16& str, StringPiece16* outPackage,
StringPiece16* outType, StringPiece16* outEntry) {
bool hasPackageSeparator = false;
bool hasTypeSeparator = false;
const char16_t* start = str.data();
const char16_t* end = start + str.size();
const char16_t* current = start;
while (current != end) {
if (outType->size() == 0 && *current == u'/') {
hasTypeSeparator = true;
outType->assign(start, current - start);
start = current + 1;
} else if (outPackage->size() == 0 && *current == u':') {
hasPackageSeparator = true;
outPackage->assign(start, current - start);
start = current + 1;
}
current++;
}
outEntry->assign(start, end - start);
return !(hasPackageSeparator && outPackage->empty()) && !(hasTypeSeparator && outType->empty());
}
std::u16string PseudoMethodBidi::text(const StringPiece16& source) {
const char16_t* s = source.data();
std::u16string result;
bool lastspace = true;
bool space = true;
for (size_t i = 0; i < source.size(); i++) {
char16_t c = s[i];
space = util::isspace16(c);
if (lastspace && !space) {
// Word start
result += k_rlm + k_rlo;
} else if (!lastspace && space) {
// Word end
result += k_pdf + k_rlm;
}
lastspace = space;
result.append(&c, 1);
}
if (!lastspace) {
// End of last word
result += k_pdf + k_rlm;
}
return result;
}
static bool lessThanEntry(const std::unique_ptr<ResourceEntry>& lhs, const StringPiece16& rhs) {
return lhs->name.compare(0, lhs->name.size(), rhs.data(), rhs.size()) < 0;
}
/**
* Converts characters so they look like they've been localized.
*
* Note: This leaves placeholder syntax untouched.
*/
std::u16string PseudoMethodAccent::text(const StringPiece16& source)
{
const char16_t* s = source.data();
std::u16string result;
const size_t I = source.size();
bool lastspace = true;
for (size_t i = 0; i < I; i++) {
char16_t c = s[i];
if (c == '%') {
// Placeholder syntax, no need to pseudolocalize
std::u16string chunk;
bool end = false;
chunk.append(&c, 1);
while (!end && i < I) {
++i;
c = s[i];
chunk.append(&c, 1);
if (isPossibleNormalPlaceholderEnd(c)) {
end = true;
} else if (c == 't') {
++i;
c = s[i];
chunk.append(&c, 1);
end = true;
}
}
// Treat chunk as a placeholder unless it ends with %.
result += ((c == '%') ? chunk : placeholder(chunk));
} else if (c == '<' || c == '&') {
// html syntax, no need to pseudolocalize
bool tag_closed = false;
while (!tag_closed && i < I) {
if (c == '&') {
std::u16string escapeText;
escapeText.append(&c, 1);
bool end = false;
size_t htmlCodePos = i;
while (!end && htmlCodePos < I) {
++htmlCodePos;
c = s[htmlCodePos];
escapeText.append(&c, 1);
// Valid html code
if (c == ';') {
end = true;
i = htmlCodePos;
}
// Wrong html code
else if (!((c == '#' ||
(c >= 'a' && c <= 'z') ||
(c >= 'A' && c <= 'Z') ||
(c >= '0' && c <= '9')))) {
end = true;
}
}
result += escapeText;
if (escapeText != u"<") {
tag_closed = true;
}
continue;
}
if (c == '>') {
tag_closed = true;
result.append(&c, 1);
continue;
}
result.append(&c, 1);
i++;
c = s[i];
}
} else {
// This is a pure text that should be pseudolocalized
const char16_t* p = pseudolocalizeChar(c);
if (p != nullptr) {
result += p;
} else {
bool space = util::isspace16(c);
if (lastspace && !space) {
mWordCount++;
}
lastspace = space;
result.append(&c, 1);
}
// Count only pseudolocalizable chars and delimiters
mLength++;
}
}
return result;
}