Reviewers: Toon Verwaest,
Description:
Improve String::WriteAscii and add PRESERVE_ASCII_NULL option.
The PRESERVE_ASCII_NULL option prevents WriteAscii from converting
\0 to white space.
[email protected]
BUG=v8:2252
TEST=test-api/StringWrite
Please review this at https://chromiumcodereview.appspot.com/10810056/
SVN Base: https://v8.googlecode.com/svn/branches/bleeding_edge
Affected files:
M include/v8.h
M src/api.cc
M test/cctest/test-api.cc
Index: include/v8.h
diff --git a/include/v8.h b/include/v8.h
index
34995d09f78723c933a4e9f6760473b31e1eaf45..5c2c8b6671827f56ddddb6ef4a71182b6d54c164
100644
--- a/include/v8.h
+++ b/include/v8.h
@@ -1064,7 +1064,8 @@ class String : public Primitive {
enum WriteOptions {
NO_OPTIONS = 0,
HINT_MANY_WRITES_EXPECTED = 1,
- NO_NULL_TERMINATION = 2
+ NO_NULL_TERMINATION = 2,
+ PRESERVE_ASCII_NULL = 4
};
// 16-bit character codes.
Index: src/api.cc
diff --git a/src/api.cc b/src/api.cc
index
e1d57e3cb2e0397dae0854cbf4ddbbe61f03acd5..cd0684b57a37f89e1de418fa64a812b25efad097
100644
--- a/src/api.cc
+++ b/src/api.cc
@@ -3847,6 +3847,9 @@ int String::WriteUtf8(char* buffer,
LOG_API(isolate, "String::WriteUtf8");
ENTER_V8(isolate);
i::Handle<i::String> str = Utils::OpenHandle(this);
+ if (options & HINT_MANY_WRITES_EXPECTED) {
+ FlattenString(str); // Flatten the string for efficiency.
+ }
int string_length = str->length();
if (str->IsAsciiRepresentation()) {
int len;
@@ -3903,11 +3906,7 @@ int String::WriteUtf8(char* buffer,
// Slow case.
i::StringInputBuffer& write_input_buffer =
*isolate->write_input_buffer();
isolate->string_tracker()->RecordWrite(str);
- if (options & HINT_MANY_WRITES_EXPECTED) {
- // Flatten the string for efficiency. This applies whether we are
- // using StringInputBuffer or Get(i) to access the characters.
- FlattenString(str);
- }
+
write_input_buffer.Reset(0, *str);
int len = str->length();
// Encode the first K - 3 bytes directly into the buffer since we
@@ -3949,8 +3948,9 @@ int String::WriteUtf8(char* buffer,
c,
unibrow::Utf16::kNoPreviousCharacter);
if (pos + written <= capacity) {
- for (int j = 0; j < written; j++)
+ for (int j = 0; j < written; j++) {
buffer[pos + j] = intermediate[j];
+ }
pos += written;
nchars++;
} else {
@@ -3963,8 +3963,9 @@ int String::WriteUtf8(char* buffer,
}
if (nchars_ref != NULL) *nchars_ref = nchars;
if (!(options & NO_NULL_TERMINATION) &&
- (i == len && (capacity == -1 || pos < capacity)))
+ (i == len && (capacity == -1 || pos < capacity))) {
buffer[pos++] = '\0';
+ }
return pos;
}
@@ -3977,28 +3978,45 @@ int String::WriteAscii(char* buffer,
if (IsDeadCheck(isolate, "v8::String::WriteAscii()")) return 0;
LOG_API(isolate, "String::WriteAscii");
ENTER_V8(isolate);
- i::StringInputBuffer& write_input_buffer =
*isolate->write_input_buffer();
ASSERT(start >= 0 && length >= -1);
i::Handle<i::String> str = Utils::OpenHandle(this);
isolate->string_tracker()->RecordWrite(str);
if (options & HINT_MANY_WRITES_EXPECTED) {
- // Flatten the string for efficiency. This applies whether we are
- // using StringInputBuffer or Get(i) to access the characters.
- str->TryFlatten();
+ FlattenString(str); // Flatten the string for efficiency.
+ }
+
+ if (str->IsAsciiRepresentation()) {
+ // WriteToFlat is faster than using the StringInputBuffer.
+ if (length == -1) length = str->length() + 1;
+ int len = i::Min(length, str->length() - start);
+ i::String::WriteToFlat(*str, buffer, start, start + len);
+ if (!(options & PRESERVE_ASCII_NULL)) {
+ for (int i = 0; i < len; i++) {
+ if (buffer[i] == '\0') buffer[i] = ' ';
+ }
+ }
+ if (!(options & NO_NULL_TERMINATION) && length > len) {
+ buffer[len] = '\0';
+ }
+ return len;
}
+
+ i::StringInputBuffer& write_input_buffer =
*isolate->write_input_buffer();
int end = length;
- if ( (length == -1) || (length > str->length() - start) )
+ if ((length == -1) || (length > str->length() - start)) {
end = str->length() - start;
+ }
if (end < 0) return 0;
write_input_buffer.Reset(start, *str);
int i;
for (i = 0; i < end; i++) {
char c = static_cast<char>(write_input_buffer.GetNext());
- if (c == '\0') c = ' ';
+ if (c == '\0' && !(options & PRESERVE_ASCII_NULL)) c = ' ';
buffer[i] = c;
}
- if (!(options & NO_NULL_TERMINATION) && (length == -1 || i < length))
+ if (!(options & NO_NULL_TERMINATION) && (length == -1 || i < length)) {
buffer[i] = '\0';
+ }
return i;
}
@@ -4017,7 +4035,7 @@ int String::Write(uint16_t* buffer,
if (options & HINT_MANY_WRITES_EXPECTED) {
// Flatten the string for efficiency. This applies whether we are
// using StringInputBuffer or Get(i) to access the characters.
- str->TryFlatten();
+ FlattenString(str);
}
int end = start + length;
if ((length == -1) || (length > str->length() - start) )
@@ -4203,8 +4221,9 @@ void v8::Object::SetPointerInInternalField(int index,
void* value) {
i::Handle<i::Foreign> foreign =
isolate->factory()->NewForeign(
reinterpret_cast<i::Address>(value), i::TENURED);
- if (!foreign.is_null())
- Utils::OpenHandle(this)->SetInternalField(index, *foreign);
+ if (!foreign.is_null()) {
+ Utils::OpenHandle(this)->SetInternalField(index, *foreign);
+ }
}
ASSERT_EQ(value, GetPointerFromInternalField(index));
}
Index: test/cctest/test-api.cc
diff --git a/test/cctest/test-api.cc b/test/cctest/test-api.cc
index
5a9df90e21fc8350c5e1d0dbc7bf5f8e8de9d1f4..71dd18cefe24edb29818558b1e8d99bcae89f35e
100644
--- a/test/cctest/test-api.cc
+++ b/test/cctest/test-api.cc
@@ -5570,6 +5570,7 @@ THREADED_TEST(StringWrite) {
v8::Handle<String> str = v8_str("abcde");
// abc<Icelandic eth><Unicode snowman>.
v8::Handle<String> str2 = v8_str("abc\303\260\342\230\203");
+ v8::Handle<String> str3 = v8::String::New("abc\0def", 7);
const int kStride = 4; // Must match stride in for loops in JS below.
CompileRun(
"var left = '';"
@@ -5780,6 +5781,29 @@ THREADED_TEST(StringWrite) {
CHECK_NE(0, strcmp(utf8buf, "abc\303\260\342\230\203"));
utf8buf[8] = '\0';
CHECK_EQ(0, strcmp(utf8buf, "abc\303\260\342\230\203"));
+
+ memset(utf8buf, 0x1, sizeof(utf8buf));
+ utf8buf[5] = 'X'; // Test that the sixth character is left untouched.
+ len = str->WriteUtf8(utf8buf, sizeof(utf8buf), &charlen,
+ String::NO_NULL_TERMINATION);
+ CHECK_EQ(5, len);
+ CHECK_EQ('X', utf8buf[5]);
+ CHECK_EQ(5, charlen);
+ CHECK_EQ(0, strncmp(utf8buf, "abcde", 5));
+ utf8buf[5] = '\0';
+ CHECK_EQ(0, strcmp(utf8buf, "abcde"));
+
+ memset(buf, 0x1, sizeof(buf));
+ len = str3->WriteAscii(buf);
+ CHECK_EQ(7, len);
+ CHECK_EQ(0, strcmp("abc def", buf));
+
+ memset(buf, 0x1, sizeof(buf));
+ len = str3->WriteAscii(buf, 0, -1, String::PRESERVE_ASCII_NULL);
+ CHECK_EQ(7, len);
+ CHECK_EQ(0, strcmp("abc", buf));
+ CHECK_EQ(0, buf[3]);
+ CHECK_EQ(0, strcmp("def", buf + 4));
}
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
