languages/dart-sdk
languages/dart-sdk
1
0
Fork 0
mirror of https://github.com/dart-lang/sdk synced 2024-10-14 18:31:32 +00:00
Code Issues Packages Projects Releases Wiki Activity

[dart2wasm] Avoid bounds check in some dart:convert functions

Browse source 
Also improves `codeUnitAt` bounds checks by using unsigned comparisons.

`package:characters` benchmarks before:

    Index Iteration: 23316 gc/ms
    String Iteration: 15329 cu/ms
    String Reversing: 6974.1 cu/ms
    String Replacing: 222.88 changes/ms

After:

    Index Iteration: 29729 gc/ms
    String Iteration: 17923 cu/ms
    String Reversing: 7270.9 cu/ms
    String Replacing: 236.28 changes/ms

Change-Id: I36832206b99ba9704e6e5863028d7eaa76412c3f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/337181
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
This commit is contained in:
Ömer Sinan Ağacan 2023-11-21 10:51:50 +00:00 committed by Commit Queue
parent bbdbd07287
commit dd8952f575
2 changed files with 61 additions and 28 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
sdk/lib/_internal/wasm/lib/convert_patch.dart
View file

@ -1688,18 +1688,18 @@ class _Utf8Decoder {
int localStart = start;
while (end - localStart > scanChunkSize) {
int localEnd = localStart + scanChunkSize;
size += _scan(bytes, localStart, localEnd, scanTable);
size += _scan(bytes, localStart, localEnd);
localStart = localEnd;
}
size += _scan(bytes, localStart, end, scanTable);
size += _scan(bytes, localStart, end);
return size;
}
int _scan(Uint8List bytes, int start, int end, String scanTable) {
int _scan(Uint8List bytes, int start, int end) {
int size = 0;
int flags = 0;
for (int i = start; i < end; i++) {
int t = scanTable.codeUnitAt(bytes[i]);
int t = scanTable.oneByteStringCodeUnitAtUnchecked(bytes[i]);
size += t & sizeMask;
flags |= t;
}
@ -1963,7 +1963,6 @@ class _Utf8Decoder {
String decode16(Uint8List bytes, int start, int end, int size) {
assert(start < end);
final String typeTable = _Utf8Decoder.typeTable;
final String transitionTable = _Utf8Decoder.transitionTable;
TwoByteString result = TwoByteString.withLength(size);
int i = start;
@ -1974,7 +1973,9 @@ class _Utf8Decoder {
// First byte
assert(!isErrorState(state));
final int byte = bytes[i++];
final int type = typeTable.codeUnitAt(byte) & typeMask;
final int type =
_Utf8Decoder.typeTable.oneByteStringCodeUnitAtUnchecked(byte) &
typeMask;
if (state == accept) {
char = byte & (shiftedByteMask >> type);
state = transitionTable.codeUnitAt(type);
@ -1985,7 +1986,9 @@ class _Utf8Decoder {
while (i < end) {
final int byte = bytes[i++];
final int type = typeTable.codeUnitAt(byte) & typeMask;
final int type =
_Utf8Decoder.typeTable.oneByteStringCodeUnitAtUnchecked(byte) &
typeMask;
if (state == accept) {
if (char >= 0x10000) {
assert(char < 0x110000);

72
sdk/lib/_internal/wasm/lib/string.dart
View file

@ -26,6 +26,13 @@ import "dart:typed_data" show Uint8List, Uint16List;
void writeIntoOneByteString(OneByteString s, int index, int codePoint) =>
s._setAt(index, codePoint);
/// Static function for `OneByteString._codeUnitAtUnchecked` to avoid making
/// `_codeUnitAtUnchecked` public, which would allow calling it in dynamic
/// invocations.
@pragma('wasm:prefer-inline')
int oneByteStringCodeUnitAtUnchecked(OneByteString s, int index) =>
s._codeUnitAtUnchecked(index);
/// Same as `writeIntoOneByteString`, but for `TwoByteString`.
@pragma('wasm:prefer-inline')
void writeIntoTwoByteString(TwoByteString s, int index, int codePoint) =>
@ -44,6 +51,12 @@ void copyRangeFromUint8ListToOneByteString(
}
}
extension OneByteStringUnsafeExtensions on String {
@pragma('wasm:prefer-inline')
int oneByteStringCodeUnitAtUnchecked(int index) =>
unsafeCast<OneByteString>(this)._codeUnitAtUnchecked(index);
}
const int _maxLatin1 = 0xff;
const int _maxUtf16 = 0xffff;
@ -1038,10 +1051,15 @@ final class OneByteString extends StringBase {
return stringFinalizeHash(hash);
}
@pragma('wasm:prefer-inline')
int _codeUnitAtUnchecked(int index) => _array.readUnsigned(index);
@override
int codeUnitAt(int index) {
RangeError.checkValueInInterval(index, 0, length - 1);
return _array.readUnsigned(index);
if (WasmI64.fromInt(length).leU(WasmI64.fromInt(index))) {
throw IndexError.withLength(index, length);
}
return _codeUnitAtUnchecked(index);
}
@override
@ -1065,7 +1083,7 @@ final class OneByteString extends StringBase {
int i = 0;
int start = 0;
for (i = 0; i < this.length; ++i) {
if (this.codeUnitAt(i) == charCode) {
if (this._codeUnitAtUnchecked(i) == charCode) {
parts.add(this._substringUnchecked(start, i));
start = i + 1;
}
@ -1117,7 +1135,7 @@ final class OneByteString extends StringBase {
return -1;
}
for (int i = start; i < len; i++) {
if (this.codeUnitAt(i) == patternCu0) {
if (this._codeUnitAtUnchecked(i) == patternCu0) {
return i;
}
}
@ -1134,7 +1152,7 @@ final class OneByteString extends StringBase {
return false;
}
for (int i = start; i < len; i++) {
if (this.codeUnitAt(i) == patternCu0) {
if (this._codeUnitAtUnchecked(i) == patternCu0) {
return true;
}
}
@ -1160,7 +1178,7 @@ final class OneByteString extends StringBase {
if (padding is! OneByteString) {
return super.padLeft(width, padding);
}
int length = this.length;
final length = this.length;
int delta = width - length;
if (delta <= 0) return this;
int padLength = padding.length;
@ -1180,7 +1198,7 @@ final class OneByteString extends StringBase {
}
}
for (int i = 0; i < length; i++) {
result._setAt(index++, this.codeUnitAt(i));
result._setAt(index++, this._codeUnitAtUnchecked(i));
}
return result;
}
@ -1189,7 +1207,7 @@ final class OneByteString extends StringBase {
if (padding is! OneByteString) {
return super.padRight(width, padding);
}
int length = this.length;
final length = this.length;
int delta = width - length;
if (delta <= 0) return this;
int padLength = padding.length;
@ -1197,7 +1215,7 @@ final class OneByteString extends StringBase {
OneByteString result = OneByteString.withLength(resultLength);
int index = 0;
for (int i = 0; i < length; i++) {
result._setAt(index++, this.codeUnitAt(i));
result._setAt(index++, this._codeUnitAtUnchecked(i));
}
if (padLength == 1) {
int padChar = padding.codeUnitAt(0);
@ -1242,7 +1260,7 @@ final class OneByteString extends StringBase {
// The German "sharp s" \xdf (ß) should be converted into two characters (SS),
// and is also marked with 0x00.
// Conversion to lower case performed by subtracting 0x20.
static const _UC_TABLE =
static const String _UC_TABLE =
"\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
@ -1262,15 +1280,19 @@ final class OneByteString extends StringBase {
String toLowerCase() {
for (int i = 0; i < this.length; i++) {
final c = this.codeUnitAt(i);
if (c == _LC_TABLE.codeUnitAt(c)) continue;
final c = this._codeUnitAtUnchecked(i);
if (c == unsafeCast<OneByteString>(_LC_TABLE)._codeUnitAtUnchecked(c))
continue;
// Upper-case character found.
final result = OneByteString.withLength(this.length);
for (int j = 0; j < i; j++) {
result._setAt(j, this.codeUnitAt(j));
result._setAt(j, this._codeUnitAtUnchecked(j));
}
for (int j = i; j < this.length; j++) {
result._setAt(j, _LC_TABLE.codeUnitAt(this.codeUnitAt(j)));
result._setAt(
j,
unsafeCast<OneByteString>(_LC_TABLE)
._codeUnitAtUnchecked(this._codeUnitAtUnchecked(j)));
}
return result;
}
@ -1279,15 +1301,18 @@ final class OneByteString extends StringBase {
String toUpperCase() {
for (int i = 0; i < this.length; i++) {
final c = this.codeUnitAt(i);
final c = this._codeUnitAtUnchecked(i);
// Continue loop if character is unchanged by upper-case conversion.
if (c == _UC_TABLE.codeUnitAt(c)) continue;
if (c == unsafeCast<OneByteString>(_UC_TABLE)._codeUnitAtUnchecked(c))
continue;
// Check rest of string for characters that do not convert to
// single-characters in the Latin-1 range.
for (int j = i; j < this.length; j++) {
final c = this.codeUnitAt(j);
if ((_UC_TABLE.codeUnitAt(c) == 0x00) && (c != 0x00)) {
final c = this._codeUnitAtUnchecked(j);
if ((unsafeCast<OneByteString>(_UC_TABLE)._codeUnitAtUnchecked(c) ==
0x00) &&
(c != 0x00)) {
// We use the 0x00 value for characters other than the null character,
// that don't convert to a single Latin-1 character when upper-cased.
// In that case, call the generic super-class method.
@ -1298,10 +1323,13 @@ final class OneByteString extends StringBase {
// characters.
final result = OneByteString.withLength(this.length);
for (int j = 0; j < i; j++) {
result._setAt(j, this.codeUnitAt(j));
result._setAt(j, this._codeUnitAtUnchecked(j));
}
for (int j = i; j < this.length; j++) {
result._setAt(j, _UC_TABLE.codeUnitAt(this.codeUnitAt(j)));
result._setAt(
j,
unsafeCast<OneByteString>(_UC_TABLE)
._codeUnitAtUnchecked(this._codeUnitAtUnchecked(j)));
}
return result;
}
@ -1371,7 +1399,9 @@ final class TwoByteString extends StringBase {
@override
int codeUnitAt(int index) {
RangeError.checkValueInInterval(index, 0, length - 1);
if (WasmI64.fromInt(length).leU(WasmI64.fromInt(index))) {
throw IndexError.withLength(index, length);
}
return _array.readUnsigned(index);
}