Skip to content

Commit

Permalink
optimize readString
Browse files Browse the repository at this point in the history
  • Loading branch information
@wenshao
wenshao committed Feb 12, 2025
1 parent 3151acc commit f6fd836
Show file tree
Hide file tree
Showing 10 changed files with 228 additions and 276 deletions.
15 changes: 10 additions & 5 deletions benchmark/src/main/java/com/alibaba/fastjson2/benchmark/BytesAsciiCheck.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -65,7 +65,7 @@ public void lambda(Blackhole bh) throws Throwable {
);
}

@Benchmark
// @Benchmark
public void direct(Blackhole bh) throws Throwable {
bh.consume(hasNegatives(bytes, 0, bytes.length));
}
Expand All @@ -76,11 +76,16 @@ public void isASCII(Blackhole bh) throws Throwable {
}

@Benchmark
public void isNonSlashASCII(Blackhole bh) throws Throwable {
bh.consume(com.alibaba.fastjson2.util.IOUtils.isNonSlashASCII(bytes, 0, bytes.length));
}

// @Benchmark
public void isLatin1(Blackhole bh) throws Throwable {
bh.consume(com.alibaba.fastjson2.util.IOUtils.isLatin1(chars, 0, chars.length));
}

@Benchmark
// @Benchmark
public void isASCIIJDK(Blackhole bh) throws Throwable {
bh.consume(com.alibaba.fastjson2.util.JDKUtils.PREDICATE_IS_ASCII.test(bytes));
}
Expand All @@ -90,17 +95,17 @@ public void indexOfSlash(Blackhole bh) throws Throwable {
bh.consume(com.alibaba.fastjson2.util.IOUtils.indexOfSlash(bytes, 0, bytes.length));
}

@Benchmark
// @Benchmark
public void indexOfSlashV(Blackhole bh) throws Throwable {
bh.consume(com.alibaba.fastjson2.util.IOUtils.indexOfSlashV(bytes, 0, bytes.length));
}

@Benchmark
// @Benchmark
public void indexOfChar(Blackhole bh) throws Throwable {
bh.consume(indexOfChar(bytes, '\'', 0, bytes.length));
}

@Benchmark
// @Benchmark
public void indexOfString(Blackhole bh) throws Throwable {
bh.consume(str.indexOf('\\'));
}
Expand Down
6 changes: 3 additions & 3 deletions benchmark/src/main/java/com/alibaba/fastjson2/benchmark/eishay/EishayParseString.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ public class EishayParseString {
}
}

@Benchmark
// @Benchmark
public void fastjson1(Blackhole bh) {
bh.consume(com.alibaba.fastjson.JSON.parseObject(str, MediaContent.class));
}
Expand All @@ -58,7 +58,7 @@ public void fastjson2Mixin(Blackhole bh) {
bh.consume(JSON.parseObject(str, MediaContent.class, JSONFactory.createReadContext(provider)));
}

@Benchmark
// @Benchmark
public void jackson(Blackhole bh) throws Exception {
bh.consume(mapper.readValue(str, MediaContent.class));
}
Expand All @@ -70,7 +70,7 @@ public void wastjson(Blackhole bh) throws Exception {
);
}

@Benchmark
// @Benchmark
public void gson(Blackhole bh) throws Exception {
bh.consume(
gson.fromJson(str, MediaContent.class)
Expand Down
2 changes: 1 addition & 1 deletion benchmark/src/main/java/com/alibaba/fastjson2/benchmark/eishay/EishayParseUTF8Bytes.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@ public class EishayParseUTF8Bytes {
}
}

@Benchmark
// @Benchmark
public void fastjson1(Blackhole bh) {
bh.consume(com.alibaba.fastjson.JSON.parseObject(utf8Bytes, MediaContent.class));
}
Expand Down
108 changes: 26 additions & 82 deletions core/src/main/java/com/alibaba/fastjson2/JSONReader.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3252,33 +3252,20 @@ public static JSONReader of(byte[] utf8Bytes, Context context) {
}

public static JSONReader of(char[] chars) {
return new JSONReaderUTF16(
createReadContext(),
return ofUTF16(
null,
chars,
0,
chars.length);
chars.length, createReadContext());
}

@Deprecated
public static JSONReader of(Context context, char[] chars) {
return new JSONReaderUTF16(
context,
null,
chars,
0,
chars.length
);
return ofUTF16(null, chars, 0, chars.length, context);
}

public static JSONReader of(char[] chars, Context context) {
return new JSONReaderUTF16(
context,
null,
chars,
0,
chars.length
);
return ofUTF16(null, chars, 0, chars.length, context);
}

public static JSONReader ofJSONB(byte[] jsonbBytes) {
Expand Down Expand Up @@ -3344,11 +3331,11 @@ public static JSONReader of(byte[] bytes, int offset, int length, Charset charse
Context context = JSONFactory.createReadContext();

if (charset == StandardCharsets.UTF_8) {
return new JSONReaderUTF8(context, bytes, offset, length);
return JSONReaderUTF8.of(bytes, offset, length, context);
}

if (charset == StandardCharsets.UTF_16) {
return new JSONReaderUTF16(context, bytes, offset, length);
return ofUTF16(bytes, offset, length, context);
}

if (charset == StandardCharsets.US_ASCII || charset == StandardCharsets.ISO_8859_1) {
Expand All @@ -3358,13 +3345,23 @@ public static JSONReader of(byte[] bytes, int offset, int length, Charset charse
throw new JSONException("not support charset " + charset);
}

private static JSONReader ofUTF16(byte[] bytes, int offset, int length, Context ctx) {
return new JSONReaderUTF16(ctx, bytes, offset, length);
}

private static JSONReader ofUTF16(String str, char[] chars, int offset, int length, Context ctx) {
return (str != null && str.indexOf('\\') == -1) || IOUtils.isNonSlashASCII(chars, offset, length)
? new JSONReaderUTF16NonSlash(ctx, str, chars, offset, length)
: new JSONReaderUTF16(ctx, str, chars, offset, length);
}

public static JSONReader of(byte[] bytes, int offset, int length, Charset charset, Context context) {
if (charset == StandardCharsets.UTF_8) {
return JSONReaderUTF8.of(bytes, offset, length, context);
}

if (charset == StandardCharsets.UTF_16) {
return new JSONReaderUTF16(context, bytes, offset, length);
return ofUTF16(bytes, offset, length, context);
}

if (charset == StandardCharsets.US_ASCII || charset == StandardCharsets.ISO_8859_1) {
Expand All @@ -3383,11 +3380,11 @@ public static JSONReader of(byte[] bytes, int offset, int length, Context contex
}

public static JSONReader of(char[] chars, int offset, int length) {
return new JSONReaderUTF16(createReadContext(), null, chars, offset, length);
return ofUTF16(null, chars, offset, length, createReadContext());
}

public static JSONReader of(char[] chars, int offset, int length, Context context) {
return new JSONReaderUTF16(context, null, chars, offset, length);
return ofUTF16(null, chars, offset, length, context);
}

public static JSONReader of(URL url, Context context) throws IOException {
Expand Down Expand Up @@ -3415,9 +3412,7 @@ public static JSONReader of(InputStream is, Charset charset, Context context) {
}

if (charset == StandardCharsets.US_ASCII) {
return JVM_VERSION == 8
? new JSONReaderASCIIJDK8(context, is)
: new JSONReaderASCII(context, is);
return new JSONReaderASCII(context, is);
}

return JSONReader.of(new InputStreamReader(is, charset), context);
Expand Down Expand Up @@ -3461,33 +3456,7 @@ public static JSONReader of(Context context, String str) {
}

public static JSONReader of(String str) {
if (str == null) {
throw new NullPointerException();
}

Context context = JSONFactory.createReadContext();
if (STRING_VALUE != null && STRING_CODER != null) {
try {
final int LATIN1 = 0;
int coder = STRING_CODER.applyAsInt(str);
if (coder == LATIN1) {
byte[] bytes = STRING_VALUE.apply(str);
if (PREDICATE_IS_ASCII != null && PREDICATE_IS_ASCII.test(bytes)) {
return new JSONReaderASCII(context, str, bytes, 0, bytes.length);
}
}
} catch (Exception e) {
throw new JSONException("unsafe get String.coder error");
}
}

final int length = str.length();
if (JVM_VERSION == 8) {
char[] chars = JDKUtils.getCharArray(str);
return new JSONReaderUTF16(context, str, chars, 0, length);
}

return new JSONReaderUTF16(context, str, 0, length);
return of(str, JSONFactory.createReadContext());
}

public static JSONReader of(String str, Context context) {
Expand All @@ -3501,7 +3470,7 @@ public static JSONReader of(String str, Context context) {
int coder = STRING_CODER.applyAsInt(str);
if (coder == LATIN1) {
byte[] bytes = STRING_VALUE.apply(str);
return new JSONReaderASCII(context, str, bytes, 0, bytes.length);
return JSONReaderASCII.of(context, str, bytes, 0, bytes.length);
}
} catch (Exception e) {
throw new JSONException("unsafe get String.coder error");
Expand All @@ -3516,36 +3485,11 @@ public static JSONReader of(String str, Context context) {
chars = str.toCharArray();
}

return new JSONReaderUTF16(context, str, chars, 0, length);
return ofUTF16(str, chars, 0, length, context);
}

public static JSONReader of(String str, int offset, int length) {
if (str == null) {
throw new NullPointerException();
}

Context context = JSONFactory.createReadContext();
if (STRING_VALUE != null) {
try {
final int LATIN1 = 0;
int coder = STRING_CODER.applyAsInt(str);
if (coder == LATIN1) {
byte[] bytes = STRING_VALUE.apply(str);
return new JSONReaderASCII(context, str, bytes, offset, length);
}
} catch (Exception e) {
throw new JSONException("unsafe get String.coder error");
}
}

char[] chars;
if (JVM_VERSION == 8) {
chars = JDKUtils.getCharArray(str);
} else {
chars = str.toCharArray();
}

return new JSONReaderUTF16(context, str, chars, offset, length);
return of(str, offset, length, JSONFactory.createReadContext());
}

public static JSONReader of(String str, int offset, int length, Context context) {
Expand All @@ -3559,7 +3503,7 @@ public static JSONReader of(String str, int offset, int length, Context context)
int coder = STRING_CODER.applyAsInt(str);
if (coder == LATIN1) {
byte[] bytes = STRING_VALUE.apply(str);
return new JSONReaderASCII(context, str, bytes, offset, length);
return JSONReaderASCII.of(context, str, bytes, offset, length);
}
} catch (Exception e) {
throw new JSONException("unsafe get String.coder error");
Expand All @@ -3573,7 +3517,7 @@ public static JSONReader of(String str, int offset, int length, Context context)
chars = str.toCharArray();
}

return new JSONReaderUTF16(context, str, chars, offset, length);
return ofUTF16(str, chars, offset, length, context);
}

final void bigInt(char[] chars, int off, int len) {
Expand Down
45 changes: 26 additions & 19 deletions core/src/main/java/com/alibaba/fastjson2/JSONReaderASCII.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -640,7 +640,7 @@ public final long readValueHashCode() {

long nameValue = 0;
for (int i = 0; offset < end; offset++, i++) {
ch = bytes[offset];
ch = bytes[offset] & 0xff;

if (ch == quote) {
if (i == 0) {
Expand Down Expand Up @@ -1474,23 +1474,7 @@ public String readString() {
offset = readEscaped(bytes, start, quote, buf);
str = new String(buf);
} else {
if (this.str != null) {
str = this.str.substring(start, offset);
} else if (STRING_CREATOR_JDK11 != null) {
str = STRING_CREATOR_JDK11.apply(Arrays.copyOfRange(bytes, start, offset), LATIN1);
} else if (ANDROID) {
str = getLatin1String(start, offset - start);
} else {
str = new String(bytes, start, offset - start, StandardCharsets.ISO_8859_1);
}
}

if ((context.features & Feature.TrimString.mask) != 0) {
str = str.trim();
}
// empty string to null
if (str.isEmpty() && (context.features & Feature.EmptyStringAsNull.mask) != 0) {
str = null;
str = subString(bytes, start, offset);
}

int ch = ++offset == end ? EOI : bytes[offset++];
Expand Down Expand Up @@ -1560,7 +1544,30 @@ protected final int readEscaped(byte[] bytes, int offset, byte quote, char[] buf
return offset;
}

protected final String subString(byte[] bytes, int start, int offset) {
String str;
if (this.str != null) {
str = this.str.substring(start, offset);
} else if (STRING_CREATOR_JDK11 != null) {
str = STRING_CREATOR_JDK11.apply(Arrays.copyOfRange(bytes, start, offset), LATIN1);
} else if (ANDROID) {
str = getLatin1String(start, offset - start);
} else {
str = new String(bytes, start, offset - start, StandardCharsets.ISO_8859_1);
}
if ((context.features & Feature.TrimString.mask) != 0) {
str = str.trim();
}
// empty string to null
if (str.isEmpty() && (context.features & Feature.EmptyStringAsNull.mask) != 0) {
str = null;
}
return str;
}

public static JSONReaderASCII of(Context ctx, String str, byte[] bytes, int offset, int length) {
return new JSONReaderASCII(ctx, str, bytes, offset, length);
return IOUtils.isNonSlashASCII(bytes, offset, length)
? new JSONReaderASCIINonSlash(ctx, str, bytes, offset, length)
: new JSONReaderASCII(ctx, str, bytes, offset, length);
}
}
Loading

0 comments on commit f6fd836

Please sign in to comment.