optimize readString

alibaba · Feb 12, 2025 · 71f84e5 · 71f84e5
1 parent 4099158
commit 71f84e5
Show file tree

Hide file tree

Showing 5 changed files with 35 additions and 51 deletions.
diff --git a/benchmark/src/test/java/com/alibaba/fastjson2/benchmark/eishay/EishayParseUTF8BytesTest.java b/benchmark/src/test/java/com/alibaba/fastjson2/benchmark/eishay/EishayParseUTF8BytesTest.java
@@ -6,7 +6,7 @@ public class EishayParseUTF8BytesTest {
     static final EishayParseUTF8Bytes benchmark = new EishayParseUTF8Bytes();
 
     public static void fastjson2() {
-        for (int j = 0; j < 10; j++) {
+        for (int j = 0; j < 5; j++) {
             long start = System.currentTimeMillis();
             for (int i = 0; i < 1000 * 1000; ++i) {
                 benchmark.fastjson2(BH);

diff --git a/core/src/main/java/com/alibaba/fastjson2/JSONReaderASCII.java b/core/src/main/java/com/alibaba/fastjson2/JSONReaderASCII.java
@@ -1474,7 +1474,16 @@ public String readString() {
                 offset = readEscaped(bytes, start, quote, buf);
                 str = new String(buf);
             } else {
-                str = subString(bytes, start, offset);
+                str = new String(bytes, start, offset - start, StandardCharsets.ISO_8859_1);
+            }
+
+            long features = context.features;
+            if ((features & Feature.TrimString.mask) != 0) {
+                str = str.trim();
+            }
+            // empty string to null
+            if (str.isEmpty() && (features & Feature.EmptyStringAsNull.mask) != 0) {
+                str = null;
             }
 
             int ch = ++offset == end ? EOI : bytes[offset++];
@@ -1544,27 +1553,6 @@ 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 IOUtils.isNonSlashASCII(bytes, offset, length)
                 ? new JSONReaderASCIINonSlash(ctx, str, bytes, offset, length)

diff --git a/core/src/main/java/com/alibaba/fastjson2/JSONReaderASCIINonSlash.java b/core/src/main/java/com/alibaba/fastjson2/JSONReaderASCIINonSlash.java
@@ -1,5 +1,9 @@
 package com.alibaba.fastjson2;
 
+import com.alibaba.fastjson2.util.IOUtils;
+
+import java.nio.charset.StandardCharsets;
+
 final class JSONReaderASCIINonSlash
         extends JSONReaderASCII {
     JSONReaderASCIINonSlash(Context ctx, String str, byte[] bytes, int offset, int length) {
@@ -11,19 +15,21 @@ public String readString() {
         char quote = this.ch;
         if (quote == '"' || quote == '\'') {
             final byte[] bytes = this.bytes;
-            int offset = this.offset;
             final int start = offset, end = this.end;
-            valueEscape = false;
-            for (; ; offset++) {
-                if (offset >= end) {
-                    throw new JSONException("invalid escape character EOI");
-                }
-                if (bytes[offset] == quote) {
-                    break;
-                }
+            int offset = IOUtils.indexOfQuote(bytes, quote, start, end);
+            if (offset == -1) {
+                throw new JSONException("invalid escape character EOI");
             }
 
-            String str = subString(bytes, start, offset);
+            String str = new String(bytes, start, offset - start, StandardCharsets.ISO_8859_1);
+            long features = context.features;
+            if ((features & Feature.TrimString.mask) != 0) {
+                str = str.trim();
+            }
+            // empty string to null
+            if (str.isEmpty() && (features & Feature.EmptyStringAsNull.mask) != 0) {
+                str = null;
+            }
 
             int ch = ++offset == end ? EOI : bytes[offset++];
             while (ch <= ' ' && (1L << ch & SPACE) != 0) {

diff --git a/core/src/main/java/com/alibaba/fastjson2/JSONReaderUTF16NonSlash.java b/core/src/main/java/com/alibaba/fastjson2/JSONReaderUTF16NonSlash.java
@@ -33,11 +33,12 @@ public String readString() {
                 str = new String(chars, start, offset - start);
             }
 
-            if ((context.features & Feature.TrimString.mask) != 0) {
+            long features = context.features;
+            if ((features & Feature.TrimString.mask) != 0) {
                 str = str.trim();
             }
             // empty string to null
-            if (str.isEmpty() && (context.features & Feature.EmptyStringAsNull.mask) != 0) {
+            if (str.isEmpty() && (features & Feature.EmptyStringAsNull.mask) != 0) {
                 str = null;
             }
 

diff --git a/core/src/main/java/com/alibaba/fastjson2/util/IOUtils.java b/core/src/main/java/com/alibaba/fastjson2/util/IOUtils.java
@@ -1709,14 +1709,6 @@ public static int digit2(byte[] bytes, int off) {
         return (d & 0xF) * 10 + (d >> 8);
     }
 
-    private static int digit2(int x) {
-        int d;
-        if ((((x & 0xF0F0) - 0x3030) | (((d = x & 0x0F0F) + 0x0606) & 0xF0F0)) != 0) {
-            return -1;
-        }
-        return (d & 0xF) * 10 + (d >> 8);
-    }
-
     public static int digit1(char[] chars, int off) {
         int d = UNSAFE.getByte(chars, ARRAY_CHAR_BASE_OFFSET + ((long) off << 1)) - '0';
         return d >= 0 && d <= 9 ? d : -1;
@@ -2075,19 +2067,16 @@ && notContains(d0 | (d1 << 8), 0x5C5C5C5C5C5C5C5CL)
     public static boolean isNonSlashASCII(byte[] bytes, int off, int len) {
         int end = off + len;
         int upperBound = off + (len & ~7);
-        long address = ARRAY_BYTE_BASE_OFFSET + off;
-        long d;
+        long addr = ARRAY_BYTE_BASE_OFFSET + off;
+        long d, x;
         while (off < upperBound
-                && ((d = UNSAFE.getLong(bytes, address)) & 0x8080808080808080L) == 0
-                && notContains(d, 0x5C5C5C5C5C5C5C5CL)
-        ) {
-            address += 8;
+                && (((d = UNSAFE.getLong(bytes, addr)) | (((x = d ^ 0x5C5C5C5C5C5C5C5CL) - 0x0101010101010101L) & ~x)) & 0x8080808080808080L) == 0) {
+            addr += 8;
             off += 8;
         }
-
         while (off++ < end) {
             byte b;
-            if (((b = UNSAFE.getByte(bytes, address++)) & 0x80) != 0 || b == '\\') {
+            if (((b = UNSAFE.getByte(bytes, addr++)) & 0x80) != 0 || b == '\\') {
                 return false;
             }
         }