Merge pull request #768 from katzyn/DataUtils

Add DataUtils.parseChecksummedMap()

h2/src/main/org/h2/mvstore/Chunk.java

@@ -259,7 +259,7 @@ public class Chunk {
         DataUtils.appendMap(buff, "block", block);
         DataUtils.appendMap(buff, "version", version);
         byte[] bytes = buff.toString().getBytes(StandardCharsets.ISO_8859_1);
-        int checksum = DataUtils.getFletcher32(bytes, bytes.length);
+        int checksum = DataUtils.getFletcher32(bytes, 0, bytes.length);
         DataUtils.appendMap(buff, "fletcher", checksum);
         while (buff.length() < Chunk.FOOTER_LENGTH - 1) {
             buff.append(' ');

h2/src/main/org/h2/mvstore/DataUtils.java

@@ -10,6 +10,7 @@ import java.io.IOException;
 import java.io.OutputStream;
 import java.nio.ByteBuffer;
 import java.nio.channels.FileChannel;
+import java.nio.charset.StandardCharsets;
 import java.text.MessageFormat;
 import java.util.ArrayList;
 import java.util.Collections;
@@ -672,6 +673,49 @@ public final class DataUtils {
         return map;
     }
 
+    /**
+     * Parse a key-value pair list and checks its checksum.
+     *
+     * @param bytes encoded map
+     * @return the map without mapping for {@code "fletcher"}, or {@code null} if checksum is wrong
+     * @throws IllegalStateException if parsing failed
+     */
+    public static HashMap<String, String> parseChecksummedMap(byte[] bytes) {
+        int start = 0, end = bytes.length;
+        while (start < end && bytes[start] <= ' ') {
+            start++;
+        }
+        while (start < end && bytes[end - 1] <= ' ') {
+            end--;
+        }
+        String s = new String(bytes, start, end - start, StandardCharsets.ISO_8859_1);
+        HashMap<String, String> map = New.hashMap();
+        StringBuilder buff = new StringBuilder();
+        for (int i = 0, size = s.length(); i < size;) {
+            int startKey = i;
+            i = s.indexOf(':', i);
+            if (i < 0) {
+                throw DataUtils.newIllegalStateException(
+                        DataUtils.ERROR_FILE_CORRUPT, "Not a map: {0}", s);
+            }
+            if (i - startKey == 8 && s.regionMatches(startKey, "fletcher", 0, 8)) {
+                DataUtils.parseMapValue(buff, s, i + 1, size);
+                int check = (int) Long.parseLong(buff.toString(), 16);
+                if (check == DataUtils.getFletcher32(bytes, start, startKey - 1)) {
+                    return map;
+                }
+                // Corrupted map
+                return null;
+            }
+            String key = s.substring(startKey, i++);
+            i = DataUtils.parseMapValue(buff, s, i, size);
+            map.put(key, buff.toString());
+            buff.setLength(0);
+        }
+        // Corrupted map
+        return null;
+    }
+
     /**
      * Parse a name from key-value pair list.
      *
@@ -721,22 +765,23 @@ public final class DataUtils {
      * Calculate the Fletcher32 checksum.
      *
      * @param bytes the bytes
+     * @param offset initial offset
      * @param length the message length (if odd, 0 is appended)
      * @return the checksum
      */
-    public static int getFletcher32(byte[] bytes, int length) {
+    public static int getFletcher32(byte[] bytes, int offset, int length) {
         int s1 = 0xffff, s2 = 0xffff;
-        int i = 0, evenLength = length / 2 * 2;
-        while (i < evenLength) {
+        int i = offset, len = offset + (length & ~1);
+        while (i < len) {
             // reduce after 360 words (each word is two bytes)
-            for (int end = Math.min(i + 720, evenLength); i < end;) {
+            for (int end = Math.min(i + 720, len); i < end;) {
                 int x = ((bytes[i++] & 0xff) << 8) | (bytes[i++] & 0xff);
                 s2 += s1 += x;
             }
             s1 = (s1 & 0xffff) + (s1 >>> 16);
             s2 = (s2 & 0xffff) + (s2 >>> 16);
         }
-        if (i < length) {
+        if ((length & 1) != 0) {
             // odd length: append 0
             int x = (bytes[i] & 0xff) << 8;
             s2 += s1 += x;

h2/src/main/org/h2/mvstore/MVStore.java

@@ -575,9 +575,9 @@ public final class MVStore {
             fileHeaderBlocks.get(buff);
             // the following can fail for various reasons
             try {
-                String s = new String(buff, 0, BLOCK_SIZE,
-                        StandardCharsets.ISO_8859_1).trim();
-                HashMap<String, String> m = DataUtils.parseMap(s);
+                HashMap<String, String> m = DataUtils.parseChecksummedMap(buff);
+                if (m == null)
+                    continue;
                 int blockSize = DataUtils.readHexInt(
                         m, "blockSize", BLOCK_SIZE);
                 if (blockSize != BLOCK_SIZE) {
@@ -586,15 +586,6 @@ public final class MVStore {
                             "Block size {0} is currently not supported",
                             blockSize);
                 }
-                int check = DataUtils.readHexInt(m, "fletcher", 0);
-                m.remove("fletcher");
-                s = s.substring(0, s.lastIndexOf("fletcher") - 1);
-                byte[] bytes = s.getBytes(StandardCharsets.ISO_8859_1);
-                int checksum = DataUtils.getFletcher32(bytes,
-                        bytes.length);
-                if (check != checksum) {
-                    continue;
-                }
                 long version = DataUtils.readHexLong(m, "version", 0);
                 if (newest == null || version > newest.version) {
                     validStoreHeader = true;
@@ -804,14 +795,8 @@ public final class MVStore {
                     end - Chunk.FOOTER_LENGTH, Chunk.FOOTER_LENGTH);
             byte[] buff = new byte[Chunk.FOOTER_LENGTH];
             lastBlock.get(buff);
-            String s = new String(buff, StandardCharsets.ISO_8859_1).trim();
-            HashMap<String, String> m = DataUtils.parseMap(s);
-            int check = DataUtils.readHexInt(m, "fletcher", 0);
-            m.remove("fletcher");
-            s = s.substring(0, s.lastIndexOf("fletcher") - 1);
-            byte[] bytes = s.getBytes(StandardCharsets.ISO_8859_1);
-            int checksum = DataUtils.getFletcher32(bytes, bytes.length);
-            if (check == checksum) {
+            HashMap<String, String> m = DataUtils.parseChecksummedMap(buff);
+            if (m != null) {
                 int chunk = DataUtils.readHexInt(m, "chunk", 0);
                 Chunk c = new Chunk(chunk);
                 c.version = DataUtils.readHexLong(m, "version", 0);
@@ -833,7 +818,7 @@ public final class MVStore {
         }
         DataUtils.appendMap(buff, storeHeader);
         byte[] bytes = buff.toString().getBytes(StandardCharsets.ISO_8859_1);
-        int checksum = DataUtils.getFletcher32(bytes, bytes.length);
+        int checksum = DataUtils.getFletcher32(bytes, 0, bytes.length);
         DataUtils.appendMap(buff, "fletcher", checksum);
         buff.append("\n");
         bytes = buff.toString().getBytes(StandardCharsets.ISO_8859_1);

h2/src/test/org/h2/test/store/TestDataUtils.java

@@ -54,24 +54,24 @@ public class TestDataUtils extends TestBase {
     private void testFletcher() {
         byte[] data = new byte[10000];
         for (int i = 0; i < 10000; i += 1000) {
-            assertEquals(-1, DataUtils.getFletcher32(data, i));
+            assertEquals(-1, DataUtils.getFletcher32(data, 0, i));
         }
         Arrays.fill(data, (byte) 255);
         for (int i = 0; i < 10000; i += 1000) {
-            assertEquals(-1, DataUtils.getFletcher32(data, i));
+            assertEquals(-1, DataUtils.getFletcher32(data, 0, i));
         }
         for (int i = 0; i < 1000; i++) {
             for (int j = 0; j < 255; j++) {
                 Arrays.fill(data, 0, i, (byte) j);
                 data[i] = 0;
-                int a = DataUtils.getFletcher32(data, i);
+                int a = DataUtils.getFletcher32(data, 0, i);
                 if (i % 2 == 1) {
                     // add length: same as appending a 0
-                    int b = DataUtils.getFletcher32(data, i + 1);
+                    int b = DataUtils.getFletcher32(data, 0, i + 1);
                     assertEquals(a, b);
                 }
                 data[i] = 10;
-                int c = DataUtils.getFletcher32(data, i);
+                int c = DataUtils.getFletcher32(data, 0, i);
                 assertEquals(a, c);
             }
         }
@@ -79,16 +79,18 @@ public class TestDataUtils extends TestBase {
         for (int i = 1; i < 255; i++) {
             Arrays.fill(data, (byte) i);
             for (int j = 0; j < 10; j += 2) {
-                int x = DataUtils.getFletcher32(data, j);
+                int x = DataUtils.getFletcher32(data, 0, j);
                 assertTrue(x != last);
                 last = x;
             }
         }
         Arrays.fill(data, (byte) 10);
         assertEquals(0x1e1e1414,
-                DataUtils.getFletcher32(data, 10000));
+                DataUtils.getFletcher32(data, 0, 10000));
         assertEquals(0x1e3fa7ed,
-                DataUtils.getFletcher32("Fletcher32".getBytes(), 10));
+                DataUtils.getFletcher32("Fletcher32".getBytes(), 0, 10));
+        assertEquals(0x1e3fa7ed,
+                DataUtils.getFletcher32("XFletcher32".getBytes(), 1, 10));
     }
 
     private void testMap() {

h2/src/tools/org/h2/dev/fs/ArchiveToolStore.java

@@ -525,7 +525,7 @@ public class ArchiveToolStore {
         }
         key[0] = cs;
         key[1] = bucket;
-        key[2] = DataUtils.getFletcher32(buff, buff.length);
+        key[2] = DataUtils.getFletcher32(buff, 0, buff.length);
         return key;
     }