fix(java-input): support Java modified UTF-8 strings (PR #2654)

* [core] fix jadx.plugins.input.java.data.ConstPoolReader.parseString() error with Kotlin Annotation of byte array as a String to follow jvms-4.4.7 rules for encoding annotation strings in class files * move decode method into utility class, add test, fix code style --------- Co-authored-by: Skylot <118523+skylot@users.noreply.github.com>
2025-10-12 00:57:00 +02:00
parent 5f1985f281
commit 0f495afc99
3 changed files with 106 additions and 3 deletions
@@ -1,6 +1,5 @@
 package jadx.plugins.input.java.data;

-import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
 import java.util.List;

@@ -23,6 +22,7 @@ import jadx.plugins.input.java.data.attributes.types.JavaBootstrapMethodsAttr;
 import jadx.plugins.input.java.data.attributes.types.data.RawBootstrapMethod;
 import jadx.plugins.input.java.utils.DescriptorParser;
 import jadx.plugins.input.java.utils.JavaClassParseException;
+import jadx.plugins.input.java.utils.ModifiedUTF8Decoder;

 public class ConstPoolReader {
 	private final JavaClassReader clsReader;
@@ -235,8 +235,7 @@ public class ConstPoolReader {

 	@NotNull
 	private String parseString(byte[] bytes) {
-		// TODO: parse modified UTF-8
-		return new String(bytes, StandardCharsets.UTF_8);
+		return ModifiedUTF8Decoder.decodeString(bytes);
 	}

 	private String fixType(String clsName) {
@@ -0,0 +1,75 @@
+package jadx.plugins.input.java.utils;
+
+import java.nio.charset.StandardCharsets;
+
+public class ModifiedUTF8Decoder {
+
+	public static String decodeString(byte[] bytes) {
+		int len = bytes.length;
+		// quick check if all chars are 7-bit
+		boolean asciiStr = true;
+		for (byte b : bytes) {
+			if ((b & 0x80) != 0) {
+				asciiStr = false;
+				break;
+			}
+		}
+		if (asciiStr) {
+			return new String(bytes, StandardCharsets.US_ASCII);
+		}
+
+		// parse modified UTF-8 according jvms-4.4.7
+		StringBuilder sb = new StringBuilder();
+		for (int i = 0; i < len; i++) {
+			int x = bytes[i] & 0xff;
+			// 4.4 ascii characters 1-127 (0 is encoded as 0xc0 0x80)
+			if ((x & 0x80) == 0) {
+				// 1 byte 7-Bit ascii (Table 4.4./4.5)
+				sb.append((char) x);
+			} else {
+				if (i + 1 >= len) {
+					throw new JavaClassParseException("Inconsistent byte array structure: too short");
+				}
+				int y = bytes[i + 1] & 0xff;
+				// 0 is encoded as 0xc0 0x80 (jvms-4.4.7)
+				if (x == 0xc0 && y == 0x80) {
+					sb.appendCodePoint(0);
+					i++;
+				} else if ((x & 0xE0) == 0xC0 && (y & 0xC0) == 0x80) {
+					// 2 byte char (Table 4.8./4.9 )
+					sb.appendCodePoint(((x & 0x1f) << 6) + (y & 0x3f));
+					i++;
+				} else if (i + 2 < len) {
+					int z = bytes[i + 2] & 0xff;
+					if ((x & 0xF0) == 0xE0 && (y & 0xC0) == 0x80 && (z & 0xC0) == 0x80) {
+						// 3 byte char (Table 4.11/4.12)
+						sb.appendCodePoint(((x & 0xf) << 12) + ((y & 0x3f) << 6) + (z & 0x3f));
+						i += 2;
+					} else if (i + 5 < len
+							&& x == 0xED // u
+							&& (y & 0xF0) == 0xA0 // v
+							&& (bytes[i + 3] & 0xff) == 0xED // x
+							&& (bytes[i + 4] & 0xF0) == 0xA0 // y
+					) {
+						// 6 byte encoded Table 4.12.
+						int u = x; // 0
+						int v = y; // 1
+						int w = z; // 2
+						x = bytes[i + 3] & 0xff;
+						y = bytes[i + 4] & 0xff;
+						z = bytes[i + 5] & 0xff;
+						if (x == 0xED && (y & 0xF0) == 0xA0) {
+							sb.appendCodePoint(0x10000 + ((v & 0x0f) << 16) + ((w & 0x3f) << 10) + ((y & 0x0f) << 6) + (z & 0x3f));
+							i += 5;
+						} else {
+							throw new JavaClassParseException("Inconsistent byte array structure: invalid 6 bytes char");
+						}
+					} else {
+						throw new JavaClassParseException("Inconsistent byte array structure: unexpected char");
+					}
+				}
+			}
+		}
+		return sb.toString();
+	}
+}
@@ -0,0 +1,29 @@
+package jadx.plugins.input.java.utils;
+
+import org.junit.jupiter.api.Test;
+
+import static jadx.plugins.input.java.utils.ModifiedUTF8Decoder.decodeString;
+import static org.assertj.core.api.Assertions.assertThat;
+
+/*
+ * TODO: find a way to enter 6-bytes char decode branch
+ */
+class ModifiedUTF8DecoderTest {
+
+	@Test
+	public void test() {
+		String str = "aÆřᛒቶ北𝄠😀🨄𐆙";
+		byte[] mUTF8Bytes = new byte[] { 97, -61, -122, -59, -103, -31, -101, -110, -31, -119, -74, -17,
+				-91, -93, -19, -96, -76, -19, -76, -96, -19, -96, -67, -19, -72,
+				-128, -19, -96, -66, -19, -72, -124, -19, -96, -128, -19, -74, -103 };
+		assertThat(decodeString(mUTF8Bytes)).isEqualTo(str);
+	}
+
+	@Test
+	public void testASCIIOnly() {
+		String str = "Hello, world!";
+		byte[] mUTF8Bytes = new byte[] { 72, 101, 108, 108, 111, 44, 32, 119, 111, 114, 108, 100, 33 };
+		assertThat(decodeString(mUTF8Bytes)).isEqualTo(str);
+	}
+
+}