core: option for control escaping of unicode characters (skylot#103)

README.md

@@ -42,20 +42,22 @@ Run **jadx** on itself:
 ```
 jadx[-gui] [options] <input file> (.dex, .apk, .jar or .class)
 options:
- -d, --output-dir        - output directory
- -j, --threads-count     - processing threads count
- -f, --fallback          - make simple dump (using goto instead of 'if', 'for', etc)
- -r, --no-res            - do not decode resources
- -s, --no-src            - do not decompile source code
-     --show-bad-code     - show inconsistent code (incorrectly decompiled)
-     --cfg               - save methods control flow graph to dot file
-     --raw-cfg           - save methods control flow graph (use raw instructions)
- -v, --verbose           - verbose output
-     --deobf             - activate deobfuscation
-     --deobf-min         - min length of name
-     --deobf-max         - max length of name
-     --deobf-rewrite-cfg - force to save deobfuscation map
- -h, --help              - print this help
+ -d, --output-dir           - output directory
+ -j, --threads-count        - processing threads count
+ -f, --fallback             - make simple dump (using goto instead of 'if', 'for', etc)
+ -r, --no-res               - do not decode resources
+ -s, --no-src               - do not decompile source code
+     --show-bad-code        - show inconsistent code (incorrectly decompiled)
+     --cfg                  - save methods control flow graph to dot file
+     --raw-cfg              - save methods control flow graph (use raw instructions)
+ -v, --verbose              - verbose output
+     --deobf                - activate deobfuscation
+     --deobf-min            - min length of name
+     --deobf-max            - max length of name
+     --deobf-rewrite-cfg    - force to save deobfuscation map
+     --deobf-use-sourcename - use source file name as class name alias
+     --escape-unicode       - escape non latin characters in strings (with \u)
+ -h, --help                 - print this help
 Example:
  jadx -d out classes.dex
 ```

jadx-cli/src/main/java/jadx/cli/JadxCLIArgs.java

@@ -1,5 +1,7 @@
 package jadx.cli;
 
+import ch.qos.logback.classic.spi.ILoggingEvent;
+import ch.qos.logback.core.Appender;
 import jadx.api.IJadxArgs;
 import jadx.api.JadxDecompiler;
 import jadx.core.utils.exceptions.JadxException;
@@ -67,6 +69,9 @@ public class JadxCLIArgs implements IJadxArgs {
 	@Parameter(names = {"--deobf-use-sourcename"}, description = "use source file name as class name alias")
 	protected boolean deobfuscationUseSourceNameAsAlias = false;
 
+	@Parameter(names = {"--escape-unicode"}, description = "escape non latin characters in strings (with \\u)")
+	protected boolean escapeUnicode = false;
+
 	@Parameter(names = {"-h", "--help"}, description = "print this help", help = true)
 	protected boolean printHelp = false;
 
@@ -117,7 +122,10 @@ private boolean process() {
 				ch.qos.logback.classic.Logger rootLogger =
 						(ch.qos.logback.classic.Logger) LoggerFactory.getLogger(Logger.ROOT_LOGGER_NAME);
 				// remove INFO ThresholdFilter
-				rootLogger.getAppender("STDOUT").clearAllFilters();
+				Appender<ILoggingEvent> appender = rootLogger.getAppender("STDOUT");
+				if (appender != null) {
+					appender.clearAllFilters();
+				}
 			}
 		} catch (JadxException e) {
 			System.err.println("ERROR: " + e.getMessage());
@@ -251,4 +259,9 @@ public boolean isDeobfuscationForceSave() {
 	public boolean useSourceNameAsClassAlias() {
 		return deobfuscationUseSourceNameAsAlias;
 	}
+
+	@Override
+	public boolean escapeUnicode() {
+		return escapeUnicode;
+	}
 }

jadx-core/src/main/java/jadx/api/IJadxArgs.java

@@ -30,4 +30,6 @@ public interface IJadxArgs {
 	boolean isDeobfuscationForceSave();
 
 	boolean useSourceNameAsClassAlias();
+
+	boolean escapeUnicode();
 }

jadx-core/src/main/java/jadx/api/JadxArgs.java

@@ -24,6 +24,8 @@ public class JadxArgs implements IJadxArgs {
 	private int deobfuscationMinLength = 0;
 	private int deobfuscationMaxLength = Integer.MAX_VALUE;
 
+	private boolean escapeUnicode = false;
+
 	@Override
 	public File getOutDir() {
 		return outDir;
@@ -149,4 +151,13 @@ public int getDeobfuscationMaxLength() {
 	public void setDeobfuscationMaxLength(int deobfuscationMaxLength) {
 		this.deobfuscationMaxLength = deobfuscationMaxLength;
 	}
+
+	@Override
+	public boolean escapeUnicode() {
+		return escapeUnicode;
+	}
+
+	public void setEscapeUnicode(boolean escapeUnicode) {
+		this.escapeUnicode = escapeUnicode;
+	}
 }

jadx-core/src/main/java/jadx/core/codegen/AnnotationGen.java

@@ -130,11 +130,11 @@ public void encodeValue(CodeWriter code, Object val) {
 			return;
 		}
 		if (val instanceof String) {
-			code.add(StringUtils.unescapeString((String) val));
+			code.add(getStringUtils().unescapeString((String) val));
 		} else if (val instanceof Integer) {
 			code.add(TypeGen.formatInteger((Integer) val));
 		} else if (val instanceof Character) {
-			code.add(StringUtils.unescapeChar((Character) val));
+			code.add(getStringUtils().unescapeChar((Character) val));
 		} else if (val instanceof Boolean) {
 			code.add(Boolean.TRUE.equals(val) ? "true" : "false");
 		} else if (val instanceof Float) {
@@ -172,4 +172,8 @@ public void encodeValue(CodeWriter code, Object val) {
 			throw new JadxRuntimeException("Can't decode value: " + val + " (" + val.getClass() + ")");
 		}
 	}
+
+	private StringUtils getStringUtils() {
+		return cls.dex().root().getStringUtils();
+	}
 }

jadx-core/src/main/java/jadx/core/codegen/ClassGen.java

@@ -348,7 +348,7 @@ private void addFields(CodeWriter code) throws CodegenException {
 			if (fv != null) {
 				code.add(" = ");
 				if (fv.getValue() == null) {
-					code.add(TypeGen.literalToString(0, f.getType()));
+					code.add(TypeGen.literalToString(0, f.getType(), cls));
 				} else {
 					if (fv.getValueType() == InitType.CONST) {
 						annotationGen.encodeValue(code, fv.getValue());

jadx-core/src/main/java/jadx/core/codegen/InsnGen.java

@@ -38,7 +38,6 @@
 import jadx.core.dex.nodes.RootNode;
 import jadx.core.utils.ErrorsCounter;
 import jadx.core.utils.RegionUtils;
-import jadx.core.utils.StringUtils;
 import jadx.core.utils.exceptions.CodegenException;
 import jadx.core.utils.exceptions.JadxRuntimeException;
 
@@ -130,8 +129,8 @@ public void declareVar(CodeWriter code, RegisterArg arg) {
 		code.add(mgen.getNameGen().assignArg(arg));
 	}
 
-	private static String lit(LiteralArg arg) {
-		return TypeGen.literalToString(arg.getLiteral(), arg.getType());
+	private String lit(LiteralArg arg) {
+		return TypeGen.literalToString(arg.getLiteral(), arg.getType(), mth);
 	}
 
 	private void instanceField(CodeWriter code, FieldInfo field, InsnArg arg) throws CodegenException {
@@ -236,7 +235,7 @@ private void makeInsnBody(CodeWriter code, InsnNode insn, Set<Flags> state) thro
 		switch (insn.getType()) {
 			case CONST_STR:
 				String str = ((ConstStringNode) insn).getString();
-				code.add(StringUtils.unescapeString(str));
+				code.add(mth.dex().root().getStringUtils().unescapeString(str));
 				break;
 
 			case CONST_CLASS:

jadx-core/src/main/java/jadx/core/codegen/RegionGen.java

@@ -255,7 +255,7 @@ private CodeWriter makeSwitch(SwitchRegion sw, CodeWriter code) throws CodegenEx
 						}
 					}
 				} else if (k instanceof Integer) {
-					code.add(TypeGen.literalToString((Integer) k, arg.getType()));
+					code.add(TypeGen.literalToString((Integer) k, arg.getType(), mth));
 				} else {
 					throw new JadxRuntimeException("Unexpected key in switch: " + (k != null ? k.getClass() : null));
 				}

jadx-core/src/main/java/jadx/core/codegen/TypeGen.java

@@ -1,7 +1,9 @@
 package jadx.core.codegen;
 
+import jadx.api.JadxArgs;
 import jadx.core.dex.instructions.args.ArgType;
 import jadx.core.dex.instructions.args.PrimitiveType;
+import jadx.core.dex.nodes.IDexNode;
 import jadx.core.utils.StringUtils;
 import jadx.core.utils.Utils;
 import jadx.core.utils.exceptions.JadxRuntimeException;
@@ -31,7 +33,16 @@ public static String signature(ArgType type) {
 	 *
 	 * @throws JadxRuntimeException for incorrect type or literal value
 	 */
+	public static String literalToString(long lit, ArgType type, IDexNode dexNode) {
+		return literalToString(lit, type, dexNode.root().getStringUtils());
+	}
+
+	@Deprecated
 	public static String literalToString(long lit, ArgType type) {
+		return literalToString(lit, type, new StringUtils(new JadxArgs()));
+	}
+
+	private static String literalToString(long lit, ArgType type, StringUtils stringUtils) {
 		if (type == null || !type.isTypeKnown()) {
 			String n = Long.toString(lit);
 			if (Math.abs(lit) > 100) {
@@ -46,7 +57,7 @@ public static String literalToString(long lit, ArgType type) {
 			case BOOLEAN:
 				return lit == 0 ? "false" : "true";
 			case CHAR:
-				return StringUtils.unescapeChar((char) lit);
+				return stringUtils.unescapeChar((char) lit);
 			case BYTE:
 				return formatByte((byte) lit);
 			case SHORT:

jadx-core/src/main/java/jadx/core/dex/nodes/ClassNode.java

@@ -43,7 +43,7 @@
 import com.android.dex.ClassDef;
 import com.android.dx.rop.code.AccessFlags;
 
-public class ClassNode extends LineAttrNode implements ILoadable {
+public class ClassNode extends LineAttrNode implements ILoadable, IDexNode {
 	private static final Logger LOG = LoggerFactory.getLogger(ClassNode.class);
 
 	private final DexNode dex;
@@ -472,10 +472,16 @@ public AccessInfo getAccessFlags() {
 		return accessFlags;
 	}
 
+	@Override
 	public DexNode dex() {
 		return dex;
 	}
 
+	@Override
+	public RootNode root() {
+		return dex.root();
+	}
+
 	public String getRawName() {
 		return clsInfo.getRawName();
 	}

jadx-core/src/main/java/jadx/core/dex/nodes/DexNode.java

@@ -28,7 +28,7 @@
 import com.android.dex.ProtoId;
 import com.android.dex.TypeList;
 
-public class DexNode {
+public class DexNode implements IDexNode {
 
 	public static final int NO_INDEX = -1;
 
@@ -210,10 +210,16 @@ public Section openSection(int offset) {
 		return dexBuf.open(offset);
 	}
 
+	@Override
 	public RootNode root() {
 		return root;
 	}
 
+	@Override
+	public DexNode dex() {
+		return this;
+	}
+
 	@Override
 	public String toString() {
 		return "DEX";

jadx-core/src/main/java/jadx/core/dex/nodes/IDexNode.java

@@ -0,0 +1,9 @@
+package jadx.core.dex.nodes;
+
+public interface IDexNode {
+
+	DexNode dex();
+
+	RootNode root();
+}
+

jadx-core/src/main/java/jadx/core/dex/nodes/MethodNode.java

@@ -44,7 +44,7 @@
 import com.android.dex.Code.CatchHandler;
 import com.android.dex.Code.Try;
 
-public class MethodNode extends LineAttrNode implements ILoadable {
+public class MethodNode extends LineAttrNode implements ILoadable, IDexNode {
 	private static final Logger LOG = LoggerFactory.getLogger(MethodNode.class);
 
 	private final MethodInfo mthInfo;
@@ -594,10 +594,16 @@ public void setRegion(Region region) {
 		this.region = region;
 	}
 
+	@Override
 	public DexNode dex() {
 		return parentClass.dex();
 	}
 
+	@Override
+	public RootNode root() {
+		return dex().root();
+	}
+
 	public MethodInfo getMethodInfo() {
 		return mthInfo;
 	}

jadx-core/src/main/java/jadx/core/dex/nodes/RootNode.java

@@ -7,6 +7,7 @@
 import jadx.core.clsp.ClspGraph;
 import jadx.core.dex.info.ClassInfo;
 import jadx.core.utils.ErrorsCounter;
+import jadx.core.utils.StringUtils;
 import jadx.core.utils.exceptions.DecodeException;
 import jadx.core.utils.exceptions.JadxException;
 import jadx.core.utils.files.DexFile;
@@ -31,6 +32,7 @@ public class RootNode {
 
 	private final ErrorsCounter errorsCounter = new ErrorsCounter();
 	private final IJadxArgs args;
+	private final StringUtils stringUtils;
 
 	private List<DexNode> dexNodes;
 	private Map<Integer, String> resourcesNames = new HashMap<Integer, String>();
@@ -41,6 +43,7 @@ public class RootNode {
 
 	public RootNode(IJadxArgs args) {
 		this.args = args;
+		this.stringUtils = new StringUtils(args);
 	}
 
 	public void load(List<InputFile> inputFiles) throws DecodeException {
@@ -194,4 +197,8 @@ public ClassNode getAppResClass() {
 	public IJadxArgs getArgs() {
 		return args;
 	}
+
+	public StringUtils getStringUtils() {
+		return stringUtils;
+	}
 }

jadx-core/src/main/java/jadx/core/utils/StringUtils.java

@@ -1,22 +1,29 @@
 package jadx.core.utils;
 
+import jadx.api.IJadxArgs;
+
 public class StringUtils {
 
-	private StringUtils() {
+	private final boolean escapeUnicode;
+
+	public StringUtils(IJadxArgs args) {
+		this.escapeUnicode = args.escapeUnicode();
 	}
 
-	public static String unescapeString(String str) {
+	public String unescapeString(String str) {
 		int len = str.length();
+		if (len == 0) {
+			return "\"\"";
+		}
 		StringBuilder res = new StringBuilder();
-
 		for (int i = 0; i < len; i++) {
 			int c = str.charAt(i) & 0xFFFF;
 			processChar(c, res);
 		}
 		return '"' + res.toString() + '"';
 	}
 
-	public static String unescapeChar(char ch) {
+	public String unescapeChar(char ch) {
 		if (ch == '\'') {
 			return "'\\\''";
 		}
@@ -25,7 +32,7 @@ public static String unescapeChar(char ch) {
 		return '\'' + res.toString() + '\'';
 	}
 
-	private static void processChar(int c, StringBuilder res) {
+	private void processChar(int c, StringBuilder res) {
 		switch (c) {
 			case '\n': res.append("\\n"); break;
 			case '\r': res.append("\\r"); break;
@@ -37,10 +44,10 @@ private static void processChar(int c, StringBuilder res) {
 			case '\\': res.append("\\\\"); break;
 
 			default:
-				if (32 <= c && c <= 126) {
-					res.append((char) c);
-				} else {
+				if (c < 32 || c >= 127 && escapeUnicode) {
 					res.append("\\u").append(String.format("%04x", c));
+				} else {
+					res.append((char) c);
 				}
 				break;
 		}