We Keep Coding

Recover unaltered java from APK

I recently lost my android game code. When I tried to retrieve it from APK in my test device, I was able to recover all assets, xml and few java files - less complex ones (using dex2java converter). However original code of the more complex java files (like renderer) couldnt be recovered properly. Lots of modifications had been made to the code (like additional while, break, continue, labels).
Please suggest how the actual JAVA code(unaltered) can be recovered from the APK.
Thanks

The simple answer is that it cannot.
Source code is not included into the compiled version. The compiled classes only contain JVM instructions. From these instructions, you can often reconstruct the logic, but rarely if ever the original source.

Decompiler Bytecode and Obfuscators

Can we completely reverse-engineer the source code from java bytecode ? Why this feature is allowed in Java and How successful are java decompilers against obfuscators.?

I know this question is old but I kept looking for a reliable answer until I found nothing.
So in this post I summarize some of my effort to obfuscate a J2EE JAR.
It seems , that by year 2014 (time of writing) there are not many options out there.
If you read this review later then things may have changed or fixed.
When I think why , I start to sense that the whole obfuscation effort gives a false sense of security. Don't get me wrong. It does add a level of security, but not as much as I would hope.
I will try to give a preview of what I found to explain myself. My recommendation are personal , others may disagree with it.
So to begin with: obfuscation in Java is the process of taking bytecode and making it less readable (using a decompiler of course) while maintaining its original functionality.What can we do, Java ,working as an interperter, must keep its bytecode exposed. You run the obfuscator as a measure of security in case the class file falls into the wrong hands. The result of the obfuscation is a reverse-mapping files and a JAR with the obfuscated classes. The reverse mapping file is used of-course to perform stack trace reading (a.k.a re-trace) or to revert the bytecode to its original shape. The runtime performance hit of an obfuscated class should not pass the 10% (but this really depends on what you do in your code).
But there is a big “but” . Obfuscation will scramble your code but it won’t make it hacker-proof. Bare in mind you only buy time and a determined hacker will find a way to reverse engineer your bytecode into its pure algorithm.
IMHO: the best way to hide a sensitive piece of code is to drown it in some huge pile of meaningless code.
Some of the hackers will try to modify your bytecode (by code injection) to help them achieve their goals. Some obfuscators offer additional level of JAR hardening , making it harder to modify.
De-obfuscators and de-compilers: my favourite Java decompiler is JD-GUI . However, when it comes to de-obfuscators I found the market pretty empty. Most of the tools ask you for a hint (what obfuscation tool was used to encrypt the source JAR) , yet none of them really deliver results (some of them even crash when trying to de-cipher the JAR). They are open source projects with low maintenance. I couldn’t even find a paid application to do a decent de-obfuscation. so enlighten me if you know something.
Free solutions
There are open source , free obfuscators which usually simply rename the classes/methods names, making it one letter method (i.e. from printUsage(String params) to a(String p) ).
They might ,as hinted here , even strip debugging information to make it a bit more difficult. (debugging information is kept at the end of every Java method bytecode and contains: line numbers, variables names ,etc.).
Its a nice effort , but an experience Java developer with a debugger can very easily deduce the purpose of each parameter while doing few live runs.
One of the nice open source obfuscators is ProGuard but there are several more tools.
Nevertheless , if you truly security fanatic you will probably want something stronger. Stronger demands more features (and more money) which leads us to the next bullet:
Paid solutions
While free products may only change classes method names , paid product will usually offer more features:
code/flow obfuscation: this will change the method code and inject empty loops/dead code/confusing switch tables and alike. Some of them may even scramble the exception table content. the obfuscation strength usually determine the output size.
Note: regarding code obfuscation: I deliberately avoided the details in my review. Some of the bytecode I saw and analyzed expose their obfuscation methods, and I wish to protect their IP. I do have an opinion about who uses better algorithms. contact me if you wish to know.
classes/method renaming : well this is the obvious , we discussed it in the free obfuscation. Some of the product will rename the class name and then recursively search for reflection usage of that class and fix those too. Paid products may even rename Spring /Wink configuration files for the same purpose (renaming in reflection).
String encryption: for every string “like this” in the code, it will encrypt it to some level and keep the key somewhere (in the class constant table/static blocks/a new method or any other mean).
debug information : stripping parts or scrambling.many of them will remove the line numbers info.
class
hardening: all kinds of methods like injecting some signing scheme into the beginning of the class/method, making sure an outsider won’t be able to easily modify the JAR and run it. Less important for Android or applets as most of them are digitally signed anyhow. some will combine hardening with water-marking to track pirated copies. But we all know anti-pirating methods by software are doomed to be hacked. Game industry suffered from it for decades until network based subscriptions arrived.
Since most products here deal with Java , some of them provides Android integration. It means it will not only obfuscate the Java (dalvik) code , but also manipulates the Android's manifest file and resources. Some offer anti debugging: remove the debug flag in android apps.
Nice GUI app to configure the various options and maybe do a re-trance on a given log file. The UI is usually used to generate a config file. with such file you can later re-play the obfuscation many times, even from command line.
Incremental build support - this is useful for large groups who release product updates/fixes frequently. You can tell the obfuscator to preserve old “obfuscation” result and randomly obfuscate only “new” code flows. this way you can be sure minimal impact on your methods signature. Without this flag , each obfuscation cycle on a JAR would yield a different output as most good tools use some level of randomness in their algorithms.
CLI and distributed builds. When you work alone then running an obfuscator is not a big issue. you need to configure the obfuscator to your relevant options and run it.However, in enterprise , when integrating obfuscator into the the build script things are a bit different. There is another level of complexity: build engine tasks (like ant/maven) and license management. The good news that all obfuscator I tested have command line API. In distributed build environment there are cluster/pool of build machines to support concurrent demand of builds. The cluster is dynamic and virtual, machines are going up or down, depending on various conditions. Some obfuscation products are based on cpuID license file or hostname. This can create quite a challenge for the build teams to integrate. Some prefer a local floating license server. Some may require public license server (but then: not all build farms have access to the public internet). Some offer multi-site license (which in my opinion is the best).
Some offer code optimizations - algebric equivalence and dropping of dead code. Its nice, but I believe that today's JDK do good job in optimizing bytecode. Its true that dead code makes you downloadable bigger, but with today's bandwidth its less than a problem. I also want to believe that in software today 20:80 thumb rule still applies. in any application 20% is probably a dead code anyway.
So who are the players I tried ?
KlassMaster by Zelix.com - one of the oldest in the industry. Yet they deliver a solid product with 3-4 releases per year. This been going for decades (since 1997). Zelix provides good email support and answered all my emails in a timely manner. They have a nice GUI client to either obfuscate a JAR or create a config file for future obfuscation. It simple and slick. nothing special here. They provided simple to read on-line documentation for all their flags. they support both “exclude” and “include” regular expressions for what the engine should obfuscate. The thing I liked about their process most is that it also adds “noise” to the exception table. It makes it a bit more confusing regarding the method exception handling. Their flow obfuscator strength is quite good and can be configured between 3 possible levels (light,medium and aggressive). Another feature I liked is the fine tuning they provide for debug info stripping (online line numbers, or online local variables or both). Klass Master doesn’t provide any
dedicated Android flags or anti-tamper methods. Their licensing model is quite simple: a text file to be placed near the KlassMaster main JAR. They also support incremental obfuscation.
JFuscator from secureTeam.net : While secureTeam also has a .Net tool , I focus on their Java tool capabilities. Their (Swing based) GUI tool seems nice but it crash when trying the simplest obfuscation task. the error was always the same: Error reading '/opt/sun-jdk1.7.0_55/jre\lib\rt.jar'. Reason: ''/opt/sun-jdk1.7.0_55/jre\lib\rt.jar': no such file or directory' . Now of course I have my Java installed in /opt/sun-jdk1.7.0_55/jre. You can image that they simply didn’t expect linux back slash structure. I contacted secureTeam.net support by email with the minor “path” problem. They asked if I am a linux user and after I replied I am , they never answered my email. I also tried their web site on-line chat : no response. So there I stopped testing. Without further results, I couldn’t examine the obfuscated bytecode quality. From their web site it seems they have anti-tamper method , String manipulation, method renaming and few other features.
GuartIt4J (by Arxan.com) : Arxan is fairly solid player in the mobile environment and as such they offer Android obfuscator which of course works well for Java. They have one of the most flexible engines.They provide code obfuscation,string encryption and alike You can define the complexity of code obfuscation. it is simply an integer. the higher - the longer your method turns out. ofcourse, you must be carefull not to exceed the JVM 64KB limit per class… As I said before one of the best strategies to hide a sensitive code is not to encrypt it , but to inject it into huge pile of garbage. This is exactly what GuardIt does. It can also explode in the same way the methods exception table. I managed to create a method with 100 exceptions in its exception table (pre-obfuscator it was 5). what they miss: their re-trace program is not part of the supplied main JAR. Nevertheless, they were kind enough to send me a sample Java program that performs re-trace given the reverse mapping file and the log. They don’t support incremental obfuscation and no flexibility regarding debug information. Debug information stripping is either all or nothing. watching the output JAR you will tons of conditions and jumps that were injected. Bare in mind , exploding the class size has its performance hit. In some methods I measured almost 50% performance hit when applying long obfuscation (no I/O in those methods). so extrapolating the code comes with a price.(from a 400 opcodes - I went up to 2200 opcodes after obfuscation). JD-GUI , my de-compiler failed to open such classes and crashed (IndexOutOfBoundException). They also supply complete class encryption . Meaning the class is encrypted with some symetrical key which demands a special (or custom written) class loader to open it in memory. This is an anti-tamper mechanism as well as hiding code. Just remember that a JVM can’t run that class without the class loader help. Its a nice feature, but the secret key and the bootstrap loader JAR are probably there. If he got the encrypted JAR the hacker will eventually get his hands and decrypt the classes. Yet this another level of obstacle the common hacker will need to pass. What I didn’t like here is the license file policy: is bounded to CPUid or need to install a floating license server.
SecureIt (by Allatori.com) : SecureIt offers all the general code obfuscation, string encryption ,renaming and such. On top of the standard obfuscation methods they also offer some kind of water-marking which is an anti-tamper/pirating method. They support Android and JavaME (who uses ME these days?!). They support incremental obfuscation. The one thing to note about configuring SecureIt: it is all command line. No GUI tool this time. Personally , I don’t mind command line tools as long as they come with good documentation. Luckily they have a very good documentation and a rich API with many flags to tune if you wish. you can re-trace with they tool (also a command line ) . They can’t obfuscate the exception table. I didn’t check their licensing mechanism.
DashO (by Preemptive.com) : DashO obfuscator will be remembered probably as the best UI tool you can get (to create your configuration). Like SecureIt they lake the exception table obfuscation but they have all the rest of the required features (as well as CLI, Spring framework and gradle/ant integration, and even an eclipse plugin) . Well, they do document a try-catch obfuscator (which is same as exception table obfuscator) , but it is only a recommendation to the engine. When I tried it , it had nil effect on the exception table. As I said , the GUI tool is superb and has a re-trace embedded into it. they also offer some kind of application signing and water-marking as an anti-tamper/pirating mechanism. DashO provides superb Android integration and also combine in their product a door for analytics uploads. You can actually track your application. Injecting crash log uploaders and reporting code to your JAR. Nevertheless that’s not the scope of obfuscation - that’s a whole different code injection product. They have a very good support. both online and by phone. Their licensing scheme is based on monthly subscription or one time purchase payment. A bit different than others. They are using a floating license server to support large environments.
I hope this helps a bit..

Can we completely reverse-engineer the source code from java bytecode ?
Not completely, because some aspects of source code, such as whitespace, local variable names, and comments, are not preserved in bytecode. Otherwise, yes -- while you can't get the exact same source code out, you can almost always get something that can at least be compiled back to the same bytecode.
Why this feature is allowed in Java
It's not so much "allowed" as it is "not prevented". And it's not prevented because doing so is impossible -- the code must be runnable to be useful; if the code is runnable, then it is analyzable; if it is analyzable, then with sufficient analysis it can be converted back to source.
How successful are java decompilers against obfuscators?
Not very. Most obfuscators I've seen (esp. ProGuard) are primarily effective in removing meaningful function and class names; obfuscating the logic itself is not typically attempted.

you can get source code from binary these days. Although the source code obtained by Java's bytecode is more readable, obfuscating will make it slightly unreadable. Its not that only Java can be reverse engineered to code. Even C/C++ these days (with Hexrays plugin for IDA Pro) can be decompiled to source. Obfuscaters will make it hard to read but not impossible. There is nothing that can save your program from an intelligent and capable reverse engineer. :).
Good luck.

Can we completely reverse-engineer the source code from java bytecode
?
The java class file is based on a spec so anyone can read into it. A tool like JD-GUI will tear into your source code easily. It is not a 'feature' per se. While 100% reverse-engineering is not possible, most of your code can be reverse engineered.
How successful are java decompilers against obfuscators?
Depends. The point of the obfuscator is to remove any meaningful names and try to introduce confusion in the code without impacting performance. Most developers are great at obfuscating code themselves :) Pro-guard is pretty good at obfuscation.

Does compiling Java to native code protect the source code?

Now, I know that...
Anything can be reverse engineered, given enough time and resources.
However, would compiling your Java code to native code with a tool like GCJ make it more difficult to decompile? I mean, given a few minutes, I can decompile a .jar using JD-GUI, and it is relatively accurate. Most of the "Java to EXE" converters are just .exe launchers for the JVM, and while there are many benefits to the JVM, I have been led to believe that security of the source code is not one of them.
Bottom line: Can you use something like GCJ to compile your Java source (or .class) files to native machine code, and if so, will that protect it from decompiling?
EDIT: Ideally, it would be something more than just obfuscation. The specific project is a commercial game, so what we are looking for is a way to make it more difficult to get to the source code to begin with, not just understand it. Also, I'm not sure that Steam accepts .jars, and we are planning on submitting it to the new Green Light project.

I wouldn't choose that approach just for source-security.
Check out some Obfuscator tools out there like ProGuard
If you want to see what those tools do to your source code, just try read the decompiled Minecraft jar if you have one on hand.
A downside to using this is, that if your code depends on using reflection, you'll have to configure the tools to ignore those functions/classes/whatever, as those will not be found at runtime otherwise.

Technically, yes. Using something like GCJ will make it harder to decompile, however keep in mind that you are losing some major benefits of using Java if you do this. Namely, you lose the ability to write cross-platform applications.
You could use an obfuscator to make the code harder to decompile AND still keep the benefits of using Java.

a Source code obfuscator like
this , this and this
makes your variables, functions, etc... unreadable by other(has no logical meaning). You should read here too!

In a pickle with obfuscating a java library

Ok I'm kinda in a predicament right now.
I have a java program that I have split into a core/outside sections.
I collated my core into a java library that the outside code (which will be publicly released) can reference.
However, I do not want the contents of this java library to be decompilable.
So I went to find a good java obfuscator.
What complicates my situation is the fact that my java library isn't exactly modular - it references/changes state of outside code (yes it's terrible but whatever)
I've tried demos of all premium obfuscators (ZKM, allatori, etc) and free ones, but they either
have very weak control flow obfuscation which is what I need
halt because of references to java library's/dependencies that are not in the jar itself but are still referenced.
Any advice?

Obfuscation does not prevent decompilation. It can always be decompiled. It just make help making your code less readable afterwards. Obfuscate only your logic, keep interfaces untouched.

If there are no obfuscators that have control flow obfuscation which meet your standards, then you will have to write your own obfuscator or submit a request to an existing vendor to improve their product.
Run both the external and internal parts of the program through the obfuscator, together at the same time. But write exclude rules for all of the external code. You should also write exclude rules for the public API of your internal code. If you don't have a public API layer on your internal code, then you are going to have a hard time, because your external code will have to refer to your internal code by obfuscated names, which will make for very unmaintainable and hard to read external source code.

I second ahanin's comment.
But, if you're looking for an obfuscator, a good one that has a pretty robust rule set is Proguard. Its used heavily in the Android space where code needs to be made as minimal (small) as posible.
Link: Proguard

How to lock compiled Java classes to prevent decompilation?

How do I lock compiled Java classes to prevent decompilation?
I know this must be very well discussed topic on the Internet, but I could not come to any conclusion after referring them.
Many people do suggest obfuscator, but they just do renaming of classes, methods, and fields with tough-to-remember character sequences but what about sensitive constant values?
For example, you have developed the encryption and decryption component based on a password based encryption technique. Now in this case, any average Java person can use JAD to decompile the class file and easily retrieve the password value (defined as constant) as well as salt and in turn can decrypt the data by writing small independent program!
Or should such sensitive components be built in native code (for example, VC++) and call them via JNI?

Some of the more advanced Java bytecode obfuscators do much more than just class name mangling. Zelix KlassMaster, for example, can also scramble your code flow in a way that makes it really hard to follow and works as an excellent code optimizer...
Also many of the obfuscators are also able to scramble your string constants and remove unused code.
Another possible solution (not necessarily excluding the obfuscation) is to use encrypted JAR files and a custom classloader that does the decryption (preferably using native runtime library).
Third (and possibly offering the strongest protection) is to use native ahead of time compilers like GCC or Excelsior JET, for example, that compile your Java code directly to a platform specific native binary.
In any case You've got to remember that as the saying goes in Estonian "Locks are for animals". Meaning that every bit of code is available (loaded into memory) during the runtime and given enough skill, determination and motivation, people can and will decompile, unscramble and hack your code... Your job is simply to make the process as uncomfortable as you can and still keep the thing working...

As long as they have access to both the encrypted data and the software that decrypts it, there is basically no way you can make this completely secure. Ways this has been solved before is to use some form of external black box to handle encryption/decryption, like dongles, remote authentication servers, etc. But even then, given that the user has full access to their own system, this only makes things difficult, not impossible -unless you can tie your product directly to the functionality stored in the "black box", as, say, online gaming servers.

Disclaimer: I am not a security expert.
This sounds like a bad idea: You are letting someone encrypt stuff with a 'hidden' key that you give him. I don't think this can be made secure.
Maybe asymmetrical keys could work:
deploy an encrypted license with a public key to decrypt
let the customer create a new license and send it to you for encryption
send a new license back to the client.
I'm not sure, but I believe the client can actually encrypt the license key with the public key you gave him. You can then decrypt it with your private key and re-encrypt as well.
You could keep a separate public/private key pair per customer to make sure you actually are getting stuff from the right customer - now you are responsible for the keys...

No matter what you do, it can be 'decompiled'. Heck, you can just disassemble it. Or look at a memory dump to find your constants. You see, the computer needs to know them, so your code will need to too.
What to do about this?
Try not to ship the key as a hardcoded constant in your code: Keep it as a per-user setting. Make the user responsible for looking after that key.

#jatanp: or better yet, they can decompile, remove the licensing code, and recompile. With Java, I don't really think there is a proper, hack-proof solution to this problem. Not even an evil little dongle could prevent this with Java.
My own biz managers worry about this, and I think too much. But then again, we sell our application into large corporates who tend to abide by licensing conditions--generally a safe environment thanks to the bean counters and lawyers. The act of decompiling itself can be illegal if your license is written correctly.
So, I have to ask, do you really need hardened protection like you are seeking for your application? What does your customer base look like? (Corporates? Or the teenage gamer masses, where this would be more of an issue?)

If you're looking for a licensing solution, you can check out the TrueLicense API. It's based on the use of asymmetrical keys. However, it doesn't mean your application cannot be cracked. Every application can be cracked with enough effort. What really important is, as Stu answered, figuring out how strong protection you need.

You can use byte-code encryption with no fear.
The fact is that the cited above paper “Cracking Java byte-code encryption” contains a logic fallacy. The main claim of the paper is before running all classes must be decrypted and passed to the ClassLoader.defineClass(...) method. But this is not true.
The assumption missed here is provided that they are running in authentic, or standard, java run-time environment. Nothing can oblige the protected java app not only to launch these classes but even decrypt and pass them to ClassLoader. In other words, if you are in standard JRE you can't intercept defineClass(...) method because the standard java has no API for this purpose, and if you use modified JRE with patched ClassLoader or any other “hacker trick” you can't do it because protected java app will not work at all, and therefore you will have nothing to intercept. And absolutely doesn't matter which “patch finder” is used or which trick is used by hackers. These technical details are a quite different story.

I don't think there exists any effective offline antipiracy method. The videogame industry has tried to find that many times and their programs has always been cracked. The only solution is that the program must be run online connected with your servers, so that you can verify the lincense key, and that there is only one active connecion by the licensee at a time. This is how World of Warcraft or Diablo works. Even tough there are private servers developed for them to bypass the security.
Having said that, I don't believe that mid/large corporations use illegal copied software, because the cost of the license for them is minimal (perhaps, I don't know how much you are goig to charge for your program) compared to the cost of a trial version.

Q: If I encrypt my .class files and use a custom classloader to load and decrypt them on the fly, will this prevent decompilation?
A: The problem of preventing Java byte-code decompilation is almost as old the language itself. Despite a range of obfuscation tools available on the market, novice Java programmers continue to think of new and clever ways to protect their intellectual property. In this Java Q&A installment, I dispel some myths around an idea frequently rehashed in discussion forums.
The extreme ease with which Java .class files can be reconstructed into Java sources that closely resemble the originals has a lot to do with Java byte-code design goals and trade-offs. Among other things, Java byte code was designed for compactness, platform independence, network mobility, and ease of analysis by byte-code interpreters and JIT (just-in-time)/HotSpot dynamic compilers. Arguably, the compiled .class files express the programmer's intent so clearly they could be easier to analyze than the original source code.
Several things can be done, if not to prevent decompilation completely, at least to make it more difficult. For example, as a post-compilation step you could massage the .class data to make the byte code either harder to read when decompiled or harder to decompile into valid Java code (or both). Techniques like performing extreme method name overloading work well for the former, and manipulating control flow to create control structures not possible to represent through Java syntax work well for the latter. The more successful commercial obfuscators use a mix of these and other techniques.
Unfortunately, both approaches must actually change the code the JVM will run, and many users are afraid (rightfully so) that this transformation may add new bugs to their applications. Furthermore, method and field renaming can cause reflection calls to stop working. Changing actual class and package names can break several other Java APIs (JNDI (Java Naming and Directory Interface), URL providers, etc.). In addition to altered names, if the association between class byte-code offsets and source line numbers is altered, recovering the original exception stack traces could become difficult.
Then there is the option of obfuscating the original Java source code. But fundamentally this causes a similar set of problems.
Encrypt, not obfuscate?
Perhaps the above has made you think, "Well, what if instead of manipulating byte code I encrypt all my classes after compilation and decrypt them on the fly inside the JVM (which can be done with a custom classloader)? Then the JVM executes my original byte code and yet there is nothing to decompile or reverse engineer, right?"
Unfortunately, you would be wrong, both in thinking that you were the first to come up with this idea and in thinking that it actually works. And the reason has nothing to do with the strength of your encryption scheme.