Jagex kills reflection apparently

[Mads]

A different thing, I've seen considered, is that they obfuscate the ID's of objects, interfaces, and items, that were used to recognize stuff for the bots. You know.. Mixing them all up, every update as part of the compilation.

[theagentd]

The question of how they killed Reflection still hangs in the air... 

It was answered in the first reply.

Oh, I missed that. Is that a freely available obfuscator or something they wrote themselves?

[JL235]

Either way, it wouldn't be that difficult to build.

Games published by our own members! Check 'em out!

[pjt33]

Is that a freely available obfuscator or something they wrote themselves?

It's Jagex. Everything is in-house except javac. The obfuscator is a source-to-source compiler, which was harder to build than JL235 thinks - and if you think it would be easy, try reading the entire JLS.

[Riven]

Odd choice. You'd think this would be a rainy-afternoon task using bytecode transformation.

[kappa]

Odd choice. You'd think this would be a rainy-afternoon task using bytecode transformation.

Indeed, or they could have just used an existing solution like ProGuard.

[Waterwolf]

Is that a freely available obfuscator or something they wrote themselves?

It's Jagex. Everything is in-house except javac. The obfuscator is a source-to-source compiler, which was harder to build than JL235 thinks - and if you think it would be easy, try reading the entire JLS.

That's strange because their string obfuscation is pretty much the same as Zelik klassmaster's. Also I really doubt it's source-to-source because most decompilers don't work with rs at all.

And kappa, proguard only obfuscates names and occasionally optimizes bytecode and that is very easy to bypass and make recompilable.

[pjt33]

Is that a freely available obfuscator or something they wrote themselves?

It's Jagex. Everything is in-house except javac. The obfuscator is a source-to-source compiler, which was harder to build than JL235 thinks - and if you think it would be easy, try reading the entire JLS.

That's strange because their string obfuscation is pretty much the same as Zelik klassmaster's. Also I really doubt it's source-to-source because most decompilers don't work with rs at all.

And kappa, proguard only obfuscates names and occasionally optimizes bytecode and that is very easy to bypass and make recompilable.

Well, maybe they've ditched the old source-to-source compiler entirely rather than updating it, but in 2006 I had a desk next to the guy who maintained it and I answered the odd question about the generics spec (as I was in the vanguard messing around with generics and giving feedback to Sun before 1.5 came out). Although now that you mention klassmaster that does ring a bell, so maybe there was slightly more NIH stuff than just javac.

[JL235]

Is that a freely available obfuscator or something they wrote themselves?

It's Jagex. Everything is in-house except javac. The obfuscator is a source-to-source compiler, which was harder to build than JL235 thinks - and if you think it would be easy, try reading the entire JLS.

Having built a couple of parsers (most of which were source to source translators), and done a fair amount of bytecode hacking, I just don't see how it could be that difficult.

If I were to build it...

• I'd use javac to parse the source code into bytecode, and that's the source all validated (no need to even open the Java Language Spec)
• Next parse the bytecode, and for this I'd try using JavaAssist, which allows you to edit compiled class files.
• If JavaAssist didn't do what I needed, then I'd build a custom bytecode parser. Given that we are taking the bytecode straight from javac, we can presume it looks like a valid program, which makes this task 10 times easier.
• Now it's parsed into your AST, apply your obfuscations (JavaAssist may even help with this, it's designed for editing Java code, but you could otherwise do it manually)
• Next output the new source/bytecode; getting an AST to print this can take time (lots of nodes to write), but is very easy to build.
• Finally you could have it re-parse the output to ensure it did output a valid program (as a sanity check). For this you could re-use javac, bytecode parsers or even spin up a JVM and manually load the classes.

Given that your just re-using existing technologies, the general structure including the parser/output stages could take maybe a week. It'll be the obfuscation stage which would take time, which could take an extra month or two (to do it right!).

[Mads]

@Economy-discussers:
Runescape's economy functions after demand and supply, right.
Obviously, with botters, there was a too large supply, causing prices to drop. However, now without bots,
supply is even bigger, because theres no more bots to process materials 
Take mining for example; before, botters would smith. They needed to buy the materials.

Now, there's an overflow of materials, causing prices to drop even more dramatically on many resources.
Nobody wants to buy meterials though, because they're expected to raise in price.

Games published by our own members! Check 'em out!

[SwampChicken]

What a fascinating thread...
Is there a place where one could see some of the java code that does the 'botting'?

[Mads]

What a fascinating thread...
Is there a place where one could see some of the java code that does the 'botting'?

https://github.com/powerbot/RSBot
RSBot is just one of the clients. It is also the most widely used, and it's open source.

[Nate]

*apparently

Thread title kills me every time I see it. :p

[Mads]

*apparently

Thread title kills me every time I see it. :p

Change shall be made! My apologies.

[ra4king]

reflection......appear-ently?

ROFLMAO