|
|
Re: Obfuscating Plug-ins [message #453574 is a reply to message #453564] |
Thu, 03 August 2006 08:32 |
Keith Harrison-Broninski Messages: 33 Registered: July 2009 |
Member |
|
|
Similarly to Patrick, I obfuscate my own RCP app and it works fine. I
use the free and open source proguard, which comes with a GUI tool you
can use to build a config file very quickly and easily, which you can
then reference from an Ant task as part of your standard build. To get
started, there is a standard config you can copy for obfuscating a
"library", which is what a plugin effectively is.
The only issues I found were memory. You need to include in the
obfuscation run a reference to all external jars that may reference your
plugin code indirectly via common superclasses, which means all other
platform plugins must be loaded into memory during the run. The
Proguard documentation suggests JVM arguments to resolve this and they
worked fine for me.
I also tried decompiling the obfuscated code to see what it looked like,
and was pleased to see that most decompilers just gave up completely on
the more complex classes. Simple classes managed to decompile, but as
required all non-public and non-protected elements were transformed into
meaningless symbols.
I might also add that I tried yguard too (another free and open source
tool), but this crashed with a NullPointerException which I found
impossible to debug.
Keith
|
|
|
Powered by
FUDForum. Page generated in 0.03237 seconds