Libgdx overcoming color limitations?

Engines, Libraries and Tools
#21

There seems to be more going on. I remember writting a simple JNI method to replace intToFloat, but the JNI bridge seemed to have inserted some magic that checks for NaN and modifies the value.

Also, this behaviour is not limited to Oracle’s VM. Dalvik exhibits the same behaviour, and that uses a modified Harmony class library. Can anyone try with Excelsior?

Who’s gonna flip the bat sign switch?

#22

It’s too bad we can’t do this on a register (like a C union is compiled to):


unsafe.putInt(register, val);
return unsafe.getFloat(register);

and that non-bulk Unsafe memory operation are not available on Android.

#23

Don’t think it’s a bug, behavior is documented in intBitsToFloat javadocs:
http://docs.oracle.com/javase/7/docs/api/java/lang/Float.html#intBitsToFloat(int)