My understanding, and the point of using reflection rather than bytecode
analysis, is that javap and similar tools based on decompilation are
questionable from a cleanroom perspective. I defer to those of you with
a better legal background than myself for this.
>From my perspective: We can't look at Sun's code -- *.java or *.class.
We can use their published specifications (which include the API
documentation) and use their reference implementation to generate
compatibility cases. The latter is what the ClassTool utility attempts
to do, WITHOUT EXPLICITLY READING Java bytecode.
I certainly agree that a much more useful and wide-ranging tool can be
built (and probably has been, see the gnu.bytecode classes referred to
earlier). I was merely trying to provide something that would not break
with the spirit of cleanroom for this project.
Wes
