On Fri, 2 Apr 2010, Bill Janssen wrote:
Andi Vajda <va...@apache.org> wrote:
At times like these it would be very handy to know just which class
file
is causing the problem.
Indeed, I wish java would tell.
Yes, it's a very poorly designed exception.
All jcc can tell is what class it attempted to load not what class is
bad since loading a class can bring in a bunch more.
I can add a catch to print out the attempted class name.
That would help, thanks. I'm still a bit mystified as to how a bad
version number could slip in there -- I'm using the same compiler for
both JCC and GoodStuff.
If you're on Windows, maybe your registry trick is biting ? javac is using
one VM, your registry trick is picking up another later.
If you're elsewhere, a possible error is using Java 5 over Java 6 or vice
versa or maybe 64-bit vs 32-bit, I'm not sure. One way to know what version
of bytecode your using is to request it via the javac -target flag. This is
why, on darwin, I hardcoded the -target flag to 1.5 in JCC's setup.py.
Note that that javac in setup.py is only used to compile the couple of java
classes used by JCC itself under the java directory tree. JCC does not
compile your own java code.
Andi..
