sam <[EMAIL PROTECTED]> wrote: > tomer: > > It is my opinion that you would loose performance if the Python > interpreter had the additional task of verifying byte code. It might be > more appropriate to have a preprocessor that did the verifying as it > compiled the byte code.
But in this case, there is no "compiled" -- just the code type's constructor: > > from types import CodeType as code > > exec code(0, 5, 8, 0, "hello moshe", (), (), (), "", "", 0, "") Yes, code() itself could do some sanity checking. But so could 'exec' -- it's never a particularly fast operation anyway... and most definitely NOT on the critical path of normal operations. I'd rather have "defense in depth", in this case. Alex -- http://mail.python.org/mailman/listinfo/python-list
