On 10/16/06, Alexei Zakharov <[EMAIL PROTECTED]> wrote:
Moreover, the spec obliges us to do so. The real question is: is it RI bug or bug in spec? Personally I am for returning non-null values in all cases.
I agree. Just tell me and I will commit the patch for EM(interpreter) and JIT modes. BTW I think that we should create interpreter.emconf file and place all VM properties needed for interpreter into this file. So, in future, when we need to add more interpreter-mode option, we will not affect the EM C++ code. -- Mikhail Fursov