I've been testing situations where sizeof(opcode_t) != sizeof(IV) != sizeof(IV).
I got most of the casting and arithmetic problems patched [1], but am running into alignment problems with the bytecode because it currently expects sizeof(opcode_t) == sizeof(IV). (Or something, I can't tell quite yet whether the problems in the assembler or the _unpack function.) In either case, I know that we'll still have alignment problems once we hit runops land. What are we going to do about this? [1] Why won't gcc allow bit-operands on pointers? Is that illegal? -- Bryan C. Warnock [EMAIL PROTECTED]
