23:34:04karloszlike, when calling a bytecode function from a bytecode function, it would be nice not to have to reinitialize the vm struct every time
23:34:57karloszi suppose with stack segments you'd still get exactly the same context switch overhead. with stack segments you'd just be allocating a new stack on every bytecode entry and you'd also basically have to reload stuff like the current bytecode vector/literals/closure
0:15:47Bikeoh i guess irc-store can just get the type from the value
0:34:04Bikeoh geez, obvious optimization i just noticed i've been missing, (make-sequence whatever (length whatever)) actually checks that the length is nonnegative
3:08:35drmeisterHere is what you need for bytecode closures...