japhb is very, very happy to see a patch that says "Fix a bunch of possible SEGVs." 17:22
Is it already known that NativeCall is failing tests on r-m? The summary error is: t/06-struct.t .......... Failed 14/19 subtests 17:55
dalek arVM: 8cfefcb | brrt++ | src/jit/emit_x64.dasc:
First operand of sp_guad* is r(obj) not int16

This caused a crash if the order of register fields in MVMSpeshOperand were reversed. JimmyZ++ for spotting it.
22:37
jnthn I would say that might fix an endian bug, but given it's the JIT for *x64*... :) 22:39