> On x86 this optimization is already applied, but since LEA loads a
> 32-bit integer, EAX can be later used as an already untagged pointer
> in %test-headers: MOV EAX, [ECX-7] => MOV EAX, [EAX], which takes one
> byte less to encode.
This only works because we're conservative on x86. Some of us still hold
hopes of enabling a precise GC at some point in the future. For a single
byte, I'd really prefer if that improvement were reverted.