resolve_scope followed by get_from_scope is exceedingly common, it is visible both in manual checks of JS2, and was also found by Keith when he more rigorously looked at the most common pairs of opcodes (it was > 3% of all bytecode if I remember correctly). Now that we've got checkpoints in bytecode ops it should be quite easy to add a shared opcode. We've found in the past that optimizing bytecode size matters a lot, both to Ramification and to real-world websites. As a bonus, this change should also reduce the interpretation overhead, making the LLint faster.
<rdar://problem/90575798>
Pull request: https://github.com/WebKit/WebKit/pull/1579