Performance data
codehag opened this issue · comments
Since one of the major motivations here is performance, do we already have some numbers that we can take a look at?
Performance is actually not a justification for this proposal at all (at least as currently presented!).
I see, so the justification here is static analysis, and ergonomics, is that right? You can close this issue in that case.
I see, so the justification here is static analysis, and ergonomics, is that right?
And also portability, yes. Performance may improve as a side effect, but the full graph is only fetched at the time of the worker invocation. Preloading this graph before the worker call is not a focus of this proposal, the other justifications are.
Ok, will close, but happy to pick up on any of these aspects as needed.