Ekleog-NEAR

As @nagisa said I don’t think there’s a way to reuse a backing `Vec` right now, and it makes sense to me given `VecAssembler` semantically has ownership of the `Vec`...

On my (real-world-ish) test case, here are the performance results: - Before any change: 28.5-29ms - By allocating only the label hashmaps that actually deserve it (ie. using this PR):...

Turns out the CLA process was easier than I expected, so that’s now done. AFAICT the remaining CI check is just someone from google needing to look at the PR,...

I just added a test by copy-pasting the test contents from our clusterfuzz UI, and ./local/tests/ci_tests.bash seems to pass, so hopefully this will be good to go :)

@jonathanmetzman Not sure whether you saw, but this is I think ready for review round 2 :)

@jonathanmetzman I don’t want to bother you so I’ll try just one last ping, I think this is ready for review :) Anyway, thank you for the work you do...

/gcbrun

Welp seems like I don’t have permission to run this. Anyway, what I can say for sure is, `./local/tests/ci_tests.bash` passes locally :)

Actually, a thought: would it be better to get the line and char out of the summary? I’d just need to end the regex at `:` instead of end-of-line; and...

We had some related discussion in https://near.zulipchat.com/#narrow/stream/295306-pagoda.2Fcontract-runtime/topic/Eth-wallet.20support/near/403646004 . In particular, I think the idea of actually having some hash-containing placeholder for the smart contract code, would help us be forward-compatible...