Gregory Popovitch

Hey @bashimao , thank for following up with this and the hard work. I think the group lookup starts at the index determined by the hash, so I'm not sure...

Actually, probably you can align the lookup, and discard the matches before the starting index, so you hit only one cache line.

Good luck!

Just a thought. `std::unordered_map` is not very efficient, `absl::flat_hash_map` would much likely be faster.

Oh, my bad, thank you @scottschurr , I thought we already did have an abseil dependency!

> Just wanted to confirm, this doesn't add any new RPCs, but simply expands the functionality of `ledger_entry`. Is that correct? Yes, `ledger_entry` can return the new xchain ledger objects....

I had a quick look at the changes, and I fully agree with the direction they are taking us towards. One thing that I found immediately surprising was `view.read()` returning...

OK, I'm fixing this now.

@smoe please try the latest phmap version... it should work now.

Better late than never. I think the issue should be fixed now.