Blaž Hrastnik

Hmm I see, this happens when you don't finish the snippet (loc has 2 placeholders so you need to press ctrl-k 3 times to finish -- [which clears the state](https://github.com/archseer/snippets.nvim/blob/190a8fc9cfc3021b77df904ca0177e8a45e02e24/lua/snippet/init.lua#L615-L624))....

About fastglobal: OTP 22 will ship with something called `persistent_term`, which is aimed at usecases like these: It's global, not garbage collected, and has no locks so it's faster than...

I wonder, have you tried to benchmark `persistent_term` based implementation first? It was specifically designed to replace mochiglobal/precompiled module type data stores, so it should perform about as fast. https://github.com/erlang/otp/pull/1989#issuecomment-430632381

Not sure why the CI is breaking. Maybe the golang bump?

> also wondering...are there tests for these changes? There are no gauntlet tests at all in this repo, but I'm working on this in https://github.com/smartcontractkit/chainlink-solana/pull/413

> but when individual TXs start to fail that would interrupt the proposal flow and make it more difficult to restart. I hope that the improved transmission logic will make...

Note: still doing some debugging, at least in the `setup.dev.flow` the digest from the input vs calculated from the state in `approveProposal` don't match

You could wrap the `onRound` listener into a promise, we do that in the tests: https://github.com/smartcontractkit/chainlink-solana/blob/805ddc7d784165bb509c75b6bda9db0ce82bb405/ts/src/index.test.ts#L94-L101

Why wasn't I pinged about this? This is because `shell.nix` is my development setup that uses `saber-overlay` to compile the entire solana toolchain from source. https://github.com/smartcontractkit/chainlink-solana/blob/a3d61f6006f9c16ecd8af969d5d7f9811c977228/flake.nix#L8-L9= Saber does offer a...

I was hoping to discuss as a parking lot item today, we can merge the current design and only handle success route for now. Failure handling will require one of...