Leszek Wiesner

> I agree, first version is safer, but it does make me slightly nervous how we maintain this well. I think we can take advantage of `max_encoded_len` now that all...

> Proposal: ~ 6 x the monthly cost of running a reference hardware machine ~ $400/month so ~ $2500. With 100 validators and no nominators the minimum value at stake...

1. We now use `dollars` / `cents` as units everywhere, shouldn't we follow this approach here? 2. There is a very big difference in the price of `JOY` vs `DOT`....

> Is it not strange this has not been an issue prior? Most of the benchmarks are already following this rule to always panic on failure. Ironically I was probably...

Unfortunately, some of the substrate's native benchmarks are also subject to this issue (they return `Err` on failure). I just ran into it again when tweaking staking parameters. Confusing error...

@mnaamani I was thinking that since the issue appears also in the substrate benchmarks and `--no-verify` "solves" it, perhaps we can use this flag by default in `./generate-weights.sh`. During actual...

> Some people are confused by absention not counting towards quorum, alternative would be to include in both approval quorum, but also to not remove from denominator in threshold. In...

During my local tests I didn't encounter any issues, everything seems to work as expected :+1:

Follow-up w.r.t. cases like `channel_update_with_assets`, where some parameters were omitted: After a few more local tests I reached following conclusions: - a margin of error of a few percent of...

> metadata: raw metadata, can be anything. > The reason raw_action is not locked to a specific set of messages is that it allows apps to start playing with distinct...