upgrade to polkadot-v0.9.40

[b-yap]

partially closes pendulum issue 219.

I have updated the weights as well, since a lot of the pallets are #![deny(warnings)].

summary of changes:

1. rename grandpa traits
2. deprecated trait store
3. hash instead of BlockId
4. deprecated from_ref..., changed to from_parts() function

Specific to clients directory:

1. Rename *-private-ipv4 to *-private-ip CLI args

Specific to testchain directory:

1. update batching api in chain spec
2. rename warp_sync to warp_sync_params
3. Extract syncing protocol from sc-network
4. Add WeightToFee and LengthToFee impls to transaction-payment Runtime API
5. KeyOwnerProofSystem of grandpa removed

extra:

1. cleanup based on this suggestion.
2. fix failing test_check_slot_position test case after updating the MAX_SLOTS_TO_REMEMBER value.

[b-yap]

@ebma the weights have increased. You can check each extrinsic's comments from the benchmark.
I can always replace from_ref_time with from_parts, if you find the numbers too big/unreasonable.

[ebma]

Thanks for adding all these changes @b-yap 🙏

I see that the weights changed indeed. Sometimes the execution time varies just a little and sometimes the weights doubled (which might look confusing at first since the weight comments now state everything in pico seconds, but since pico is 1e-3 smaller than nano and the values mostly have three additional 0s now, they are pretty similar.) But it could be that the fact that you were now using --execution=wasm also amplified this variation. It's great though that you managed to use that parameter as it's preferred and should be more accurate AFAIK. I only did not use it before because it did not work for some reason.

Maybe we can fix some of the CI warnings but besides that it's good to go 👍

[ebma]

Wow, since when does it work with wasm? That's great. I remember when I created all the benchmarks a while ago this parameter was not working and unfortunately I was not able to figure out why. Great that it works now

[ebma]

Why this comment? I need to change the amount when running benchmarks with the testchain but use the 1000 otherwise? Can we change it such that it always uses 100_00000u32 then so that it works in both cases?

[b-yap]

changed amount to the fixed 1000_0000u32.into()