v0.9.19 Dependency Update

[notlesh]

What does it do?

Updates our dependencies to substrate v0.9.19.

Generated update issue: #1428
Frontier branch: https://github.com/PureStake/frontier/commits/moonbeam-polkadot-v0.9.19

TODO:

• Frontier update PR
• Frontier patches
• Crowdloan-rewards update PR
• Nimbus update PR
• Make Moonbeam compile
• CI green / tests passing
• Benchmarking changes
• Grandpa revert: Grandpa revert procedure paritytech/substrate#11162
• Review use of polkadot_primatives::v2 (instead of v0)
• Review moonbeam-foundation/crowdloan-rewards@c3fb5aa
• merge master 🙄
• cargo test fails (see my comment below)

[notlesh]

cargo test is failing:

error[E0308]: mismatched types
   --> pallets/parachain-staking/src/tests.rs:776:13
    |
776 |                     error: 2,
    |                            ^ expected array `[u8; 4]`, found integer

For more information about this error, try `rustc --explain E0308`.

[girazoki]

As I said I reviewed that the types that we changed from v0 or v1 to v2 have not changed (HeadData, Id, AccountId). But please feel free to confirm

[tgmichel]

cargo test is failing:

error[E0308]: mismatched types
   --> pallets/parachain-staking/src/tests.rs:776:13
    |
776 |                     error: 2,
    |                            ^ expected array `[u8; 4]`, found integer

For more information about this error, try `rustc --explain E0308`.

ModuleError was changed in paritytech/substrate#10242

The legacy implementation still available in https://github.com/paritytech/substrate/blob/f4157f44220f3382d8787c3cd348e2ea7aabb048/primitives/runtime/src/legacy/byte_sized_error.rs#L29-L38

[crystalin]

Why is this one changed ? This looks like a breaking change

[notlesh]

We've had to change these indices frequently, often because the ordering of the array changes (a new event is emitted, events are split into multiple ones, are combined, etc.)

I think we need a better way to express these expectations, something like "ExtrinsicSuccess" is followed by "SomeDepositEvent" but not "any error" would be awesome.

I'll try to see what caused this particular case...