feat: upgrade to Polkadot v0.9.19

[wischli]

fixes KILTProtocol/ticket#1946

• Adds LengthToFee for pallet_transaction_payment --> required refactor of our WeightToFee impl (see Substrate PR)
• Upgrades from Polkadot v0.9.17 to Polkadot v0.9.19
  • In v0.9.18, a database migration happens, which also integrates an IPFS server into Substrate (new remark pallet) and adds remark_with_index (PR) which basically enables serving smaller files (< 50kb)
  • However, we do not need to add any migration to our runtime

paritytech/substrate#10779 is a database migration that results in backwards incompatible changes to the database.
paritydbusers: please take a look at paritytech/polkadot#5168 (comment) to avoid issues with your database.

• Adds Benchmark Subcommands to both clients (main PR, block cmd)
  • Documentation only exist in rust docs or corresponding PRs, see below for an extract

Benchmark Cmd Refactor

Sub-commands concerned with benchmarking. The pallet benchmarking moved to the `pallet` sub-command

USAGE:
    substrate benchmark <SUBCOMMAND>

OPTIONS:
    -h, --help       Print help information
    -V, --version    Print version information

SUBCOMMANDS:
    block       Benchmark the execution time of historic blocks
    help        Print this message or the help of the given subcommand(s)
    overhead    Benchmark the execution overhead per-block and per-extrinsic
    pallet      Benchmark the extrinsic weight of FRAME Pallets
    storage     Benchmark the storage speed of a chain snapshot

Pallet Benchmark

We already know this command

cargo run --release -p kilt-parachain --features=runtime-benchmarks -- benchmark pallet \\
   --chain=dev --steps=50 --repeat=2 --pallet=pallet-tips --extrinsic="*" --execution=wasm \\
   --wasm-execution=compiled --heap-pages=4096

Storage Benchmark

# Rocks DB
cargo run --release -p kilt-parachain --features=runtime-benchmarks -- benchmark storage \\
 --chain=dev --state-version=1

# Parity DB
cargo run --release -p kilt-parachain --features=runtime-benchmarks -- benchmark storage \\
 --chain=dev --state-version=1 --db paritydb-experimental

Block Benchmark

Requires database snapshot

Let's say you are on a Substrate chain and want to verify that the first 3 blocks did not use more weight than declared which would otherwise be an issue.
To test this with a dev node, first create one with a temp directory:

$ substrate --dev -d /tmp/my-dev --execution wasm --wasm-execution compiled

And wait some time to let it produce 3 blocks. Then benchmark them with:

$ substrate benchmark-block --from 1 --to 3 --dev -d /tmp/my-dev \\
  --execution wasm --wasm-execution compiled --pruning archive

Overhead benchmark

We don't provide support yet as its not supported yet in the cumulus parachain node template.

ToDo

• Check for missing migrations --> We don't need any, Kusama/Polkadot only run a crowdloan migration which is a result from an incorrect previous migration
• Handle runtimes better in parachain client --> Not optimal at all, but improvements need to be handled in a separate PR

Checklist:

• I have verified that the code works
  • No panics! (checked arithmetic ops, no indexing array[3] use get(3), ...)
• I have verified that the code is easy to understand
  • If not, I have left a well-balanced amount of inline comments
• I have left the code in a better state
• I have documented the changes (where applicable)

[trusch]

Lgtm :)