Add benchmarking support for digest items

[notlesh]

This PR provides a mechanism for pushing digest items into the benchmark overhead and benchmark extrinsic pipelines. It is a potential fix for #12142. The rationale for it is thoroughly described there, but briefly: this solves the dependency problem where an inherent requires a pre-runtime digest to be present.

Note that I tested a similar fix through Moonbeam against an older version of Substrate (v0.9.26) but have not tested the code in this PR. I'll be happy to do so if there is support for it in the first place.

Companions

• polkadot: Reflect benchmarking fn signature change polkadot#5958
• cumulus: not needed
• nimbus: not required, but could be used to demonstrate that this works

TODO

• no-op companion PRs
• test - can be accomplished by implementing in nimbus which is where this all started
• any documentation required (minor breaking changes)

polkadot companion: paritytech/polkadot#5959

[ggwpez]

Nice.

It still needs companions and a fmt. In the future we can think about putting all of these args in a config struct to have fewer companion requirements.

[notlesh]

I am working on implementing benchmark overhead/extrinsic in nimbus because I'll be able to test this change there (the problematic inherent/digest pair lives in that repo).

However, I believe this is mergeable as-is.
Nevermind, Cumulus needs a companion apparently.

I believe this will build if CI will run against my polkadot companion (am I missing something?)

[ggwpez]

I believe this will build if CI will run against my polkadot companion (am I missing something?)

Looks like it. But it also needs a Cumulus comp, since we also expose these commands there…

[notlesh]

I believe this will build if CI will run against my polkadot companion (am I missing something?)

Looks like it. But it also needs a Cumulus comp, since we also expose these commands there…

I don't think it does unless it's compiling a branch other than master (I can't tell from the CI log). Cumulus master looks like it doesn't implement BenchmarkCmd::Overhead or BenchmarkCmd::Extrinsic.

It looks to me like the CI failure here is coming out of polkadot, example:

/builds/parity/mirrors/substrate/extra_dependencies/polkadot/cli/src/command.rs:555:14
1950    |
1951555 | ...                   cmd.run(client.clone(), inherent_data, &ext_factory)
1952    |                           ^^^ --------------  -------------  ------------ supplied 3 arguments
1953    |                           |
1954    |                           expected 4 arguments

Is cumulus not being built against my polkadot companion?

[ggwpez]

bot merge

I don't think it does unless it's compiling a branch other than master…

Yea you are right. Cumulus just ignores these commands (currently).
PS: You need to tick a bot in the comp MR.

[paritytech-processbot]

Error: Github API says "Allow edits from maintainers" is not enabled for paritytech/polkadot#5958. The bot would use that permission to push the lockfile update after merging this PR. Please check https://docs.github.com/en/github/collaborating-with-pull-requests/working-with-forks/allowing-changes-to-a-pull-request-branch-created-from-a-fork.

[notlesh]

I recreated my companion PR from my personal repo so I could enable edits from maintainers. Sorry for the mess...

New PR: paritytech/polkadot#5959

[bkchr]

bot merge

[paritytech-processbot]

Error: Github API says #12159 is not mergeable

[ggwpez]

bot merge