feat: refactor signing in order to more easily be able to dryrun

[daniel-savu]

Co-authored-by: Sander Bosma sanderbosma@gmail.com
Co-authored-by: Daniel Savu savudani04@yahoo.ro

create_signed is refactored, adding support for dry_running an extrinsic before sending it.

[cla-bot-2021]

User @savudani8, please sign the CLA here.

[daniel-savu]

Looks like the clippy error is unrelated to this PR. Maybe it has to do with the used toolchain?

[jsdw]

Looks like the clippy error is unrelated to this PR. Maybe it has to do with the used toolchain?

Yup, I suspect owing to a new version of rust and such being released yesterday. I'll look into it!

[jsdw]

A fix for the clippy error is on it's way (#548)

I need to stare at the code a little more but offhand it looks good to me; thankyou for the PR!

[jsdw]

If you merge up from master, the clippy error should go away now!

[jsdw]

Since we only need the bytes, how about this just takes in a Vec<u8>, so we don't need the various lifetimes and generics (I'd go with a &[u8] if possible but that may be trickier)?

The SignedSubmittableExtrinsic could then return a &[u8] (give the users the option to clone or not) from the encoded() function to pass in here.

[jsdw]

In fact, I think if we go with &[u8] here, we can use hex::encode to turn that into hex for the RPC params.

[jsdw]

As suggested in the other comment, I reckon this can return a &[u8], giving users the option to clone or not as needed. I'd add some docs too; something like:

/// Returns the SCALE encoded extrinsic bytes.

[jsdw]

Looking great; a couple of small suggestions but otherwise I am really liking this separation; thank you for the contribution!

[sander2]

I haven't followed this pr too closely but in the initial commit it was possible to do create_signed().dry_run(). Now instead you need to do

let signed = create_signed();
client.rpc().dry_run(signed)

which in my opinion is a little bit less elegant. @jsdw would you be opposed to adding dry_run back into SignedSubmittableExtrinsic just as a wrapper to the one in rpc.rs?

[jsdw]

I'd be perfectly up for having a dry_run function on SignedSubmittableExtrinsic that calls the RPC method with the bytes it has available!

[daniel-savu]

For some reason the tests are not finishing execution on my machine. I followed the GitHub workflow, including the substrate binary here, but I get this:

test client::chain_subscribe_blocks has been running for over 60 seconds
test client::chain_subscribe_finalized_blocks has been running for over 60 seconds
test client::dry_run_fails has been running for over 60 seconds
test client::dry_run_passes has been running for over 60 seconds
test client::fetch_block has been running for over 60 seconds
test client::fetch_block_hash has been running for over 60 seconds
test client::fetch_keys has been running for over 60 seconds
test client::fetch_read_proof has been running for over 60 seconds
test client::fetch_system_info has been running for over 60 seconds
test client::insert_key has been running for over 60 seconds
test client::test_iter has been running for over 60 seconds
test events::balance_transfer_subscription has been running for over 60 seconds

[jsdw]

One thing to do to fix the CI tests is to prepend "0x" to the hex string generated by hex::encode() (I always forget about that).

As for local tests not working, my only guess is that the substrate binary isn't on your PATH or something like that, so all of the tests are waiting for it to run (I assume no integration tests complete).

[jsdw]

Looks good to me!