Introduce prove RPC method

[rakanalh]

Description

• Add the RPC method
• Write a test

Linked Issues

• Fixes RPC for manual proof production, in case we need to hand trigger stuff #1285

[codecov]

Codecov Report

Attention: Patch coverage is 76.78883% with 133 lines in your changes missing coverage. Please review.

Project coverage is 78.1%. Comparing base (dcfda45) to head (0851ac5).
Report is 1 commits behind head on nightly.

Files with missing lines	Patch %	Lines
crates/prover/src/proving.rs	88.4%	39 Missing ⚠️
crates/prover/src/rpc.rs	57.3%	38 Missing ⚠️
crates/prover/src/da_block_handler.rs	50.8%	30 Missing ⚠️
crates/prover/src/errors.rs	0.0%	20 Missing ⚠️
crates/fullnode/src/da_block_handler.rs	73.3%	4 Missing ⚠️
crates/common/src/da.rs	95.6%	2 Missing ⚠️

Additional details and impacted files

Files with missing lines	Coverage Δ
crates/common/src/utils.rs	87.9% <ø> (-1.9%)	⬇️
crates/prover/src/lib.rs	50.0% <ø> (ø)
crates/prover/src/runner.rs	90.1% <100.0%> (+0.4%)	⬆️
...ule-system/sov-modules-rollup-blueprint/src/lib.rs	100.0% <ø> (ø)
...eign-sdk/rollup-interface/src/state_machine/stf.rs	26.3% <ø> (ø)
crates/common/src/da.rs	95.8% <95.6%> (-0.4%)	⬇️
crates/fullnode/src/da_block_handler.rs	72.7% <73.3%> (-1.3%)	⬇️
crates/prover/src/errors.rs	0.0% <0.0%> (ø)
crates/prover/src/da_block_handler.rs	75.4% <50.8%> (-6.3%)	⬇️
crates/prover/src/rpc.rs	50.7% <57.3%> (+43.3%)	⬆️
... and 1 more

[eyusufatik]

everything else LGTM

[yaziciahmet]

Why do we need this functionality? It seems pretty risky and not so useful to me tbh.

With the current rpc implementation, anyone can invoke a proving process. If someone is set out to bankrupt us, he can invoke proving the moment they see a sequencer commitment proof in DA, and force us to prove only 1 commitment, causing us a lot of money from both proving and paying fee to DA side.

Even if this problem was to be solved, I can't see the use of this functionality.

UPDATE:

I learned that the prover is not public to the internet. So a random person can not invoke this rpc. But I am still skeptical of this feature as this still relies on the human (us) intervention, and a small mistake can cause a lot. If we want to prove a certain commitment right away without waiting for sampling, it makes more sense to lower the sampling config number and restart the prover for one time.

[yaziciahmet]

Also a question, what happens if prover is busy proving a commitment at l1 block 100, but there are commitments at 150 and 200. Someone invoked the commitment at l1 block 200? Do we break the assumption of ordered proofs on DA? If so it might be problematic on the light client and bridge side.

[eyusufatik]

Also a question, what happens if prover is busy proving a commitment at l1 block 100, but there are commitments at 150 and 200. Someone invoked the commitment at l1 block 200? Do we break the assumption of ordered proofs on DA, if so it might be problematic on the light client and bridge side.

prove rpc is useful when we do proof sampling and we find some interesting sequencer commitments that we want to prove.

So if there is no proof sampling we wont use the RPC

[yaziciahmet]

prove rpc is useful when we do proof sampling and we find some interesting sequencer commitments that we want to prove.
So if there is no proof sampling we wont use the RPC

@eyusufatik

UPDATE:
I learned that the prover is not public to the internet. So a random person can not invoke this rpc. But I am still skeptical of this feature as this still relies on the human (us) intervention, and a small mistake can cause a lot. If we want to prove a certain commitment right away without waiting for sampling, it makes more sense to lower the sampling config number and restart the prover for one time.

[eyusufatik]

great refactor in the latest commits.

[eyusufatik]

I learned that the prover is not public to the internet. So a random person can not invoke this rpc. But I am still skeptical of this feature as this still relies on the human (us) intervention, and a small mistake can cause a lot. If we want to prove a certain commitment right away without waiting for sampling, it makes more sense to lower the sampling config number and restart the prover for one time.

the thing is, that's also something we have to implement :) and it's not worth the time spent on shutting down the prover and restarting etc.

having this RPC is really helpful while we still have proof sampling.