feat(starknet_l1_provider): add start_block

[giladchase]

• replace preivous API with the agreed upon start_block.
• added state setter for use of start_block; commit_block will have a different setter.

[reviewable-StarkWare]

This change is 

[matan-starkware]

Reviewed 8 of 8 files at r2, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @giladchase)

---

crates/starknet_l1_provider/src/lib.rs line 0 at r2 (raw file):
In separate PRs can you move stuff out of lib.rs?

---

crates/starknet_l1_provider/src/lib.rs line 51 at r2 (raw file):

                Err(L1ProviderError::unexpected_transition(self, new_state))
            }
        }

Might be worth a docstring explaining this.

Code quote:

        match new_state {
            ProviderState::Propose | ProviderState::Validate => Ok(new_state),
            ProviderState::Uninitialized | ProviderState::Pending => {
                Err(L1ProviderError::unexpected_transition(self, new_state))
            }
        }

[giladchase]

Reviewable status: 3 of 9 files reviewed, 2 unresolved discussions (waiting on @matan-starkware)

---

crates/starknet_l1_provider/src/lib.rs line 51 at r2 (raw file):

Previously, matan-starkware wrote…

Might be worth a docstring explaining this.

Good point.
How about strong-typing it? i added a new "external" type for start_block, which will be the one sent through clients.
This has the added benefit of hiding the internal Pending/Uninitialized states from the users, which probably should have been the case to begin with 😅

---

crates/starknet_l1_provider/src/lib.rs line at r2 (raw file):

Previously, matan-starkware wrote…

In separate PRs can you move stuff out of lib.rs?

Yep i remember you asked, i have a commit ready after this is merged, state is moved into l1_provider

[matan-starkware]

Reviewed 5 of 5 files at r3, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @giladchase)

---

crates/starknet_l1_provider/src/l1_provider_tests.rs line 103 at r3 (raw file):

#[test]
#[should_panic(expected = "Uninitialized L1 provider")]

Given that this (and get_txs) are public functions which already return a result type, should we just return an error from them instead?

Code quote:

#[should_panic(expected = "Uninitialized L1 provider")]

---

crates/starknet_l1_provider/src/l1_provider_tests.rs line 125 at r3 (raw file):

    l1_provider.start_block(BlockNumber(1), ValidateSession).unwrap();
    l1_provider.start_block(BlockNumber(1), ValidateSession).unwrap();

IDK if this is actually a good idea but would be kinda fun to replace all of this with cartesian_product - https://docs.rs/itertools/latest/itertools/trait.Itertools.html#method.cartesian_product

Code quote:

    l1_provider.start_block(BlockNumber(1), ProposeSession).unwrap();
    l1_provider.start_block(BlockNumber(1), ValidateSession).unwrap();

    l1_provider.start_block(BlockNumber(1), ProposeSession).unwrap();
    l1_provider.start_block(BlockNumber(1), ProposeSession).unwrap();

    l1_provider.start_block(BlockNumber(1), ValidateSession).unwrap();
    l1_provider.start_block(BlockNumber(1), ValidateSession).unwrap();

[giladchase]

Reviewable status: 8 of 11 files reviewed, 2 unresolved discussions (waiting on @matan-starkware)

---

crates/starknet_l1_provider/src/l1_provider_tests.rs line 103 at r3 (raw file):

Previously, matan-starkware wrote…

Given that this (and get_txs) are public functions which already return a result type, should we just return an error from them instead?

Yeo, doing it in the commit_block stack.
The test is removed there because this flow is becomes a special case of the "get/validate called without start_block".

---

crates/starknet_l1_provider/src/l1_provider_tests.rs line 125 at r3 (raw file):

Previously, matan-starkware wrote…

IDK if this is actually a good idea but would be kinda fun to replace all of this with cartesian_product - https://docs.rs/itertools/latest/itertools/trait.Itertools.html#method.cartesian_product

NOICE!
I can't we're using cartez in code in way that actually improves readability.

[github-actions]

Artifacts upload workflows:

• Started at 2025-01-15T13:53:06Z
• Started at 2025-01-15T16:12:50Z