Bump the LEZ dependency from the `lez-core-v0.2.0` tag to `v0.2.0-rc6` across
the workspace and all guest manifests (still resolving via the renamed
`lee_core`/`lee` packages), and regenerate the lockfiles to match.
rc6 moved the clock program out of `nssa` into a separate system-programs crate
(gated behind the guest-building `artifacts` feature), so adapt the tests:
- Import `ClockAccountData` and `CLOCK_01_PROGRAM_ACCOUNT_ID` from `clock_core`
instead of `nssa`, and build clock data via `ClockAccountData::to_bytes()`
rather than hand-encoding the Borsh layout.
- `V03State::new()` no longer auto-creates the clock account, so AMM tests seed
the canonical 1-block clock explicitly before ops that read it.
- `advance_clock` now writes the clock account directly via
`force_insert_account` (the clock can no longer be ticked with a real
transaction), matching how upstream rc6 state-machine tests seed accounts.
- Add the `clock_core` dependency to integration_tests/benchmark.
Add the first end-to-end coverage of the oracle's RecordTick path, which
previously existed only as native unit tests:
- amm_twap_observations_accumulate_across_swaps_and_yield_time_weighted_average:
drives swaps + RecordTick across simulated time, then checks the cumulative
accumulator and the consulted time-weighted average.
- amm_twap_record_tick_sampling_guard_skips_calls_below_min_interval: exercises
the min-interval sampling guard through the real instruction path.
Running RecordTick through the zkVM surfaced that committing the oracle-owned
~100 KiB observations account costs ~50.9M cycles — over the 2^25 (~33.5M)
public-execution limit — so the instruction aborted on chain. Reduce
OBSERVATIONS_CAPACITY 6396 -> 2048 (~16.8M cycles, ~half the limit); window
coverage is unchanged, only sampling resolution.
Add programs/benchmark, a standalone crate (excluded from the workspace so CI
and the Makefile skip it) that runs the guest ELF through the RISC Zero
executor and reports the per-instruction cycle split, reproducing the on-chain
pass/fail at the limit. Its cost-vs-capacity sweep still spans to 6396, guarding
against bumping capacity back into the over-budget range.
