Execution test fails when changing the constant periods_per_cycle #3742
Comments
Leo-Besancon
added a commit
that referenced
this issue
Apr 4, 2023
* Initial commit for network restart: Archive read / write Needs testing * Skeleton of the snapshot creation workflow * Removed archiving tests * fmt * cleaned a bit the FinalState init * Cleaned (check / clippy) * Added Pos deserialization * Added final state serialization to file * Refactoring + create final_state dir if it does not exist * fixed path issue * Reverted some commits of local config * Don't reset ledger if from snapshot (will be changed for testnet21) Indeed, we will always keep the ledger from now on. * Compute and check hash of recovered final_state * Don't replace old ledger * Changed timestamps to account for LAST_START_PERIOD But no finalized blocks are made anymore, I'll have to check the concensus rules... * fmt * Fixed some final_cursor issues create_snapshot and --from_snapshot should work well in sandbox mode! * clippy allow lint in case the constant is not 0 * Changed LAST_START_PERIOD from const to static ref + Ledger rollback based on genesis_timestamp * Fixed cycle_history issue and block_header deserialize assumption * fmt * Fixed hash computation in --from_snapshot * cleaned the code * Rethought the cycle logic when LAST_STAR_PERIOD is not 0. * Keep previous roll_counts * fmt * clippy * Fixed initial_rolls when snapshotting * Added feed_cycle_state_hash in --from_snapshot Don't know if I should do it or not :/ * fmt * Set up PoS_final_state after bootstrap * fmt * Added debug info * Fix Mutex lock issue * change config for tests * Added bootstrap of initial pos_state info - initial_cycle - initial_rolls - initial_ledger_hash * Change desynchronisation checks * Added cycle_history debug info * Boostrapable last_start_period value * Final_state in ledger * Fixed final state in ledger The dump / commit logic is probably false right now, need to investigate the best way to handle it * doctest pass * Fix bootstrap message deser * disable desync detection + bootstrap fix * Fixed deployment (lazy_static parsing fail) + fmt * logs for cycle info changes * Fixed execution cycle lookback * Tried to fix the draws * Updated deferred credit logic * Fixed deferred credit hash for snapshot + removed info!() * Make test pass + clean info!() logs * Removed final_state_path setting, no longer used * Added documentation, restored periods_per_cycle at 128 * Refactoring + Only bootstrap the initial_state once * Update massa-final-state/src/final_state.rs Co-authored-by: Ben <benphawke@gmail.com> * fmt * Cleaned the FinalState::from_snapshot() method Created Ser/Deser helpers for FinalStateRaw. * Better handle final_state in RocksDB * Added a TODO for get_pinned * Use range() on the BTreeMap to better filter the slot range. * Fixed a regretion in the previous commit * Doc comments lints * Doc comment for final_state_data in ledger * cleaned final_state_raw_deserializer args * Used get_pinned_cf for final_state_hash * fmt + _args => args * Changed final_state dispatch in main * Interpolate the cycles (untested) * Small fixes and lints * Delete intiial_state logic (bootstrap, deser, ...) + iniitial_cycle for pos_state * fmt * clippy fix * update snapshot logic without get_final_state * Set period_per_cycle to 10 for tests * Fix feed_cycle_hash * fmt * Feed selector during downtime interpolation * feed_selector and compute draws logic added * clippy fix * Don't dump the final_state on finalize() * Added comments for periods_per_cycle * small refactoring and docs update for pos_final_state * Make CI pass Cf #3742 * Small refactor of final_state * Add TODO for commented imports * Refactor final_state downtime interpolation * Removed an unneeded rocks_db call * Included (x-1) > Excluded (x) in bounds * fmt * Updated deferred_credits logic, TODO: TESTS Inspired from #3740 * Small cleanup / renames * Updated roll_sell test to test the new deferred_credits impl * Cleanup after changes: no more last_start_period needed in pos_final_state * TODO: CHANGE BACK CONST * Fixed block_factory next_slot if before last_start_period * Update final_state.rs * Changed rng_seed push to avoid loops * info > debug for all denuncuations * Revert "info > debug for all denuncuations" This reverts commit 2b28b33. * Make PR Ready! * Fix merge with GRPC * Improved some docs * Fix execution test after merge * New pass over documentation * Final doc improvements * Apply suggestions from code review Co-authored-by: Ben <benphawke@gmail.com> * Applied most of the suggestions in review * fmt * fixed stats feature flags * fmt --------- Co-authored-by: Ben <benphawke@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug
In particular, the test
tests::scenarios_mandatories::tests::events_from_switching_blockcliquepanics with aCycleUnavailableerror after a block miss.To Reproduce
In
massa\massa-models\src\config\constants.rs, change:to
Then run the tests:
cargo test --features testingExpected behavior
The test should pass.
The text was updated successfully, but these errors were encountered: