KOVACS Krisztian
KOVACS Krisztian
In particular, tokio isn't pinned to 1.15 now. Pinning tokio was making it hard for us to use `goose` in projects which use other dependencies that require a newer version...
``` hash/pedersen_hash time: [13.247 µs 13.546 µs 13.891 µs] change: [-42.552% -40.380% -38.164%] (p = 0.00 < 0.05) ``` End-to-end sync tests performed on the first 80k blocks of the...
Just like we do checkpoint sync as a first step of a P2P sync this change implements a special "bulk" sync mode for feeder gateway sync. In bulk mode we...
Block commitment signatures on Starknet 0.13.2 will be just a signature over the block hash. To verify pre-0.13.2 blocks we'll need to include a pre-calculated hash of all pre-0.13.2 blocks...
Pathfinder currently does iterate over the range of blocks in the filter and checks the per-block bloom filter for potential matches. Unfortunately loading/checking the bloom filters is fairly slow: for...
The new specification adds the (recursive) `CONTRACT_EXECUTION_ERROR` type for representing stack traces. We don't currently have detailed stack trace information in the JSON-RPC code because the execution layer converts blockifier's...
The semantics of the method are not quite clear, so based on a Slack discussion to clarify the specification here's some extra information: - Any combination of `class_hashes`, `contract_addresses` and...
We should throw the new error over JSON-RPC 0.8.0 _and keep the old one for previous versions_. The message for the error should be the following: `The transaction's resources don't...
Looks like we're keeping a read transaction (in `ExpectedDeclarations::spawn()`) open for the whole duration of the class definitions sync. This causes SQLite to be unable to merge WAL records back...
To add support for the `starknet_getMessagesStatus` method (https://github.com/starkware-libs/starknet-specs/pull/223) we'd need to revamp our Ethereum client. We need access to events, both for getting historical stuff and ideally a Websocket subscription...