* grandpa: stricter gossip message filtering
* gossip: remove filtered message on send_message
* gossip: add test for tracking of broadcast attempts
* grandpa: only restrict gossip if we're connected to more than 5 authorities
* grandpa: add test for progressive gossip
* grandpa: add test for gossip filtering on local non-authority node
* grandpa: fix doc
* gossip, grandpa: tabify
* grandpa: relax filtering logic for global messages
* Apply breaking changes of new libp2p versions
* Oops, forgot to update version
* Fix tests
* Fix imports with WASM
* Fix WASM for real
* Update core/network/src/debug_info.rs
Co-Authored-By: Roman Borschel <romanb@users.noreply.github.com>
* Fix compilation
* Implement Debug for some structures
`NetworkConfiguration`, `TransportConfig`, `NodeKeyConfig` and `Secret`.
Needs a new release of the `rust-libp2p` crate. This PR is just a
reminder.
* Explicitly separate `std` and `core`
* Add manual implementation for Secret
* core/finality-grandpa: Pass Grandpa msg sender up to UntilImported
* core/finality-grandpa: Track senders to maybe later request blocks
* core/finality-grandpa: Make BlockStatus pub only within crate
* core/finality-grandpa: Abstract NetworkBridge with BlockSyncRequester
* core/finality-grandpa: Pass BlockSyncRequester to UntilImported
* core/finality-grandpa: Track block number of pending within UntilImported
* core/finality-grandpa: Request block sync on long wait
* core/finality-grandpa: Adjust unit tests to previous changes
* core/finality-grandpa: Fix line length
* core/finality-grandpa: Add comment explaining in & out vote combination
* core/finality-grandpa: Log after, not before, timeout expired
The UntilImported component should log whenever waiting for a specific
block to be imported surpassed a defined constant timeout. Without this
patch the code would log whenever the current time was below the
timeout.
* core/finality-grandpa: Collect senders as HashSet for deduplication
* Revert "core/finality-grandpa: Track senders to maybe later request blocks"
This reverts commit 61ac9dd715612d5fdbf7b8f00b84e450f282ade0.
* Revert "core/finality-grandpa: Pass Grandpa msg sender up to UntilImported"
This reverts commit afdc9646a6c314f99a9d19242f1878f85980e70d.
* core/network/sync: Ask for block from all peers if none provided
When requesting an explicit fork sync, try to sync from all known peers,
when no specific peers were provided.
* core/network/sync: Request specific fork sync from peers ahead or on par
When making an explicit fork sync request without specifying any peers,
make sure to only request it from the locally known peers that are
either ahead or on a par compared to the block number we are looking
for.
* grandpa: fix tests
* grandpa: fix warnings
* grandpa: add test for block sync request on until_imported
* grandpa: rename Environment field inner to client
* grandpa: fix minor nits
* grandpa: minor nits in until_imported
* grandpa: copy docs for set_sync_fork_request
* grandpa: remove stale TODO on UntilImported
* consensus-pow: add difficulty data to auxiliary
* Timestamp api
* Implement FinalityProofProvider for ()
* Add DifficultyApi
* Remove assumption that Difficulty is u128
* Use a separate trait for add instead of hard-code it as Saturating
* Some convenience functions to work with PowVerifier
* Try to fix mining unstability
* Fix generic resolution
* Unused best_header variable
* Fix hash calculation
* Remove artificial sleep
* Tweak proposer waiting time
* Revert sleep removal
The reason why it was there is because when mine_loop returns, it means an error
happened. In that case, we'd better sleep for a moment before trying again,
because immediately trying would most likely just fail.
* Pass sync oracle to mining
So that it does not mine when major syncing
* Expose build time as a parameter
Instead of hardcode it as previously 100ms.
* Update lock file
* Fix compile
* Support skipping check_inherents for ancient blocks
For PoW, older blocks are secured by the work, and can mostly be considered to
be finalized. Thus we can save both code complexity and validation time by
skipping checking inherents for them.
* Move difficulty fetch function out of loop
To make things faster
* Remove seed from mining
Each engine can use its own Rng source.
* Better comments
* Add TotalDifficulty definition for U256 and u128
* Update core/consensus/pow/src/lib.rs
Co-Authored-By: André Silva <andre.beat@gmail.com>
* Rename TotalDifficulty::add -> increment
* Use SelectChain to fetch the best header/hash
* Update lock file
* Update all dependencies
* Upgrade dependencies whenever “easy”
“easy” means that there are no major changes required.
* Fix build and bump paste dependency to 0.1.6
* Remove dead code
* Re-add = dependency for futures-preview
* Add missing std features for runtime-io
* Remove git dependencies
as updated versions have been published to crates.io
* try to debug bug
* For sr-io, "std" should imply "no_oom" and "no_panic_handler".
Otherwise, rustc complains (correctly) about duplicate lang items.
* Add missing "runtime-io/std" features
* Fix compilation errors
* Prevent duplicate lang items
Rust does not allow duplicate lang items. When compiled without the
`std` feature, `sr-io` defines two lang items. Therefore, `sr-io`
compiled without `feature = "std"` must not be linked with `std`.
However, `pwasm-utils` and `wasmi-validation` both bring in `std` unless
compiled with `default-features = "false"`. This caused a duplicate
lang item error. Building both with `default-features = "false"`
prevents this error. When building with `feature = "std"`, they should
both be built with the `std` feature, so this feature needs to be
explicitly depended on.
* Bump `impl_version`
* Make tests pass
Three tests used 1 less gas than they had previously.
* Try to un-break build
* Add a Cargo.lock file
* Revert offchain code
* Revert "Revert offchain code"
This reverts commit d216d08cc6ca0344614669c1d24cde3aa5c0d4e2.
* Don’t try to send a body with a GET request
without adding a Transfer-Encoding or Content-Length header.
This has always been wrong, but hyperium/hyper#1925 hid the bug until
hyper was upgraded to 0.12.35.
* Change some more GET requests to POST requests
* Fix excess line width and remove an `extern crate`
* Delete commented-out extern crate
Co-Authored-By: Sergei Pepyakin <sergei@parity.io>
* Fix regression in Cargo.toml files
dev-dependencies need `default-features = false`, too.
* Bump parity-wasm dependency
* Bump `futures-preview`
* Apply suggestions from code review
Co-Authored-By: Bastian Köcher <bkchr@users.noreply.github.com>
* Update Cargo.lock files
* Apply suggestions from code review
Co-Authored-By: Bastian Köcher <bkchr@users.noreply.github.com>
* Update core/service/src/chain_ops.rs
Co-Authored-By: Sergei Pepyakin <sergei@parity.io>
* async txpool API
* Update core/rpc/src/author/mod.rs
Co-Authored-By: Tomasz Drwięga <tomusdrw@users.noreply.github.com>
* Update core/transaction-pool/graph/src/pool.rs
Co-Authored-By: Tomasz Drwięga <tomusdrw@users.noreply.github.com>
* Pool -> Pool + ValidatedPool
* removed lost block_on when importing xt from network
* fix grumbles
* alias for future::Executor in rpc
* removed executor from Author RPCs
* Pool + SharedValidatedPool -> Pool
* fix compilation after merge
* another fix
* another fix
* Add some chainspec tests and make sure we validate it.
* Manual implementation of Extension + Forks definitions.
* Move chain spec to separate crate.
* Allow using ChainSpec with extensions.
* Renames.
* Implement Extension derive.
* Implement Extension for Forks.
* Support specifying fork blocks.
* make for_blocks work
* Support forks correctly.
* Add a bunch of docs.
* Make fork blocks optional.
* Add missing docs.
* Fix build.
* Use struct for check_block params.
* Fix tests?
* Clean up.
The current `debug_assert_eq` is clearly wrong as the code handles zero
response blocks. What should have been expressed is that we expect 0 or
1 response blocks.
* always fetch epoch from runtime
* node integration tests don't test light nodes
* give stand-in full node a FULL role
* rejig babe APIs
* introduce next-epoch-descriptor type
* overhaul srml-BABE epoch logic
* ensure VRF outputs end up in the right epoch-randomness
* rewrite `do_initialize` to remove unnecessary loop
* begin accounting for next epoch in epoch function
* slots passes header to epoch_data
* pass slot_number to SlotWorker::epoch_data
* begin extracting epoch-change logic into its own module
* aux methods for block weight
* aux methods for genesis configuration
* comment-out most, refactor header-check pipeline
* mostly flesh out verifier again
* reinstantiate babe BlockImport implementation
* reinstate import-queue instantiation
* reintroduce slot-worker implementation
* reinstate pretty much all the rest
* move fork-choice logic to BlockImport
* fix some, but not all errors
* patch test-runtime
* make is_descendent of slightly more generic
* get skeleton compiling when passing is_descendent_of
* make descendent-of-builder more succinct
* restore ordering of authority_index / slot_number
* start fiddling with tests
* fix warnings
* improve initialization architecture and handle genesis
* tests use correct block-import
* fix BABE tests
* fix some compiler errors
* fix node-cli compilation
* all crates compile
* bump runtime versions and fix some warnings
* tweak fork-tree search implementation
* do backtracking search in fork-tree
* node-cli integration tests now work
* fix broken assumption in test_connectivity
* babe tests fail for the right reasons.
* test genesis epoch logic for epoch_changes
* test that epochs can change between blocks
* First BABE SRML test
* Testing infrastructure for BABE
Also includes a trivial additional test.
* Apply suggestions from code review
Co-Authored-By: Bastian Köcher <bkchr@users.noreply.github.com>
* A little more test progress
* More work on BABE testing
* Try to get the tests working
* Implement `UintAuthorityId`-based test mocks
* Fix compilation errors
* Adjust to upstream changes
* Block numbers are ignored in BABE epoch calculation
* authority_index() should ignore invalid authorities
* Fix compile error
* Add tests that session transitions happen
* Check if BABE produces logs
It currently does not.
* Fix test suite
This was really nasty, due to a type confusion that showed up as an
off-by-1 buffer error.
* Add additional tests
Most of these were derived from the current output, so they are only
useful to guard against regressions.
* Make the tests more readable
Also bump impl_version.
* Fix excessive line width
* Remove unused imports
* Update srml/babe/src/lib.rs
Co-Authored-By: André Silva <andre.beat@gmail.com>
* try to fix imports
* Fix build errors in test suite
* tests did not pass
* Try to get at least one digest to be output
Currently, the code emits either no digests (if I don’t call
`Session::rotate_session()` or two digests (if I do), which is wrong.
* More tests
They still don’t work, but this should help debugging.
* fix silly error
* Don’t even try to compile a broken test
* remove broken check_epoch test and add one for genesis epoch
* Check that the length of the pre-digests is correct
* Bump `impl_version`
* use epoch_for_descendent_of even for genesis
* account for competing block 1s
* finish srml-babe docs
Co-Authored-By: André Silva <andre.beat@gmail.com>
* address grumbles
* grandpa: don't request justification for consensus changes on full node
* sync: finalize justification/finality proof requests in-order
* sync: ignore result of try_finalize_root on justification import
With the *authority-discovery* module an authoritative node makes itself
discoverable and is able to discover other authorities. Once discovered, a node
can directly connect to other authorities instead of multi-hop gossiping
information.
1. **Making itself discoverable**
1. Retrieve its external addresses
2. Adds its network peer id to the addresses
3. Sign the above
4. Put the signature and the addresses on the libp2p Kademlia DHT
2. **Discovering other authorities**
1. Retrieve the current set of authorities
2. Start DHT queries for the ids of the authorities
3. Validate the signatures of the retrieved key value pairs
4. Add the retrieved external addresses as ~reserved~ priority nodes to the
peerset
* node/runtime: Add authority-discovery as session handler
The srml/authority-discovery module implements the OneSessionHandler in
order to keep its authority set in sync. This commit adds the module to
the set of session handlers.
* core/network: Make network worker return Dht events on poll
Instead of network worker implement the Future trait, have it implement
the Stream interface returning Dht events.
For now these events are ignored in build_network_future but will be
used by the core/authority-discovery module in subsequent commits.
* *: Add scaffolding and integration for core/authority-discovery module
* core/authority-discovery: Implement module logic itself
* Initial implementation, some redundancy is awkward and there is
some useless computation (but there is a pending pr for that).
Next are tests.
* Minimal tests and fix extend child.
* implement iterator for change child trie.
* prune child trie.
* Fix pruning test.
* bump spec version.
* Avoid empty child trie (could also be checked before)
* tabs.
* Fix child digest overriding each others.
* Restore doc deleted on merge.
* Check correct child value on extrinsics build.
* Revert runtime version update.
Arkadiy Paronyan