* Keep node information about disconnection

* Fix line widths

* init store for slots

* fix: add check_equivocation to Aura/Babe

* fix tests

* fix: add pruning bound

Co-Authored-By: André Silva <[email protected]>

* use saturating_sub

Niklas Adolfsson authored May 14, 2019 and Svyatoslav Nikolsky committed May 14, 2019

Fixes warning: the item `HashSet` is imported redundantly

* Randomly fuzz the PSM

* Fix test

* Run it moar

* Move ProtocolMsg to service.rs

* Fix line width

* Fix line width again

* Correct whitespace

* Another whitespace correction

* do not try to sync from light node

* use Roles' utility methods

* fix light client strtup (build_select_chain -> Option)

* fixed node template compilation

* Update core/service/src/lib.rs

Co-Authored-By: Benjamin Kampmann <[email protected]>

* Make the behaviour in libp2p generic

* Fix indentation

* Fix bad merge

* GrandpaLightBlockImport

* extract authorities in AuraVerifier

* post-merge fix

* restore authorities cache

* license

* new finality proof draft

* generalized PendingJustifications

* finality proof messages

* fixed compilation

* pass verifier to import_finality_proof

* do not fetch remote proof from light import directly

* FinalityProofProvider

* fixed authorities cache test

* restored finality proof tests

* finality_proof docs

* use DB backend in test client

* justification_is_fetched_by_light_client_when_consensus_data_changes

* restore justification_is_fetched_by_light_client_when_consensus_data_changes

* some more tests

* added authorities-related TODO

* removed unneeded clear_finality_proof_requests field

* truncated some long lines

* more granular light import tests

* only provide finality proof if it is generated by the requested set

* post-merge fix

* finality_proof_is_none_if_first_justification_is_generated_by_unknown_set

* make light+grandpa test rely on finality proofs (instead of simple justifications)

* empty_finality_proof_is_returned_to_light_client_when_authority_set_is_different

* missing trait method impl

* fixed proof-of-finality docs

* one more doc fix

* fix docs

* initialize authorities cache (post-merge fix)

* fixed cache initialization (post-merge fix)

* post-fix merge: fix light + GRANDPA tests (bad way)

* proper fix of empty_finality_proof_is_returned_to_light_client_when_authority_set_is_different

* fixed easy grumbles

* import finality proofs in BlockImportWorker thread

* allow import of finality proofs for non-requested blocks

* limit number of fragments in finality proof

* GRANDPA post-merge fix

* BABE: pos-merge fix

Pierre Krieger authored May 12, 2019 and Arkadiy Paronyan committed May 12, 2019

* Move the is_offline and is_major_syncing logic to service.rs

* Move the ImportQueue to service.rs

* Remove stop() and abort()

* Add some more documentation to sync.rs

* bump crates to v2.0.0

* fix crate authors

* bump node runtime spec_version

* remove unused import for no-std

* proof on child trie

* higher level api for child storage proof

* boilerplate for proof from light fetch

* actually check proof on light fetch

* Do not break former encoding

* tabify

* tabify2

* Add child trie root tx to full_storage_root transaction.

* Shorten long lines.

* Temp rename for audit

* Make full_storage a trait method

* Name back and replace some code with full_storage where it looks fine.

* fix indentations, remove unused import

* flush child root to top when calculated

* impl +1

* finalize srml modules in reverse order

* update runtime versions

* fixed gossip related tests deadlock

* cleaning sync tests log

* fixed typo in logs

* send GRANDPA messages without holding validator lock

* Revert "fixed gossip related tests deadlock"

This reverts commit d3fe55e4.

* first impl

* rename origin::inherent to none

* fix

* fix

* Apply suggestions from code review

Co-Authored-By: thiolliere <[email protected]>

* comment

* better error

* doc

* (add unsigned module 🤦)

* doc

* fix

* implement for node-template as well

* add validated unsigned to executor

* fix

* fix

* bump version

* testing xt

* remove extraneous logic

* licence

* impl test

* Rewrite the PSM

* Fix disconnecting from reserved peers

* Minor adjustements

* Address review

* Reputation changes adjustements

* More adjustements

* Adjust all reputations

* More fixes and adjustments

* Improve proof

* Remove the possible panic

* Make sure reputation reaches 0

* move SelectChain trait out of client

* Extend SelectChain, move longest chain implementation into it

* Bring SelectChain into service

* implement LongestChain SelectChain

* implement longest chain for node

* update Cargo.lock's

* in between erroring tests

* deprecate ::backend and ::import_lock

* Remove unneded space

Co-Authored-By: gnunicorn <[email protected]>

* Remove unneded space

Co-Authored-By: gnunicorn <[email protected]>

* Fixes test compilation

* remove todo

* re-enable client test

* add doc

* fixing tests

* Clarify SelectChain Interface, intended implementation and usage

* minor components cleanups

* minor cleanups

* Update lock files

* Implement cleaner interface for SelectChain

* addressing comments

* Updating tests

* bump node runtime impl version

* address grumbles

* Imply forced authoring on --dev

* Combine conditionals

* Add some benchmarks for ed25519

* We don't need the Debug impl

* Add multiple message sizes

`impl_runtime_apis!` is not only implementing the apis, it also calls
them internally and thus generates the warning. So, we just allow
depracted calls in the generated code.

* Remove the Incoming enum

* Use tokio in the protocol background thread

* Some internal protocol logic simplifications

* Merge the protocol thread with the network thread

* More the status_sinks logic to Service

* Remove FromNetworkMsg

* Use saturating_pow for max digest interval

* Impl version

* ignore light nodes in ConsensusGossip

* fixed method name

* temporary disabled penalty when block is announced

* remove traces of BLOCK_ANNOUNCE_REPUTATION_CHANGE

- Add convert
- Add btree_set for without_std

Gavin Wood authored May 06, 2019 and Sergey Pepyakin committed May 06, 2019

* Reformulate random seed to be more random

- First 80 random values come from cycling the incomplete series (
  instead of filling with zeroes)
- Calculate random material each usage (use a single amalgamated
  ring buffer to store them for avoiding 81 lookups each time)
- Mutate inputs by hashing each with:
  - its index (into the 81)
  - an additional "subject" key provided by caller

This keeps the final output low-influence while still allowing
it to be used as the seed to independent contexts. (Hashing the
result to give the final seed is no better than using parent_hash).

* Docs

* Bump runtime

* Update notes

* Remove feature(alloc)

* Update srml/system/src/lib.rs

Co-Authored-By: gavofyork <[email protected]>