* add stub for Client.best_chain_containing_block_hash
* add fn blockchain::Backend::leaf_hashes
* fix typo
* sketch out Client.best_chain_containing_block_hash
* fix indent
* Blockchain.leaf_hashes -> Blockchain.leaves
* add unimplemented! stub impls for Blockchain.leaves
* start impl of Blockchain.leaves for in-memory client db
* Client.best_chain_containing...: check canonical first and make compile
* first rough attempt at maintaining leaf list in in-memory db
* fix tab indent
* add test best_chain_containing_single_block
* add failing test best_chain_containing_with_fork
* pub use client::blockchain; in test-client to prevent circular dep in client tests
* best_chain_containing_with_single_block: improve and test leaves
* far improve in-memory Backend::leaves impl
* test blockchain::Backend::leaves more thoroughly
* handle more edge cases in blockchain::Backend::leaves impl for in memory
* fix test best_chain_containing_with_fork (two distinct test blocks had same hash)
* make best_chain_containing_block_hash pass existing tests
* improve docstring for Blockchain::leaves
* Client.best_chain_containing: some cleanup. support max_block_number
* best_chain_containing: remove broken outcommented fast check for best = canonical
* remove blank line
* best_block_containing: return None if target_hash not found
* best_chain_containing: add unreachable! at end of function
* improve tests for best_chain_containing
* renames
* more elaborate test scenario for best_containing
* best_containing: fix restriction of search through maybe_max_number
* best_containing: tests for restriction of search
* get rid of unnecessary clones
* replace Client::new_in_mem by new_with_backend which is useful for testing backends
* add test_client::new_with_backend for testing different backend impls
* add test for in_mem::Backend::leaves
* remove unused imports
* in_mem test_leaves: simplify
* flesh out tests for in_mem leaves impl
* remove tests for leaves from client which are now covered in implementing module
* improve comment
* add Client.new_in_mem again
* unwrap in test_client::new_with_backend
* make test_client::BlockBuilderExt work not just with in-mem backend
* make test client ext not just work with in mem backend
* add failing Backend.leaves test for client-db
* update Cargo.lock
* replace KeccakHasher with Blake2Hasher
* refactor
address grumble https://github.com/paritytech/substrate/pull/740#discussion_r217822862
* refactor using NumberFor
address grumble https://github.com/paritytech/substrate/pull/740#discussion_r217823341
* add test that exposes possible problem
* update docstring for Client.best_containing
* extract test for Backend.leaves for reuse
* improve test blockchain_header_and_hash_return_blocks_from_canonical_chain_given_block_numbers
* extract test_blockchain_query_by_number_gets_canonical to easily test multiple impls
* remove whitespace
* remove todo
* Client.best_containing: pre-empt search loop when target in canonical
* best_containing: prevent race condition by holding import lock
* add todo
* extract leaf list update code into function
* add comment
* client-db: use in-memory-kvdb for tests
* use BTreeSet to store leaves for in-mem which is faster and simpler
* add docstring
* add comments and fix formatting
* add initial raw version of LeafSet
* remove Client::update_leaves which has been superceded by LeafSet
* use LeafSet in in-mem backend
* keccak -> blake2
* don't reexport codec traits in primitives
addresses https://github.com/paritytech/substrate/pull/740#discussion_r219538185
* fix rebase mistake
* improve LeafSet and use it in state-db
* correct Transfer nonces to fix ApplyExtinsicFailed(Stale)
* use given backend in canoncal test
* kill dead tree-route code in util
* fix warnings
* tests for leafset
* reorganizations in in_mem backend
* fix reorganization canon block logic
* DB commit and safe reversion on write error
* fix style nits
* state-db: typo fixes
* use fmt::Result for 'fmt'
* state-db: typo fixes
* state-db: use the same pattern to pass the parameter
* state-db: follow the same style
* state pruning after window even when nothing is finalized
* rename and alter canonicalization delay
* fix constant
* address grumbles
* add comment on canonicality vs finality
* finalization for in_mem
* fetch last finalized block
* pruning: use canonical term instead of final
* finalize blocks in full node
* begin to port light client DB
* add tree-route
* keep number index consistent in full nodes
* fix tests
* disable cache and finish porting light client
* add AsMut to system module
* final leaf is always best
* fix all tests
* Fix comment and trace
* removed unused Into call
* add comment on behavior of `finalize_block`
* move `tree_route` to client common
* tree_route tests
* return slices in TreeRoute
* apply finality up to parent
* add `finalize_block` call
* adjust formatting
* finality notifications and add last finalized block to chain info
* exhaustive match and comments
* fix sync tests by using non-instant finality
* finalization for in_mem
* fetch last finalized block
* pruning: use canonical term instead of final
* finalize blocks in full node
* begin to port light client DB
* add tree-route
* keep number index consistent in full nodes
* fix tests
* disable cache and finish porting light client
* add AsMut to system module
* final leaf is always best
* fix all tests
* Fix comment and trace
* removed unused Into call
* add comment on behavior of `finalize_block`
* Move module metadata from json string to custom metadata
* Revisit the metadata structures
1. Move the structures into the metadata crate.
2. Switch to using Cow/MaybeOwnedArray to support Encode/Decode
* Adapt to new metadata structures
* Convert event json metadata to new metadata structures
* Convert storage json metadata to new metadata structures
* Convert runtime metadata from json to new metadata structs
* Implements new metadata structures in client and runtime
* Fixes errors on `no_std`
* Fixes errors after rebasing master
* Do not use `Cow` anymore in metadata
Also replace `String` with our own type definition `StringBuf`.
This fixes compilation on `no_std`.
* Wrap `RuntimeMetadata` in `RuntimeMetadataVersioned` to support versioning
* Move metadata into `srml` and make core unaware of the implementation
* First effort
* API versioning
* Introduce validate_transaction
* Introduce the API plus fixes.
* Docs
* Typo
* Add longevity parameter to transaction validity info.
* Remove unneeded script
* Rename Substrate Demo -> Substrate
* Rename demo -> node
* Build wasm from last rename.
* Merge ed25519 into substrate-primitives
* Minor tweak
* Rename substrate -> core
* Move substrate-runtime-support to core/runtime/support
* Rename/move substrate-runtime-version
* Move codec up a level
* Rename substrate-codec -> parity-codec
* Move environmental up a level
* Move pwasm-* up to top, ready for removal
* Remove requirement of s-r-support from s-r-primitives
* Move core/runtime/primitives into core/runtime-primitives
* Remove s-r-support dep from s-r-version
* Remove dep of s-r-support from bft
* Remove dep of s-r-support from node/consensus
* Sever all other core deps from s-r-support
* Forgot the no_std directive
* Rename non-SRML modules to sr-* to avoid match clashes
* Move runtime/* to srml/*
* Rename substrate-runtime-* -> srml-*
* Move srml to top-level
snd