Hopefully this can be merged if the tests pass.

When messing with the signature, we need to make sure that we acutally
mess-up the signature. As the generated private/public key is random,
the signature is random as well. It can happen that `bytes[0] ==
bytes[2]` which makes the test fail. We fix this problem by just
inverting the bytes at `0` and `2`.

* Fix call to backend in child_storage_hash

* child storage test.

This will hopefully reduce merge conflicts in the future

* block_import: switch to Box<dyn Any> for intermediates representation

* Use Cow and return Error instead of Option

* Remove unused error

* Distinguish NoIntermediate/InvalidIntermediate

This pulls in the latest release of *ring* and also removes some
duplicate dependencies.

* Adding Serde:Deserialise to UncheckExtrinsics

to be able to use the ChainApi RPC from the client side

* Update primitives/runtime/src/generic/unchecked_extrinsic.rs

Co-authored-by: Bastian Köcher <[email protected]>

* Initial version.

* Fix tests.

* Refactor using StorageValueRef.

* Add tests and apply review suggestions.

* Bump runtime.

Co-authored-by: Gavin Wood <[email protected]>

* Expose information about the extrinsic in the metadata

This pr exposes some information about the extrinsic used in the runtime
via metadata. The following information are exposed:

- Version of the extrinsic
- List of all signed extensions used by the extrinsic.

* Increment `spec_version`

It changes the way we extended the CLI functionalities of substrate to allow more flexibility. (If this was not clear, here is another version: it changes the `sc_cli` API to allow more flexibility).

This touches a few important things:
 - the startup of the async task with tokei:
    This was in node and node-template and I moved it to substrate. The idea is to have 1 time the code that handles unix signals (SIGTERM and SIGINT) properly. It is however possible to make this more generic to wait for a future instead and provide only a helper for the basic handling of SIGTERM and SIGINT.
 - increased the version of structopt and tokei
 - no more use of structopt internal's API
 - less use of generics

Related to #4643 and https://github.com/paritytech/cumulus/pull/42: the implementation of "into_configuration" and "get_config" are similar but with better flexibility so it is now possible in cumulus to have the command-line arguments only of the run command for polkadot if we want

Related to https://github.com/paritytech/cumulus/issues/24 and https://github.com/paritytech/cumulus/issues/34 : it will now be possible to make a configuration struct for polkadot with some overrides of the default parameters much more easily.

* Add `get_global` for `Sandbox`

This pr adds `get_global` to retrieve a `global` variable from an
instantiated sandbox wasm blob.

* Bump `spec_version`

* Update primitives/wasm-interface/src/lib.rs

Co-Authored-By: Sergei Pepyakin <[email protected]>

* `get_global` -> `get_global_val`

Co-authored-by: Sergei Pepyakin <[email protected]>
Co-authored-by: Gavin Wood <[email protected]>

* adding license fields to all crates

* Apply suggestions from code review

Co-Authored-By: Bastian Köcher <[email protected]>

Co-authored-by: Bastian Köcher <[email protected]>

* Rpc api for offchain storage

* Replace Vec<u8> to Bytes, replace Mutex to RwLock

* Remove pub

* Modify copyright year

* Add SS58 network identifier for Centrifuge

* Bump version

* Print extrinsic failed error always, not just on import

Before we printed the error of a transaction only when importing a
block, this pr changes it to print the error also at building the block.

* Increment `impl_version`

* WASM runtime switch to import memory

Up to now runtimes have exported their memory. To unify it with
sandboxing, this pr switches runtimes to import memory as well.

From a functional perspective, exporting/importing memory makes no
difference to the runtime.

To provide backwards compatibility, WASM exported memory is still supported.

* Revert debug stuff

* Revert some stuff

* Fixes for libp2p 0.15

* More work

* Update libp2p

* Update to libp2p 0.15

* transaction pool changes

* service & network changes

* address review

* reduce future pool

* instant/manual seal

unbounded queues are evil

Apply suggestions from code review

Co-Authored-By: Robert Habermeier <[email protected]>

add fork tests, docs, remove todos

moar docs

Update client/consensus/manual-seal/src/rpc.rs

Co-Authored-By: Robert Habermeier <[email protected]>

remove unbound generic, parameter, docs, deps, code style changes

Apply suggestions from code review

Co-Authored-By: Tomasz Drwięga <[email protected]>

code style chnges

remove unused deps, remove dep renames, check if block is empty before importing, use ? for error propagation

fix tests

log errors for instant seal

use debug

code style changes, updated copyright dates

use txpool::Pool instead of BasicPool, code style changes

fixed tests

* fix tests

* requested changes from review

* check inherents len

* rebase

* Introduces `author_hasKey` and `author_hasSessionKeys` rpc endpoints

Both endpoints can be used to check if a key is present in the keystore.

- `hasKey` works on with an individual public key and key type. It
checks if a private key for the given combination exists in the
keystore.
- `hasSessionKeys` works with the full encoded session key blob stored
on-chain in `nextKeys`. This requires that the given blob can be decoded
by the runtime. It will return `true`, iff all public keys of the
session key exist in the storage.

Fixes: https://github.com/paritytech/substrate/issues/4696

* Update client/rpc-api/src/author/error.rs

Co-Authored-By: Nikolay Volf <[email protected]>

* Indentation

Co-authored-by: Nikolay Volf <[email protected]>

* async/await in tx graph

* review notes

* remove unused typedef

* Reduction.

* Reformation.

* add locked timer stuff

* fix issues and introduce full pool

* arrange together

* fix benches

* fix new_light

* Add revalidation test case

* review fixes

* review fixes

* use just ready future

* address review

* consensus, pow: intermediate separation and fail

* Fix compiles

* Update primitives/consensus/common/src/block_import.rs

Co-Authored-By: Robert Habermeier <[email protected]>

* Update primitives/consensus/common/src/block_import.rs

Co-Authored-By: Robert Habermeier <[email protected]>

* Document what None means for `fork_choice` in block import params

Co-authored-by: Robert Habermeier <[email protected]>

Bastian Köcher authored Jan 22, 2020 and Sergey Pepyakin committed Jan 22, 2020

This moves the wasm-allocator (`FreeingBumpHeapAllocator`) into its own
crate `sp-allocator`. This new crate can theoretically provide multiple
different allocators. Besides moving the allocator, this pr also makes
`FreeingBumpHeapAllocator` compile on `no_std`.

* Support `u128`/`i128` in runtime interface

This implements support for `u128`/`i128` as parameters/return value in
runtime interfaces. As we can not pass them as identity, as for the
other primitives types, we pass them as an pointer to an `[u8; 16]` array.

* Remove some unsafe code usage

* Make debug builds more usable

This pr makes debug builds more usable in terms of `cargo run -- --dev`.

1. `--dev` activates `--execution native`, iff `--execution` is not
given or no sub `--execution-*` is given.
2. It was probably a mistake to compile WASM in debug for a debug build.
So, we now build the WASM binary always as `release` (if not requested
differently by the user). So, we trade compilation time for a better
debug experience.

* Make sure we only overwrite default values

* Make it work

* Apply suggestion

Some fixes after: https://github.com/paritytech/substrate/pull/4502

This removes the unwanted `expect`s from `MultiSigner`. Instead we
convert from full to compressed in `TryFrom` and can return an error on
invalid input.

Bastian Köcher authored Jan 19, 2020 and Shawn Tabrizi committed Jan 19, 2020

By using `cfg(doc)`, we can generate docs for modules which are only
available from `no_std`. This drastically improves the documentation for
the developers.

Adds a table to the rustdoc that shows how each individual type is
passed between the wasm and the host side.

* Drive by fix of doc of `Value`.

* Apply suggestions from code review

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

Co-authored-by: Bastian Köcher <[email protected]>
Co-authored-by: André Silva <[email protected]>

* Expose proof generation and verifying api.

* tabs to spaces

* bring back license comment

* Revert "tabs to spaces"

This reverts commit 4c3f72f9ef76b6a9f8988ed15b1bab17a9e51d2f.

* Formatting and docs nits

* Bump deps versions

* Upadte Cargo.lock

* into -> in

* Add typedefs for storage types

* Fix after merge

* DigestItem::ChangesTrieSignal

* introduce changes_trie::State

* introduce config activation block

* ChangesTrieSignal::as_new_configuration

* moved well_known_cache_keys to client

* extracted DbChangesTrieStorage to separate file

* change meaning of none in blockchain cache

* changes trie config (FULL) cache draft

* eliminating const ChangesTrieConfiguration

* delay pruning

* continue elimination

* do not prune CT config from cache

* removed redundant code

* fix some TODOs

* introduce ConfigurationRange

* use Configuration range in build

* build skewed digest

* remove debug print

* extracted surface iterator

* key_changes works with skewed digests

* fix client build

* add test for NeverPrune

* fix TODO

* fixed some TODOs

* more tests

* fixing TODOs

* fixed compilation

* update runtime version

* git rid of large tuple

* too long lines

* config_activation_block -> zero

* obsolete TODO

* removed unjustified expect

* update TODOs with issue number

* new CT pruning algorithm

fixed cache + multiple blocks finalization

track CT configuraiton on light clients

support CT configuration change revert

revert CT config test

new CT pruning algorithm

fixed cache + multiple blocks finalization

track CT configuraiton on light clients

support CT configuration change revert

revert CT config test

* BlockIdOrHeader isn't really required

* removed debug leftovers + some docs

* more docs

* more post-merge fixes

* more post-merge fixes

* revertes some unnecessary changes

* reverted unnecessary changes

* fix compilation + unnecessary changes

* (restart CI)

* fix cache update when finalizing multiple blocks

* fixed tests

* collect_extrinsics -> set_collect_extrinsics

* restore lost test

* do not calculate block number twice

* Update primitives/blockchain/src/error.rs

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

* map_err -> unwrap_or

* document get_at Result

* delete abandoned file

* added weight for set_changes_trie_config

* prefer_configs -> fail_if_disabled

* Update client/api/src/backend.rs

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

* Update client/db/src/changes_tries_storage.rs

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

* CommitOperation+merge -> CommitOperations

* fixed test compilation

* merged two different CTRange structs

* lost file

* uggrade db from v0 to v1 (init CT cache + add column)

* fix after merge

Co-authored-by: cheme <[email protected]>
Co-authored-by: Gavin Wood <[email protected]>

* Implement runtime version checks in `set_code`

Check that the new runtime code given to `set_code` fullfills some
requirements:

- `spec_name` matches
- `spec_version` does not decreases
- `impl_version` does not decreases
- Either `spec_version` and `impl_version` increase

* Make tests almost work

* Some fixes after master merge

* Fix tests

* Add missed file

* Make depedency check happy?

* Remove leftover `sc-executor`

* AHHHHH

* Reset debug stuff

* Remove some 'static

* More 'static

* Some docs

* Update `Cargo.lock`