* Do not call `initialize_block` before any runtime api
Before this change we always called `initialize_block` before calling
into the runtime. There was already support with `skip_initialize` to skip
the initialization. Almost no runtime_api requires that
`initialize_block` is called before. Actually this only leads to higher
execution times most of the time, because all runtime modules are
initialized and this is especially expensive when the block contained a
runtime upgrade.
TLDR: Do not call `initialize_block` before calling a runtime api.
* Change `validate_transaction` interface
* Fix rpc test
* Fixes and comments
* Some docs
* Move alloc primitive (not used in /pallets)
* Move to alternative location as not shared
* moved crates to different dir
* ren sp_chain_spec to sc_chain_spec_primatives
* merged sc-chain-spec and moved allocation up one.
* no no_std
* nudge
* Bump CI
* Make it possible to override maximum payload of RPC
* Finish it.
* remove todo.
* Update client/cli/src/commands/run_cmd.rs
* Apply suggestions from code review
Co-authored-by: David <dvdplm@gmail.com>
* Apply suggestions from code review
Co-authored-by: David <dvdplm@gmail.com>
* Incorporate suggestions
* Thread rpc_max_payload from configuration to trace_block
* Try obey line gitlab/check_line_width.sh
* update state rpc tests
* Improve readbility
* Apply suggestions from code review
* Apply suggestions from code review
Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com>
Co-authored-by: Zeke Mostov <32168567+emostov@users.noreply.github.com>
Co-authored-by: David <dvdplm@gmail.com>
This instructs the Aura runtime api to skip initialize block, when
requesting the authorities. This is important, as we don't want to use
the new authorities that should be used from the next block on.
Besides that, it removes the caching stuff. The cache is not available
on full nodes anyway. In the future we should store the authorities
probably in the aux store.
* Decommit instance memory after a runtime call on Linux
* Update documentation for the test
* Remove unfinished comment
* Use saturating_sub.
Also update the doc comment.
* Precise RSS tracking in the test
Instead of tracking RSS for the whole process we just look at the particular mapping that is associated with the linear memory of the runtime instance
* Remove unused import
* Fix unused imports
* Fix the unused imports error for good
* Rollback an accidental change to benches
* Fix the test
* Remove now unneeded code
* Transaction pool: Remove futures-diagnose and thread pool
This pr removes `futures-diagnose` as this isn't used anymore. Besides
that the pr also removes the thread pool that was used to validate the
transactions in the background. Instead of this thread pool we now spawn
two separate long running tasks that we use to validate the
transactions. All tasks of the transaction pool are now also spawned as
essential tasks. This means, if any of these tasks is stopping, the node
will stop as well.
* Update client/transaction-pool/src/api.rs
* Migrate ProfilingLayer to tracing registry API
* Remove the `current_span` field from `BlockSubscriber`.
* Bump the `tracing-subscriber` version
* Fix Gitlab CI
* validation extension in sp_io
* need paths
* arc impl
* missing host function in executor
* io to pkdot
* decode function.
* encode primitive.
* trailing tab
* multiple patch
* fix child trie logic
* restore master versionning
* bench compact proof size
* trie-db 22.3 is needed
* line width
* split line
* fixes for bench (additional root may not be needed as original issue was
with empty proof).
* revert compact from block size calculation.
* New error type for compression.
* Adding test (incomplete (failing)).
Also lacking real proof checking (no good primitives in sp-trie crate).
* There is currently no proof recording utility in sp_trie, removing
test.
* small test of child root in proof without a child proof.
* remove empty test.
* remove non compact proof size
* Missing revert.
* proof method to encode decode.
* Transaction pool: Ensure that we prune transactions properly
There was a bug in the transaction pool that we didn't pruned
transactions properly because we called `prune_known`, instead of `prune`.
This bug was introduced by:
https://github.com/paritytech/substrate/pull/4629
This is required to have stale extrinsics being removed properly, so
that they don't fill up the tx pool.
* Fix compilation
* Fix benches
* ...
`ready_at` returns when we have processed the requested block. However,
on startup we already have processed the best block and there
are no transactions in the pool on startup anyway. So, we can set `updated_at`
to the best block on startup.
Besides that `ready_at` now returns early when there are no ready nor
any future transactions in the pool.
* Update wasmtime to 0.27
A couple of notes:
- Now we are fair about unsafeness of runtime creation via an compiled artifact.
This change was prompted by the change in wasmtime which made
`deserialize` rightfully unsafe. Now `CodeSupplyMode` was hidden and
the `create_runtime` now takes the blob again and there is now a new
fn for creating a runtime with a compiled artifact.
- This is a big change for wasmtime. They switched to the modern backend
for code generation. While this can bring performance improvements, it
can also introduce some problems. In fact, 0.27 fixed a serious issue
that could lead to sandbox escape. Hence we need a proper burn in.
This would require a change to PVF validation host as well.
* Filter regalloc logging
Squirrel