Initial Substrate Header Chain Implementation #296
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
tomusdrw
reviewed
Aug 20, 2020
I've been struggling with getting the generic types between the storage and verifier traits to play nice with each other. As a way to continue making progress I'm moving everything to the pallet. This way I hope to make progress towards a functional pallet.
Fixes a problem with how the scheduled change was counted as well as a SCALE encoding issue
There's no guarantee that the consensus digest item will be the last one in a header, which is how it was previously being checked. Thanks to Andre for pointing me to the Grandpa code that does this.
When a header that finalizes a chain of headers is succesfully imported we also want to mark its ancestors as finalized.
This should be added back at some later point in time, but right now the code doesn't allow for this behaviour.
This is better than using `get().is_some()` since we skip decoding the storage value
|
Based off your review comments, here is a list of things left for other PRs:
This is on top of the changes I mentioned I wanted to address in the PR description. |
HCastano
changed the title
tomusdrw
approved these changes
Sep 23, 2020
There was a problem hiding this comment.
Looks good! Given we have a list of things that are yet to be done, I'm down to merge as-is and parallelize efforts on this.
@HCastano please create a set of issues that we (Slava and I) can assign to and create a bunch of smaller PRs.
This was referenced Sep 23, 2020
|
@tomusdrw I've created some issues which I think can be picked up independently from each other. I didn't create issues for everything since some of these things are either too small, or too abstract to be nicely put into an issue. |
svyatonik
pushed a commit
that referenced
this pull request
Jul 17, 2023
* rename parachain-{upgrade -> system}
* Merge message-broker into parachain-system
* Remove message-broker and clean up
* Update docs
* Test upward messages sending
And also update the relay-sproof-builder so that it allows to set the
relay dispatch queue size for the given parachain.
* Test horizontal message sending
* Remove old inherent definitions
serban300
pushed a commit
to serban300/parity-bridges-common
that referenced
this pull request
Mar 27, 2024
* Add pallet template from Substrate Dev Hub * Clean up un-needed stuff from template * Sketch out dispatchable interface * Introduce notion of finality chain * Add dependencies which were removed during a rebase * Sketch out idea for finality header-chain pallet * Sketch out ChainVerifier trait * Add storage parameter to verifier * Write out some things I think I need for finality verification * Add some pseudocode for marking finalized headers * Remove parity_scale_codec duplicate * Move verification logic into pallet I've been struggling with getting the generic types between the storage and verifier traits to play nice with each other. As a way to continue making progress I'm moving everything to the pallet. This way I hope to make progress towards a functional pallet. * Start doing verification around authority set changes * Remove commented BridgeStorage and ChainVerifier traits * Create Substrate bridge primitives crate * Add logic for updating scheduled authority sets * Introduce notion of imported headers * Implement basic header ancestry checker * Add mock runtime for tests * Add testing boilerplate * Add some storage read/write sanity tests * Add some basic header import tests * Add tests for ancestry proofs * Create helper for changing authority sets * Fix authority set test Fixes a problem with how the scheduled change was counted as well as a SCALE encoding issue * Correctly check for scheduled change digests There's no guarantee that the consensus digest item will be the last one in a header, which is how it was previously being checked. Thanks to Andre for pointing me to the Grandpa code that does this. * Mark imported headers as finalized when appropriate When a header that finalizes a chain of headers is succesfully imported we also want to mark its ancestors as finalized. * Add helper for writing test headers * Add test helper for scheduling authority set changes * Bump Substrate pallet and primitives to rc6 * Remove Millau verifier implementation * Add some doc comments * Remove some needless returns * Make Clippy happy * Split block import from finalization * Make tests compile again * Add test for finalizing header after importing children * Create a test stub for importing future justifications * Start adding genesis config * Reject justifications from future We should only be accepting justifications for the header which enacted the current authority set. Any ancestors of that header which require a justification can be imported but they must not be finalized. * Add explanation to some `expect()` calls * Start adding GenesisConfig * Plug genesis config into runtime * Remove tests module * Check for overflow when updating authority sets * Make verifier take ownership of headers during import * Only store best finalized header hash Removed the need to store the whole header, since we store it was part of the ImportedHeaders structure anyways * Add some helpers to ImportedHeader * Update ancestry checker to work with ImportedHeaders * Update ancestry tests to use ImportedHeaders * Update import tests to use ImportedHeaders * Clean up some of the test helpers * Remove stray dbg! * Add doc comments throughout * Remove runtime related code * Fix Clippy warnings * Remove trait bound on ImportedHeader struct * Simplify checks in GenesisConfig * Rename `get_header_by_hash()` * Alias `parity_scale_codec` to `codec` * Reword Verifier documentation * Missed codec rename in tests * Split ImportError into FinalizationError * Remove ChainVerifier trait This trait was a remenant of the original design, and it is not required at the moment. Something like it should be added back in the future to ensure that other chains which conform to this interface can be used by higher-level bridge applications. * Fix the verifier tests so they compile * Implement Deref for ImportedHeader * Get rid of `new` methods for some Substrate primitives * Ensure that a child header's number follows its parent's * Prevent ancestry checker from aimlessly traversing to genesis If an ancestor which was newer than the child header we were checking we would walk all the way to genesis before realizing that we weren't related. This commit fixes that. * Remove redundant clones * Ensure that old headers are not finalized Prevents a panic where if the header being imported and `best_finalized` were the same header the ancestry checker would return an empty list. We had made an assumption that the list would always be populated, and if this didn't hold we would end up panicking. * Disallow imports at same height as `best_finalized` * Fix Clippy warnings * Make NextScheduledChange optional * Rework how scheduled authority set changes are enacted We now require a justification for headers which _enact_ changes instead of those which _schedule_ changes. A few changes had to be made to accomodate this, such as changing when we check for scheduled change logs in incoming headers. * Update documentation for Substrate Primitives * Clarify why we skip header in requires_justification check * Add description to assert! call * Fix formatting within macros * Remove unused dependencies from runtime * Remove expect call in GenesisConfig * Turn FinalityProof into a struct * Add some inline TODOs for follow up PRs * Remove test which enacted multiple changes This should be added back at some later point in time, but right now the code doesn't allow for this behaviour. * Use `contains_key` when checking for header This is better than using `get().is_some()` since we skip decoding the storage value * Use initial hash when updating best_finalized * Add better checks around enacting scheduled changes * Rename finality related functions * Appease Clippy
serban300
pushed a commit
to serban300/parity-bridges-common
that referenced
this pull request
Apr 8, 2024
* Add pallet template from Substrate Dev Hub * Clean up un-needed stuff from template * Sketch out dispatchable interface * Introduce notion of finality chain * Add dependencies which were removed during a rebase * Sketch out idea for finality header-chain pallet * Sketch out ChainVerifier trait * Add storage parameter to verifier * Write out some things I think I need for finality verification * Add some pseudocode for marking finalized headers * Remove parity_scale_codec duplicate * Move verification logic into pallet I've been struggling with getting the generic types between the storage and verifier traits to play nice with each other. As a way to continue making progress I'm moving everything to the pallet. This way I hope to make progress towards a functional pallet. * Start doing verification around authority set changes * Remove commented BridgeStorage and ChainVerifier traits * Create Substrate bridge primitives crate * Add logic for updating scheduled authority sets * Introduce notion of imported headers * Implement basic header ancestry checker * Add mock runtime for tests * Add testing boilerplate * Add some storage read/write sanity tests * Add some basic header import tests * Add tests for ancestry proofs * Create helper for changing authority sets * Fix authority set test Fixes a problem with how the scheduled change was counted as well as a SCALE encoding issue * Correctly check for scheduled change digests There's no guarantee that the consensus digest item will be the last one in a header, which is how it was previously being checked. Thanks to Andre for pointing me to the Grandpa code that does this. * Mark imported headers as finalized when appropriate When a header that finalizes a chain of headers is succesfully imported we also want to mark its ancestors as finalized. * Add helper for writing test headers * Add test helper for scheduling authority set changes * Bump Substrate pallet and primitives to rc6 * Remove Millau verifier implementation * Add some doc comments * Remove some needless returns * Make Clippy happy * Split block import from finalization * Make tests compile again * Add test for finalizing header after importing children * Create a test stub for importing future justifications * Start adding genesis config * Reject justifications from future We should only be accepting justifications for the header which enacted the current authority set. Any ancestors of that header which require a justification can be imported but they must not be finalized. * Add explanation to some `expect()` calls * Start adding GenesisConfig * Plug genesis config into runtime * Remove tests module * Check for overflow when updating authority sets * Make verifier take ownership of headers during import * Only store best finalized header hash Removed the need to store the whole header, since we store it was part of the ImportedHeaders structure anyways * Add some helpers to ImportedHeader * Update ancestry checker to work with ImportedHeaders * Update ancestry tests to use ImportedHeaders * Update import tests to use ImportedHeaders * Clean up some of the test helpers * Remove stray dbg! * Add doc comments throughout * Remove runtime related code * Fix Clippy warnings * Remove trait bound on ImportedHeader struct * Simplify checks in GenesisConfig * Rename `get_header_by_hash()` * Alias `parity_scale_codec` to `codec` * Reword Verifier documentation * Missed codec rename in tests * Split ImportError into FinalizationError * Remove ChainVerifier trait This trait was a remenant of the original design, and it is not required at the moment. Something like it should be added back in the future to ensure that other chains which conform to this interface can be used by higher-level bridge applications. * Fix the verifier tests so they compile * Implement Deref for ImportedHeader * Get rid of `new` methods for some Substrate primitives * Ensure that a child header's number follows its parent's * Prevent ancestry checker from aimlessly traversing to genesis If an ancestor which was newer than the child header we were checking we would walk all the way to genesis before realizing that we weren't related. This commit fixes that. * Remove redundant clones * Ensure that old headers are not finalized Prevents a panic where if the header being imported and `best_finalized` were the same header the ancestry checker would return an empty list. We had made an assumption that the list would always be populated, and if this didn't hold we would end up panicking. * Disallow imports at same height as `best_finalized` * Fix Clippy warnings * Make NextScheduledChange optional * Rework how scheduled authority set changes are enacted We now require a justification for headers which _enact_ changes instead of those which _schedule_ changes. A few changes had to be made to accomodate this, such as changing when we check for scheduled change logs in incoming headers. * Update documentation for Substrate Primitives * Clarify why we skip header in requires_justification check * Add description to assert! call * Fix formatting within macros * Remove unused dependencies from runtime * Remove expect call in GenesisConfig * Turn FinalityProof into a struct * Add some inline TODOs for follow up PRs * Remove test which enacted multiple changes This should be added back at some later point in time, but right now the code doesn't allow for this behaviour. * Use `contains_key` when checking for header This is better than using `get().is_some()` since we skip decoding the storage value * Use initial hash when updating best_finalized * Add better checks around enacting scheduled changes * Rename finality related functions * Appease Clippy
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR introduces the first version of the Substrate bridge pallet. The idea behind this pallet is that it mainly only cares about two things:
Using this simple interface it can accept headers from a different Substrate chain and keep track of their status. Higher level applications that build on this will mainly only care about headers that have been finalized.
Some notable things which will be addressed in follow-ups include:
chain_spec)I removed some of the code to wire the pallet into the runtime because of the changes introduced in #341. I'll add that back in later.
Note, that while this PR looks kinda big, I think a decent chunk of the line count comes from the Verifier tests. These are kinda verbose at the moment, but that's also something that can be improved in the future.