-
Notifications
You must be signed in to change notification settings - Fork 2.6k
Kill the light client, CHTs and change tries. #10080
Conversation
This should be unblocked now. #9684 has landed and now the diff looks a fair bit smaller. |
b0e1364
to
8596e34
Compare
self.storage_changes_root(parent_hash) | ||
.expect("Invalid `parent_hash` given to `changes_root`.") | ||
/// Always returns `None`. This function exists for compatibility reasons. | ||
fn changes_root(&mut self, _parent_hash: &[u8]) -> Option<Vec<u8>> { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bkchr Is there a way to remove this without breaking compatibility with existing chains? E.g. when loading polkadot genesis wasm, we check that the host function must exist, even though it is never called.
@@ -69,7 +69,6 @@ struct InnerValue<V> { | |||
/// Current value. None if value has been deleted. | |||
value: V, | |||
/// The set of extrinsic indices where the values has been changed. | |||
/// Is filled only if runtime has announced changes trie support. | |||
extrinsics: Extrinsics, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should we just remove this Extrinsics
field (I remember thinking of some optimization to it, but not having it is a better optimization).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thinks this is also involved in supporting transactions in the runtime. I.e. to revert storage changes when extrinsic fails. In any case I'd rather do complex logic changes and optimisations in a separate PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
+1 for separate, IIRC transaction stuff just merge extrinsic when commiting (it is just additional computation).
@andresilva @bkchr @sorpaas @adoerr Would be fantastic to see this land soon, anything we can do to help clear up the remaining doubts? :) |
@joao-paulo-parity Not sure why the companion check fails here with |
@arkpar I think we can ignore. Just let us merge this and then update the companion using the old school way ;) |
* master: (27 commits) Bump rustversion from 1.0.4 to 1.0.5 (#10243) Kill the light client, CHTs and change tries. (#10080) tuple to struct event variants (#10206) Bump thiserror from 1.0.26 to 1.0.30 (#10240) Warn about usage of pallet collective set members call. (#10156) Bump git2 from 0.13.22 to 0.13.23 (#10238) Add group name in task metrics (#10196) Bump proc-macro-crate from 1.0.0 to 1.1.0 (#10237) Bump parity-util-mem from 0.10.0 to 0.10.2 (#10236) Bump substrate-bip39 from 0.4.2 to 0.4.4 (#10213) Upgrade jsonrpsee to v0.4.1 (#10022) expose substrate-cli service (#10229) Intend to reactivate cargo-unleash check (#10167) CI: build docs with deps (#9884) use CountedMap in pallet-bags-list (#10179) Move all example pallets under `examples` folder. (#10215) Upgrade wasm builder (#10226) upgrade ss58-registry with additional networks. (#10224) move wiki -> docs (#10225) new remote-ext mode: (#10192) ...
* Update cargo.lock * rust version 2021 * Reflect paritytech/substrate#10080 * all builds * more changes related to DigestItem * service changes
* Remove light client, change tries and CHTs * Update tests * fmt * Restore changes_root * Fixed benches * Cargo fmt * fmt * fmt
* Remove light client, change tries and CHTs * Update tests * fmt * Restore changes_root * Fixed benches * Cargo fmt * fmt * fmt
This PR removes:
polkadot companion: paritytech/polkadot#4191