-
Notifications
You must be signed in to change notification settings - Fork 684
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Inconsistent Transferable Balance Calculation Logic #1833
Comments
IN any case, an item missing (and rather high priority) in #226 is then migrate |
@ahmadkaouk Just to point that if the account cannot be reaped (which should be the case when there's a hold and/or freeze) or we don't want it to be reaped, the transferable balance calculation is this: |
Spendable balance indeed seems like it should be I will work on this. |
@ahmadkaouk I have a couple of questions:
After initial review, it seems was originated by the change in behavior of Prior implementation: polkadot-sdk/substrate/frame/balances/src/lib.rs Lines 1070 to 1083 in c699876
New implementation: polkadot-sdk/substrate/frame/balances/src/impl_fungible.rs Lines 44 to 70 in 495d24d
It seems only frozen and free are used in the prior implementation, I don't see usage of reserved anywhere. Thanks. |
#12951 did alter the functionality of existing reserves by design:
That Frontier is now showing some inconsistencies in its |
@gavofyork I have a clarifying question about the Here's the doc comment describing the behavior of the method: polkadot-sdk/substrate/frame/support/src/traits/tokens/fungible/regular.rs Lines 80 to 85 in c4211b6
Here is the polkadot-sdk/substrate/frame/balances/src/impl_fungible.rs Lines 44 to 70 in 495d24d
Is line 54 ( Otherwise, even when |
They're both correct. If Perhaps the confusion stems from the fact that we consider frozen in relation to the total, but end up calculating reducible by subtracting what is untouchable from what is free. |
Ah, thank you for clarifying. Visual aid:
|
@ahmadkaouk could you please confirm if updating the Is there any test case I can use to reproduce this? |
Yes, and you can add the ED into that, too:
so without anything frozen:
(or if we don't care about ED because we have other providers or will reap and don't have any consumers, then it's just |
@liamaharon, yes I believe that updating Here's a summary of the scenario that triggers the problem:
Balance info for the new account at the end:
Note that ED is 0 in this scenario. |
Hey @ahmadkaouk can try the |
We kind of need a unit test here in the SDK to prevent this from happening again and checking that #2292 does solve the issue. |
What helped me understand this convo and digram(s) better is to specify that |
Hey @liamaharon, just wanted to let you know that I tested the |
Part of #226 Related #1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by #1296, needs the `Unbalanced::decrease_balance` fix
@ahmadkaouk fyi this is finally merged and will be included in the next release. |
Part of #226 Related #1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by #1296, needs the `Unbalanced::decrease_balance` fix
Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix
Part of paritytech/polkadot-sdk#226 Related paritytech/polkadot-sdk#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech/polkadot-sdk#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75ac49786a7246531cf729b25c208cd38e6)
* Migrate fee payment from `Currency` to `fungible` (#2292) Part of paritytech/polkadot-sdk#226 Related paritytech/polkadot-sdk#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech/polkadot-sdk#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75ac49786a7246531cf729b25c208cd38e6) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0fcd37e4e8c14aeb83b5c9e680981e16079) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (#2292) Part of paritytech/polkadot-sdk#226 Related paritytech/polkadot-sdk#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech/polkadot-sdk#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75ac49786a7246531cf729b25c208cd38e6) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0fcd37e4e8c14aeb83b5c9e680981e16079) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (paritytech#2292) Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (paritytech#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (paritytech#2292) Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (paritytech#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (paritytech#2292) Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (paritytech#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (paritytech#2292) Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (paritytech#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (paritytech#2292) Part of paritytech#226 Related paritytech#1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by paritytech#1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (paritytech#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
* Migrate fee payment from `Currency` to `fungible` (#2292) Part of #226 Related #1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by #1296, needs the `Unbalanced::decrease_balance` fix (cherry picked from commit bda4e75) * Upgrade `trie-db` from `0.28.0` to `0.29.0` (#3982) - What does this PR do? 1. Upgrades `trie-db`'s version to the latest release. This release includes, among others, an implementation of `DoubleEndedIterator` for the `TrieDB` struct, allowing to iterate both backwards and forwards within the leaves of a trie. 2. Upgrades `trie-bench` to `0.39.0` for compatibility. 3. Upgrades `criterion` to `0.5.1` for compatibility. - Why are these changes needed? Besides keeping up with the upgrade of `trie-db`, this specifically adds the functionality of iterating back on the leafs of a trie, with `sp-trie`. In a project we're currently working on, this comes very handy to verify a Merkle proof that is the response to a challenge. The challenge is a random hash that (most likely) will not be an existing leaf in the trie. So the challenged user, has to provide a Merkle proof of the previous and next existing leafs in the trie, that surround the random challenged hash. Without having DoubleEnded iterators, we're forced to iterate until we find the first existing leaf, like so: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } println!("RECONSTRUCTED TRIE {:#?}", trie); // Create an iterator over the leaf nodes. let mut iter = trie.iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. let mut prev_key = None; for element in &mut iter { if element.is_ok() { let (key, _) = element.unwrap(); prev_key = Some(key); break; } } assert!(prev_key.is_some()); // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. assert!(prev_key.unwrap() <= challenge_hash.to_vec()); // The next element should exist (meaning there is no other existing leaf between the // previous and next leaf) and it should be greater than the challenged hash. let next_key = iter.next().unwrap().unwrap().0; assert!(next_key >= challenge_hash.to_vec()); ``` With DoubleEnded iterators, we can avoid that, like this: ```rust // ************* VERIFIER (RUNTIME) ************* // Verify proof. This generates a partial trie based on the proof and // checks that the root hash matches the `expected_root`. let (memdb, root) = proof.to_memory_db(Some(&root)).unwrap(); let trie = TrieDBBuilder::<LayoutV1<RefHasher>>::new(&memdb, &root).build(); // Print all leaf node keys and values. println!("\nPrinting leaf nodes of partial tree..."); for key in trie.key_iter().unwrap() { if key.is_ok() { println!("Leaf node key: {:?}", key.clone().unwrap()); let val = trie.get(&key.unwrap()); if val.is_ok() { println!("Leaf node value: {:?}", val.unwrap()); } else { println!("Leaf node value: None"); } } } // println!("RECONSTRUCTED TRIE {:#?}", trie); println!("\nChallenged key: {:?}", challenge_hash); // Create an iterator over the leaf nodes. let mut double_ended_iter = trie.into_double_ended_iter().unwrap(); // First element with a value should be the previous existing leaf to the challenged hash. double_ended_iter.seek(&challenge_hash.to_vec()).unwrap(); let next_key = double_ended_iter.next_back().unwrap().unwrap().0; let prev_key = double_ended_iter.next_back().unwrap().unwrap().0; // Since hashes are `Vec<u8>` ordered in big-endian, we can compare them directly. println!("Prev key: {:?}", prev_key); assert!(prev_key <= challenge_hash.to_vec()); println!("Next key: {:?}", next_key); assert!(next_key >= challenge_hash.to_vec()); ``` - How were these changes implemented and what do they affect? All that is needed for this functionality to be exposed is changing the version number of `trie-db` in all the `Cargo.toml`s applicable, and re-exporting some additional structs from `trie-db` in `sp-trie`. --------- Co-authored-by: Bastian Köcher <git@kchr.de> (cherry picked from commit 4e73c0f) * Update polkadot-sdk refs * Fix Cargo.lock --------- Co-authored-by: Liam Aharon <liam.aharon@hotmail.com> Co-authored-by: Facundo Farall <37149322+ffarall@users.noreply.github.com>
As the fix is merged, we can close this? |
Part of #226 Related #1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by #1296, needs the `Unbalanced::decrease_balance` fix
Part of #226 Related #1833 - Deprecate `CurrencyAdapter` and introduce `FungibleAdapter` - Deprecate `ToStakingPot` and replace usage with `ResolveTo` - Required creating a new `StakingPotAccountId` struct that implements `TypedGet` for the staking pot account ID - Update parachain common utils `DealWithFees`, `ToAuthor` and `AssetsToBlockAuthor` implementations to use `fungible` - Update runtime XCM Weight Traders to use `ResolveTo` instead of `ToStakingPot` - Update runtime Transaction Payment pallets to use `FungibleAdapter` instead of `CurrencyAdapter` - [x] Blocked by #1296, needs the `Unbalanced::decrease_balance` fix
An inconsistency has been identified in the Polkadot SDK concerning the logic used for computing transferable balances. This divergence may mislead users about the actual spendable balance, potentially leading to transaction failures due to inadequate fee coverage.
The inconsistency arises from the utilization of different formulas:
transferable = free - max(frozen, reserved)
.transferable = free - (frozen - reserved)
.This originates from the
CurrencyAdapter
, which relies on the outdatedCurrency
trait as illustrated here. The new transferable balance is computed utilizing thereducible_balances
function defined here.The discrepancy is particularly noticeable in the
eth.getBalance
method in Frontier, where the new formula is employed. Under certain circumstances, a positive balance may be displayed, indicating available funds. However, transactions may fail due to an insufficient balance to cover the fees, given that the old formula is utilized for fee calculations.The text was updated successfully, but these errors were encountered: