External stacking locks #57
Conversation
maurolacy
force-pushed
the
external-stacking-locks
branch
from
efce10a
to
de4fdfb
Compare
maurolacy
force-pushed
the
external-stacking-locks
branch
from
de4fdfb
to
cef8ba0
Compare
There was a problem hiding this comment.
Looks like a good start. Add some detailed comments, but it is really minor stuff.
Updating the external-staking (and mock-external-staking) contracts is an important step for tests. The real one should just send back a message in the same tx to commit_tx. When there is IBC, we can change it, but at least you upgrade the vault <-> external staking communication to be "IBC safe"
| )?; | ||
|
|
||
| // Rollback user's max_lien | ||
| let mut user = self.users.load(ctx.deps.storage, &tx.user)?; |
There was a problem hiding this comment.
Yeah, this is kind of why I wouldn't apply it in the beginning, just lock and only apply on the success case.
Calculating maximum is harder when reducing (are there others higher?) vs raising (compare my new value with the max). This is an argument in favor of approach 2 above
maurolacy
force-pushed
the
external-stacking-locks
branch
3 times, most recently
from
8c5f3f3
to
4a251b6
Compare
maurolacy
force-pushed
the
external-stacking-locks
branch
from
3cefb74
to
1b59f93
Compare
| @@ -71,4 +81,24 @@ impl VaultApiHelper { | |||
| }; | |||
| Ok(wasm) | |||
| } | |||
|
|
|||
| pub fn commit_tx(&self, tx_id: u64) -> Result<WasmMsg, StdError> { | |||
There was a problem hiding this comment.
Good to add here, but question - isn't such a helper class now auto-generated?
Maybe a worthwhile PR to remove these and replace with mt autogen types (please do independent from this PR)
| } | ||
| } | ||
|
|
||
| pub fn next_tx_id(&self, store: &mut dyn Storage) -> StdResult<u64> { | ||
| let id: u64 = self.tx_count.may_load(store)?.unwrap_or_default() + 1; |
There was a problem hiding this comment.
Note: this may be the best use of update. You can use it as a return result directly.
(But it may require initializing the count to 0 in instantiate)
There was a problem hiding this comment.
Copied this from somewhere else. Will refactor if possible.
Using pending txs list
Using Lockable liens
Locked range fails with ParseError
Use enum instead of struct for generality / extensibility
maurolacy
force-pushed
the
external-stacking-locks
branch
from
73c8eea
to
8fd2445
Compare
Alternative to #54, but using
Lockablefor synchronisation.Proves that
Lockableis a useful abstraction around this.TODO:
commit_tx.rollback_tx.