implement wait() method for QiTx response #317
Conversation
There was a problem hiding this comment.
The wait mechanism looks good to go.
I would like to avoid changing any of the interfaces that force passing in unnecessary inputs like hash and zone when they are not needed. I know that this was done because Qi transactions where not producing correct QIP10 hashes and this was the only way to get this to work but we need to test again against latest go-quai and remove whatever is not needed.
src/providers/abstract-provider.ts
Outdated
| } | ||
| case 'qiTransaction': { | ||
| assert(zone != null, 'zone is required for qiTransaction event', 'MISSING_ARGUMENT'); | ||
| assert(hash != null, 'hash is required for qiTransaction event', 'MISSING_ARGUMENT'); | ||
| return { type: _event, tag: _event, hash, zone }; | ||
| } | ||
| } | ||
| } | ||
|
|
||
| if (isHexString(_event, 32)) { | ||
| const hash = _event.toLowerCase(); | ||
| zone = toZone(hash.slice(0, 4)); | ||
| zone = zone ?? toZone(hash.slice(0, 4)); | ||
| return { type: 'transaction', tag: getTag('tx', { hash }), hash, zone }; | ||
| } |
There was a problem hiding this comment.
Why did we add an extra hash argument to the getSubscription method if we are already handling hash events for regular transactions? Cant we just use that same pattern?
There was a problem hiding this comment.
I have re implemented the Qi Tx wait() logic under the assumption that the tx input hash is permuted following QIP10.
src/providers/abstract-provider.ts
Outdated
| @@ -1674,8 +1686,8 @@ export class AbstractProvider<C = FetchRequest> implements Provider { | |||
| return this._wrapBlock(params, network); | |||
| } | |||
|
|
|||
| async getTransaction(hash: string): Promise<null | TransactionResponse> { | |||
| const zone = toZone(this.shardFromHash(hash)); | |||
| async getTransaction(hash: string, _zone?: Zone): Promise<null | TransactionResponse> { | |||
There was a problem hiding this comment.
Let's revisit this once go-quai and the genallocs get updated. I believe the core team is going to enforce QIP10 hashes so we should be able to rely on zone derived from hash and drop the zone input.
There was a problem hiding this comment.
I have re implemented the Qi Tx wait() logic under the assumption that the tx input hash is permuted following QIP10.
alejoacosta74
force-pushed
the
qitx-wait.v2
branch
from
5be3d89 to
095346e
Compare
No description provided.