Fix flaky MultiAssetReturnCollateral test#6504
Conversation
c7cc8d9 to
bf80309
Compare
| -- Find the UTxO with tokens at wallet1 (for collateral) | ||
| txinCollateralWithTokensM <- | ||
| -- Find the UTxO with tokens at wallet2 (for collateral) | ||
| (txinCollateralWithTokens, collateralTxOut) <- retryUntilJustM epochStateView (WaitForBlocks 10) $ |
There was a problem hiding this comment.
The mint transaction may not be confirmed within a single block, causing findLargestMultiAssetUtxoWithAddress to return Nothing and the test to fail spuriously
When this happens? Why it may not get confirmed within a single block?
There was a problem hiding this comment.
Hard to say exactly, but maybe a block gets forged before the tx has time to enter the mempool. We then wait for that block and query the node state, but it still reflects the pre-tx state. The tx only gets included in the next block, which updates the node state but by then the test has already failed.
There was a problem hiding this comment.
This test should use integrationRetryWorkspace instead of integrationWorkspace
d36573e to
014595c
Compare
|
This PR is stale because it has been open 45 days with no activity. |
Summary
retryUntilJustMpolling (up to 10 blocks) for the multi-asset UTxO inhprop_collateral_with_tokensfindLargestMultiAssetUtxoWithAddressto returnNothingand the test to fail spuriouslyTest plan
Collateral With Multiassetstest on aarch64-linux (previously failing: check run 69208753482)🤖 Generated with Claude Code