walletFactory heap is still growing after u18 #10794
Assignees
Labels
Comments
mhofman
added
bug
|
Added this to U19 for nearterm tracking. Its a P2 already, so we can easily prioritize this out if needed. |
|
We're seeing some evidence that #10756 may not be fully fixed. The smart wallet is still growing some empty WeakRef for vpids in slotToVal. |
|
@mhofman fyi I will roll this over into U20 since its still open. |
mergify bot
pushed a commit
that referenced
this issue
Mar 13, 2025
refs: #10955 refs: #10794 ## Description The fast-usdc contract is leaking (#10955). One of the leak source is from the vow watch-utils allocating a heap Map to track vow resolution results, but that doesn't cleanup once settled. ### Security Considerations None ### Scaling Considerations Removes a heap leak ### Documentation Considerations None ### Testing Considerations Verified manually using [mhofman/fu-impact](dc-fu-impact...mhofman/fu-impact). I don't see how to unit test this besides creating an intrusive `Map` mock, and we're not structured to easily test heap size growth. ### Upgrade Considerations All vats using vow watch-utils would need to upgrade to pick up this fix. That includes any vat that uses orchestration's `zoeTools`. Using the regex `(?<!Promise\.)(\ballSettled\b|\ballVows\b|\ball\b)\(`, I have found: - fast-usdc - network - orchestration - smart-wallet
This was referenced Mar 13, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
and others
Describe the bug
Investigating the wallet factory leak in u17 and before (#10706), we identified an issue (#10757) and implemented a workaround for it (#10756), however from preliminary observations of the wallet factory heap in u18 on devnet, there is still a remaining leak.
To Reproduce
Observe the
uncompressedSizeofheap-snapshot-saveslog entries for the walletFactory vat.Expected behavior
Heap size flat over time
The text was updated successfully, but these errors were encountered: