You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Have you read the CONTRIBUTING guide on posting bugs, and CODE_OF_CONDUCT?
yes I read the things
This issue exists in the latest tap version
I am using the latest tap
Description
Hey!
We recently upgraded a large codebase from Node version 18.18.2 to 18.19.0 and encountered an odd issue. Some of our test suites are randomly getting stuck until a timeout forces them to terminate. We have confirmed that no other dependencies were altered during this process.
I've included a repository with a reproducible example.
To replicate the problem, simply run the 'make' command. After a while, you'll notice that the test suite gets stuck, waiting for the timeout.
Specifically, the destroy function is receiving a series of ids that are not present in domains (which typically contains only one id). Interestingly, when I manually change the id passed to domains.get through the debugger, the test suite completes successfully.
Additionally, it's worth noting that when we roll back to version 18.18.2, this issue seems to be resolved.
I feel like this might related to an issue of ts-node which doesn't play nicely with node 18.19.x and 20.x: TypeStrong/ts-node#1997
Hey,
Thanks for your answer,
If it could be helpful I've tried also disabling ts plugin (the repro I've submitted doesn't contain ts) and the problem persists.
I'll be switching to https://github.com/tapjs/tsimp instead of ts-node, but there's still a few kinks to comb out. In the meantime, yes, tsx is an option.
Is there an existing issue for this?
Have you read the
CONTRIBUTING
guide on posting bugs, andCODE_OF_CONDUCT
?This issue exists in the latest tap version
Description
Hey!
We recently upgraded a large codebase from Node version 18.18.2 to 18.19.0 and encountered an odd issue. Some of our test suites are randomly getting stuck until a timeout forces them to terminate. We have confirmed that no other dependencies were altered during this process.
I've included a repository with a reproducible example.
To replicate the problem, simply run the 'make' command. After a while, you'll notice that the test suite gets stuck, waiting for the timeout.
While attempting to debug the Node process in this stuck state, I discovered that the system seems to be looping at this particular line: https://github.com/tapjs/async-hook-domain/blob/9885b0d917abd053abf619ba8d37a9a57397ba0a/src/index.ts#L169.
Specifically, the destroy function is receiving a series of ids that are not present in
domains
(which typically contains only one id). Interestingly, when I manually change theid
passed todomains.get
through the debugger, the test suite completes successfully.Additionally, it's worth noting that when we roll back to version 18.18.2, this issue seems to be resolved.
Reproduction
https://github.com/pmarchini/tap-fastify-repro
Environment
EDIT: I've checked, and the async-hooks that I see during debugging (while stuck) are probably related to the timer.
The text was updated successfully, but these errors were encountered: