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
When I run my tests with datatest-stable (0.2.9), using nextest as the runner, a semi-random number of tests are marked as leaky.
This seems to only depend on how many tests the crate has, and not what the tests are doing.
I have an example repo here: https://github.com/zaneduffield/datatest_leak_issue
That repo has 1000 test files, which are used with a no-op test function.
When I run cargo nextest run on it, about 40% of the tests are marked as 'leaky'.
One issue that may be related is how slow this is: ~17s on my machine compared to ~0.2s with cargo test.
The text was updated successfully, but these errors were encountered:
Maybe this is related to AV getting in the way (https://nexte.st/book/antivirus-gatekeeper). Even if that is the case, I'm not sure why the tests would be considered leaky.
Thanks for the report. It does seem like antivirus getting in the way. Have you tried the new "dev drive" feature on Windows?
Leak detection works by monitoring the stdout and stderr file descriptors after the file has closed. If antivirus or something else is keeping the fds open, this issue will be hit. You can bump up the timeout with the "leak-timeout" parameter.
I think this isn't really a datatest-stable issue, more of a nextest issue. Moving it there.
sunshowers
changed the title
Tests are considered 'leaky' when run with nextest
[windows] lots of tests running in parallel may cause "leaky" warnings
Apr 26, 2024
When I run my tests with
datatest-stable
(0.2.9), usingnextest
as the runner, a semi-random number of tests are marked as leaky.This seems to only depend on how many tests the crate has, and not what the tests are doing.
I have an example repo here: https://github.com/zaneduffield/datatest_leak_issue
That repo has 1000 test files, which are used with a no-op test function.
When I run
cargo nextest run
on it, about 40% of the tests are marked as 'leaky'.One issue that may be related is how slow this is: ~17s on my machine compared to ~0.2s with
cargo test
.The text was updated successfully, but these errors were encountered: