Description
I've been trying to use Web Test Runner with our Bazel build system. This means Web Test Runner may be invoked multiple times, executing multiple tests concurrently, with isolated instances of WTR.
This worked great overall, but we've noticed some instability/flakiness with the Puppeteer/web/test-runner-chome launcher and it was extremely hard to debug and required lots of patching of @web/test-runner-core to actually see what is going on. The problem is that, even in non TTY mode, with static logging option enabled, the dynamic terminal logic was swallowing the errors— likely due to the regular clears() or dynamic logging that is still happening even with staticLogging: true.
It would be great to improve the logic around logging here, to make it more CI/non TTY friendly.