feat(ci): Add support for test_timeout option

[MarcoPolo]

Description

Node and browser tests take a little bit longer to spin up in the interop tests. The multidim-interop can tell the tests to increase the timeout when testing against these targets, but the test itself needs to use the passed in value.

This updates the test code to handle the new test_timeout option. See libp2p/test-plans#107

Notes

Links to any relevant issues

Open Questions

Change checklist

• I have performed a self-review of my own code
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• A changelog entry has been made in the appropriate crates

[mxinden]

Following up on a recent conversation, is the configuration parameter needed? The alternative would be to have a static large timeout for all tests. I would assume we optimize running-time for the happy case only. I would expect most pull requests to pass the tests (happy case) and only a small subset to fail. As a developer I care about the majority of my pull requests to be tested quickly. I don't care so much that CI runs longer on buggy pull requests, given that they are rare.

I am not deeply involved here. Please feel free to ignore. Change looks good to me. Feel free to add the send-it label.

[mergify]

This pull request has merge conflicts. Could you please resolve them @MarcoPolo? 🙏

[thomaseizinger]

I think having the timeout configurable is overall a reasonable feature. IMO, one thing that we should keep in the back of our mind is that ideally, all of the different test binaries that we will be building should have the same or at least consistent interface. That would allow us to build shared abstractions between them and makes things overall easier to understand.

[MarcoPolo]

same or at least consistent interface

Agreed. Right now this is still in the evolving phase, but soon I'll write up what this interface is.

Following up on a recent conversation, is the configuration parameter needed?

This is not needed. We have the reasonable default of 10s. This lets us extend this timeout explicitly when we need to. I rather make it explicit that NodeJS and the Browser tests need more time. It would be surprising if Go/Rust hit these timeouts (Maybe the CI worker is overloaded or affected by a noisy neighbor?).

This is really a small point, and doesn't have much bearing either way.

[MarcoPolo]

Rebased off master. Sorry for doing this right after your approval @thomaseizinger

[thomaseizinger]

Rebased off master. Sorry for doing this right after your approval @thomaseizinger

No worries. Unless there were conflicts, there is no need as the merge bot will update the PR! :)