fix browser: ignore context cancellation during target attachment - Close #5504

[LBaronceli]

What?

Fixes a race during browser shutdown where Browser.onAttachedToTarget could attempt to create new pages or sessions after the browser context had already been canceled.

The change ensures that context cancellation during teardown is handled gracefully and does not surface as a fatal error or panic.

Why?

Under high concurrency (e.g. repeated test runs or low GOMAXPROCS), Target.attachedToTarget events can still be delivered while the browser is shutting down.

In this situation, attempting to create new pages leads to panics or test failures, even though shutdown is expected behavior.

This change aligns Browser.onAttachedToTarget with existing defensive handling elsewhere in the browser codebase by treating context cancellation as a non-fatal condition.

Checklist

• I have performed a self-review of my code.
• I have commented on my code, particularly in hard-to-understand areas.
• I have added tests for my changes.
• [] I have run linter and tests locally (make check) and all pass.
• Note: make check reports existing lint issues in unrelated packages that are already present on master. This PR intentionally does not address them to keep the fix narrowly scoped.

Checklist: Documentation (only for k6 maintainers and if relevant)

Please do not merge this PR until the following items are filled out.

• I have added the correct milestone and labels to the PR.
• I have updated the release notes: link
• I have updated or added an issue to the k6-documentation: grafana/k6-docs Context canceling causing browser integration tests to fail very often #5504 if applicable
• I have updated or added an issue to the TypeScript definitions: grafana/k6-DefinitelyTyped#NUMBER if applicable

Related PR(s)/Issue(s)

Closes #5504

[CLAassistant]

All committers have signed the CLA.

[ankur22]

@LBaronceli we've had some discussions and I've played around with your change, and indeed it helps us avoid a panic which can cause various other issues in the integration tests (data races) as well as when running k6 normally (panic causing the test to fail unnecessarily).

Could you take a look at the error that the linter is pointing to and resolve it? I believe we should be able to work with something like:

err := ...
if err is context canceled {
    log
    continue
}
if err not nil {
    panic
}

WDYT?

[LBaronceli]

Hey Ankur,

Sure! made sure the if statements aren't nested, and added a debug log in case the continue is triggered by the Canceled or DealineExceeded errors.

Thank you for you patience and help with this one!

[ankur22]

Brilliant! Thanks for digging into this @LBaronceli and helping us with the fix 🙇 🚀