fix(runner): fail tests on background context errors

[DudaGod]

I check pending WebdriverIO session-manager work before Testplane finalizes each test. A context update failure now fails the test even when it occurs after the last browser command. If the test has already failed, I keep that failure as the primary error and append the context error to its stack. I also mark the session broken so the browser pool closes it instead of reusing a potentially stale context.

The check is optional so Testplane remains compatible with currently published WebdriverIO versions that do not expose it. This PR should be paired with webdriverio#54 for the new behavior.

Validation:

• Testplane build, typecheck and lint passed.
• Testplane unit suite: 3595 passed, 1 pending.
• With both packages linked into a local reproduction, the two context-error tests failed normally (including an event after the final command), while an unrelated worker passed. No unhandled rejection occurred.

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/gemini-testing/testplane@1335

commit: 3b777d8

[github-actions]

✅ Testplane E2E run succeed

Report

[github-actions]

✅ Testplane browser-env run succeed

Report