* fix(qa-channel): add timeout to qa-bus state guarded fetch
* test(qa-channel): add executable negative control for bus-state hang
* test(qa-channel): oxfmt negative-control bus-client test
* fix(qa-channel): sync SDK facade type with getQaBusState options parameter
Co-Authored-By: nebulacoder-v8.0 <noreply@zte.com.cn>
* test(qa-channel): add standalone production-path proof for bus-state timeout
Proof exercises the real getQaBusState code path against loopback
TCP peers that accept but never return HTTP headers:
- Negative control: fetchWithSsrFGuard without timeoutMs stays pending
- Positive control: getQaBusState with timeout rejects with TimeoutError
- Valid response: normal server still resolves within the timeout floor
Co-Authored-By: nebulacoder-v8.0 <noreply@zte.com.cn>
* test(qa-channel): log node version and head SHA in bus-state timeout proof
Co-Authored-By: nebulacoder-v8.0 <noreply@zte.com.cn>
* fix(qa-channel): drop test timeout from public Plugin SDK facade type
Keep the `timeoutMs` option internal to `getQaBusState` for test-only
short floors. The public facade callers do not need it and the SDK
surface should not expose test-only parameters.
Co-Authored-By: nebulacoder-v8.0 <noreply@zte.com.cn>
* chore(qa-channel): remove standalone proof script per review
The negative, timeout, and responsive cases are already durably covered in
bus-client.test.ts. The executed live output remains in the PR body as
evidence.
Co-Authored-By: nebulacoder-v8.0 <noreply@zte.com.cn>
* chore: trigger CI after rebase onto main
* chore: trigger CI after rebase onto main
* test(qa-channel): keep state timeout internal
---------
Co-authored-by: nebulacoder-v8.0 <noreply@zte.com.cn>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Co-authored-by: Peter Steinberger <peter@steipete.me>
d11f304519