test(e2e): dump full URL/trail/request diagnostic when setup times out #104

Merged

nrupard merged 1 commit from fix/e2e-bearer-diagnostics into main

2026-06-05 18:18:47 +02:00

nrupard commented

2026-06-05 18:17:21 +02:00

Owner

Summary

Two consecutive E2E runs reported (none) for /api/v1 requests after a successful SPA login. The previous error wording only said "no Bearer", which is uninformative when ZERO requests fire: we can't tell whether the SPA bounced through bunyip incorrectly, landed on a non-hydrating page, hits a different API base, or has stopped using Bearer auth entirely.

This PR replaces the terse timeout with a full diagnostic dump:

postLoginUrl (where loginViaSpa thought it landed)
finalUrl (where the page is at timeout)
urlTrail (every top-frame navigation since setup started)
All /api/v1 request URLs observed
ALL request URLs observed (any host, any path)

Lists capped at the last 30 entries. No behavior change beyond error formatting.

Test plan

Post-merge run: if setup still times out, the error log identifies which host/path the SPA actually hit, so the next iteration is precise instead of speculative.

## Summary Two consecutive E2E runs reported `(none)` for `/api/v1` requests after a successful SPA login. The previous error wording only said "no Bearer", which is uninformative when ZERO requests fire: we can't tell whether the SPA bounced through bunyip incorrectly, landed on a non-hydrating page, hits a different API base, or has stopped using Bearer auth entirely. This PR replaces the terse timeout with a full diagnostic dump: - `postLoginUrl` (where `loginViaSpa` thought it landed) - `finalUrl` (where the page is at timeout) - `urlTrail` (every top-frame navigation since setup started) - All `/api/v1` request URLs observed - ALL request URLs observed (any host, any path) Lists capped at the last 30 entries. No behavior change beyond error formatting. ## Test plan - [ ] Post-merge run: if setup still times out, the error log identifies which host/path the SPA actually hit, so the next iteration is precise instead of speculative.

nrupard added 1 commit

2026-06-05 18:17:21 +02:00

test(e2e): dump full URL/trail/request diagnostic when setup times out

Create release / Create release from merged PR (pull_request) Has been skipped

Details

Check / fmt + clippy + compile + tests (pull_request) Successful in 1m24s

Details

b30dc74690

Two consecutive CI runs reported \`(none)\` /api/v1 requests observed after a successful SPA login. The previous error message only said "no Bearer found", which is uninformative when ZERO requests fire - we can't tell whether the SPA bounced through bunyip incorrectly, landed on a page that doesn't hydrate, hits a different API base, or stopped using Bearer entirely.

Replace the terse error with a full diagnostic dump on timeout:
- postLoginUrl (where loginViaSpa thought it landed)
- finalUrl (where the page is at timeout)
- urlTrail: every top-frame navigation since setup started
- All /api/v1 request URLs observed
- ALL request URLs observed (any host, any path)

Each list capped at the last 30 entries for readability. Next failure tells us EXACTLY what the SPA did instead of guessing.

#PMS-140