fix(e2e): isolate error-tracker tests from admin page-load noise

[MathieuLamiot]

Description

Closes #101

Two error-tracker-cap.spec.ts tests were failing intermittently in CI on PRs that change class-sybgo.php (#91, #92): known signature (expects count=1, gets 2) and dashboard widget (expects count=3, gets 4).

Root cause: the two beforeEach hooks ran in the wrong order — CLEAR_QUERY fired before loginAsAdmin. Navigating the WP admin during login runs Sybgo::init() (via plugins_loaded), which may emit a PHP notice captured by Error_Tracker. That stray row landed in the DB after the clear, inflating exact-count assertions in tests that don't fill the cap (tests 1–3, 5–6 fill cap=5 so an extra row is absorbed; tests 4 and 7 assert counts of 1 and 3 so they break).

Type of change

• Bug fix (non-breaking change which fixes an issue).

Detailed scenario

What was tested

• Reviewed the beforeEach hook execution order against the Playwright docs (hooks fire in declaration order)
• Traced the init-time PHP notice path: Sybgo::init() → init_abilities() → plugin code runs during admin page load → Error_Tracker::handle_error() → row inserted with report_id=0
• Confirmed PR Extract Cron_Manager from class-sybgo #90 (no init_abilities() changes) passes E2E while Extract Ability_Manager as a WP7 ability registration utility #91/Extract Admin_Manager as a pure admin registration utility #92 (which add init_abilities()) fail the same two tests
• Verified tests 1–3, 5–6 are unaffected (they fill cap=5, extra row is evicted/rejected)

How to test

1. Check out this branch, run bin/dev-up.sh, then npx playwright test in tests/e2e/
2. All tests in error-tracker-cap.spec.ts should pass, including the two previously failing
3. Run 3 consecutive times to confirm stability

Affected Features & Quality Assurance Scope

• tests/e2e/specs/error-tracker-cap.spec.ts — error tracker daily-limit tests only
• No production code changed

Technical description

Documentation

No documentation update needed.

Fix 1 — beforeEach order: swap so loginAsAdmin runs first, then CLEAR_QUERY. This ensures init-time PHP notices from the admin page load are cleaned up before each test asserts.

Fix 2 — dashboard widget test: the test body itself calls widget.goto() which triggers another WP admin load. Added CLEAR_QUERY inline after goto() (before triggering the 3 test errors), and a page.reload() to give the widget fresh DB data.

New dependencies

None.

Risks

None. Test-only change; no production code modified.

Mandatory Checklist

Code validation

• I validated all the Acceptance Criteria. If possible, provide screenshots or videos.
• I triggered all changed lines of code at least once without new errors/warnings/notices.
• I implemented built-in tests to cover the new/changed code.

Code style

• I wrote a self-explanatory code about what it does.
• I protected entry points against unexpected inputs.
• I did not introduce unnecessary complexity.
• Output messages (errors, notices, logs) are explicit enough for users to understand the issue and are actionable.

Unticked items justification

None.

Additional Checks

• In the case of complex code, I wrote comments to explain it.
• When possible, I prepared ways to observe the implemented system (logs, data, etc.).
• I added error handling logic when using functions that could throw errors (HTTP/API request, filesystem, etc.)