Dashboard CSS consolidation: dedupe inline styles between testing.html and testing-session.html

[jonathaneoliver]

Summary

While fixing #337, an audit of inline <style> blocks in content/dashboard/testing.html and content/dashboard/testing-session.html surfaced ~210 lines of CSS that's duplicated identically in both pages, plus a small handful of drift items that should be reconciled.

The shared stylesheet content/dashboard/testing-session-refactored.css is loaded by both pages — that's where these rules belong.

Cleanup wins (zero behavioral change)

• .abrchar-* block (~70 lines) — duplicated byte-for-byte in both pages. Consolidate to shared.
• .events-timeline-wrap + vis-timeline rules (~100 lines) — duplicated byte-for-byte. Consolidate.
• .chart-expand-btn / .chart-expand-icon — duplicated. Consolidate.
• @keyframes chart-pulse-dot + button[data-action="pause-bitrate-chart"].chart-live rules — duplicated. Consolidate.

Port-then-consolidate (testing.html missing rules from testing-session.html)

• .abrchar-ruler-marker.wire-active-1s .abrchar-ruler-line { background: #059669; }
• .abrchar-ruler-marker.wire-active-1s .abrchar-ruler-label { color: #047857; }

These exist only in testing-session.html. Verify whether testing.html actually needs them (does its ABR characterization UI render wire-active-1s markers?). If yes, port; either way, consolidate.

Drift to reconcile (visual decision needed)

• .chart-paused-badge — testing.html uses rgba(220, 38, 38, 0.85) red sans-serif "PAUSED"; testing-session.html uses rgba(30, 41, 59, 0.72) dark monospace. Pick one palette and apply on both.

Out of scope (intentional differences, leave alone)

• .chart-wrap differs significantly (testing-session.html has white background + border + padding for single-card density; testing.html is bare for multi-session dashboard density). Don't unify.
• Page-shell rules: .page-header, .session-tabs, .network-log-* (testing.html-only); .player-card, .player-frame, .stats-grid (testing-session.html-only).

Acceptance

• Both pages render identically for the consolidated rules (visually verify on :21000).
• Net diff is mostly subtractive (delete from inline blocks, single addition to shared CSS).
• No behavioral or visual regression on either page.

Origin

Audit performed during PR for #337 (fix/337-testing-html-buttons).

[jonathaneoliver]

Points: 3 (Fibonacci)
Effort: ~2–3 hours
Priority: P3 — no behavioral change, no user-facing regression risk; pure housekeeping with no urgency
Value: low — improves maintainability (one source of truth for shared CSS) but zero visible user impact
Rationale: All the mechanical work (move ~210 lines, reconcile one badge palette) is bounded and low-risk. The only judgment call is the .chart-paused-badge design decision, which is cosmetic. Rates 3 points rather than 2 because of the wire-active-1s port-or-confirm step and the required visual verification pass on both pages at :21000. No split needed.