test: stabilize upload-button visual tests by disabling hover transitions (CP: 25.1)

[web-padawan]

This PR cherry-picks changes from the original PR #11598 to branch 25.1.

---

Summary

Aura defines two transitions on the shared :is(vaadin-button, vaadin-upload-button, …) selector at packages/aura/src/components/button.css:7,86-88:

• scale 180ms on press ([active])
• opacity 100ms, background-color 100ms on the ::before hover overlay

The upload-button visual tests' hover and active scenarios trigger these via sendMouseToElement / mousedown and capture immediately after, so the screenshot lands somewhere on the 100–180 ms ramp. Different Playwright versions land at different points along that ramp, producing baseline drift on every bump.

Same Playwright-timing cascade pitfall as #11587/#11588/#11591/#11595.

Changes

• New packages/upload/test/visual/not-animated-styles.js registering transition: none !important on vaadin-upload-button :host, :host::before, :host::after.
• Imported from the upload-button visual tests for all 3 themes (aura, lumo, base).

!important is required because Aura is adopted into adoptedStyleSheets after registerStyles, and both rules sit at specificity 0,1,0.

Baseline impact

Three Aura *-hover.png upload-button baselines shifted to the fully-opaque hover overlay end state (sub-LSB diff, MAX_DELTA = 1, button center byte-identical). The active, default, focus-ring, disabled, icon and accent baselines stay byte-identical, as do all Lumo and base baselines. The vaadin-upload (legacy) and upload-file-list-thumbnails tests don't render vaadin-upload-button, so they don't import this stylesheet.

Test plan

• Two consecutive yarn update:aura --group upload runs in Docker produce byte-identical baselines (md5 confirmed).
• yarn test:{aura,lumo,base} --group upload — all 38/38 (aura), 39/39 (lumo), 39/39 (base) pass against the new baselines.
• CI visual-tests workflow stays green.

🤖 Generated with Claude Code

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud