test(web): wait out the self-dismissing WRL infobar before the chrome baseline

The WRL→STEP migration notice auto-dismisses after 10 s
(pcb_edit_frame.cpp ShowMessageFor); when the timer fired between the
pre-hide GL-box baseline and the post-restore comparison, the canvas
shifted by the bar's height and the ±3 px restore-exactness check could
never pass (CI 30289317464, web-firefox — pure timing, the wx dark
form-control bump on that run was innocent). Gate the baseline on the
notice being HIDDEN — its DOM node outlives the dismiss, so absence
never happens.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y32etYBCKV6t1qDoLoGmgF

tests/web/chrome-toggle.spec.ts

@@ -66,6 +66,15 @@ test.beforeAll(async ({ browser }: { browser: Browser }) => {
   // Loading overlays (boot + lib fat-load, both `inset-0 z-30`) must be gone
   // before geometry is trusted.
   await expect(page.locator('div.inset-0.z-30')).toHaveCount(0, { timeout: 180000 });
+  // The WRL→STEP migration notice self-dismisses after 10 s
+  // (pcb_edit_frame.cpp ShowMessageFor) and shifts the GL canvas when it
+  // goes — wait it out so the pre-hide baseline and the post-restore box
+  // are measured in the same layout. It is shown synchronously during the
+  // file load the title poll above already gates on, and its DOM node
+  // outlives the dismiss, so wait for hidden, not absent. The un-timed
+  // "older version of KiCad" infobar stays up on both sides of the
+  // comparison, so it cancels out.
+  await expect(page.getByText(/WRL 3D model/)).toBeHidden({ timeout: 30000 });
 });
 
 test.afterAll(async () => {