The pictures UI test waits for the worker that checks their sizes (#743)

crates/ui-egui/tests/smoke.rs

@@ -505,7 +505,8 @@ fn pictures_of_the_original_go_with_the_draft_then_into_the_project() {
     // "Nothing — start fresh" is picked: the request box offers pictures (not for websites).
     h.get_by_label("Add pictures of the original (optional: screenshots help it look right)");
     h.state_mut().add_pictures(vec![pic.clone(), notes, pic.clone()]);
-    h.run_steps(3);
+    // Their sizes are checked on a worker first (#720).
+    assert!(wait(&mut h, |a| !a.state.make.pictures.is_empty()), "checked on a worker, then added");
     assert_eq!(h.state().state.make.pictures, vec![pic.clone()], "pictures only, once each");
     h.get_by_label("original.webp");
     h.get_by_label("Remove original.webp").click();

docs/pitfalls.md

@@ -103,6 +103,11 @@ same temp directory, and game detection walked up and found the game. Give each
 For a refactor that must keep output identical, dump the outputs for every combination before and
 after and compare them byte for byte (done for every brief across 162 combinations in R1.1).
 
+**A step moved to a worker, its test still counting frames.** #720 moved the picture size check
+off the UI thread; a UI test that added pictures, ran three frames and then looked kept passing on
+fast machines and failed in CI on unrelated PRs. When work moves to a worker, grep the tests that
+call it and make them `wait` for the result, not `run_steps(n)`.
+
 ## Building and running
 
 **A model catalogue queried in the wrong project.** OpenCode V2's raw `api get /api/model` does