[Bug]: Chromium: page.mouse.move() never resolves when a dragstart handler hides or moves the drag source

Finding: In an external report against Playwright 1.62.1 (code reportedly unchanged on main) with Chromium 151.0.7922.34 on Ubuntu 24.04/WSL2, page.mouse.move() after page.mouse.down() never resolves if the draggable source's dragstart handler hides it (display:none) or moves it (transform). The test then runs into its timeout. Source: https://github.com/microsoft/playwright/issues/43256 Reported evidence: The reporter's table shows that with a no-op dragstart handler, Chromium resolves. With display:none or translateY(100px) on the source, Chromium stays pending after 5 s in both headless shell and headed mode, while the page logs dragstart and dragend. Firefox and WebKit resolve in all three cases. The page stays responsive: a second CDP session can evaluate and requestAnimationFrame keeps running. Reporter's real-suite observation: The reporter says a Playwright e2e spec measured a drag handle's bounding box before an async data load grew the list rows, then pressed 0.25 px inside the next row's top edge. On dragstart the row got transform: scale(0.99) (React Aria's data-dragging). The spec hung in about 1 of 9 drags on a loaded machine. This is the reporter's own suite; the reporter's attribution of the hang to the transform moving the row edge is their interpretation and was not independently confirmed. Reporter hypothesis (not maintainer-confirmed): In crDragDrop.ts, DragManager.interceptDragCausedByMove enables Input.setInterceptDrags, dispatches the move, probes the page for an un-prevented dragstart, then awaits the Input.dragIntercepted promise without racing it against progress or a timeout. The reporter says Blink can abort the drag after dragstart if the source is no longer under the drag origin, and dispatch dragend instead. In that case Chromium never sends Input.dragIntercepted and the await never settles. The reporter also attributes the real-suite hang to the transform moving the row edge past Blink's integer drag origin; this is hypothesis, not observed fact. Proposed fix, not implemented in the sources: make the page-side probe also listen for dragend and report no drag when one follows dragstart, and/or race the await with progress so the hang becomes an action timeout. No fix commit or release is cited, so the status is unfixed as far as the supplied data shows. Workaround: none is stated in the source; no workaround is confirmed. Limitations: This is a single external report. KFA did not reproduce it. The Blink explanation is the reporter's analysis and was not verified against Chromium source. There is no maintainer response in the supplied data, and affected versions beyond 1.62.1 and Chromium 151 are unknown.