2. TREE BACK FROM AN OPEN THREAD: THE ASK STAYS VISIBLE (Hypha HOLD 2) Today: Back closes the thread and the ask disappears from view at all 3 sizes. Its DB state stays "asked", but nothing shows it until another place opens. Fix: after Back closes the thread, the ask stays visible, in view and tappable (Seal it on top, hit by a real click) in the place the person lands, at all 3 sizes. A second Back still defers it ("later"), as with Nursery. Say where it lands and how. 3. GC PAGE ENTRY 74 (honest layer) New docs/gc.html entry 74 (lambda 74.000) in a new page-note PR stacked on #267 (or on #267's branch; say which). This is the only FreeLattice file you touch. Correct entry 73 on three points: (a) the 1280 canvas is 677x169, both shown (CSS) and drawn (buffer), not 677 by 320; (b) entry 73 misquoted Hypha's walk as seeing a "bitmap 677 by 320"; she saw 677x169; (c) entry 73 left out the two Tree gaps above (Gathering Seal it, Back from a thread). Browser and version only. Entry 73 stays untouched. ---------------------------------------------------------------------- 4. MEASURE WITH REAL CLICKS, HYPHA'S WAY Use Hypha's method from her notes (headless Chrome via Playwright, real clicks and taps; viewports 1280x800, 390x844 and 640x400 with device_scale_factor 2, not a CSS zoom). Seal it counts as hittable only if elementFromPoint at its centre is Seal it (or inside it) AND a real page.mouse click / tap there seals. Say which doors you opened with a JS click (Hypha used a JS click for hidden Tree doors) and which with a real click. Tree: ask open, Gathering open: Seal it box, centre element, real click result, at 3 sizes. Ask open, thread open, Back: where she lands, ask visible, Seal it box, real click result; then Back again: deferred. Re-confirm every 043h and 043g PASS above. Name browser and version. 5. THE SMOKE New scripts/smoke-tree-ask-last-v0.1.js on the Tree: the ask is not clipped and Seal it is on top with the Gathering open; Back from a thread keeps the ask shown; a second Back defers. Prints SMOKE_OK tree ask last v0.1. All earlier Tree smokes re-run after the last edit; say which, if any, still fail and why. 6. REPORT New head sha on Tree #123, the new page-note PR link and head, every removed line and where its job went, md5s of every locked file (mind-workspace.js and particle-clay.js Tree == FL main 3e3e8445), all smokes, the measurements from 4, shot paths, and entry 74 with t, lambda, epsilon, delta, omega, sigma so Celeste can recompute psi. Hypha walks it again before merge.