Record what the layout actually costs, and what of it is necessary

The random tree at depth 7 measured against the two revisions before it: the
GPU pass is a tenth of a millisecond everywhere, and every difference is the
CPU laying out. A repaint of one leaf costs 1313 draws and 6.5 ms where the
pre-#16 code cost one draw, because `redraw` escalates to the top size reader
before knowing whether the size changed at all. Drawing first and escalating
only on a different size takes that to one draw and takes a resize below the
pre-#16 number, but misplaces four widgets in the resize case, so it is the
owner's to call.

Also records the measure-by-drawing multiplier -- 1.3x the drawn widgets at
depth 4, 5.2x at depth 8 -- and that scrolling is now part of the random
trees.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

docs/IRIS_EXTRACTION_HANDOFF.md

@@ -9,8 +9,8 @@ Canonical `main` is **`ca2b4b2`** (#17, the headless rig). Sixteen slices are
 in.
 
 **#18 `split/18-position-chain`** is open and finished apart from one decision:
-worktree `/home/bob/repos/iris-pr18`, head `4178dfb`, twelve commits, 49 tests
+worktree `/home/bob/repos/iris-pr18`, head `cdec293`, thirteen commits, 50
-passing, fmt and clippy clean. It is LAYOUT.md §2's position chain, generalised
+tests passing, fmt and clippy clean. It is LAYOUT.md §2's position chain, generalised
 to boxes. `/home/bob/repos/ai-app-2` is on `rustify`, worktree clean.
 
 **The one thing waiting on the owner.** `tabs` at 1920x1200 is no longer
@@ -115,10 +115,62 @@ A chain is irrelevant at an example's couple of hundred primitives; a
 transcript's glyphs are tens of thousands, which is the regime `chain_cost`
 measures.
 
+### How much of that work is necessary
+
+Measured 2026-09-14 on a random tree at seed 1, depth 7 -- 1061 widgets, 839
+of them drawn, 10,872 primitive instances -- against the same rig at `43ce8c7`
+(the last commit before #16 sized a widget by drawing it) and at `f942385`
+(#16 itself). The rig is a `Harness` load counting widget draws, text shapes
+and instance writes per frame, plus the GPU pass through timestamp queries;
+it lives in the scratch worktrees `/home/bob/repos/iris-size-{old,new}` and is
+not in the repository, because the counters it reads are patched into
+`iris-core`.
+
+| per frame | before #16 | #16 | #18 head |
+| --- | --- | --- | --- |
+| cold layout | 19.3 ms, 839 draws | 30.9 ms, 3317 | 23.6 ms, 2755 |
+| repaint one leaf | 0.000 ms, 1 draw | 8.5 ms, 1683 | 6.5 ms, 1313 |
+| scroll one scroller | 0.001 ms, 1 draw | 8.9 ms, 1683 | 6.5 ms, 1313 |
+| resize the output | 3.2 ms, 839 draws | 31.8 ms, 6940 | 25.2 ms, 5512 |
+| GPU pass | 0.129 ms | 0.114 ms | 0.115 ms |
+
+**The GPU is not the subject.** The pass is a tenth of a millisecond at every
+revision and every phase; all of this is the CPU laying out.
+
+Three separate costs, in the order they are worth fixing:
+
+- **A repaint escalates to the top size reader and redraws its whole
+  subtree.** `redraw` asks `mark_readers` first, so a widget whose drawing
+  changed but whose *size* did not still redraws everything from the highest
+  widget that ever read it. At depth 8 a one-rectangle repaint costs 4825
+  draws and 24 ms, which is more than building the tree from nothing. The
+  measured alternative is to draw the widget in its own box first, compare
+  the `Size` it returns with the one it had, and escalate only when it
+  differs: repaint falls to 1 draw and 0.000 ms, a scroll to 11 draws, and a
+  resize to 532 draws and 2.3 ms -- below the pre-#16 numbers on every line.
+  **It is not correct as written**: the resize case in `tests/generated.rs`
+  fails at seed 1 with four widgets misplaced, because drawing the widget
+  clears the mark that was the only thing stopping a later parent draw from
+  reusing it into a different box. The patch is kept at
+  `/tmp/escalate.patch`. This is the owner's call, since it changes when a
+  widget is drawn.
+- **A container measures a child by drawing it in a box it will not keep.**
+  `Span::draw` places each child without an exact `size_hint` at
+  `UiRegion::FULL` to read its length, then places it again at the box it
+  worked out, and the second place redraws whatever the first drew. It
+  compounds with nesting: draws per drawn widget are 1.3x at depth 4, 3.3x at
+  7 and 5.2x at 8. Wrapping text is shaped at the trial width and then again
+  at the real one -- 332 shapes for 152 text widgets on a cold frame.
+- **Redundant draws rewrite primitives.** A cold frame writes 48,050
+  instances for a tree that holds 10,872. What reaches the GPU is 10,872,
+  since a layer uploads whole, so this is CPU cost only -- but a one-leaf
+  repaint still uploads 5,863 instances where the pre-#16 code uploaded 106.
+
 ## The random trees
 
 `iris::random` grows a seeded tree -- spans in every direction holding two to
-four children, stacks, padding with each of its four sides its own number,
+four children, stacks, scrolls on either axis, padding with each of its four
+sides its own number,
 rects with varying opacity, text both wrapping and overflowing, a declared size
 over half of it, stopping at a depth. `examples/random.rs` draws one
 (`IRIS_SEED`, `IRIS_DEPTH`). `tests/generated.rs` grows each seed twice -- once