You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit f69fa5c
Browse filesBrowse the repository at this point in the historyBrowse files
docs(adr): state the measured --depth cost of dropping the frontier, on the route where it engaged
The earlier sentence rested on snapshot -i runs, which carry no presented depth and never engaged
the frontier on main. Re-measured with --depth 2 and --depth 3 (n=15 per cell) on the recursive
tier: form 135 raw nodes and scrolled catalog 279 raw nodes walked in full to present 4 or 6,
acquisition p50 within ±4 % and p95 within ±7 % of main, mixed sign.
Copy file name to clipboardExpand all lines: docs/adr/0004-ios-snapshot-backend-strategy.md
+10-3Lines changed: 10 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -316,9 +316,16 @@ the walk has already taken — the fold would read reported geometry, and a turn
316
316
device's native space could be cut at the wrong presented depth before the pass ever ran. With the
317
317
frontier gone, no acquisition-time decision reads geometry, the walk is bounded only by raw traversal
318
318
depth and the node cap, and the visibility fold and the presented-depth cut both happen inside
319
-
`SnapshotPresentation`, on the normalized array, at `maximumDepth`. Measured on the recursive tier
320
-
(form, catalog; 7 warm captures each) the frontier bought nothing: acquisition p50 within 4 % and p95
321
-
within 10 % of `main`, in the branch's favour, with identical presented trees. De-duplication drops a repeated node and re-parents its children onto that node's
319
+
`SnapshotPresentation`, on the normalized array, at `maximumDepth`. The frontier only ever engaged
320
+
when a presented depth was set (`snapshot --depth N`; `snapshot -i` carries none), so that is the
321
+
route whose cost the change could move. Measured there on the recursive tier, 15 warm captures per
322
+
cell against `main` on the same simulator: the form (135 raw nodes, 4 or 6 presented) and the
323
+
scrolled catalog list (279 raw nodes, 4 or 6 presented) walk their whole raw tree on this branch
324
+
where `main` pruned it at the presented depth, and acquisition p50 stays within ±4 % and p95 within
325
+
±7 % of `main` with mixed sign. The walk runs over an `XCUIElementSnapshot` tree the platform has
326
+
already materialized in one call, which is where acquisition time goes; the node construction the
327
+
frontier saved is not measurable on these trees. Screens that recover to private AX are untouched
328
+
by it. The bound that remains is the raw node cap. De-duplication drops a repeated node and re-parents its children onto that node's
322
329
own parent, so identical rows collapse under one addressable owner instead of splitting a subtree
323
330
across two nodes with the same identity. Scoped captures remain broad because depth is relative to the
0 commit comments