Context
The 13.2 notes (.github/RELEASE_NOTES/v13.2.0.md) were last swept up to d75cc68543 by #19482 (PR #19484, merged 2026-10-09). Since then dev has 37 first-parent merges up to c09b6c7107; 8 of them edit the notes themselves. Step 5 of #14800's cut order (6097885253) is the notes re-review, and this is its sweep.
One chapter is now wrong, not only behind. § Green in emulation, broken by hand says that a tab torn out and carried over another window parks "clear of the target, in the corner of that display's work area it covers least" (#19374). It also says that "a parked vessel re-shown as the pointer leaves the target now lands under the held pointer" (#19458). #19539 (#19533, ADR 0029 §2.8.6) retired that park for a pointer drag:
- Over another window of the group, that window carries the drag as the source's tab-header proxy, and the riding vessel retires on the claim.
- Leaving for the desktop opens a fresh vessel under the held pointer.
- Only a native title-bar drop still parks behind its target.
#19542 deleted the dormant pointer-path park code. #19556 (Ada, open) makes real Chrome follow the same rule, so the notes describe the rule the release ships. I found this at ~12:56Z and promised the ticket in a lane-claim; this files it.
The Problem
The notes are the release's public account. They narrate a behavior the release no longer has, and they contradict the docking guide's fourth station, which #19539 already rewrote to explain the close. The merges since the last sweep, among them the B rule itself, the cross-window fixes and four VDOM repairs, are not dispositioned at all.
The Architectural Reality
- The park sentences: the third bullet under "Each repair moved a measurement off the emulated plane", and the last sentence of the paragraph after it.
- Authority for the rule: ADR 0029 §2.8.6 (
learn/agentos/decisions/0029-docking-design.md) and #19539's body. The history of the three hand-found defects (size, plane, z-order) stays true and stays.
- Merges to disposition, first-parent
d75cc68543..c09b6c7107:
- VDOM: #19483, #19499, #19509, #19525
- Dock: #19516, #19526, #19529, #19539, #19542, #19551, #19552
- Workstation: #19532, #19534, #19536, #19546
- One each: tab #19545, data #19454, SEO #19503, release #19555
- Plus tests, CI, one dependency group and the three 13.2 blog posts
The Fix
- Rewrite the park sentences to the claim-and-retire rule, linking #19539 and #19542, and keep the title-bar park as the native drop's.
- Disposition every first-parent merge in
d75cc68543..<PR base> as a chapter line with its PR link, an upgrade row, or the Full Changelog. Each line is written from its PR, never from the subject alone.
Acceptance Criteria
Out of Scope
#19556's code fix; the FM drop-zone check (cut step 4); the film take (#15252).
Related
#14800 (epic), #19482 (previous sweep), #19533 / #19539, #19538 / #19542, #19556, ADR 0029.
Decision Record impact: none (the notes follow ADR 0029 §2.8.6).
Sweeps: live latest-open sweep, latest 20 open issues at 15:49Z and latest 5 at 15:50Z, no equivalent · A2A in-flight sweep: 30 messages across read states, no competing claim (mine at ~12:56Z is the earlier one) · MC sweep: "13.2 release notes Green in emulation park vessel pointer drag retired", only my own 12:56Z record of this defect · Own-assignment sweep: #14800 (this epic), #15000 (unrelated).
Origin Session ID: 9ea1c6d0-80b1-4552-b07d-004bb1dba240
Retrieval Hint: "13.2 release notes iteration 7 sweep d75cc68543 pointer park retired claim"
Context
The 13.2 notes (
.github/RELEASE_NOTES/v13.2.0.md) were last swept up tod75cc68543by #19482 (PR #19484, merged 2026-10-09). Since thendevhas 37 first-parent merges up toc09b6c7107; 8 of them edit the notes themselves. Step 5 of #14800's cut order (6097885253) is the notes re-review, and this is its sweep.One chapter is now wrong, not only behind. § Green in emulation, broken by hand says that a tab torn out and carried over another window parks "clear of the target, in the corner of that display's work area it covers least" (#19374). It also says that "a parked vessel re-shown as the pointer leaves the target now lands under the held pointer" (#19458). #19539 (#19533, ADR 0029 §2.8.6) retired that park for a pointer drag:
#19542 deleted the dormant pointer-path park code. #19556 (Ada, open) makes real Chrome follow the same rule, so the notes describe the rule the release ships. I found this at ~12:56Z and promised the ticket in a lane-claim; this files it.
The Problem
The notes are the release's public account. They narrate a behavior the release no longer has, and they contradict the docking guide's fourth station, which #19539 already rewrote to explain the close. The merges since the last sweep, among them the B rule itself, the cross-window fixes and four VDOM repairs, are not dispositioned at all.
The Architectural Reality
learn/agentos/decisions/0029-docking-design.md) and #19539's body. The history of the three hand-found defects (size, plane, z-order) stays true and stays.d75cc68543..c09b6c7107:The Fix
d75cc68543..<PR base>as a chapter line with its PR link, an upgrade row, or the Full Changelog. Each line is written from its PR, never from the subject alone.Acceptance Criteria
Out of Scope
#19556's code fix; the FM drop-zone check (cut step 4); the film take (#15252).
Related
#14800 (epic), #19482 (previous sweep), #19533 / #19539, #19538 / #19542, #19556, ADR 0029.
Decision Record impact: none (the notes follow ADR 0029 §2.8.6).
Sweeps: live latest-open sweep, latest 20 open issues at 15:49Z and latest 5 at 15:50Z, no equivalent · A2A in-flight sweep: 30 messages across read states, no competing claim (mine at ~12:56Z is the earlier one) · MC sweep: "13.2 release notes Green in emulation park vessel pointer drag retired", only my own 12:56Z record of this defect · Own-assignment sweep: #14800 (this epic), #15000 (unrelated).
Origin Session ID: 9ea1c6d0-80b1-4552-b07d-004bb1dba240 Retrieval Hint: "13.2 release notes iteration 7 sweep d75cc68543 pointer park retired claim"