v1.46.1: re-check of v1.46.0 — HP-1460-01, -02, -03

HP-1460-01: v1.46.0 stopped overwriting attachments, but picking a free name
and taking it were two steps. Two uploads racing between them agreed on the
same name, both answered 200, and one set of bytes replaced the other;
files/migrate had the same check-then-copy gap. reserve_filename now claims the
name with O_CREAT|O_EXCL as it picks it, and both paths use it. It also splits
the extension off the RAW name and budgets the stem against MAX_FILENAME
including the collision tag — a maximal name lost its '.pdf' and then grew past
the limit, so the view sanitised the request back to a different name and the
attachment 404'd for good.

HP-1460-02: cleanup lived in an 'except Exception', which CancelledError walks
past, only one tmp_path was tracked, promotion had no finally, and the
collector only walks marker folders — an aborted transfer stranded a .upload-*
that nothing would ever remove. An outer finally owns every temporary, a second
'file' part is refused, promotion failure cleans up, and sweep_upload_temps
runs at setup, daily, and inside the commit-scoped collector. Chunks are
batched to 1 MB per disk task instead of one per 64 KB.

HP-1460-03: the layout event reached the static card and not the full one, so
two full cards diverged until a reload. The full card subscribes now and
re-reads ONLY the layout, keyed on its revision. Two hazards handled: it
records revisions it produced itself, and the reaction is deferred ~200 ms
because the event can beat the reply to our own write over the same socket;
positions dragged but not yet sent are flushed and merged on top, so a fix for
a stale UI cannot become a lost drag.

Tests: smoke_layout_sync (fails on a v1.46.0 build), four pure tests for atomic
reservation incl. 20-thread concurrency and the length boundary, a backend test
walking every failing exit path of an upload, and — as the report asked — an
HA-harness test that a repair issue disappears with its space.
Docs: CHANGELOG.md + CHANGELOG.ru.md + ARCHITECTURE.md + TESTING.md + STATUS.md.

docs/CHANGELOG.md

@@ -1,5 +1,31 @@
 # Changelog
 
+## v1.46.1 — 2026-07-28 (re-check of v1.46.0: HP-1460-01 … -03)
+- **Two uploads of the same file name can no longer collide (HP-1460-01).**
+  v1.46.0 stopped overwriting attachments, but choosing a free name and taking
+  it were two steps: two uploads racing between them agreed on the same name,
+  both reported success, and one set of bytes replaced the other. The name is
+  now claimed atomically as it is chosen — twenty simultaneous uploads of
+  `manual.pdf` produce twenty files. The same helper is used when rebinding
+  moves files, which had the same gap.
+  Also fixed there: a name at the length limit lost its extension, and the
+  collision suffix pushed it past the limit, so the attachment was stored under
+  a name the server would not serve back — a permanent 404 on a file the UI
+  reported as attached.
+- **An interrupted upload no longer leaves a temporary file forever
+  (HP-1460-02).** Cleanup ran in an `except Exception`, which a cancelled
+  request walks straight past, and the collector only ever looks inside marker
+  folders — so an aborted transfer left a `.upload-*` in place with nothing able
+  to remove it. Every exit path now cleans up, a request carrying two files is
+  refused outright, and abandoned temporaries are swept at startup and daily.
+  Uploads also write in 1 MB batches instead of one disk task per 64 KB.
+- **Two full cards side by side keep the same positions (HP-1460-03).** v1.46.0
+  taught the static card to follow position changes and left the full one
+  behind, so dragging an icon in one window did not move it in another until a
+  reload. It follows now, without disturbing a drag of its own: a revision
+  arriving mid-drag is merged rather than applied over the top, and a card does
+  not re-read what it just wrote itself.
+
 ## v1.46.0 — 2026-07-28 (full external audit of v1.45.4: HP-1454-01 … -10)
 
 **Security**