fix(web): lane inline-create opens the split pane on desktop, stays put on mobile (IDEA-2298)

[xarmian]

What

The board lane + opens a Trello-style draft card (TASK-1676); Enter created the item and then navigated the whole page away to .../{item}?new=1. TASK-1676 predates the split pane (PLAN-2105), so nothing ever revisited that destination.

That left the one create gesture that already knows its title as the only card-open path bypassing the pane:

• Clicking an existing card opens the split pane (onItemOpen → ?item=); creating one left the board entirely.
• ?new=1 exists to drop you into the title editor of a fresh "Untitled" item (createNewItem). Here the title was just typed — so it re-opened the title editor with that title select-alled. That's the specific thing that felt wrong.
• No viewport branching at all. The lane + is fully present at ≤768px, so mobile got ejected off the board too, making a second add a Back navigation.

Behavior now

	Desktop	Mobile
Item created + card appears in lane	✅	✅
Opens	split pane (?item=)	nothing — tap the card if you want it
Pathname changes	never	never
?new=1	never	never

quickCreateInColumn's third param becomes reveal (caller intent) rather than navigate (a destination). The local-index upsert is unconditional so the card always lands in its lane; revealing means openItemPane(item) — the same entry point a card click uses — gated on !viewport.isMobile. The page owns what revealing means, so BoardView stays unaware the pane exists.

The composer closes on submit on every viewport. On desktop the pane takes focus, so keeping it open for Trello-style rapid entry would fight it; uniform close is the deliberate call for now, revisitable on feedback.

The nav-guard's Save-all keeps passing reveal: false — saving drafts on the way out must never open anything.

Testing

New web/e2e/lane-inline-create.spec.ts pins both destinations plus the two invariants (pathname unchanged, no ?new=1).

Two traps worth recording, since they made the first draft of both tests vacuous:

1. The e2e server serves the embedded build — a source edit without make install tests the stale binary.
2. More substantively, the created card renders off the synchronous local-index upsert, so it is not a sync point for the navigation that follows. Both tests initially passed against the reverted fix because the URL assertions raced an unresolved goto (the desktop pathname assertion passed vacuously; only the ?item= check bit). Desktop now waits on ?item= landing. Mobile can't poll an absence, so it proves the negative positively — it re-opens the lane composer, which only works if the board is still mounted, and awaiting that click gives a would-be navigation time to land. It also encodes the real UX win: add a second card immediately instead of navigating back.

Verified by mutating the fix back out and rebuilding: both fail for the right reasons, pass on the fix.

Also note: the docs collection empty-states to "No docs yet" with no lanes at all, so the spec seeds an anchor doc before there's a + to click.

Gates

• npm run check — 0 errors (6 pre-existing warnings)
• npm run test — 490 passed
• new spec — 2/2
• pane e2e suite — 64/65 on two runs. Each run failed a different spec, both passing in isolation; the second was confirmed by screenshot as BUG-2334 (cross-test SSE "E2E Admin created" toasts covering the graph detail card). Neither is reachable from this diff — nothing here touches collab or graph.
• Codex review — CLEAN (CONVE-735), two rounds, the second focused on whether the new spec's assertions actually bite.

Out of scope

ListView groups by status but has no per-group + at all, so inline create remains board-only. Extending it there is a separate item.

Closes IDEA-2298.

https://claude.ai/code/session_01E2fRi12n8rARczvdEa2LYT