feat(ui): route mouse-wheel to preview in stacked (below) layout (#1626 follow-up)

[jdidion]

Note: This PR was largely generated by Claude and has not been completely reviewed by me (the human). Feel free to defer review until this warning is removed.

Follow-up to the merged #1626 (preview-below orientation), addressing a CodeRabbit review finding on that PR.

What

In the below/stacked preview layout, the mouse wheel always moved the session-list cursor — it never scrolled the preview. Wheel-to-preview routing was dual-layout only (X-based: "is the wheel over the right column?"). Stacked mode has no right column — the preview is the lower region — so scroll input never reached it.

This adds Y-based routing for stacked mode: a wheel event at or below the horizontal divider scrolls the preview (previewScrollOffset), matching the dual layout's X-based behaviour. Single layout is unchanged (no preview target).

How

• stackedPreviewTopY() computes the divider's screen-Y by mirroring renderStackedLayout's height split — it reuses the shared stackedListHeight helper plus a new contentChromeTop() that both this and the render path use, so the routing Y can't drift from what's actually drawn.
• The wheel handler switches on layout mode: dual routes by X (unchanged), stacked routes by Y (new), single falls through to list-scroll.

Scope

This is the wheel-scroll half of CodeRabbit's finding. Divider resize in stacked mode stays on the < / > keys (added in #1626), not a mouse drag: the existing divider-drag path (isOnDivider / setPreviewPctFromMouseX) is X-axis only, and a parallel Y-drag would duplicate fragile layout math for redundant polish (the keyboard path already covers stacked resize). Flagged in the commit as a possible follow-up if you want mouse-drag parity.

Verification

go build ./..., go vet, gofmt clean; internal/ui layout/preview/mouse tests pass.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved mouse-wheel behavior when viewing stacked layouts so wheel input scrolls the PREVIEW region rather than moving the session list cursor.
  • PREVIEW scrolling now responds to cursor position within the preview pane (using the pane’s true on-screen location).
  • Preview area detection now correctly accounts for the title/help/header rows and optional update/maintenance/debug banners.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 577d0059-0fa5-430e-ab16-d74aad93bfd5

📥 Commits

Reviewing files that changed from the base of the PR and between 7debcdf and 7acaa3d.

📒 Files selected for processing (1)

• internal/ui/home.go

💤 Files with no reviewable changes (1)

• internal/ui/home.go

---

📝 Walkthrough

Walkthrough

Adds layout-aware helpers for locating the stacked PREVIEW pane and routes mouse-wheel events there in stacked layouts, while preserving existing dual and single layout behavior.

Changes

Stacked PREVIEW wheel routing

Layer / File(s)	Summary
Preview geometry helpers internal/ui/home.go	Computes the main content origin and stacked PREVIEW pane position using update, maintenance, and debug layout flags.
Preview wheel routing internal/ui/home.go	Routes stacked-layout wheel events within the PREVIEW region to previewScrollOffset; dual-layout X-based routing and single-layout list scrolling remain unchanged.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• asheshgoplani/agent-deck#1583: Both changes modify mouse-event handling around the PREVIEW area in internal/ui/home.go.
• asheshgoplani/agent-deck#1626: Both changes modify stacked PREVIEW layout geometry in internal/ui/home.go.

Suggested reviewers: asheshgoplani

🚥 Pre-merge checks | ✅ 5 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Remote_parity	⚠️ Warning	preview_scroll_test only covers local sessions; renderRemotePreview ignores previewScrollOffset, and no t.Skip documents RemoteSession N/A.	Add a RemoteSession wheel/preview-scroll test (or a documented t.Skip if intentionally N/A) and, if needed, apply previewScrollOffset in remote preview rendering.
Test_coverage_per_surface	⚠️ Warning	TUI feature added stacked-layout wheel routing, but tests only cover dual and single layouts; no test exercises stacked wheel behavior or the new Y helper.	Add internal/ui tests for stacked/below wheel events over preview vs list and the divider Y helper/offset reset path; no web/CLI/remote surfaces are applicable.

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title follows Conventional Commits and accurately describes the main change to mouse-wheel routing in stacked layout.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

👋 Thanks for opening this — a couple of intake fields would help the maintainer's validation pipeline (and any reviewer) understand it faster. Nothing here blocks you and nothing is closing: this is a heads-up, and editing the description re-runs this check automatically.

• Missing required sections: ## What problem does this solve?, ## Why this change, ## User impact, ## AI disclosure, ## What actually bothered you, ## Checklist.
• No AI-disclosure box is checked — pick exactly one (Human-written / AI-assisted / AI-authored). AI is welcome here with equal standing; this is a signal, not a filter.
• ## What actually bothered you is empty — one real sentence of human intent is the one field intake cannot accept blank. If an agent opened this PR, quote what the human asked for.

The full field-level contract is in .github/INTAKE.md. AI-authored PRs are welcome here with equal standing.

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@internal/ui/home.go`:
- Line 1102: Remove the unused panelTitleLines constant declaration from the
surrounding function in internal/ui/home.go, leaving the remaining panel layout
logic unchanged.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 5460e40b-60a6-49cc-bd3a-98db6b289c82

📥 Commits

Reviewing files that changed from the base of the PR and between 580e772 and 7debcdf.

📒 Files selected for processing (1)

• internal/ui/home.go

[jdidion]

@asheshgoplani heads-up on CI, and a request to re-run when you get a chance.

The code side is done and golangci is green. But the Full test suite (PR gate) (-race) has gone red on three runs of this branch, and I'm fairly confident it's pre-existing -race flakiness, not this PR:

• This PR is a one-line change on top of the merged feat(ui): add preview-below orientation toggle for portrait terminals #1626 (removes an unused const in home.go; the substantive change is the ~60-line stacked-mouse-wheel routing, which touches only internal/ui).
• The gate passed the Full suite on an earlier push of this same branch, then failed on a subsequent push, and failed again on an empty re-trigger commit (no code change at all).
• The failing packages change every run — internal/tmux (softkill), then internal/session, then internal/statedb + internal/session — which is the signature of resource-contention flakiness under -race, not a deterministic failure in one test. None of these packages are touched by this PR.
• main's own go-test runs show the same: mostly green, with intermittent failures (e.g. a failure on 2026-07-26). So the gate flakes on main too, independent of this branch.
• I couldn't reproduce a matching failure locally (my box hits an unrelated env-specific bind: invalid argument in TestIssue1421_CleanStaleSSHSockets, which wouldn't fail on your Linux runners).

Could you re-run the gate (you have admin; I can't re-run from the fork)? If it's a known-flaky suite you already manage, happy to defer to whatever your standard is. And if you'd rather I split anything out or adjust scope, just say.

Same disclosure as before: largely Claude-generated, diff reviewed by me.