fix(development): make driveless /dashboard/development a GLOBAL view

[2witstudios]

Summary

Fixes a scope bug in the Development surface (merged in #2015): the driveless /dashboard/development route redirected to a resolved "current/last" drive, opening a seemingly-random drive instead of showing everything. Per the "Routing correction (2026-07-12)" section of the Development surface spec, it's now a real GLOBAL command center — all machines across all drives the user can access, grouped by drive. /dashboard/{driveId}/development (per-drive) is unchanged.

• packages/lib/services/machines/machine-list: new listMachinesAcrossDrives, grouping the existing per-drive listMachinesInDrive result by drive and dropping drives with no visible machines (avoids empty group headers). Per-drive failures are isolated (Promise.allSettled) so one bad drive can't blank the whole global view.
• GET /api/machines: a request with no driveId now returns { drives: [{ driveId, driveName, machines }] } instead of 400 — same app-admin gate, same per-page canUserViewPage filtering as the per-drive path (a machine withheld by a page-level grant never appears in either mode).
• machine-list-runtime: listAllMachines(), sourcing accessible drives from listAccessibleDrives — the same drive universe GET /api/drives already hands DriveSwitcher.
• useDriveMachines: new useAllMachines(enabled) hook, same SWR/polling convention as useDriveMachines.
• DevelopmentSidebar: branches on driveId presence — drive-scoped list unchanged, new grouped-by-drive list in global mode. MachineTreeSection routes via a centralized buildMachineHref (next to parseSelectedMachineId in development-route.ts), so global-mode machines route to /dashboard/development/{machineId} rather than crossing into a specific drive's route tree.
• New /dashboard/development/layout.tsx + [machineId]/page.tsx: the global twin of the drive-scoped detail region, sourced from useAllMachines(). Deliberately its own route tree (not folded into the drive-scoped layout) — Next.js remounts a segment when the route tree changes, so crossing into /dashboard/{driveId}/development would tear down the global layout's keep-alive host and every terminal it's keeping warm. (Known, documented tradeoff: navigating from global into a specific drive's Development view still crosses that boundary and disconnects every warm terminal, not just the one being left — mirrors an already-accepted tradeoff elsewhere in the surface and is recoverable, not data loss.)
• Extracted useStickyMachineIds, useDrainPendingSession, DetailState/DetailNotice, and resolveDisplayedMachine out of the drive-scoped layout into lib/development/ so both layouts share identical logic instead of duplicating it.
• Removed the now-dead resolveActiveDriveId redirect helper (and its test).

MachineTree/MachineView/Machine-page tabs are untouched, per the sub-task's scope.

Merged origin/master to resolve a real conflict: #2017 (workspace-owns-pane-grid) changed the same [driveId]/development/layout.tsx this PR refactors (pending-session workspace lookup moved from a flat map to selectActiveWorkspace). Ported that into the shared use-drain-pending-session.ts hook.

Convergence pass (8-angle automated review + Codex): fixed the Promise.all → Promise.allSettled correctness bug above, a stale doc comment in sidebar-routes.ts, deduped fetcher/globalFetcher and the sidebar's list-notice guard chains, replaced a GLOBAL_SCOPE magic string with undefined, and replaced a raw NUL byte Codex flagged in use-sticky-machine-ids.ts with a proper escape. See PR comments for full detail on what was fixed vs. deliberately left alone (and why).

Test plan

• packages/lib/services/machines/machine-list.test.ts — grouping, per-page filtering, and per-drive failure isolation for listMachinesAcrossDrives
• apps/web/api/machines/route.test.ts — global-mode branch: admin gate, filtering, empty list, auth propagation
• apps/web/app/dashboard/development/__tests__/layout.test.tsx — new global layout (flattening, sticky ids, admin gate, detail states)
• apps/web/app/dashboard/[driveId]/development/__tests__/layout.test.tsx — unchanged behavior, still green after extracting shared hooks + the master merge
• DevelopmentSidebar.test.tsx — drive-scoped tests still green + global-mode tests (grouped rendering, routing to /dashboard/development/{machineId}, admin gate, empty/error states)
• development-route.test.ts — driveless URL prefix + new buildMachineHref, including a parse/build round-trip test
• displayed-machine.test.ts — new, for the extracted resolveDisplayedMachine
• bun run typecheck (full monorepo, including next build) — green
• bun run lint (web + lib) — green, no new warnings
• Full apps/web vitest suite — 12,800+ passed; a handful of pre-existing failures are unrelated (DB-role provisioning tests needing bun run test's DB setup, one pre-existing timezone flake)

🤖 Generated with Claude Code

https://claude.ai/code/session_01PZNYjLCR2FxEcvxZbsfDqp

[coderabbitai]

Warning

Review limit reached

@2witstudios, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 20 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: dacbfd34-5482-4fcd-a242-fafdbcf8f6aa

📥 Commits

Reviewing files that changed from the base of the PR and between b7859dc and f475d16.

📒 Files selected for processing (23)

• apps/web/src/app/api/machines/__tests__/route.test.ts
• apps/web/src/app/api/machines/route.ts
• apps/web/src/app/dashboard/[driveId]/development/layout.tsx
• apps/web/src/app/dashboard/development/[machineId]/page.tsx
• apps/web/src/app/dashboard/development/__tests__/layout.test.tsx
• apps/web/src/app/dashboard/development/layout.tsx
• apps/web/src/app/dashboard/development/page.tsx
• apps/web/src/components/layout/left-sidebar/DevelopmentSidebar.tsx
• apps/web/src/components/layout/left-sidebar/__tests__/DevelopmentSidebar.test.tsx
• apps/web/src/components/layout/left-sidebar/sidebar-routes.ts
• apps/web/src/hooks/useDriveMachines.ts
• apps/web/src/lib/development/DetailState.tsx
• apps/web/src/lib/development/__tests__/development-route.test.ts
• apps/web/src/lib/development/__tests__/displayed-machine.test.ts
• apps/web/src/lib/development/__tests__/resolve-active-drive.test.ts
• apps/web/src/lib/development/development-route.ts
• apps/web/src/lib/development/displayed-machine.ts
• apps/web/src/lib/development/resolve-active-drive.ts
• apps/web/src/lib/development/use-drain-pending-session.ts
• apps/web/src/lib/development/use-sticky-machine-ids.ts
• apps/web/src/lib/machines/machine-list-runtime.ts
• packages/lib/src/services/machines/__tests__/machine-list.test.ts
• packages/lib/src/services/machines/machine-list.ts

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch pu/development-global-view

---

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.}

[chatgpt-codex-connector]

💡 Codex Review

PageSpace/apps/web/src/lib/development/use-sticky-machine-ids.ts

Line 22 in 18dbf66

*/

Escape the NUL separator in source

This template literal contains an actual NUL byte rather than the textual \u0000 escape used by the previous inline hook. The runtime string is the same, but embedding NUL makes this TypeScript file look binary to some editors/search/diff scripts and can break tooling around the repo; replace the raw byte with \u0000 so the separator stays source-safe.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[2witstudios]

Addressed in 210effc: replaced the raw NUL byte in use-sticky-machine-ids.ts's key template literal with a proper JS escape sequence. Same runtime string, now source-safe.

[2witstudios]

Follow-up convergence pass on this PR — ran an 8-angle automated review (correctness ×3, reuse, simplification, efficiency, altitude, CLAUDE.md conventions) plus resolved a real merge conflict against master (#2017 changed the same layout.tsx this PR refactors). Summary of what changed, pushed in f475d16:

Correctness fix (real bug): listMachinesAcrossDrives used Promise.all across every accessible drive — one drive's scan (or its per-page visibility check) throwing would 500 the entire global view, hiding every other drive too. Switched to Promise.allSettled + drop the failed drive, matching this file's own existing "swallow and hide" philosophy for per-page permission-check failures (which already do this at the machine level). New tests cover both failure points.

Merge conflict resolved: master's #2017 (workspace-owns-pane-grid) changed [driveId]/development/layout.tsx's pending-session workspace lookup from a flat workspaces map to selectActiveWorkspace — the same file this PR had refactored to extract shared hooks. Ported that change into the extracted use-drain-pending-session.ts so both the drive-scoped and new global layout pick it up correctly.

Cleanup (converged on by 2-3 finders independently):

• Completed the hook extraction this PR started: isKnownMachine/displayedMachineId derivation was left duplicated across both layouts — now a shared resolveDisplayedMachine.
• Centralized the machine detail-URL builder (buildMachineHref) next to the existing parseSelectedMachineId, so parse/build can't drift apart — MachineTreeSection goes back to a driveId prop instead of an ad hoc basePath string.
• Deduped the near-identical guard-chain early-returns in DriveMachineList/GlobalMachineList into a shared resolveListNotice.
• Deduped the byte-for-byte identical fetcher/globalFetcher in useDriveMachines.ts into one generic machinesFetcher<T>.
• Dropped a GLOBAL_SCOPE magic-string sentinel in favor of undefined (the hook's own "no scope" idiom).
• Fixed a stale doc comment in sidebar-routes.ts that still described the driveless route as "which redirects" / "no driveless twin" — both now false.

Documented, not fixed (explicitly out of scope for this PR): navigating from the global Development view into a specific drive's Development view crosses route trees and disconnects every warm terminal across every drive, not just the one being left. This mirrors an already-accepted tradeoff elsewhere in the surface (switching drives already tears down the other drive's terminals — "a PTY stream must not outlive its drive context") and is recoverable (reconnects on remount), not data loss. A real fix would mean hoisting one MachineKeepAliveHost above both route trees, which is a bigger change than this bug-fix PR warrants. Noted in the global layout's doc comment for whoever picks it up next.

Not changed (considered and deliberately left alone): the per-drive→per-machine canUserViewPage fan-out (already an accepted, documented tradeoff at the per-drive level from #2015, just widened across drives — the file's own comment already anticipates revisiting this "if a drive ever holds enough machines for it to matter"); the 30s poll interval on useAllMachines (this was explicitly directed by the sub-task spec — "same SWR/fetch convention" as useDriveMachines); isAdmin being computed independently client-side in a few places (a pre-existing, 15+-files-wide codebase convention, not something to centralize in this PR); useStickyMachineIds' in-render setState pattern (inherited verbatim from the already-shipped per-drive layout, not introduced by this PR).

All targeted tests green (232+ tests), full monorepo typecheck+build green, lint green.