[nabalone]

This does NOT address TT-7748 in which we will split up the hotkeys. This is an intermediate step that I need to get the Internalize General Resource flow working, should cause no regressions. I don't want the General Resource flow to have to wait on TT-7748

[Copilot]

🟢 Approval recommended

No active issues remain after a full review that verified the hotkey-gating changes against the surrounding code.

No open issues 🎉

What changed in this PR

This PR adds an opt-out hotkeys prop (default true) to WSAudioPlayer, allowing a player to subscribe none of its keyboard shortcuts so that a second, off-screen player can yield all keys to the visible one. It's a low-risk refactor split out of the larger add-resource wizard PR (#741) to ease review. Because the default is true, all existing screens behave exactly as before; the only false caller arrives with #741.

Changes:

• Added a hotkeys prop to WSAudioPlayer and gated every hotkey subscription effect (transport, record, region nav, copy/cut/paste) on it, including adding hotkeys to each effect's dependency array.
• Forwarded the hotkeys prop from MediaRecord to its WSAudioPlayer.
• Added two tests covering no subscriptions when hotkeys={false} and unsubscription when the prop transitions to false.

File	Description
web/​src/​components/​WSAudioPlayer.tsx	Adds the hotkeys prop (default true) and short-circuits all four hotkey subscription effects when it is false, with hotkeys added to their dependency arrays.
web/​src/​components/​MediaRecord.tsx	Adds a hotkeys prop (default true) and forwards it to the embedded WSAudioPlayer.
web/​src/​components/​WSAudioPlayer.hotkeys.test.tsx	Adds two tests verifying no subscriptions when hotkeys={false} and unsubscription when the prop turns false.

Verification notes: I confirmed via a content search that all subscribe(...) calls in WSAudioPlayer.tsx reside inside effects now guarded by hotkeys, that each affected effect's dependency array includes hotkeys, and that MediaRecord contains a single WSAudioPlayer instance which receives the forwarded prop. The new tests use defaultProps with allowRecord: true, so the assertions are meaningful.

_{Review effort: Balanced}

[nabalone]

Why the unmount unsubscribe list in the device-listener effect was deleted (17fc1b6)

That cleanup unsubscribed PLAY_PAUSE, ALT_PLAY_PAUSE, HOME, END, BACK, AHEAD, TIMER, RECORD, LEFT and RIGHT when the player unmounted. It is left over from before TT-7283 (20959ef9), when a single effect owned all the subscriptions. TT-7283 moved each group into its own effect with its own cleanup:

• transport (incl. ALT_PLAY_PAUSE): L1349, L1357-L1364
• record: L1381
• region nav: L1388-L1391

React runs those cleanups on unmount, so for an active player the deleted list unsubscribed keys that were already being unsubscribed. Nothing changes there.

For a passive player (hotkeys={false}), the list was harmful. Those effects return early and subscribe nothing, but the leftover list still unsubscribed every key. HotKeyContext.unsubscribe removes a key regardless of who registered it, so unmounting a passive player stripped the active player's shortcuts (Devin finding 2). Deleting the list fixes that with no ref or guard. It is covered by the new test "does not unsubscribe active keys when a passive player unmounts", which fails on the old code.

Possible follow-up (out of scope here): make HotKeyContext track ownership, e.g. unsubscribe(key, cb) only removes the key if cb is still the registered callback, so components sharing a key can't clobber each other.

🤖 Generated with Claude Code

[Copilot]

🟢 Approval recommended

Every hotkey subscription is correctly gated on the new default-true prop with matching dependency arrays, the forced-unsubscribe removal is justified by per-effect cleanup, and Jest tests cover the new behavior, leaving no open issues.

No open issues 🎉

_{Review effort: Balanced}