fix(orbit): hysteresis on Catch Up visibility so the button stays clickable (#527)

* fix(orbit): keep Catch Up button visible long enough to click

Reported during testing: the Catch Up button "appears for a moment,
then disappears too fast to click". The single-stage debounce only
filtered the show direction (require ≥ 3 s sustained over-threshold
before showing), but on a real-world high-latency session the genuine
drift fluctuates between ~1 s and ~8 s in lockstep with both sides'
chunked `currentTime` updates — so the button vanished as soon as
drift dipped briefly under 3 s, even though the drift baseline was
still 5–8 s.

Add a hide-side hysteresis: once visible, the button stays visible
until drift has been under a tighter 1 s threshold for ≥ 1 s.
Otherwise the noisy 1–3 s drift valleys keep the button up so the
user can actually click it.

Constants left as locals; if testing wants different floors we can
extract.

* docs(changelog): add catch-up hysteresis bullet

* fix(orbit): show Catch Up when host paused + serialise seek-then-pause

Two follow-ons after the hysteresis change:

**1. Show Catch Up even when the host is paused.** The visibility
   gate required `state.isPlaying === true`, but a guest who joined
   while the host was paused still benefits from Catch Up if their
   sync to the host's paused position failed (engine drop, brief
   network blip during initial-sync, manual local seek). Drop the
   `state.isPlaying` gate — the only signal that should matter is
   "drift between us and the host's reported state" — which
   `computeOrbitDriftMs` already computes correctly in the paused
   case (no time-extrapolation, just `guestPos - hostPos`).

**2. Defer the play-state mirror after `player.seek`.** Reported
   symptom: "I press Catch Up, the player pauses; when I press play,
   it resumes from the *old* position and the waveform jumps back."
   Root cause: `player.seek` debounces `audio_seek` via
   `setTimeout(0)`, while `player.pause` / `player.resume` fire their
   invokes synchronously. Pause arrives at the engine first, leaving
   it paused at the pre-Catch-Up position with the seek queued
   behind. When resume happens later, the engine plays from the old
   spot. Push the play-state mirror behind a 200 ms `setTimeout` so
   the seek's invoke lands first.

CHANGELOG.md

@@ -232,6 +232,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 * **Double-clicking the inline play button on an album-track row now suggests/enqueues to the host's queue**, matching the existing double-click-on-row behaviour. Previously the button stopped propagation on click, so its double-click never reached the row's orbit-aware handler — clicking it twice just bounced the "double-click to add" hint toast.
 * **Track preview is now hidden + blocked during an Orbit session.** The preview path runs through the same Rust audio engine as the shared playback, so starting a preview as a guest would clobber the host's track in the local player. The preview button is hidden across all surfaces (album / artist / favourites / playlist / random-mix track lists) via a global `[data-orbit-active]` CSS rule, and `previewStore.startPreview` no-ops as a defensive guard for keyboard shortcuts and any programmatic callers.
 * **Audio reliably starts on join, even after a slow first cold-start** (PR [#526](https://github.com/Psychotoxical/psysonic/pull/526)). Two cases were leaving the guest silent on join: (a) the initial sync's 5 s ready-poll timed out (slow Navidrome warmup), the next pull tick took the cheap "track already loaded" shortcut and fired `seek` + `resume` on a stuck engine — both no-oped against a track that never started; (b) `playTrack`'s optimistic `isPlaying: true` write masked a later `audio_play` rejection, so the guest tick recorded a "successful" sync but the engine had silently fallen back to paused. Both are now handled: the shortcut is gated on engine state matching the host's expected state, and a recovery check at the top of the pull tick resets the anchor whenever the engine is paused while the host is still playing the same track — the next 500 ms fast-poll fires a fresh `playTrack` and audio kicks in.
+* **Catch Up button stays clickable on high-latency sessions** (PR [#527](https://github.com/Psychotoxical/psysonic/pull/527)). On a real-world high-latency session the genuine drift fluctuates between ~1 s and ~8 s in lockstep with both sides' chunked `currentTime` updates — the previous single-stage debounce hid the button as soon as drift briefly dipped below 3 s even though the baseline drift was still 5–8 s, so the button "appeared and vanished too fast to click". Two-stage hysteresis now: show after drift > 3 s for 3 s, hide only after drift < 1 s for 1 s.
 
 ### Context menu — render above the floating player bar