- Replace AudioGenerateResponse with AudioTriggerResponse to handle both 200 (cached) and 202 (async job) responses
- Add pollAudioStatus() in APIClient that polls /api/audio/status every 2s until done or failed
- Update generateAudio() and prefetchNext() in AudioPlayerService to poll instead of expecting a synchronous URL response
- Bump chapter list font sizes: number xs→sm, title sm→base, date/badge xs→sm, row padding py-2→py-2.5
- AudioPlayerService: replace loading/generating states with a single
.generating state; presign fast path before triggering TTS; fix URL
resolution for relative paths; cache cover artwork to avoid re-downloads;
fix duration KVO race (durationObserver); use toleranceBefore/After:zero
for accurate seeking; prefetch next chapter unconditionally (not just when
autoNext is on); handle auto-next internally on playback finish
- APIClient: fix URL construction (appendingPathComponent encodes slashes);
add verbose debug logging for all requests/responses/decoding errors;
fix sessions() to unwrap {sessions:[]} envelope; fix BrowseResponse
CodingKey hasNext → camelCase
- AudioGenerateResponse: update to synchronous {url, filename} shape
- Models: remove redundant AudioStatus enum; remove CodingKeys from
UserSettings (server now sends camelCase)
- Views: fix alert bindings (.constant → proper two-way Binding); add
error+retry UI to BrowseView; add pull-to-refresh to browse list;
fix ChapterReaderView to show error state and use dynamic WKWebView
height; fix HomeView HStack alignment; only treat audio as current
if the player isActive; suppress CancellationError from error UI
