Recordings were just a thin named pointer over an already-captured
session's events, so the whole separate feature (recordings table, API
routes, pages, UI) is gone: any completed session can now be named and
replayed directly. Replaying no longer creates a session or duplicates
events of its own - it just bumps the source session's playCount/lastPlayedAt.
Also renames play_sessions/playSession(s) to sessions/session(s) throughout
the schema, queries, API routes, and UI for consistency, and updates the
README to match the new flow.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
getSessionTimeline() grouped/ordered by a bare `bucket` identifier,
assuming SQLite would resolve it against the SELECT list's alias -
it doesn't in this generated query, so every /sessions/:id page and
/api/stats/sessions/:id/timeline request failed with "no such column:
bucket". Fixed by reusing the actual bucket expression object in
groupBy/orderBy instead of referencing it by name. Verified against a
seeded session - previously every request, now correct bucketed data.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W8WkFF5ppURBAB918593Eb
Next.js app with a browser-side buttplug/buttplug-wasm control layer
(server never touches real-time device commands), SQLite storage via
Drizzle, single-secret auth, recordings/replay with device remapping,
a usage stats dashboard, Docker deployment, and Gitea CI.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W8WkFF5ppURBAB918593Eb