Free-text annotations on any channel or video, searchable from the
global filter box.
## Storage
New `notes` table keyed by (target_kind, target_id):
- target_kind: "channel" or "video"
- target_id: "platform/handle" for channels, video ID for videos
- empty body deletes the row, so the table only holds real notes
DB methods: get_note, set_note (upsert-or-delete), get_all_notes.
## API
- GET /api/notes → { "video:<id>": body, "channel:<plat>/<handle>": body }
fetched once on page load, held in memory for indicators + search
- POST /api/notes/:kind/:id { body } → upsert/delete
## Web UI
- 📝 button on every video card; turns accent-colored when a note
exists. Opens an edit-in-place modal.
- 📝 Add/Edit note… sub-action under each channel in the sidebar.
- Global search now also matches note bodies, so you can find a clip
by what you wrote about it.
## Desktop UI
- Note field in the video detail panel (bottom dock). Loads lazily
from the DB when the selection changes; persists on focus-loss.
## Restore
Backup import merges the notes table too (keep-later-timestamp,
same as watched/positions). Guarded with a table-exists check so
pre-notes backups still import cleanly. RestoreSummary gains
notes_added; both UIs show it.
4 new tests (2 notes DB behavior, restore merge already covered the
shape). 79 pass total.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
YouTube increasingly requires a per-video Proof-of-Origin token bound
to each video ID before handing back format URLs. Without one, yt-dlp
sees empty formats and downloads fail. The upstream solution is the
Brainicism bgutil-ytdlp-pot-provider Python plugin paired with a
long-running server that mints tokens via BotGuard.
We integrate the Rust port (jim60105/bgutil-ytdlp-pot-provider-rs) so
the server is a single static binary, no Node.js required.
## Architecture
- bgutil-pot binary lives alongside the bundled deno + yt-dlp at
~/.local/share/yt-offline/bin/bgutil-pot.
- The matching Python plugin is pip-installed into the bundled venv:
python -m pip install bgutil-ytdlp-pot-provider.
- Downloader spawns the server lazily on first job (port 4416,
loopback only) and kills it on Drop.
- Each yt-dlp invocation gets
--extractor-args "youtubepot-bgutilhttp:base_url=http://127.0.0.1:4416"
appended in spawn_job when use_pot_provider is on and the server
child is alive.
## UX
- Settings → "POT token provider" row with toggle + Install/Update
button (mirrors the bundled yt-dlp row).
- Disabled unless use_bundled_ytdlp is also on (plugin lives in that
venv).
- Refuses install if the bundled yt-dlp venv isn't there yet, with a
helpful 428 error.
## Lifecycle
- Lazy spawn: ensure_pot_server() runs at the top of start() so the
server is up before yt-dlp queries the plugin.
- Drop: kill_server() on Downloader drop so we don't leave an orphaned
process bound to 4416.
- start_pot_provider_update() kills the running child first so the
install can overwrite the binary in place (no ETXTBSY).
3 new unit tests cover URL/extractor-args formatting; 77 total pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The page-bottom footer holding the URL input + quality picker + flags
was a fixed 50-ish-pixel strip eating screen space on every screen,
even though most of the time users aren't starting a new download.
Move all of it into the ⬇ Downloads modal as a "new download" section
above the active jobs list. Same controls (URL, quality, Fast/Live/
Clips), same element IDs so existing helpers (previewDownload,
fullScan, dlLive, refreshTwitchClipsVisibility, updateDlMode) keep
working unchanged. URL input auto-focuses on modal open so the
paste-and-Enter flow is one motion.
Footer shrinks to just the AGPL §13 source notice (required for
deployment compliance, ~12px tall now).
updateDlMode is hardened with null-checks since the labels only exist
while the modal is open.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Backup direction has shipped for a while; this adds the import side
so users can actually round-trip their snapshot.
Database::restore_from_backup attaches the uploaded file and merges
each table with per-table semantics:
- watched / positions / channel_options: keep the later timestamp
- video_flags: bitwise-OR each flag (favourite + bookmark across two
devices both stick)
- folders: insert names that don't exist yet
- channel_assignments: insert when (platform, handle) is unassigned;
folder_id is re-resolved by name so backup/live ID drift is fine
- settings: deliberately skipped (password hash + source_url are
per-deployment state, not user data)
Re-importing the same snapshot is a no-op. Schema validation rejects
non-yt-offline SQLite files up front with a 400.
POST /api/restore/db takes raw .db bytes (100 MB cap, magic-header
check) and returns a RestoreSummary with per-table added counts.
After success, the in-memory watched/positions/flags caches are
refreshed and the library version is bumped so cached /api/library
responses revalidate.
Desktop: 📂 Import library backup… button in Settings, mirrors the
existing 💾 save button. Triggers rescan after the merge so channel
folder assignments take effect.
Web UI: 📂 Import library snapshot… button next to the download
button. Hidden file input + confirm dialog + result summary.
Includes 4 new unit tests in src/database.rs covering the watched +
positions merge, flag OR-ing, idempotency on re-run, and schema
rejection.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
When a job fails, today the user sees a raw stderr line — often
something like "ERROR: [youtube] dQw4w9WgXcQ: Sign in to confirm
you're not a bot" that doesn't tell a non-expert what to do.
New src/error_class.rs scans the failed job's log buffer for the
nine well-known yt-dlp fingerprints and returns:
- an ErrorClass enum (rate-limited, members-only, geo-blocked,
not-found, codec-missing, disk-full, network-error, bad-cookies,
other)
- a one-line human-readable suggested action paired with each class
Auto-populated in Job::drain when state transitions to Failed.
Exposed on JobSnapshot as error_class + error_hint. Rendered in
the web UI Downloads modal (red badge + hint box) and the desktop
downloads panel (colored label + hint text).
Other (no fingerprint matched) is filtered out of the badge so the
existing raw last_line keeps doing the talking for unknown failures.
Includes 10 unit tests covering each pattern + the walks-from-end
priority case.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Tartube parity 1.3 — the grid now sits under a row of filter chips that
narrow what's shown without changing the sort or sidebar view.
Dimensions:
- Watch: All / Unwatched / In progress / Watched
- Date: All / Today / Week / Month / Year / Older (against upload_date)
- Size: All / < 100 MB / 100 MB – 1 GB / > 1 GB
- 💬 Subs toggle: only videos with subtitle tracks
- 🔖 Chapters toggle: only videos with chapter markers
Filters AND together and persist to localStorage alongside view/sort.
They apply across every grid mode (channel, continue, recent, smart
folders) so a single chip selection follows you between views.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The fixed bottom #jobs bar ate up to 40vh of screen real estate even
when only one job was running, and on mobile it competed with the URL
input. Replace it with a ⬇ button in the header that:
- shows a badge with active+queued count (turns accent-colored when >0)
- opens a modal listing all jobs when clicked (or with the 'd' shortcut)
- repaints live from the same WS/polling snapshot that drove the old bar
The footer URL input + quality picker stay where they are — that's the
entry point, not the status display.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
WebSocket progress
- New `/ws/progress` route using axum 0.7's `ws` feature. Upgrades to
a WebSocket that streams the same `ProgressResponse` payload as the
HTTP polling endpoint, pushed instantly whenever a job moves.
- `WebState.progress_tx` is a `tokio::sync::broadcast::Sender<String>`
(capacity 16, lossy — Lagged subscribers resync on the next tick).
- A background tokio task ticks every 500 ms while any download is
running or queued, every 5 s when idle. Skips the JSON encode when
no subscribers are listening, so a single-user install pays zero
CPU for the broadcast when the browser tab is closed.
- JS replaces the per-interval HTTP poll with a WebSocket connection.
The polling path stays as fallback for mobile carriers / reverse
proxies that strip WS — when the socket closes or fails to open,
schedulePoll() resumes immediately and a reconnect attempt fires
after 5 s.
- Initial snapshot is sent right after subscribe so the UI is live
before the next tick.
Mobile-responsive web UI
- Existing `@media(max-width:640px)` block tightened:
- Hide the H1, AGPL footer notice, sort dropdown, and library-size
counter — they don't fit and aren't actionable.
- Header buttons get min 34×34 hit targets (Apple HIG / Material).
- `.card-foot` wraps and grows to 34×34 buttons so the five action
icons (play / watched / favourite / bookmark / waiting) stay
tappable instead of cramming into thumb-unfriendly 22 px chips.
- Modals go full-screen (no 10 px page-margin) so the iPhone safe
area doesn't crop content.
- Footer (URL bar + download button) wraps onto two rows so neither
squeezes the other off-screen.
- Bulk-actions toolbar wraps and gets bigger buttons.
- Per-job stdout preview hides on small screens.
- New `@media(max-width:380px)` collapses the grid to single column
for very narrow phones.
55 unit tests still pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
🎲 button in the header picks a random video from
{unwatched + downloaded + has video_url} and opens it directly in the
player. Falls through to "any downloaded video" when everything is
already watched, with a status message that lets the user know.
Tartube has nothing like it — small bit of extra "surpass" beyond
plain parity.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Reloading the tab no longer blows away the user's context. The active
view (channels / continue / recent / favourites / bookmarks / waiting /
specific channel / specific playlist), the current filter input, and
the chosen sort all survive a hard refresh via localStorage.
- New `saveUiState()` serializes `{view, search, sort}` after every
interaction that could change them: each setX() view switch, search
oninput, sort onchange.
- `currentViewToken()` collapses the seven `showX` booleans + the two
index integers into a single string. Channel-view tokens reference
the channel by name (`channel:<name>|<playlistIdx>`) instead of by
array index, so a library re-order doesn't strand the user on the
wrong channel.
- `restoreUiState()` runs once after the first library load completes —
the channel resolution needs the `library` array populated, so the
init block was wrapped in an async IIFE to sequence the two steps.
Channels deleted between sessions fall back to "All" cleanly because
`findIndex` returns -1 and we just leave the show* booleans at their
default.
55 unit tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Four shortcuts: `/` focuses the filter input, `r` triggers a library
rescan, `Esc` closes the topmost modal, `?` opens a help dialog
listing the bindings. All disabled while the focus is inside a text
input / textarea / contentEditable element so typing a video title
doesn't accidentally fire actions — except Esc, which always closes
modals (even from an input).
A new `?` button next to the existing rescan / maintenance / settings
icons makes the shortcuts discoverable; the help dialog is plain HTML
table with monospace key chips.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Bulk tagging
- Desktop bulk-mode toolbar grows three new buttons next to ✓ Watched
and ○ Unwatched: ★ Favourite, 🔖 Bookmark, ⏳ Waiting. Each applies
the corresponding flag to every video in the current bulk selection.
- New `App::bulk_set_flag` mirrors the existing `bulk_mark_watched`:
iterates selection, persists each flag via `Database::set_video_flag`,
mirrors into the in-memory `flags` bundle, busts the cards-cache so
the smart-folder counts refresh on the next render.
- Web bulk-actions row gets the same three buttons. New `bulkFlag(flag)`
fires the POSTs in parallel via Promise.all and reloads the library
once they complete.
- Tagging in bulk only ever sets a flag — no toggle. Removing flags in
bulk is rarely what users want; the per-card buttons handle un-flag.
Channel-name search
- Both UIs already filter by video title + id; extend to match the
channel name too. Searching "Linus" now surfaces every video from a
channel called "Linus Tech Tips" without typing the title.
- Description matching deliberately skipped — would require reading the
.description sidecar per-video per-keystroke. Mentioned in the code
comment as a future "load descriptions into the search index on
rescan" pass if real users want it.
- New JS helper `matchesSearch(v, chName, q)` centralises the four
call sites in `currentVideos()` (continue / recent / smart-folder /
default).
55 unit tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Two small follow-ups to the folder hierarchy that round out the
folder-as-a-profile workflow and give the user a one-click way to keep
an offsite copy of their library state.
Per-folder check
- New `POST /api/folders/:id/check` mirrors the existing
/api/scheduler/run endpoint but scoped to one folder's members.
Each channel's stored DownloadOptions (quality / rate limit /
match-filter / extra args) applies just like a scheduled re-check.
Returns 409 when downloads are already running.
- Web folder-manager modal: each non-empty folder gets a "⬇ Check"
button next to Rename / Delete. Status line reports the count of
channels queued.
- Desktop folder-manager window: same — a small ⬇ icon-button per row,
visible only for folders with at least one member. Same options-
honouring re-check path as the right-click "Check for new videos"
action.
- This effectively gives us "Profiles" lite: organize channels into a
folder, then trigger a batch re-check whenever you want.
DB backup
- New `GET /api/backup/db` reads `<channels_root>/yt-offline.db` and
streams it back with `Content-Disposition: attachment` so the
browser downloads instead of rendering. Filename includes the unix
timestamp so downloads don't clobber. Returns 404 when running in
in-memory mode (no DB file to dump).
- We deliberately don't bundle config.toml + cookies.txt here:
config is short to recreate and cookies.txt carries live session
credentials that shouldn't fly over the wire unprompted.
- Web settings modal gains a "⬇ Download library snapshot" button in
a new Backup section, with a hint explaining what's covered (watched
/ favourites / bookmarks / waiting / channel-options / folders).
55 unit tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Users can now organize channels into named folders independent of the
underlying platform groupings. v1 ships single-level (no nesting) since
that's where ~90% of the organizational value lives — nested folders +
drag-to-reparent + per-folder cascading options are deferred to v2 once
we have real-world feedback on what the workflow needs.
Data layer
- New `folders` SQLite table: id, name (UNIQUE), position, created_at.
- New `channel_assignments` table: (platform, handle) → folder_id with
ON DELETE CASCADE so deleting a folder cleanly unfiles its channels.
- Database methods: create_folder, rename_folder, delete_folder (with
PRAGMA foreign_keys ON for the cascade), list_folders,
set_channel_folder, get_all_channel_assignments.
- New `FolderRecord` struct (id/name/position) is `Serialize` so it
rides on the /api/library response.
Library
- `Channel` gains `folder_id: Option<i64>` (None = Unfiled).
- New `library::apply_channel_folders` mirror of `apply_channel_options`
populates it from the bulk DB map after every scan / rescan.
Web API
- `POST /api/folders` create with `{name}` → `{id}`.
- `POST /api/folders/:id/rename` rename with `{name}`.
- `DELETE /api/folders/:id` cascade-deletes assignments.
- `POST /api/channels/:platform/:handle/folder` upsert/clear via
`{folder_id: <i64|null>}`. All endpoints bump the library ETag.
- `/api/library` response now carries a `folders: [...]` array so the
client can render the sidebar grouping without a second round trip.
Web UI
- Sidebar gains a "📁 Folders" section above the platform groupings.
Each folder lists its member channels (with platform icon prefix);
channels with `folder_id` are pulled out of their platform section
to avoid duplication. Unfiled channels keep platform grouping.
- "manage" / "new" button next to the Folders heading opens a modal
with the full folder list (member counts inline) plus rename /
delete buttons and a Create input.
- New per-channel sub-action "📁 Move to folder…" opens a small
picker modal with one row per folder + "— Unfiled —". Current
assignment highlighted.
Desktop UI
- Sidebar refactored to build a `Vec<SidebarItem>` ordered list
(FolderHeader / FolderManageBtn / PlatformHeader / Channel) and
iterate it once, instead of the previous flat
`for i in 0..library.len()` loop. Folder section renders above
the platform sections; channels with `folder_id` are skipped from
the platform-grouped loop so they appear in exactly one place.
- New `folder_manager_window` + `move_to_folder_window` egui windows
cover create / rename (via the create-buffer field) / delete and
per-channel folder picking. Right-click context menu on a channel
gains "📁 Move to folder…".
- App struct gains: `folders`, `show_folder_manager`,
`folder_create_buffer`, `show_move_to_folder`, `move_to_folder_target`.
Tartube spec checklist
- Folder hierarchy box ticked (with the v2 deferrals noted in the
inline annotation).
55 unit tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Three Tartube-parity items shipped together since they touch the same
schema and rendering paths.
Per-video state flags
- New `video_flags` SQLite table with `bookmark`/`favourite`/`waiting`/
`archive` columns (one row per video that has at least one flag set).
- `Database::set_video_flag` validates the flag name against an
allow-list before interpolating into the SQL — column names can't be
parameterised in SQLite, but the allow-list keeps it injection-free.
- `Database::get_video_flags` bulk-loads into a new
`VideoFlagsBundle { bookmark, favourite, waiting, archive: HashSet<String> }`
so subsequent reads are O(1) in-memory hits.
- `WebState.flags` (Mutex<VideoFlagsBundle>) + `App.flags`
(VideoFlagsBundle, GUI is single-threaded) populated at startup.
- `POST /api/videos/:id/flags/:flag` toggles a flag; the in-memory
bundle is mirrored immediately so the next /api/library reflects the
change without a rescan. Library ETag is bumped.
- `VideoInfo` JSON gains `bookmark` / `favourite` / `waiting` /
`archive` boolean fields.
- `archive` is wired through everywhere but no UI gate yet — it's
reserved for the future auto-delete feature.
Smart folders
- Both UIs gain "★ Favourites", "🔖 Bookmarks", "⏳ Waiting" sidebar
entries between Continue Watching and Channels. Each is hidden when
its set is empty, so a fresh install isn't polluted with zero-count
rows.
- Desktop: new `SidebarView::Favourites / Bookmarks / Waiting`
variants. `compute_cards` filters the library by the matching
`self.flags.<set>` HashSet and returns the resulting cards using the
same rendering path as the other views.
- Web: `showFavourites/Bookmarks/Waiting` flags + smart-folder branch
in `currentVideos()`. Refactored the view-mode selectors through a
new `resetViewSelectors()` helper so adding three more booleans
didn't quadruple every `setX` function.
Per-card flag toggles
- Web: each video card grows three new action buttons (☆/★ favourite,
🔖 bookmark, ⏳ waiting) next to the existing watched toggle.
Optimistic UI — flip in-memory immediately, undo on server error.
- Desktop: same three buttons next to Watched in the card row, deferred
through a `toggle_flag_card: Option<&'static str>` to keep the
borrow checker happy. New `App::toggle_video_flag` helper persists
the flip to SQLite and invalidates the card cache.
Comments capture
- `DownloadOptions` gains `fetch_comments: bool`. When set, the
downloader emits `--write-comments` so yt-dlp embeds the full
comment tree into info.json.
- New `GET /api/comments/:id` reads the array out of info.json and
returns a slimmed-down version (id, author, text, likes, parent,
time) — strips the multi-MB extras yt-dlp can otherwise inline.
- Web player modal grows a 💬 button that opens a separate modal
with a threaded viewer: replies indented by parent id, "N comments"
header, helpful message when nothing's captured pointing at the
channel-options toggle.
- Channel-options dialog in both UIs gains a "Fetch comments"
checkbox + hint about the slowdown.
Tartube spec checklist
- Three more boxes ticked: per-video state flags, smart folders,
comments capture. Score now: 5 ahead, 5 behind, 1 tied + matching
on 4 parity items. (Half done.)
55 unit tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Closes the single largest gap with Tartube documented in docs/tartube-spec.md.
Users can now attach a small set of yt-dlp overrides to any individual
channel; the overrides apply automatically during scheduled re-checks
and the right-click "Check for new videos" action.
Data layer
- New `channel_options` SQLite table keyed on (platform, handle), with
`options_json TEXT` blob + updated_at. PK ensures one row per channel.
- Database gains get_channel_options / set_channel_options /
delete_channel_options / get_all_channel_options. The bulk getter is
used by the library scanner so each rescan hydrates options without
per-channel SQL round-trips.
- `DownloadQuality` derives Serialize/Deserialize so it can roundtrip
through the options blob.
New module: src/download_options.rs
- `DownloadOptions` struct with 9 fields:
quality (Option<DownloadQuality>) — per-channel quality cap
audio_only (bool) — force --extract-audio chain
limit_rate_kb (Option<u32>) — --limit-rate
min_filesize_mb / max_filesize_mb — yt-dlp filesize filters
date_after (Option<String>) — --dateafter YYYYMMDD
match_filter (Option<String>) — yt-dlp --match-filter passthrough
subtitle_langs (Vec<String>) — --sub-langs CSV
extra_args (Vec<String>) — raw yt-dlp passthrough (Tartube's
`extra_cmd_string`)
- `apply(&mut Command)` appends the right flags conditionally.
- `is_empty()` lets the API layer collapse no-op writes into DELETEs so
we don't keep useless rows around.
- `from_json()` falls back to defaults on malformed rows so a schema
drift doesn't take a channel offline.
- 8 new unit tests cover apply-emits-correct-flags, JSON roundtrip,
corrupt-fallback, and is_empty.
Library
- `Channel` gains `download_options: DownloadOptions` (default empty).
- New `library::apply_channel_options(channels, map)` helper that the
caller invokes after `scan_channels` to hydrate from the bulk DB load.
- All five `scan_channels` call sites updated.
Downloader
- `Downloader::start` gains a new `channel_options: Option<&DownloadOptions>`
parameter. Applied last so per-channel overrides win over global
defaults. The explicit "submit from URL bar" flow passes None
(channel unknown until yt-dlp resolves it). Scheduled re-checks +
right-click checks pass the channel's stored options. Channel-options'
`quality` field overrides the hard-coded DownloadQuality::Best for
re-checks.
Web API
- New routes on /api/channels/:platform/:handle/options:
GET — fetch (returns defaults when no row exists)
POST — upsert (empty body collapses to DELETE)
DELETE — clear overrides
- All three update the in-memory library snapshot immediately so the
next re-check sees the change without waiting for a rescan, and bump
the library_version ETag so polling clients pick up the new state.
Web UI
- Sidebar each channel's expanded section gains a "⚙ Channel options…"
entry next to "Check for new videos".
- openChannelOptions() fetches the current options + renders a modal
with one input per field. Save / Clear / Cancel buttons. The Clear
button confirms via window.confirm() because it's destructive.
- All form fields validate locally before POSTing; empty numeric
fields map to null.
Desktop UI
- New `channel_options_window` egui::Window opened from the existing
channel right-click context menu (new "⚙ Channel options…" item).
- Form mirrors the web side: ComboBox for quality, checkbox for
audio-only, TextEdit (singleline + multiline) for the other fields.
- Save / Clear / Cancel buttons; saves go straight to SQLite and the
live `App::library` so the next scheduled-check or right-click runs
with the new overrides.
Documentation
- `docs/tartube-spec.md` checklist: per-target download options ticked
with a note describing the v1 scope.
55 unit tests pass (47 + 8 new).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Four nice-to-have items, plus surfacing the upload date in both UIs
since the user explicitly asked.
Upload date display
- Desktop: card meta row now includes a `YYYY-MM-DD` cell between the
ID and the duration; detail panel header shows it as `📅 YYYY-MM-DD`.
- Web: video player modal title gains a date+duration subline so the
viewer has context without opening the metadata pane (card meta
already shows the date).
- New `format_upload_date()` helper on the desktop side; the web side
reuses the existing `fmtDate()` JS.
Extract HTML_UI to include_str!
- `LOGIN_HTML` and `HTML_UI` moved out of `src/web.rs` into
`src/web_ui/login.html` and `src/web_ui/index.html`. The Rust file
drops from 2915 lines to 1806; editors now syntax-highlight the
HTML/CSS/JS, and UI diffs no longer overwhelm the Rust diff view.
- `include_str!` bakes them into the binary at compile time, so there
is no runtime fs lookup and no extra deploy artifacts.
Twitch clips-only mode
- `classify_twitch` now recognises `twitch.tv/<user>/clips` as a
Channel URL (was Unknown), so yt-dlp's TwitchClips extractor is
reachable from our normal download pipeline.
- Both UIs gain a "Clips only" toggle visible only when the URL is
a Twitch channel (not a specific VOD/clip). On submit, the URL is
rewritten to `twitch.tv/<user>/clips` before dispatch.
- 1 new platform test covers the clips-listing classification.
DB connection pool (r2d2_sqlite)
- `Database` now wraps an `r2d2`-managed SQLite pool. File-backed
databases get up to 4 connections; the in-memory fallback is capped
at 1 (since each in-memory connection is a fresh empty DB by default
in SQLite).
- `WebState.db` drops its outer `Mutex` — concurrent handlers each
check out their own connection from the pool. Static-file fetches
through `/files/` no longer serialize on the watched/positions/
password lookups that the auth middleware performs.
- Pool init failures are converted to `rusqlite::Error` so the
existing `Result<T>` return types still work.
47 unit tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
