- pick the produced file safely: Path("") is Path("."), which passes .exists(),
so a missing yt-dlp filepath skipped the staging fallback and crashed in
with_name ("PosixPath('.') has an empty name") on non-YouTube sources
- unwrap a playlist/multi_video extraction to the entry that carries the media
- run_ffmpeg: drain stderr while it runs (a full 64KB pipe blocked ffmpeg
forever and hung the worker thread), 1s cancel polling, stall watchdog and a
bounded wait on exit
- escape apostrophes/backslashes in the ffmpeg concat list (pre-0.51.0 names)
- sweep orphaned HLS segment dirs at startup (nothing else ever reaped them)
The atomic claim added in trigger_job only covered the manual path; a scheduled
fire sets running via _job without claiming, so a manual trigger racing it could
still double-run, and the claim (set outside _job) could leak if _job threw
before its own cleanup. Move _claim_running to the top of _job — the one point
both the scheduled fire and the manual trigger pass through — and release it in
_job's finally, so exactly one run wins whichever path it comes from and a
failure can never wedge the job "running". trigger_job's _is_running check is now
just a best-effort fast answer for the UI. Unit-test _claim_running (and
is_messageable_user).
- C-3.7: fold a per-link password_version into the signed watch grant (migration
0059 adds the column) and bump it on every password change, so rotating a share
link's password invalidates outstanding grants at once instead of letting them
live out the 6h TTL. make_grant/check_grant take the version; update_link bumps.
- C-3.8: the share recipient picker and share-by-email now reuse the messageable
filter (not-demo AND active AND not-suspended) instead of only excluding demo,
so suspended/deactivated addresses aren't leaked or reachable as targets.
- C-3.9: GET /keys/{user_id} adopts get_thread's MB2 guard — a non-messageable
target with no shared history returns the same "User not found" as a missing id,
so it can't be probed to enumerate users or fetch suspended users' keys.
- C-3.12: _register_account catches IntegrityError on the insert (the select-then-
insert TOCTOU) and treats it as the already-registered no-op.
- C-3.13: trigger_job claims the running flag atomically under _activity_lock
(compare-and-set) instead of a bare check, closing the double-start window.
- Move _messageable/is_messageable_user into a shared app/userscope.py so downloads
doesn't import the messages route module.
The set-once key guard made a lost passphrase a permanent wall (409 on re-setup,
KeyGate offered only unlock/setup). Add an escape hatch:
- POST /api/messages/keys/reset drops the user's MessageKey so a new passphrase
can be set. Because every conversation key is derived from the keypair,
replacing it orphans all stored ciphertext in both directions, so the now-dead
`user` messages are deleted too (clean slate; system messages untouched).
Password accounts must re-verify (a hijacked session must not silently reset
messaging and read future messages); Google-only accounts rely on the session
+ the client danger-confirm. Rate-limited per user.
- KeyGate gains a "Forgot passphrase?" path in unlock mode → a danger reset view
(spells out the permanent history loss; a current-password field for password
accounts) → resets and returns to a fresh setup.
- e2ee.resetLocal wipes this device's key + derived conversation keys + the
IndexedDB copy; api.resetMessageKey; HU/EN strings.
Re-add the 127.0.0.1 callbacks to the dev host allowlist (now registered in the
Google console alongside the localhost ones). Deriving the redirect_uri from the
request host keeps the whole round-trip — login, Google callback, and the
session/state cookie — on one origin; a cross-host fallback (127.0.0.1 → the
fixed localhost callback) would silently lose the OAuth state, since a session
cookie set on 127.0.0.1 is never sent to localhost. Clarify that in the comment
and cover all four hosts in the unit test.
- OAuth-cancel feedback now keys on the live session, not link_uid: ANY
signed-in cancel (add-another-account via /auth/login, link, or upgrade)
redirects to /?oauth=cancelled and gets a neutral "Google sign-in was
cancelled" toast — the add-account path set no marker and was silent, and
the shared ?link= message wrongly called an upgrade "account linking"
- drop 127.0.0.1 from the dev OAuth host allowlist: only the localhost
callbacks are registered in the Google console, so a 127.0.0.1 origin now
falls back to the fixed callback instead of a redirect_uri_mismatch
- Welcome: extract the shared goResend handler (banner CTA + sign-in link) and
merge the two adjacent signin-only blocks
- link/upgrade OAuth cancel now redirects to /?link=cancelled (a channel App.tsx
surfaces as a toast) instead of the pre-auth /?login=oauth_cancelled banner a
signed-in user never sees; a plain sign-in cancel is unchanged
- the login banner (suspended / oauth_cancelled) is now state, cleared on the
first deliberate action like the verify banner — no more stale lingering
- add a "Didn't get the verification email?" resend entry on the sign-in screen
so a never-arrived verification mail isn't a dead end (was only reachable from
the expired-link banner)
- unit-test _oauth_redirect_uri (dev-host derive, unknown-host fallback, prod
ignores the Host header — the injection guard)
- collapse the 5-deep title/button ternaries into Record<Mode,string> maps
A Google login started on the Vite dev server (:5173) always came back to the
fixed :8080 callback, because login/link/upgrade passed the static
settings.oauth_redirect_url. Derive the redirect_uri from the request origin
instead, but only in local dev and only for a known dev-host allowlist —
production still uses the fixed configured URL and never trusts the Host header
(host-injection guard). Set the Vite proxy to changeOrigin:false so it forwards
the real Host:localhost:5173 (it was rewriting it to the 127.0.0.1:8080 target).
Both callbacks must be registered in the Google console.
- password reset also sets email_verified (the mail proves the mailbox),
rescuing an account that never clicked the verify link; is_active (admin
approval) is deliberately left untouched
- add POST /auth/verify/resend (rate-limited, off-response-path, anti-enum)
+ a "resend" mode/CTA on the verifyInvalid banner, so an expired/lost
verification link is no longer a dead end (re-register was a silent no-op)
- OAuth cancel/deny now redirects to /?login=oauth_cancelled with a friendly
banner instead of a raw 400 JSON page; pop the oauth_link markers BEFORE the
token exchange so an aborted link flow can't poison the next clean sign-in
Follow-up to user prod testing of 0.51.0:
- re-download now clears job.display_name so the card title + Content-Disposition filename are
re-derived from the freshly-cleaned title (previously an auto-name from an older un-cleaned
title — still carrying a "… views · … reactions |" prefix — survived the rebuild).
- on-disk AND served names are now a strict shell-safe slug: only [A-Za-z0-9._-] survive; every
other char (the "!" the user saw, plus ( ) [ ] & $ ' " ; | ? * …) collapses to "_". The id no
longer uses [brackets] (glob metacharacters). Shared _slug helper for sanitize + display_filename.
- GC gains a 5th pass: sweep on-disk files/dirs no ready asset owns (old accented channel dirs, a
poster from a superseded naming scheme left after a re-download), skipping the dot system trees
and anything newer than an hour so an in-flight download is never touched.
- Tests: slug strips !/$/()[]; rel_path is bracket-free; orphan-sweep keeps owned/system/fresh files.
- redownload now ALWAYS rebuilds into a fresh private asset instead of ever reusing a shared
cached file: the cache sig is salted with job_id + current asset_id, so the fork is distinct
from both the dedup pool and the job's own current asset. This resolves the co-held silent
no-op (finding #1) — a shared asset is left intact for its other holders while THIS job
re-fetches its own copy. Resolving the fork BEFORE releasing the old hold (and the unique
salted key) also removes the IntegrityError-rollback-after-file-delete window (finding #2).
- format_sig gains an optional `salt` for that fork identity.
- titles: don't consume the space after `|`, so a "stats | #hashtags" remainder keeps the
leading whitespace _TRAILING_HASHTAGS needs — and fall back to the prefix+hashtag-stripped
text, so neither the stats nor the SEO hashtags are resurrected when a title empties (#3).
- worker: early-fill asset.title falls back to the video id so a title that cleaned to empty
doesn't show a blank card while downloading (#5).
- Tests for the salt fork and the hashtag-remainder case.
(Finding #4 — quota on re-download — intentionally not changed: a re-download reuses the
existing job and re-fetches ~the same bytes, so calling check_enqueue would wrongly reject on
the job-count cap while footprint is unchanged.)
- redownload: never clobber a co-held (deduped) asset. Drop the force-reset-to-pending;
_release_asset already deletes+recreates a solo asset (genuine re-fetch), while a shared
ready asset is left intact and the re-queued job re-points to it via _finish_from_cache —
so one user's re-download can't delete another's file mid-serve (finding #1).
- redownload: reset queue_pos to the tail so a re-download doesn't front-run newer queued
jobs on the worker's `ORDER BY queue_pos` claim (finding #3). Made service.next_queue_pos public.
- formats: video-only mp4 selector prefers every video-only option before any progressive
`best`, so a "Video only" download stays audio-free whenever a video-only stream exists (#2).
- titles: a title that is only the engagement prefix no longer resurrects the stripped stats
(return the stripped text / empty, not the raw) (#4).
- worker: probe the file's duration when yt-dlp gave none, so a long re-encode shows real
progress instead of a frozen 0% (#5); tighten the ensure_browser_playable docstring to its
actual mp4-profile scope (#6).
Mini-epic "Downloads correctness & UX" (S1–S3), on top of the iOS-AV1 + 720p fixes.
S1 — naming & titles:
- ASCII-fold on-disk dir/file names AND the served (Content-Disposition) filename: no
spaces (underscore-joined), no accents (á→a, ő→o, ß→ss), pure ASCII — portable and free
of the `?` mojibake accented names show in non-UTF-8 tools. (Forward-only; existing files
clean up via re-download.)
- Strip a leading social engagement prefix ("1.6M views · 66K reactions | …") from titles.
S2 — universal browser playability (generalizes the iOS re-encode):
- ensure_browser_playable guarantees the stored file is H.264 + AAC/MP3 in mp4 — the only
combo every browser decodes. Re-encode just the offending stream(s) (video for AV1/VP9/HEVC,
audio for Opus/etc), or remux when only the container is wrong. Gated on the mp4 compat
profile (an explicit vcodec / non-mp4 custom profile opts out).
S3 — force re-download (preserves share links):
- POST /downloads/{job_id}/redownload keeps the job row so its public DownloadLink survives
and resolves to the freshly-downloaded file; deletes the old file and re-fetches under the
current rules. A failed re-download leaves the job in error (the intended broken-link case).
- Library "Re-download" action (icon + confirm), api method, hu/en strings.
Tests: naming/title-strip (test_naming), browser_transcode_flags matrix; existing
download_filename expectation updated for the underscore policy.
Shared /watch links showed a "broken play button" on iPad (any iOS
browser — all forced onto WebKit) while playing fine on desktop Chrome.
Root cause: the stored mp4's video codec was AV1 (confirmed via ffprobe),
which WebKit can't decode. The existing H.264 guard was only a
format_sort *preference*, so a non-YouTube source (Facebook) that ships
AV1 as its best video-only stream slipped through.
- B1: for the mp4 compat profile, prefer H.264 at the SELECTION level
with a progressive fallback (bestvideo[vcodec^=avc1]+bestaudio /
best[vcodec^=avc1] / anything) so an avc1 progressive is chosen over
an AV1-only video-only stream.
- B2: re-encode the rare AV1/VP9-only result to H.264+AAC (+faststart)
in the worker before storing, gated on the compat profile; corrects
the asset codec metadata to match.
- Bump _SIG_VERSION 2->3 so cached AV1 assets re-derive.
- Default download preset is now 720p (migration 0058 re-seeds builtins
720p-first; ProfileEditor BLANK 1080->720).
- Tests for the selector, compat opt-outs, and the transcode predicate.
The demo account is shared, so writing deep_requested onto its subscription polluted
shared state (harmless today — the scheduler's deep paths are human-only — but a
future scheduler path trusting the flag would leak quota). Guard the flag WRITE at the
source (not user.is_demo), which also makes deep_turned_on always False for demo, so
the redundant is_demo check on the immediate backfill is removed.
/code-review (S1, finding 1): sanitize truncated the title at 120 CHARACTERS, but a
path component's hard limit is 255 BYTES and rel_path packs channel + date + id + ext
into that same component — so a long accented/multi-byte title (Hungarian titles are
common here) could already overflow to ENAMETOOLONG, and the new _a{id} suffix shrank
the margin further. sanitize now truncates on a UTF-8 boundary by byte length, so the
title (≤120B) + channel (≤80B) + fixed parts stay comfortably under 255B. Two tests:
byte-cap on accented input, and every rel_path component ≤255B for a pathological
long-accented channel+title+asset_id.
A MediaAsset is unique on (source_kind, source_ref, format_sig), so the same video in
two formats is two assets — but rel_path keyed only on video_id, so they computed the
same path: the second download silently overwrote the first, and deleting either
removed the shared file. rel_path now takes an optional asset_id and appends "_a{id}"
(which Plex ignores — the .nfo sidecar carries the metadata); the worker passes the
asset id. Only affects NEW downloads (existing assets keep their stored rel_path).
Covered by a new test (two ids → two paths; omitted → the plain name, unchanged).
The demo account could spend real YouTube quota where the human-only guard was
missing (the channel_detail enrichment already had it):
- channels PATCH: an immediate run_recent_backfill when deep_requested is turned on;
- discovery: the channel-detail enrich on the Discover tab;
- feed get_video_detail: the off-catalog videos.list lookup for a video we don't
store (the DB-hit path stays free; demo now gets a 404 for unknown videos).
The asset-busy and claim-fail branches wrote status='queued' unconditionally, so a
pause/cancel the route had just written was clobbered back to queued and the job
silently revived. New _requeue_if_running does the requeue as a conditional UPDATE
(...WHERE status='running'), a no-op once the route has moved the job off 'running'.
C-3.5: /auth/request-access had no rate limiter (unlike register/login/reset/demo)
and answered "approved" for an allow-listed email — a public enumeration oracle for
the allow-list, and an unthrottled invite/admin-mail firehose. It now rate-limits per
client IP and returns a uniform "pending" for every outcome (allowed, pending, new,
throttled); an already-allowed user can still just sign in with Google.
C-3.11: _consume_token was a read-check-write, so two concurrent requests could both
consume the same verify/reset token. It's now a single conditional UPDATE
(...WHERE used_at IS NULL AND expires_at >= now RETURNING user_id) — only the request
that flips used_at gets a user back.
- Backend: add channel description to the shared _channel_summary projection, so both
the subscribed list and the discovery list carry it (no migration; description is
already stored from the channels.list call)
- New AboutCell: one truncated line in the table + full text in a portalled, always-on
glass hover/focus overlay (wider than Tooltip, flips above the anchor near the bottom,
closes on scroll/resize)
- About column (after Channel) on both the subscribed and discovery tables; hidden in the
mobile card fallback
Lazy enrichment only covers viewed titles, so an actor's count/results were
partial. Add a background backfill: each plex_sync rich-fetches a bounded batch
(60) of not-yet-enriched movies/shows and persists their FULL cast, stamping
people_enriched_at so each is fetched once — the whole library is covered over
successive runs (migration 0057 adds the marker). Extracted the enrichment into a
shared app/plex/people.py (cast_from_meta / enrich_row_people / prune_orphan_people
+ the person-image host/proxy) so the info-page lazy path and the batch share ONE
implementation — no duplication; the info page also stamps the marker so the batch
skips it. Verified: an 8-item batch grows those movies' cast_names 3->24 and fills
plex_people (32->224); the info-page cast is intact after the _rich_meta refactor.
(Plex cast-explore S1b)
Guard the prune's NOT IN against the classic NULL trap: a JSON-null cast array
element would surface as SQL NULL in the referenced set, making name NOT IN
(…, NULL) evaluate NULL for every row → silently prune nothing. Filter NULLs out
of the subquery. (Plex cast-explore S4 review)
The plex_people headshot cache can outlive its subjects — when a title leaves the
Plex library, cast/crew who appear in nothing else are orphaned. prune_orphan_people
deletes rows whose name no longer appears in ANY movie/show cast_names/directors,
run at the end of every plex_sync (best-effort; a jsonb_typeof='array' guard skips
scalar/null JSONB values). The filterable cast_names self-clean with their row;
this cleans the only per-person cache. Verified: an injected orphan is dropped, real
people kept. (Plex cast-explore S4)
The actors currently in the filter now render as cards above the poster grid —
headshot + name + in-scope title count + a remove x — so you SEE who you're
exploring, not just chips in the rail (generalises the old search-only person
cards to the applied filter). New GET /plex/people/cards (photo from the
plex_people cache, count live against cast_names); PlexBrowse fetches + renders
them; i18n en/hu. Verified: John Goodman (11) / Douglas M. Griffin (1) / Bradley
Cooper (14) with photos + counts. (Plex cast-explore S3)
Multiple selected actors now combine as OR (any) by default — click several cast
and see films featuring ANY of them — instead of the old hardcoded AND that
emptied the grid. An AND (all) toggle in the ACTIVE filter section (shown for 2+
actors, mirroring the genre match toggle) narrows to films with ALL of them.
Backend actor_mode param mirrors genre_mode. Verified: OR John Goodman|Douglas
M. Griffin = 11 films; AND = 1 (10 Cloverfield Lane). (Plex cast-explore S2)
The cheap section listing that drives the main sync only stores the top ~3 cast,
so any cast member below that was unfilterable (clicking them AND-emptied the
grid). Now the movie/show info fetch (/item, /show) — which already pulls the full
rich cast — persists it onto cast_names + people_text (so every displayed cast
member becomes filterable) and caches each headshot in the new plex_people table
(migration 0056). Idempotent (only grows a thinner stored set); best-effort (never
500s a read). Verified: 10 Cloverfield Lane 3->8 cast after viewing; plex_people
populated with photos. (Plex cast-explore S1a)
Pick any number of channels; they OR together and every other filter applies
across the lot — "the last week's unwatched from these three".
`channelId`/`channelName` are replaced outright by `channelIds: string[]`, no
legacy reader (standing rule: no back-compat unless asked — a saved view written
against the old field simply loses its channel filter). The wire and the share
URL move with it: repeated `channel_ids` params, `?channel=a,b`. Backend ORs via
`Video.channel_id.in_()`; the count endpoint shares the same params object.
Two things the migration exposed:
- SavedViewsWidget fed DB blobs straight into serializers that now index an
array — the FeedFilters type says nothing about what an older build wrote, so
the blob is coerced where it enters. It white-screened the app before this.
- Chips carried channelName to label themselves. Without it they resolve names
from the channel cache, so they now fall back to the id, not a shared
"This channel" that would render N identical chips while the cache loads.
Only English and Hungarian remain. Drops the de locale files, the LANGUAGES
entry, the Google-locale mapping in auth, and the German welcome-message
template. A legacy "de" preference falls back to English gracefully. Content-
language detection and subtitle languages are untouched (media, not UI).
- apply_channel_details: (branding.get("channel") or {}) so a "channel": null
in brandingSettings can't AttributeError-abort the enrich batch.
- topicLabels: drop empty labels so a malformed topic URL can't render a blank chip.
- fix: reserve the scrollbar gutter (scrollbar-gutter:stable) on the channel
page's scroll container, so switching between the tall Videos tab and the
short About tab no longer shifts the banner/avatar/buttons a few px sideways
(the vertical scrollbar was appearing/disappearing between the two tabs).
- About tab now shows Country (flag + localized name), Language, Topics
(topicCategories → readable chips), and Keywords (brandingSettings keywords
parsed into chips, quoted multi-word tags kept whole). country/language/topics
were already stored; keywords is new (migration 0055_channel_keywords, mapped
in apply_channel_details, returned by channel_detail).
- Discovery → channel page: the Channel-manager "Discover from playlists" tab now
links each channel name to its in-app channel page (ChannelLink onView), so a
discovered channel's About/videos can be inspected BEFORE subscribing (was
subscribe-only, plain text before).
Note: the channel info-card epic itself was already delivered in v0.19.0; this is
the About-tab enrichment follow-up + the header-shift fix. external_links stays
empty by design (YouTube removed the field from the Data API ~2023).
- scheduler _run_changed: count a run as "changed" only on a truthy NUMERIC
value — status-string no-op dicts like {"skipped":"disabled"} (Plex off, the
default) or {"skipped":"no demo account"} were flooding the trail every interval.
- audit.record: truncate target_id to the column width (128) like summary[:255],
so an over-long target (e.g. a 128+ char demo email) can't abort the mutation
the audit row is committed alongside.
- config set/reset: redact URL userinfo (user:pass@) from logged non-secret
values — a proxy setting can embed credentials that shouldn't land in the trail.
- config set: skip the audit row when a non-secret value didn't actually change
(no-op re-save shouldn't add noise); secrets are write-only so always logged.
- audit page title (pageMeta): add the missing "audit" case so the browser tab
reads "Audit log · Siftlode" instead of falling back to "Feed".
- channels CB3: sort discovery rows NULLs-last (title is None) to match the DB
ORDER BY exactly — coercing NULL title to "" floated untitled channels to the top.
- AdminUsers SC1: track in-flight rows in a Set, not a single id — per-row disabling
now lets the admin start concurrent row actions, and a shared id was cleared by
whichever settled first (re-enabling a still-pending row → possible double-submit).
- ChatThread CT4: reset the incoming-count ref when partnerId changes — the Messages
page reuses one ChatThread across conversation switches, so a same-count switch
could skip the open-marks-read badge refresh.
- messages MB3: push a bare {type:"unread"} (drop the now-dead count query — the
client re-fetches on the signal and ignored the number anyway).
- api.ts: extract the shared keepalive `beacon()` helper (saveProgressBeacon +
plexProgressBeacon were near-verbatim copies).
Verified each deferred per-subsystem item against current source, then fixed the
real ones (tradeoffs left as-is, see siftlode-ops/CODE-HYGIENE.md closeout):
- search BUG-3: don't finalise shorts_probed on un-enriched stubs (enrichment
failure/omitted rows) — restores the scheduler's enriched_at guard so a real
Short can't leak into the feed and never get reclassified.
- downloads GC: 4th pass reaps orphaned errored, fileless, unreferenced assets
(they carry no TTL, so the expiry passes never cleared them).
- downloads B6: quota/edit errors now return a structured {code,reason,limit}
the trilingual client localises, instead of hardcoded English + dot-decimal GB.
- channels BB1: reset-backfill re-arms deep sync on every subscriber (bulk update)
instead of 404ing when the admin isn't personally subscribed.
- channels BB2: enrich the stub on BOTH the normal and "already exists" desync
path (nest the insert try inside the client `with`).
- channels CB3: drop the redundant second _discovery_rows read (identity-mapped
rows are already enriched; re-sort in Python for the title tie-break).
- playlists PC1: read the live playlist once in push() and share it with plan_push
+ push_playlist (halves read-quota, closes the second TOCTOU window).
- playlists PC2/PC5: batch the cover thumbnails in one window query (was N+1);
rename combines count+duration into one aggregate + a single cover lookup.
- messages MB2: get_thread only resolves a messageable partner OR one we already
share a thread with (closes the user-enumeration oracle).
- messages MB3: push a live unread-count to the user's other tabs after mark-read.
- messages MC4: list_conversations uses DISTINCT ON + grouped COUNT instead of
loading the whole message history into memory.
- config AB3: per-spec allow_empty so smtp_user/youtube_api_proxy can be blanked
to disable them rather than snapping back to the env default.
purge_user's revoke used `decrypt(refresh) or tok.access_token`, but access_token is
stored ENCRYPTED — so a token row without a refresh token would send ciphertext to
Google's revoke endpoint and silently fail. Decrypt it. (Pre-existing; surfaced by the
review of the OAuth-scope fix.)
Correcting the earlier scope-union-only fix. The real damage from a logout→login isn't
just the stored `scope` field — verified via a live token refresh that the stored refresh
token itself had been REPLACED with a base-only one: a plain sign-in requests only
BASE_SCOPES and Google handed back a fresh base-only access+refresh token, which
_store_token adopted verbatim, destroying the read/write grant (the old refresh token
stays valid on Google's side, so overwriting it is what loses access).
_store_token now detects when an exchange would NARROW our YouTube scopes and, in that
case, keeps the WHOLE existing grant (refresh token, access token, scopes) untouched.
Broader-or-equal exchanges (first grant, read→write upgrade) adopt + union as before.
Unit-verified across base-relogin / upgrade / new-user / base-user cases.
Two user-reported signin bugs:
1) A plain re-login (or logout→login) wiped the user's YouTube grant: login requests
only BASE_SCOPES and Google's returned `scope` can list just those, but _store_token
wrote it verbatim — dropping a previously-granted youtube.readonly/youtube scope, so
can_read flipped to false and the feed demanded a reconnect. The underlying grant
(refresh token) survives such a login, so UNION the scopes instead of narrowing; they
only shrink on full disconnect (purge deletes the token row).
2) Admin 'new access request' email: (a) it named the wrong menu ('Settings → Account'
instead of Users → Access requests); (b) the requester address was invisible so the
'reply reaches them' note looked wrong (Reply-To is in fact set to the requester) —
now spelled out + a deep-link straight to the approve view (?admin=access-requests,
handled in App); (c) it emailed only the static env ADMIN_EMAILS — now notifies the
ACTUAL admins (active role=admin users) unioned with env, so a UI-promoted admin (who
CAN approve — the approve UI is role-gated) is notified too.
resolved_user_id already accesses ws.session unguarded just above (SessionMiddleware
covers the WS scope), so the '"session" in ws.scope' fallback was dead code — read
ws.session directly for consistency (review follow-up).
Loose ends to finish the auth security round:
- register + password-reset-request had an enumeration TIMING oracle: an already-
registered email skipped the create path (hash + row writes + email scheduling) and
responded measurably faster than a new one. Move the whole lookup+create (register)
and lookup+token+email (reset) into a background task with its own DB session, so the
endpoint returns in the same time for any valid email regardless of whether it exists.
Verified: existing vs new now ~equal (was 34ms vs 82ms on register); accounts/tokens
still created off-path.
- messages_ws re-implemented resolved_user_id's per-tab wallet-gated account resolution.
Generalize resolved_user_id to take any HTTPConnection (Request OR WebSocket) and call
it from the WS — one shared, wallet-gated resolution. Behavior-identical (unit-checked).
Post-review hardening (both fail-safe, not bugs):
- _client_ip normalizes IPs via ipaddress and unwraps IPv4-mapped IPv6, so a plain
TRUSTED_PROXY_IPS=10.10.0.1 also matches a peer surfaced as ::ffff:10.10.0.1 (a
Docker/IPv6 self-host footgun that would otherwise silently collapse everyone into
one rate-limit bucket).
- entrypoint.sh + _client_ip docstring: explicit warning never to add uvicorn
--proxy-headers, which would rewrite request.client from the forgeable XFF and defeat
the trust check.
_client_ip trusted the first X-Forwarded-For hop unconditionally, so anyone able to
reach the app port could forge XFF and dodge the login/register/reset/demo rate limits.
Now trust XFF ONLY when the request's socket peer is a configured reverse proxy
(settings.trusted_proxy_ips, e.g. the VPS Caddy's WireGuard peer IP), and take the
RIGHTMOST entry — the client our proxy actually saw and appended, immune to a client
pre-seeding a fake XFF. A request from any other peer (hitting the port directly) is
keyed on its real socket IP, so XFF can't be forged to bypass the limits.
New TRUSTED_PROXY_IPS env (empty default = no proxy, use the direct peer). Documented in
.env.example, docs/self-hosting.md, README. Unit-verified against spoof-through-proxy and
direct-bypass cases.
Adversarial re-review of the session-epoch work surfaced:
- WS auth (messages_ws) skipped the epoch check, so a revoked-but-unexpired cookie
could still open the live push channel after a reset/logout-others. Now mirrors
current_user: loads the user once, rejects a stale-epoch cookie before connecting.
- set_password + logout_others re-stamped the cookie BEFORE db.commit(); a failed
commit would strand the current session at a newer epoch than the DB and wrongly
401 it. Commit first, then re-stamp.
- Welcome verify effect could double-POST the single-use token (StrictMode/remount)
and flip the banner to a false 'invalid'. Fire-once useRef guard.
Left as-is (low value, documented): the Plex image proxy authenticates without a DB
load / epoch check (poster/art fetches only); adding one would cost a DB hit per image.
Signed client-side session cookies had no server-side kill switch: logout + password
reset couldn't evict a stolen/copied cookie (valid until expiry). Add User.session_epoch
(migration 0053), record it in the cookie at every login, and reject in current_user any
cookie whose recorded epoch is behind the account's current one.
- Bump the epoch on: password reset (kills ALL sessions — a reset is a compromise response),
password change + a new 'Log out other sessions' action (both re-stamp the CURRENT cookie
so the caller stays signed in, evicting only the others).
- Per-account epoch map in the session so one account's revocation doesn't evict the other
signed-in accounts in the same browser wallet.
- Missing epoch (pre-SA4 cookie) is treated as 0, so the first bump revokes grandfathered
sessions too.
- New POST /auth/logout-others + a Settings → Account 'Active sessions' button (trilingual).
Secret email tokens now ride the URL fragment (#reset=/#verify=), never the query
(?reset=/?token=): a fragment isn't sent to the server, so the token can't leak into
proxy/access logs or a Referer header.
- Reset: link → /#reset=; the SPA reads the token from location.hash and POSTs it
(unchanged /password-reset/confirm).
- Verify: link → /#verify=; new POST /auth/verify (token in body). The legacy GET
/auth/verify?token= is kept so pre-deploy emails in flight still work until they
expire. The SPA reads the fragment token, POSTs it, shows ok/invalid.
- Welcome: read secret tokens from the fragment, status flags from the query; strip
both after capture so nothing lingers in history.
Closes the Plex #9 backend backlog (frontend was the prior follow-up). Behavior-
neutral cleanup + two two-way-sync bug fixes; PW2 verified a non-bug.
fix(plex): watch-sync PW1 pagination + PW3 lost-update guard
- PW1 watch_history: the history is a GLOBAL cross-account feed read as a single
500-row page, then filtered to the owner — on a busy family server the owner's
recent views can sit past page 1 and be missed. Now paginate (desc) until the
since-cutoff, bounded by max_pages (daily reconcile is the backstop; logs if hit).
- PW3 push_state_to_plex: it flagged synced_to_plex=True unconditionally after a
push, so a local edit landing between the push and the flag-set was marked clean
and never pushed (lost update). Capture the row's freshest timestamp when the
push is scheduled (_push_watch) and only settle the flag if the row is unchanged.
- PW2 was flagged as "accountID hardcoded to 1" but is NOT a bug: on a PMS account 1
is reserved for the owner/admin and an owner link always holds the admin token —
added a clarifying comment so it isn't re-flagged.
chore(plex): backend dedup
- PC2 unified_library + facets shared a ~13-field filter Query signature + p-dict →
one LibraryFilters dataclass injected via Depends(); as_dict() feeds the builders.
- PS-C1 sync.py collection upsert dup (resync_collection ≈ _sync_collections) →
_upsert_collection().
- PS-C2 sync.py 8-field filterable-metadata block dup (_apply_item ≈ _sync_shows) →
_apply_facet_fields().
- PW-C1 _repush_dirty read duration from the mirrored PlexItem.duration_s instead of
a per-row plex.metadata() round-trip.
- PW-C2 rk→id map built inline in two places → _rk_to_id() helper.
- plex.py: delete the no-op _enabled() (never wired as a Depends/called; the real
gate is sysconfig plex_enabled).
- PlexBrowse.tsx: collapse toggleEpisodeWatched into toggleWatched — they were
byte-identical (both take a PlexCard).
- Delete 6 dead plex.json keys (loadMore, playerSoon, filter.library,
playlist.up/down/remove) from en/hu/de — verified 0 refs (the live
playlist.removeShow/removeSeason keys are kept).
Behavior-neutral. tsc green, ruff clean on touched files, localdev boots, Plex renders.
stream.py enforced a hardcoded _MAX_SESSIONS = 4 and never read the
DB-overridable plex_max_transcodes ConfigSpec, so the Configuration → Plex →
"Max concurrent transcodes" knob was dead (same env-vs-DB drift class as the
Downloads/Admin fixes). _enforce_cap(cap) now takes the cap from
sysconfig.get_int(db, "plex_max_transcodes") (>=1 so playback can't be capped to
zero). Bumped the config default 1→4 so the effective default matches the old
hardcoded cap (no behavior change for non-overriders; the knob now works).
peter