Release names say what their files are — 2160p, HDR, 10-bit, Dual-Audio —
and nothing showed it. naming::quality reads those words, whatever they
are separated by, and Title::quality keeps what more than half a title's
files state, so one odd special does not mark a show down. The facts line
of a title's page ends with them: 4K, 1080p, HDR, 10-bit, Dual audio.
Release groups' names are never read.
TVDB has no films, so ani.zip has no fanart or logo for one, and most of
the titles in a real library with nothing wide to draw were films. ani.zip
does name the TMDB entry of every AniList match. With a TMDB key stored
and AniList chosen, an entry left without a backdrop or a logo is asked
for /{movie,tv}/{id}/images: one request, by id, so no title name goes to
TMDB. A bare backdrop and a PNG logo in the viewer's language are
preferred, and a failure there only costs the art.
The TMDB key field is now offered with either provider, labelled optional
with AniList. TMDB as the provider fetches its logo the same way, and
backdrops at w1280 rather than w780. A forced run re-enriches hand-picked
titles too, so saving a key and matching everything again reaches them.
With AniList chosen, a title AniList has no match for is now searched on
TVmaze, keyless, before it is given up on. One provider for the whole library
left live-action shows in a share of anime unmatched, and switching to TMDB
would have dropped every AniList match, chain and shelf.
The service holds a chain of providers. A match is stored under the provider
that gave it, and enrichment and refreshes go to that provider; a miss is
stored under the last provider asked, and is_fresh only trusts a miss from the
end of the chain, so AniList misses stored before this are asked again at once.
A provider failure still ends the lookup instead of falling through.
The hand matcher searches every provider in the chain and labels each result
with its source. TVmaze is never offered in the settings and is not asked
about titles known to be films.
ani.zip sent each series' TVDB clear logo and every episode's running
time from the start, and neither was kept. Enrichment now stores the logo
in title_metadata.logo_url (kept across a refresh of the same entry, like
the fanart) and episode runtimes in episode_metadata.runtime, TVDB's
first and AniDB's length after it. AniList's duration lands in
TitleDetails::runtime; TMDB episode runtimes are kept too.
Schema v13 adds the two columns and marks AniList titles un-enriched
without deleting anything. TitleDetails::revision lets details written
before runtime existed refetch by id while the old ones stay on screen,
instead of blanking them in the migration as v10 did.
Clear logos are padded to a square, so each client crops them to their
visible pixels before drawing. The logo replaces the name on the
library's banner and the title page (not on a light theme's title page,
where lettering made for dark art can vanish). Episode rows show how long
they run, in place of the file size on Android and iOS; a film's runtime
joins its line of facts.
Plain-text subtitles (SRT) were drawn in mpv's bare default, thin white
text that bright scenes swallow. Settings → Playback now picks a size,
an outline or a dark box, and white or yellow text. Styled (ASS) tracks
keep their own look and take only the size.
The box options are mpv 0.38's, so they are set best effort: an older
libmpv logs and plays on. The style syncs with the other playback
settings, and the Android bridge keeps the stored style instead of
resetting it on save.
Anime releases often carry two subtitle tracks in one language, the
dialogue and a "Signs & Songs" track, and the second is as often as not
the one marked default or forced. mpv, picking by language, played it.
pstr_core::subtitles reads a track's name and forced flag for whether it
is the dialogue, and every client asks it once per file, after mpv has
made its own pick, whether to switch to a track in the same language
that is. A subtitle track the viewer chose by name stays, and subtitles
that are off stay off.
crawl_for_news crawls the shares due a look that were crawled here
before, and returns per title how many files first appeared during it and
the first of them. A share crawled for the first time is left out, since
every file in it is new.
The bridge's open_subtitles opens and publishes them and names the one to
turn on. The native player holds them until mpv has the episode loaded,
adds them with sub-add, and releases them instead when another episode
was loaded in the meantime.
A transport failure's text is reqwest's "error sending request for url
(…)", and the cause sits further down its source chain. Error::describe
appends the causes and drops the URL, so the reason fits on a share row.
The bridge's open_stream started no fetch until mpv asked, so every
open was the head and then the tail, one after the other. It now warms
both edges as the desktop does, and the resume block too when the
episode resumes.
warm_episode fetches another episode's opening blocks into the cache
without playing it. Both phones call it for the next episode in the
last two minutes or once the credits start, as the desktop already
did, so autoplay starts from the cache.
ShareStore::rename changes only the name; the id that the catalog, watch
state and offline files are keyed by stays the same. Sync carries the
change as it does any other change to the share's register. The desktop
renames in the share's row. The phones rename from the share's menu.
Before, the name given at add time was permanent short of removing the
share.
The library filter moves from the desktop into pstr_core::browse and
gains a Downloaded filter: titles with at least one episode kept on the
device. The bridge's arrangement() takes the filter. Android and iOS get
a row of filter chips above the grid. The choice is remembered per
device, and shelves stay hidden under a filter, as on the desktop.
The bridge gains stop_crawl(). A running crawl watches for it and returns
without an error. Shares already read stay stored, and a queue of new
shares stops instead of going on to the next. The activity banner shows
Stop while a crawl runs, because listing a large share can take minutes
and holds the connection the player wants.
No client crawled a share unless someone pressed Refresh, so a new
episode stayed out of the library until then, and nothing said how stale
the library was.
Catalog schema v11 adds share_crawls: when each share's last crawl
finished and was attempted, how many files it stored and how many were
new, and why the last attempt failed. A failure keeps the last good
crawl's numbers.
Every client now crawls, quietly, any share not looked at for six hours:
the desktop at launch and every half hour while open, the phones at
launch and on their existing catch-up timer. A quiet crawl says only
news ("anime: 3 new files") and leaves failures on the share's row
rather than in a toast.
Share rows on all three show "120 files - crawled 2 h ago", or the last
failure until a crawl succeeds.
Continue watching listed only part-watched episodes, so finishing one -
the most ordinary moment to want the next - dropped the show from the
shelf. Title::continue_with now reads the episode touched last: a
part-watched one is resumed, a finished one offers the episode after it
("Next: S01E04"), and one reset to the start (what leaving the shelf
does) keeps the title off. A few seconds of an episode are passed over.
The phones computed their own shelf from episode records. The bridge now
exports the engine's choice as TitleRecord.continue_index, and Play on a
title lands there on every client.
Undo on the phones re-saved a watch state with a fresh time, which both
reordered the shelf and could not express "never played". A new
restore_watch_state puts the state back as it was read.
The age check and the one retry behind a refresh (B22) lived only in the
Android bridge. The desktop opened streams, offline copies and crawls
against whatever session the share was opened with, so a window left open
for a few hours failed to play or recrawl until it was restarted.
The logic moves to pstr-core as SessionAges, owned by SharedLibrary and
reached through SharedLibrary::reauthenticating. Both engines go through
it for play, warm, offline copy and (desktop) crawl.
Sync carried only watch positions, so a second device started empty. The
per-installation history file now also carries per-key registers
(pstr-core::profile): every share with its link and custom password, the
playback, appearance and metadata settings, the TMDB key, hand-picked
matches and per-title track choices. Newer updated_at wins per key, a
removal travels as an empty register, and a first sync stamps at 0 so a
fresh install takes the library the other devices have instead of
overwriting it with defaults. Volume, mute and the device-only settings
(cache budget, Wi-Fi only, background audio, hardware decoding) stay out.
The secrets live only in the end-to-end encrypted Drive file, never in
shares.json. FORMAT_VERSION stays 1; an older build ignores the registers.
Adding a share stored it before opening the link, so a first failure left a
broken entry and every retry said "already configured". add_verified opens
the link first and stores nothing on failure, and only the new share is
crawled afterwards, so another broken share cannot fail the add.
Android drops picture-in-picture: the window sized the video wrongly and
background audio already covers leaving the app.
iOS sizes the Metal drawable on every layout, so the picture no longer
stretches after a rotation or a Split View resize. Settings show a save at
once rather than after the match run, and a failed save reverts and says
why. Both phone clients gain an activity banner naming what is running, and
the add-share, metadata and change-match forms show progress and errors in
place.
- Desktop shows Proton's CAPTCHA instead of failing the sign-in: a wry webview
in a child process of the app (`pstr-app::verify`), confined to
https://*.proton.me, hands the token back and the sign-in is retried. Linux
packages and CI gain WebKitGTK 4.1; the Windows zip and MSI ship
WebView2Loader.dll, without which a -gnu build does not start.
- The desktop keyring splits a secret too long for one entry across several,
so a Proton session with many key salts survives Windows' 1280-unit cap.
Short secrets stay one entry, so every existing secret still reads.
- `WatchSync` runs against a small `HistoryDrive` trait and is tested end to
end against an in-memory Drive: two devices, newer wins, no re-upload of an
unchanged history, a missing file written again, an unreadable own file
never overwritten, a newer-format file left alone.
- An installation's file unwritten for 90 days is retired by the next device
to sync, after its entries are merged and carried in that device's file.
- Removing a share keeps its watch state: sync would bring it back from the
other devices anyway, and re-adding the share picks up where it left off.
- Watch history syncs once more on desktop exit (bounded to 3 s) and, on
Android, through a WorkManager job enqueued when the app leaves the screen.
- The second-factor prompts say security keys are not supported.
A Proton account is now optional alongside public links. Signing in (password,
then a second-factor code and a mailbox password where the account has them;
a CAPTCHA gate is solved in the Android app) stores one session blob in the OS
credential store under `proton-account`, writing back every refreshed token
since refresh tokens are single-use. With an account, folders of the viewer's
own Drive — My Files, other devices, shared with me — can be added as shares
through the same `ShareClient` surface as a public link.
Watch history syncs through a `proton-stream` device in that Drive: one
`watch-history/<installation>.json` per installation, so no two devices ever
write the same file. Each sync reads the others and merges per episode, newer
`updated_at` wins, with a deterministic tie-break. It runs at launch, after
sign-in, whenever the player closes and every five minutes. A session that can
no longer refresh signs the app out instead of failing on every launch.
The CLI gains `login`, `logout`, `account`, `browse` and `sync`.
Also in this change:
- Android offline downloads no longer look stuck (B64): two transfer at once,
through a copy stream with no read-ahead, with a smoothed transfer rate.
- Choosing a metadata match by hand now enriches the title too, instead of
keeping the old match's episodes and artwork.
The catalog schema is unchanged.
Bring the phase 2 library to Android. Search, sorting and franchise folding
move from the desktop app into pstr_core::browse, so both clients find,
fold and order the same titles the same way.
The bridge gains AndroidEngine::arrangement, which returns the grid's tiles
and up to three shelves as keys into the library's titles. Each TitleRecord
now carries its display name, wide art, format, season, studios, directors,
tags, airing state and franchise. The library bar has a sort menu with the
franchise toggle, folded tiles show a +N badge, and shelves sit above the
grid. The title page shows the same facts, tags and franchise row as the
desktop. Settings -> Appearance gains Title names.
The default title name is now English, which is what Android always showed.
A folded franchise now sorts by its lead's displayed name rather than by its
alphabetically first member.
The AniList query now keeps the start date, season, banner, popularity,
status, next airing episode, tags, main studios, director and story
relations, as TitleDetails in one JSON column (schema v10). Matches
stored without them are fetched again by id, never searched, so
hand-picked matches stay; refreshing the same entry keeps its episodes
and fanart. Enrichment records the entries it walked. node_first_seen
remembers when each file appeared, across recrawls.
pstr_core::franchise groups titles that share a provider id. The grid
folds each franchise into its lead with a +N badge, and the title page
lists the whole franchise in release order. pstr_core::shelves picks
rows by airing status, director, studio and genre. A row needs three
franchises, so one franchise's titles never make a row on their own.
Desktop:
- Covers are the default grid (Settings > Appearance > Library tiles).
- The hero and title band use fanart or banner decoded at 1920 px.
- New sorts: recently added, newest, highest rated, most popular.
- Titles can show the share's, English or romaji name; search matches
all of them.
Android carries TitleDetails through a hand-picked match; its screens
are unchanged.
AniList's per-season lookup searched for "<title> 2nd Season" and the
scorer picked the base entry again, so Attack on Titan seasons 2-4 and
Jigokuraku season 2 were captioned with season one's names. Episode
names came from streamingEpisodes, which is partial and numbered however
a streaming site counted. Bleach S17 and Mushishi S02/S03 had none.
AniList now walks SEQUEL relations from the match and takes each
entry's episodes from ani.zip, which carries TVDB's season, number and
absolute number. An episode row keeps all three plus its place in the
AniList entry. EpisodeGuide::get matches whichever numbering a file
uses and never answers from another season.
Episodes and fanart are stored as one enrichment. The episode list is
replaced even when empty, and the enrichment is recorded as asked, so a
film is not re-asked every run. Storing a different match drops the old
entry's episode rows. AniList's 4.75:1 bannerImage is no longer stored
as the 16:9 backdrop; ani.zip's 1920x1080 fanart is used instead.
Schema v9 adds the columns and deletes the AniList episode rows and
banners. Matches are kept. Desktop, Android and the CLI now share
service::plan and MetadataService::work. The settings text names
ani.zip. Audit and roadmap in docs/METADATA.md; bug entry B61.
- Settings > Storage chooses the streaming cache's budget, 1 to 16 GiB,
and shows what it holds against it. The bridge gains
set_stream_cache_budget, which the next connection is built with and
an open one evicts down to at once. Android keeps the choice in its own
settings, like the desktop keeps it in desktop.json, and tells the
engine as it is created. Unset stays at the shared 4 GiB default.
- Settings > Playback can turn hardware decoding off for a device whose
decoder shows green or torn frames. pstr_mpv sets hwdec before each
loadfile, so it applies from the next episode.
A track picked on one episode now carries over to the next by its title as
well as its language, so e.g. "Signs & Songs" vs "Full" subtitles in the same
language stick. Catalog migration V8 adds the title columns; the Android
bridge keeps a stored title only while its language is unchanged.
storedPalette() reads and resolves the theme file without building the
engine, so onCreate can set the window background and system-bar icon
polarity before anything draws. Previously the background was a hex literal
in themes.xml and the bar icons were pinned dark, which meant every cold
start flashed one flavour's colour and Latte drew white icons on a
near-white page.
Adds the Material scheme mapping, type ramp and shape tokens the palette
feeds, plus unit tests for both.
nvrl