A book the index does not hold cannot be found by any positive filter, and
every negated filter - which resolves to every document in the index - drops
it from the grid instead. That is why issue #1989 reads as half a working
filter: Absent found nothing while -Absent removed exactly the absent books.
The index is only ever written as a whole, and a rebuild that fails is
deliberately swallowed so a bad index cannot fail a good scan, so a short
index stayed short until something else changed the library. Count the
index against the library once per run and rebuild when it is short.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
Only a newly imported book ever recorded a supplement, so a title that gained a
PDF after its first import never got one, and a title that lost its PDF went on
claiming one - which in issue #1973 is why three titles Audible has no PDF for
were still being asked for.
Sync from updateBook as well, and give Book set-semantics for the one supplement
Audible reports per title. The duplicate guard compared the incoming url to
itself, so it happened to mean 'this book already has a supplement' and a url
that had changed was silently ignored.
A supplement is dropped only when Audible says outright that no supplement url is
available. A missing url says nothing by itself: episodes come from the catalog,
which is never asked for pdf_url, so there it means 'not asked'. A PDF already
downloaded is left alone either way, since the file is on disk and the library
should go on saying so.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
Microsoft.EntityFrameworkCore.Sqlite 10.0.11 is the first 10.0.x servicing
release to depend on SQLitePCLRaw 2.1.12, which drops the pre-3.50.2 SQLite
build flagged by GHSA-2m69-gcr7-jv3q (CVE-2025-6965). DataLayer's Sqlite
reference is where that chain enters the solution, so bumping it clears
NU1903 from all 23 projects that transitively saw it.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
Seven references across six projects, all now on published versions.
What the app gains is from Dinah.Core: OsSecretStore.Create bounds how
long it waits for the backend, so a Linux or macOS start with a keyring
that never answers falls through to the portable master key path instead
of hanging there. IdentityTokenStorageWiring.ResolveSecretStore is the
caller, and it runs at startup whenever no key file or env var is set.
AudibleApi 11.0.3.1 carries no code change for us - it is the release
where its nuspec finally declares the Dinah.Core floor its own code needs.
Dinah.Core.WindowsDesktop and Dinah.EntityFrameworkCore move to 10.2.4.1
as well, keeping every Dinah package on one version.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
Dinah.Core 10.2.2.2 gives SecretString a Redacted property, so a
destructured secret carries its length without the logger being told
anything. That was the only thing AsScalar was buying - safety never
depended on it - so the registration and its explanation go away, and the
knowledge lives in the type instead of in this file.
MaskedLogEntryPolicy stays: an ILogMasked has no equivalent property, and
without the policy one logged as {@Account} is still written out property
by property.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
AudibleApi 11 holds token, key, and cookie values in a SecretString
rather than a string, so nothing public exposes plaintext for a reflective
logger to find. Picking it up is a breaking upgrade: the seven package
references move, and the nine places that read a secret now call Reveal().
Two of those needed thought rather than a mechanical edit. Mkb79Auth
exports to and imports from audible-cli's JSON format, which is plaintext
by definition, so the cookie projections reveal explicitly in both
directions and the file format is unchanged. And the account's own
DecryptKey stays a plain string here: converting it is separate work.
This is the dependency bump only. The log leak it enables fixing - an
AuthenticationRequiredException carrying a live Account, whose address and
activation bytes Serilog.Exceptions writes into a shared log - is still
open, and none of the account-side masking has landed yet.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
getProposedDownloadFilePath looked for the book's audio file and fell back to
the Books directory itself when it found none. That lookup matches on the
product id appearing in the path, so it finds nothing for a library whose
folder and file templates omit <id>, and nothing for a title marked downloaded
whose files are not on this machine. Those PDFs landed in the library root,
where they also shared one namespace and so could collide with each other.
Fall back to the folder template instead - the same folder the audiobook itself
would go in - and create it, since nothing else does on the PDF-only path.
Also give MockLibraryBook a three-field version: ToVersionString formats to at
least three fields, so the two-field default threw as soon as anything rendered
a naming template for a mock book.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
A license denial left no trace: BookStatus stayed NotLiberated, so every
liberate run asked again. Only the GUI's bad-book dialog could mark a title
Error, and license denials take their own path and never reach that dialog, so
a headless install had no way at all to stop the retries. A cron schedule then
re-requested the same refused licenses every run and printed the same warning
block for each one, which is both wasted API traffic and the log noise reported
in issue #1947.
Record the refusal instead, with a wait that doubles per consecutive failure:
one day for an eligibility denial (up to 30), six hours for a title Audible has
no audio for yet such as an unreleased preorder (up to 7 days), one hour when
the denial names GenericError, which the GUI already reads as an outage
(up to 12). Nothing is permanent - every kind is attempted again on its own.
Only failures attributable to Audible are recorded. A dropped connection, a
decrypt error or a full disk keeps being retried on the next run as before.
Naming a title, --force, and setting a download status all clear the record:
asking for a title explicitly overrides the wait.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
Records every successful audiobook download in the library database (a new
DownloadHistory table) and, when the user opts in, stops downloading once the
rolling 24 hour window is full.
The history lives in the database rather than a file under LibationFiles
because in Docker only the database is on a volume; a file there is discarded
on every container restart.
The limit is checked immediately before each book downloads, never at queueing
time, so a full queue stays full and the user can raise or disable the limit
mid-run. When nothing in the queue can proceed the queue pauses and re-checks
every 15 seconds, recomputing settings, history and clock from scratch, so a
queue left running for days drip-feeds itself as downloads age out. The CLI
never waits: it skips covered titles and reports a count.
Co-authored-by: rmcrackan <rmcrackan@gmail.com>
DbContexts.GetContext() runs ApplyMigrations on every context creation,
calling Database.Migrate(). Migrate() acquires the EF migration lock
before checking whether anything is pending. For SQLite that lock is a
persisted row in __EFMigrationsLock with no timeout (SQLite has no
connection-scoped lock that frees on disconnect). A process killed after
acquiring the lock but before releasing it - even during an otherwise
no-op Migrate() - orphans the row, and every later Migrate() then spins
forever in SqliteHistoryRepository.AcquireDatabaseLock(). The row lives
in the db file, so restarts never clear it, matching the "scan hangs,
reboot doesn't help" reports in #1729.
Guard Migrate()/MigrateAsync() behind GetPendingMigrations(), which only
reads __EFMigrationsHistory and never acquires the lock. The steady-state
path (schema already current) no longer touches the lock at all; first
run and real migrations are unchanged.
Category names were accidently removed during the last DB migration in d67692355f.
Re-add the Names column and modify CategoryImporter to upsert the names on update.
Add GetUnliberated_Flat_NoTracking() which queries only unliberated books/episodes and does not load the entire library.
Fix UnLiberated() query to only return products or episodes (not parents) (#1564)
- Use the main display grid control to display deleted books
- Added search functionality for deleted books. This required creating a temporary search index in the `InProgress` folder. The products grid control now uses an instance of `ISearchEngine` to filter its grid entries. The main grid uses a singleton instance of `MainSearchEngine`, which merely wraps `SearchEngineCommands.Search()`. The TrashBinDialogs use `TempSearchEngine`.
- Users can now batch select `Everyting` as well as `Audible Plus Books`
Avalonia:
- Refactor main grid context menus to no longer require reflection
- Update all project runtime targets
- Update all dependencies
- NOTE: Using Npgsql.EntityFrameworkCore.PostgreSQL RTM build from MyGet
- Delete unused pubxml files (they were made redundant by recent workflow changes)
- Replace Libation.sln with Libation.slnx
- Add `LIBATION_FILES_DIR` environment variable to specify LibationFiles directory instead of appsettings.json
- OptionsBase supports overriding setting
- Added `EphemeralSettings` which are loaded from Settings.json once and can be modified with the `--override` command parameter
- Added `get-setting` command
- Prints (editable) settings and their values. Prints specified settings, or all settings if none specified
- `--listEnumValues` option will list all names for a speficied enum-type settings. If no setting names are specified, prints all enum values for all enum settings.
- Prints in a text-based table or bare with `-b` switch
- Added `get-license` command which requests a content license and prints it as a json to stdout
- Improved `liberate` command
- Added `-force` option to force liberation without validation.
- Added support to download with a license file supplied to stdin
- Improve startup performance when downloading explicit ASIN(s)
- Fix long-standing bug where cover art was not being downloading
- Added `MockLibraryBook` which contains factories for easily creating mock LibraryBooks and Books
- Added mock Configuration
- New `IPersistentDictionary` interface
- New `MockPersistentDictionary` class which uses a `JObject` as its data store
- Added `public static Configuration CreateMockInstance()`
- This method returns a mock Configuration instance **and also sets the `Configuration.Instance` property**
- Throws an exception if not in debug
- Updated all chardonnay controls to use the mocks in design mode. Previously I was using my actual database and settings file, but that approach is fragile and is unfriendly towards anyone else trying to work on it.
Supporting postgres simplifies deployments to environments such as kubernetes. Since sqlite doesn't work well on nfs shares it can be easier for databases to have a dedicated db set up that applications can connect to. Sqlite is easier for most deployments though, so this will default to that if the settings haven't been updated to support it.
This change does the following:
- Separate out SQLite from the DataLayer and adds a Postgres assembly for migrations as well
- Add a configuration setting for a postgres connection string that will be used if it is there, otherwise reverts to the original sqlite string
- Add a copydb command for the cli to bootstrap the postgres db
- A convenience script to update migrations for both dbs at the same time
