Per review feedback from cvium: failing to read/backup an entity due to
corrupt underlying data is a significant event that should be surfaced
as an error, not silently downgraded to a warning.
* Prevent ffmpeg from hanging extracting subtitles
Add `RunSubtitleExtractionProcess` to unify the external
_ffmpeg_ process handling and error management.
Add a `-nostdin` flag that prevents _ffmpeg_ from reading from
_stdin_ and blocking on an inherited stdin handle (e.g. when
Jellyfin runs as a service under NSSM), which otherwise hangs
subtitle extraction forever when _ffmpeg_ blocks on any
keyboard-interaction read until the timeout (30 minutes).
Close the redirected _stdin_ to ensure immediage EOF.
Drain the _stderr_ to a string and log it, to ensure we don't block
the _ffmpeg_ process on errors that exceed the pipe length.
Pass `-y` to _ffmpeg_ to ensure it overwrites any existing output file
without prompting for confirmation.
* Address review comments
Make sure we always drain stderr.
Make sure the timeout also honors the cancellationToken.
Make sure when we get cancelled we don't log it as a ffmpeg error.
* show production companies instead of networks
* keep both production companies and networks
* fix whitespace
* fix nullable type
* networks first, then production companies
BaseItemDto.Genres, .Tags, and .ProviderIds are plain auto-properties with
no default initializer, so they deserialize to null when a client omits
them from a partial POST /Items/{itemId} body. The OpenAPI spec documents
every BaseItemDto field as optional, but ItemUpdateController.UpdateItem
fed these three properties straight into Distinct()/Select()/ToList()
without a null check, so a request that (for example) only sets Tags
throws ArgumentNullException("source") once it reaches the unguarded
Genres line, before Tags is even processed.
Guard all three assignments with the same "if (request.X is not null)"
pattern already used for the neighboring Studios/Taglines/ProductionLocations
fields in this method, so omitted fields are left unchanged instead of
crashing the request.
Adds ItemUpdateControllerTests covering the reported repro (only Tags
supplied) and a companion case asserting existing Genres/ProviderIds are
preserved when omitted from the payload.
Signed-off-by: zerafachris <christopher.zerafa@blocklabs.io>
A single row with malformed KeyframeTicks JSON (e.g. a truncated array
from an interrupted write) currently aborts the entire backup, because
the try/catch in BackupService.CreateBackupAsync only wraps
serialization of an already-materialized entity, not the enumeration
itself. EF Core throws JsonReaderException from MoveNextAsync() while
materializing the corrupt row, which propagates past that catch block.
Switch to manual enumerator iteration so MoveNextAsync() failures can
be caught per-row, logged as a warning identifying the affected table,
and skipped, allowing the remaining rows and the rest of the backup to
complete.
Fixes#17216
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
SubtitleEncoder.ConvertSubtitles parsed subtitles with libse's static
Subtitle.Parse, which iterates a statically cached list of shared
SubtitleFormat instances. Format parsers keep mutable per-parse state on
the instance, so concurrent subtitle requests corrupted each other's
output (cues mixed across streams and languages, truncated files) or
failed with NullReferenceException when format detection broke down and
Subtitle.Parse returned null.
Parse through the injected ISubtitleParser instead. SubtitleEditParser
instantiates a fresh format parser per call, so requests no longer share
state. Its Parse method now returns the libse Subtitle directly (the
SubtitleTrackInfo flattening was unused since the SubtitleEdit writer
rework) so the writers keep full fidelity such as ASS styling.
KAUN