aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorTim Eisele <Tim_Eisele@web.de>2026-07-21 14:52:59 +0200
committerGitHub <noreply@github.com>2026-07-21 14:52:59 +0200
commited61acc19a958ceaf8de6f6f0f481b2b5370463a (patch)
treef154b5c001e78d4147a69d01c64f41355e83660f
parent9b9b609c83404faeff485919b18cb88ce6cc4db6 (diff)
Fix subtitle encoding for local files (#17281)
* Fix subtitle encoding * Add short-circuit * Use IsTextFormat * Update MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs Co-authored-by: Bond-009 <bond.009@outlook.com> --------- Co-authored-by: Bond-009 <bond.009@outlook.com>
-rw-r--r--MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs34
-rw-r--r--tests/Jellyfin.MediaEncoding.Tests/Subtitles/SubtitleEncoderTests.cs244
2 files changed, 201 insertions, 77 deletions
diff --git a/MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs b/MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs
index c568a74b01..bd516f0a9f 100644
--- a/MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs
+++ b/MediaBrowser.MediaEncoding/Subtitles/SubtitleEncoder.cs
@@ -163,28 +163,36 @@ namespace MediaBrowser.MediaEncoding.Subtitles
return (stream, fileInfo);
}
- private async Task<Stream> GetSubtitleStream(SubtitleInfo fileInfo, CancellationToken cancellationToken)
+ internal async Task<Stream> GetSubtitleStream(SubtitleInfo fileInfo, CancellationToken cancellationToken)
{
- if (fileInfo.Protocol == MediaProtocol.Http)
+ if (fileInfo.IsExternal && MediaStream.IsTextFormat(fileInfo.Format))
{
var result = await DetectCharset(fileInfo.Path, cancellationToken).ConfigureAwait(false);
var detected = result.Detected;
- if (detected is not null)
+ var stream = fileInfo.Protocol == MediaProtocol.Http
+ ? await _httpClientFactory.CreateClient(NamedClient.Default)
+ .GetStreamAsync(new Uri(fileInfo.Path), cancellationToken)
+ .ConfigureAwait(false)
+ : AsyncFile.OpenRead(fileInfo.Path);
+
+ // Short-circuit when the file is already UTF-8/ASCII.
+ if (detected is null
+ || string.Equals(detected.EncodingName, "utf-8", StringComparison.OrdinalIgnoreCase)
+ || string.Equals(detected.EncodingName, "ascii", StringComparison.OrdinalIgnoreCase)
+ || string.Equals(detected.EncodingName, "us-ascii", StringComparison.OrdinalIgnoreCase))
{
- _logger.LogDebug("charset {CharSet} detected for {Path}", detected.EncodingName, fileInfo.Path);
+ return stream;
+ }
- using var stream = await _httpClientFactory.CreateClient(NamedClient.Default)
- .GetStreamAsync(new Uri(fileInfo.Path), cancellationToken)
- .ConfigureAwait(false);
+ _logger.LogDebug("charset {CharSet} detected for {Path}", detected.EncodingName, fileInfo.Path);
- await using (stream.ConfigureAwait(false))
- {
- using var reader = new StreamReader(stream, detected.Encoding);
- var text = await reader.ReadToEndAsync(cancellationToken).ConfigureAwait(false);
+ await using (stream.ConfigureAwait(false))
+ {
+ using var reader = new StreamReader(stream, detected.Encoding);
+ var text = await reader.ReadToEndAsync(cancellationToken).ConfigureAwait(false);
- return new MemoryStream(Encoding.UTF8.GetBytes(text));
- }
+ return new MemoryStream(Encoding.UTF8.GetBytes(text));
}
}
diff --git a/tests/Jellyfin.MediaEncoding.Tests/Subtitles/SubtitleEncoderTests.cs b/tests/Jellyfin.MediaEncoding.Tests/Subtitles/SubtitleEncoderTests.cs
index 2d0fa29c9a..48850b2f67 100644
--- a/tests/Jellyfin.MediaEncoding.Tests/Subtitles/SubtitleEncoderTests.cs
+++ b/tests/Jellyfin.MediaEncoding.Tests/Subtitles/SubtitleEncoderTests.cs
@@ -21,81 +21,110 @@ namespace Jellyfin.MediaEncoding.Subtitles.Tests
private const int StreamCount = 8;
private const int CueCount = 500;
- public static TheoryData<MediaSourceInfo, MediaStream, SubtitleEncoder.SubtitleInfo> GetReadableFile_Valid_TestData()
+ // A Greek line that requires a non-UTF-8 legacy encoding to reproduce the bug. The accented
+ // characters (ά, έ, ή, ί, ό, ύ, ώ) share the same code points in windows-1253 and iso-8859-7,
+ // so a Greek-vs-Greek charset misdetection still round-trips correctly.
+ private const string GreekText = "Καλημέρα κόσμε, αυτό είναι ένας υπότιτλος.";
+
+ static SubtitleEncoderTests()
{
- var data = new TheoryData<MediaSourceInfo, MediaStream, SubtitleEncoder.SubtitleInfo>();
+ // Mirrors Jellyfin.Server startup so legacy code pages (e.g. Greek windows-1253) are available.
+ Encoding.RegisterProvider(CodePagesEncodingProvider.Instance);
+ }
- data.Add(
- new MediaSourceInfo()
- {
- Protocol = MediaProtocol.File
- },
- new MediaStream()
- {
- Path = "/media/sub.ass",
- IsExternal = true
- },
- new SubtitleEncoder.SubtitleInfo()
- {
- Path = "/media/sub.ass",
- Protocol = MediaProtocol.File,
- Format = "ass",
- IsExternal = true
- });
+ // Enough Greek text to give the charset detector a strong, unambiguous signal.
+ private static string BuildGreekSrt()
+ {
+ var builder = new StringBuilder();
+ for (var i = 1; i <= 8; i++)
+ {
+ builder.Append(i.ToString(CultureInfo.InvariantCulture)).Append('\n');
+ builder.Append("00:00:0").Append(i.ToString(CultureInfo.InvariantCulture))
+ .Append(",000 --> 00:00:0").Append((i + 1).ToString(CultureInfo.InvariantCulture)).Append(",000\n");
+ builder.Append(GreekText).Append('\n');
+ builder.Append("Η γρήγορη καφέ αλεπού πηδάει πάνω από το τεμπέλικο σκυλί.\n\n");
+ }
- data.Add(
- new MediaSourceInfo()
- {
- Protocol = MediaProtocol.File
- },
- new MediaStream()
- {
- Path = "/media/sub.ssa",
- IsExternal = true
- },
- new SubtitleEncoder.SubtitleInfo()
- {
- Path = "/media/sub.ssa",
- Protocol = MediaProtocol.File,
- Format = "ssa",
- IsExternal = true
- });
+ return builder.ToString();
+ }
- data.Add(
- new MediaSourceInfo()
+ public static TheoryData<MediaSourceInfo, MediaStream, SubtitleEncoder.SubtitleInfo> GetReadableFile_Valid_TestData()
+ {
+ var data = new TheoryData<MediaSourceInfo, MediaStream, SubtitleEncoder.SubtitleInfo>
+ {
{
- Protocol = MediaProtocol.File
+ new MediaSourceInfo()
+ {
+ Protocol = MediaProtocol.File
+ },
+ new MediaStream()
+ {
+ Path = "/media/sub.ass",
+ IsExternal = true
+ },
+ new SubtitleEncoder.SubtitleInfo()
+ {
+ Path = "/media/sub.ass",
+ Protocol = MediaProtocol.File,
+ Format = "ass",
+ IsExternal = true
+ }
},
- new MediaStream()
{
- Path = "/media/sub.srt",
- IsExternal = true
+ new MediaSourceInfo()
+ {
+ Protocol = MediaProtocol.File
+ },
+ new MediaStream()
+ {
+ Path = "/media/sub.ssa",
+ IsExternal = true
+ },
+ new SubtitleEncoder.SubtitleInfo()
+ {
+ Path = "/media/sub.ssa",
+ Protocol = MediaProtocol.File,
+ Format = "ssa",
+ IsExternal = true
+ }
},
- new SubtitleEncoder.SubtitleInfo()
{
- Path = "/media/sub.srt",
- Protocol = MediaProtocol.File,
- Format = "srt",
- IsExternal = true
- });
-
- data.Add(
- new MediaSourceInfo()
- {
- Protocol = MediaProtocol.Http
+ new MediaSourceInfo()
+ {
+ Protocol = MediaProtocol.File
+ },
+ new MediaStream()
+ {
+ Path = "/media/sub.srt",
+ IsExternal = true
+ },
+ new SubtitleEncoder.SubtitleInfo()
+ {
+ Path = "/media/sub.srt",
+ Protocol = MediaProtocol.File,
+ Format = "srt",
+ IsExternal = true
+ }
},
- new MediaStream()
{
- Path = "/media/sub.ass",
- IsExternal = true
- },
- new SubtitleEncoder.SubtitleInfo()
- {
- Path = "/media/sub.ass",
- Protocol = MediaProtocol.File,
- Format = "ass",
- IsExternal = true
- });
+ new MediaSourceInfo()
+ {
+ Protocol = MediaProtocol.Http
+ },
+ new MediaStream()
+ {
+ Path = "/media/sub.ass",
+ IsExternal = true
+ },
+ new SubtitleEncoder.SubtitleInfo()
+ {
+ Path = "/media/sub.ass",
+ Protocol = MediaProtocol.File,
+ Format = "ass",
+ IsExternal = true
+ }
+ }
+ };
return data;
}
@@ -113,6 +142,55 @@ namespace Jellyfin.MediaEncoding.Subtitles.Tests
Assert.Equal(subtitleInfo.IsExternal, result.IsExternal);
}
+ public static TheoryData<Encoding> GetSubtitleStream_NonUtf8LocalFile_TestData()
+ {
+ return
+ [
+ // Greek legacy encodings – the exact scenario reported in issue #17267.
+ Encoding.GetEncoding("windows-1253"),
+ Encoding.GetEncoding("iso-8859-7"),
+ // Wide encoding with a BOM.
+ new UnicodeEncoding(bigEndian: false, byteOrderMark: true),
+ ];
+ }
+
+ [Theory]
+ [MemberData(nameof(GetSubtitleStream_NonUtf8LocalFile_TestData))]
+ public async Task GetSubtitleStream_NonUtf8LocalFile_ConvertedToUtf8(Encoding sourceEncoding)
+ {
+ var cancellationToken = TestContext.Current.CancellationToken;
+ var srt = BuildGreekSrt();
+ var path = Path.GetTempFileName();
+ try
+ {
+ await File.WriteAllTextAsync(path, srt, sourceEncoding, cancellationToken);
+
+ var fixture = new Fixture().Customize(new AutoMoqCustomization { ConfigureMembers = true });
+ var subtitleEncoder = fixture.Create<SubtitleEncoder>();
+
+ var fileInfo = new SubtitleEncoder.SubtitleInfo
+ {
+ Path = path,
+ Protocol = MediaProtocol.File,
+ Format = "srt",
+ IsExternal = true
+ };
+
+ using var stream = await subtitleEncoder.GetSubtitleStream(fileInfo, cancellationToken);
+ using var reader = new StreamReader(stream, new UTF8Encoding(false));
+ var text = await reader.ReadToEndAsync(cancellationToken);
+
+ // The Greek text must survive round-trip and contain no replacement characters.
+ Assert.Contains(GreekText, text, StringComparison.Ordinal);
+ Assert.DoesNotContain('�', text);
+ Assert.DoesNotContain('?', text);
+ }
+ finally
+ {
+ File.Delete(path);
+ }
+ }
+
[Fact]
public void ConvertSubtitles_SequentialCalls_AreDeterministic()
{
@@ -130,6 +208,44 @@ namespace Jellyfin.MediaEncoding.Subtitles.Tests
}
[Fact]
+ public async Task GetSubtitleStream_Utf8LocalFile_PreservesContent()
+ {
+ var cancellationToken = TestContext.Current.CancellationToken;
+ var srt = BuildGreekSrt();
+ var path = Path.GetTempFileName();
+ try
+ {
+ await File.WriteAllTextAsync(path, srt, new UTF8Encoding(false), cancellationToken);
+
+ var fixture = new Fixture().Customize(new AutoMoqCustomization { ConfigureMembers = true });
+ var subtitleEncoder = fixture.Create<SubtitleEncoder>();
+
+ var fileInfo = new SubtitleEncoder.SubtitleInfo
+ {
+ Path = path,
+ Protocol = MediaProtocol.File,
+ Format = "srt",
+ IsExternal = true
+ };
+
+ using var stream = await subtitleEncoder.GetSubtitleStream(fileInfo, cancellationToken);
+
+ // An already-UTF-8 file must be short-circuited and served directly from disk,
+ // not read into memory and re-encoded (which would produce a MemoryStream).
+ Assert.IsNotType<MemoryStream>(stream);
+
+ using var reader = new StreamReader(stream, new UTF8Encoding(false));
+ var text = await reader.ReadToEndAsync(cancellationToken);
+
+ Assert.Contains(GreekText, text, StringComparison.Ordinal);
+ }
+ finally
+ {
+ File.Delete(path);
+ }
+ }
+
+ [Fact]
public async Task ConvertSubtitles_ConcurrentCalls_MatchSequentialBaseline()
{
const int Iterations = 10;