windows/QuickLook
1
0
Fork 0
mirror of https://github.com/QL-Win/QuickLook.git synced 2025-09-12 18:39:45 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Double encoding detection #471 #600 #769 #954

Browse Source
This commit is contained in:
ema
2024-12-24 11:22:05 +08:00
parent 596ff6d9d4
commit 36807702e2
1 changed files with 57 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

59
QuickLook.Plugin/QuickLook.Plugin.TextViewer/TextViewerPanel.cs
View File

@@ -24,6 +24,7 @@ using QuickLook.Common.Helpers;
using QuickLook.Common.Plugin; using QuickLook.Common.Plugin;
using System; using System;
using System.IO; using System.IO;
using System.Linq;
using System.Reflection; using System.Reflection;
using System.Text; using System.Text;
using System.Threading.Tasks; using System.Threading.Tasks;
@@ -177,8 +178,8 @@ public class TextViewerPanel : TextEditor, IDisposable
var bufferCopy = buffer.ToArray(); var bufferCopy = buffer.ToArray();
buffer.Dispose(); buffer.Dispose();
var encoding = CharsetDetector.DetectFromBytes(bufferCopy).Detected?.Encoding ?? var result = CharsetDetector.DetectFromBytes(bufferCopy);
Encoding.Default; var encoding = result.DoubleDetectFromResult(bufferCopy); // Fix issues
var doc = new TextDocument(encoding.GetString(bufferCopy)); var doc = new TextDocument(encoding.GetString(bufferCopy));
doc.SetOwnerThread(Dispatcher.Thread); doc.SetOwnerThread(Dispatcher.Thread);
@@ -199,3 +200,57 @@ public class TextViewerPanel : TextEditor, IDisposable
}); });
} }
} }
file static class DetectionExtensions
{
public static Encoding DoubleDetectFromResult(this DetectionResult result, byte[] buffer)
{
// Determine the highest confidence encoding, or fallback to ANSI
var encoding = result.Detected?.Encoding ?? Encoding.Default;
// When mixing encodings, one of the encodings may gain higher confidence
// In this case, we should return to encodings UTF8 / UTF32 / ANSI
// https://github.com/QL-Win/QuickLook/issues/769
if (encoding != Encoding.UTF8 && encoding != Encoding.UTF32 && encoding != Encoding.Default)
{
if (result.Details.Any(detail => detail.Encoding == Encoding.UTF8))
{
encoding = Encoding.UTF8;
}
else if (result.Details.Any(detail => detail.Encoding == Encoding.UTF32))
{
encoding = Encoding.UTF32;
}
else if (result.Details.Any(detail => detail.Encoding == Encoding.Default))
{
encoding = Encoding.Default;
}
}
// When the text is too short and lacks a BOM
// In this case, we should fallback to an encoding if it is not recognized as UTF8 / UTF32 / ANSI
// https://github.com/QL-Win/QuickLook/issues/471
// https://github.com/QL-Win/QuickLook/issues/600
// https://github.com/QL-Win/QuickLook/issues/954
if (buffer.Length <= 50)
{
if (encoding != Encoding.UTF8 && encoding != Encoding.UTF32 && encoding != Encoding.Default)
{
if (!Encoding.UTF8.GetString(buffer).Contains("\uFFFD"))
{
encoding = Encoding.UTF8;
}
else if (!Encoding.UTF32.GetString(buffer).Contains("\uFFFD"))
{
encoding = Encoding.UTF32;
}
else if (!Encoding.Default.GetString(buffer).Contains("\uFFFD"))
{
encoding = Encoding.Default;
}
}
}
return encoding;
}
}