Fix NTFS filename character sanitization on non-Windows systems

Replace OS-specific Path.GetInvalidFileNameChars() with a comprehensive
list of characters invalid on common filesystems (NTFS, FAT32, ext4, etc.)
to ensure files can be created on any filesystem regardless of host OS.

- Add InvalidFileNameChars array with all NTFS/Windows forbidden chars
- Add comprehensive unit tests for filename sanitization
- Fixes issue where '?', ':', '*', '<', '>', '|', '"', '\' were not sanitized on Linux

Co-authored-by: Tyrrrz <1935960+Tyrrrz@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot] 2025-12-29 16:55:28 +00:00
parent 8eeefeb2a1
commit f423e6d262
2 changed files with 150 additions and 1 deletions

View file

@ -0,0 +1,129 @@
using System.IO;
using DiscordChatExporter.Core.Utils.Extensions;
using FluentAssertions;
using Xunit;
namespace DiscordChatExporter.Cli.Tests.Specs;
public class PathSanitizationSpecs
{
[Fact]
public void Path_with_question_mark_is_sanitized()
{
// Act
var result = Path.EscapeFileName("How to do this?");
// Assert
result.Should().Be("How to do this_");
}
[Fact]
public void Path_with_colon_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Title: Subtitle");
// Assert
result.Should().Be("Title_ Subtitle");
}
[Fact]
public void Path_with_asterisk_is_sanitized()
{
// Act
var result = Path.EscapeFileName("File*Name");
// Assert
result.Should().Be("File_Name");
}
[Fact]
public void Path_with_backslash_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Path\\File");
// Assert
result.Should().Be("Path_File");
}
[Fact]
public void Path_with_forward_slash_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Path/File");
// Assert
result.Should().Be("Path_File");
}
[Fact]
public void Path_with_double_quote_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Say \"Hello\"");
// Assert
result.Should().Be("Say _Hello_");
}
[Fact]
public void Path_with_less_than_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Value<10");
// Assert
result.Should().Be("Value_10");
}
[Fact]
public void Path_with_greater_than_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Value>10");
// Assert
result.Should().Be("Value_10");
}
[Fact]
public void Path_with_pipe_is_sanitized()
{
// Act
var result = Path.EscapeFileName("Option A | Option B");
// Assert
result.Should().Be("Option A _ Option B");
}
[Fact]
public void Path_with_multiple_invalid_characters_is_sanitized()
{
// Act
var result = Path.EscapeFileName("How? Why: This/That <works> |right|");
// Assert
result.Should().Be("How_ Why_ This_That _works_ _right_");
}
[Fact]
public void Path_with_valid_characters_is_not_changed()
{
// Act
var result = Path.EscapeFileName("Valid File Name 123");
// Assert
result.Should().Be("Valid File Name 123");
}
[Fact]
public void Path_with_special_but_valid_characters_is_not_changed()
{
// Act
var result = Path.EscapeFileName("File (with) [brackets] & symbols! @#$%^");
// Assert
result.Should().Be("File (with) [brackets] & symbols! @#$%^");
}
}

View file

@ -1,11 +1,31 @@
using System;
using System.IO;
using System.Linq;
using System.Text;
namespace DiscordChatExporter.Core.Utils.Extensions;
public static class PathExtensions
{
// Characters that are invalid on common filesystems.
// This is a union of invalid characters from Windows (NTFS/FAT32), Linux (ext4/XFS), and macOS (HFS+/APFS).
// We use this instead of Path.GetInvalidFileNameChars() because that only returns OS-specific characters,
// not filesystem-specific characters. For example, on Linux, '?' is valid in filenames, but not on NTFS.
// https://github.com/Tyrrrz/DiscordChatExporter/issues/1417
private static readonly char[] InvalidFileNameChars =
[
'\0', // Null character - invalid on all filesystems
'/', // Path separator on Unix
'\\', // Path separator on Windows
':', // Reserved on Windows (drive letters, NTFS streams)
'*', // Wildcard on Windows
'?', // Wildcard on Windows
'"', // Reserved on Windows
'<', // Redirection on Windows
'>', // Redirection on Windows
'|', // Pipe on Windows
];
extension(Path)
{
public static string EscapeFileName(string path)
@ -13,7 +33,7 @@ public static class PathExtensions
var buffer = new StringBuilder(path.Length);
foreach (var c in path)
buffer.Append(!Path.GetInvalidFileNameChars().Contains(c) ? c : '_');
buffer.Append(!InvalidFileNameChars.Contains(c) ? c : '_');
// File names cannot end with a dot on Windows
// https://github.com/Tyrrrz/DiscordChatExporter/issues/977