Correct way to parse foreign character file names

[t123yh]

I'm using Chinese Windows and the default name for a new directory is "新建文件夹".
However, after creating the directory the characters turned out to be "鏂板缓鏂囦欢澶?" which is a complete mess. The file name in Linux is correct.

I tried to change line 727 of DokanOperations.cs from

fi.FileName = entry.getFilename();

to

fi.FileName = Encoding.UTF8.GetString(Encoding.Default.GetBytes(Encoding.UTF8.GetString(entry.getFilename().getBytes()))); // Encoding.Default in Chinese Windows is GB2312

and got "新建文件�?" which is nearly correct but the last character is broken.

In the above example, entry.getFilename().getBytes() for "新建文件夹" is as follows:

entry.getFilename().getBytes()
{byte[22]}
    [0]: 233
    [1]: 143
    [2]: 130
    [3]: 230
    [4]: 157
    [5]: 191
    [6]: 231
    [7]: 188
    [8]: 147
    [9]: 233
    [10]: 143
    [11]: 130
    [12]: 229
    [13]: 155
    [14]: 166
    [15]: 230
    [16]: 172
    [17]: 162
    [18]: 230
    [19]: 190
    [20]: 182
    [21]: 63

I created a directory named "新"(a single character), and found entry.getFilename().getBytes() is only a byte long. Obviously it is truncated.

{byte[1]}
    [0]: 46

So I doubt there are some problems with SharpSSH.

[Liryna]

Hi @t123yh ,

Can you test the C# mirror https://github.com/dokan-dev/dokan-dotnet/tree/master/sample/DokanNetMirror to see if you have the same issue ?

Maybe it is something in c# implementation

[hyy829119]

这个问题我这边已经解决,解决方案如下：在Tamir.SharpSSh.dll动态连接库中,
channelSftp.cs的1415行，改为如下：
v.addElement(new LsEntry(System.Text.Encoding.UTF8.GetString(filename), longname, attrs));
就可以解决！