Bug: C++WinRT does not correctly handles Chinese characters using `Windows.Foundation.Uri`

[HO-COOH]

Version

2.0.240405.15

Summary

I have a uri with Chinese character and query string that needs to be parsed with Windows.Foundation.Uri. Using C++WinRT, it always returns me an empty result. But equivalent C# code does returns me expected result.

I have tried changing the file encoding to either UTF-8 and UTF-8 BOM, nothing worked.

Reproducible example

With C++WinRT, create a C++WinRT console application:

    winrt::Windows::Foundation::Uri uri{ LR"(myapp://open?file="C:/我.txt")" };
    winrt::Windows::Foundation::WwwFormUrlDecoder parsed{ uri.Query() };
    for (auto entry : parsed)
    {
        std::wcout << entry.Name().data() << L'\t' << entry.Value().data() << L'\n';
    }

With C#, create a UWP project:

            var uri = new Uri("myapp://open?file=\"C:/我.txt\"");
            var parsed = new WwwFormUrlDecoder(uri.Query);
            foreach (var entry in parsed)
            {
                Debug.WriteLine(entry);
            }

Expected behavior

C# result is expected:

Actual behavior

C++WinRT result:

Additional comments

Repro here

[sylveon]

Are you sure the bug is not with the console output? Check the debugger maybe?

[HO-COOH]

Are you sure the bug is not with the console output? Check the debugger maybe?

The break point inside the for loop is not hit. That should be clear enough.

[kennykerr]

For API questions I suggest: https://docs.microsoft.com/en-us/answers/topics/windows-api.html

[HO-COOH]

@kennykerr I don't know, but both examples uses the same Windows.Foundation.WwwFormUrlDecoder class which makes me wondering. I will ask there too.

[DefaultRyan]

It uses the same Windows.Foundation.WwwFormUrlDecoder class, but System.Uri is not the same as Windows.Foundation.Uri. The doc page for Windows.Foundation.Uri calls out some of these potential differences in the Remarks https://learn.microsoft.com/en-us/uwp/api/windows.foundation.uri?view=winrt-26100#remarks, including potentially relevant statements about percent-encoding non-ASCII characters.

That's as far as I got, but I'm confident that this is where you're seeing behavior differences.