MOUSEINPUT is of the wrong type

[pentamassiv]

Summary

It is the same issue as microsoft/windows-rs#2708. Even after updating the metadata in microsoft/windows-rs@b62b802, the issue still persists.

In version 0.51, windows::Win32::UI::Input::KeyboardAndMouse::MOUSEINPUT.mouseData was of type i32. Starting in version 0.52, it is of type u32. According to the documentation, the type is DWORD. However it also states:

If dwFlags contains MOUSEEVENTF_WHEEL, then mouseData specifies the amount of wheel movement. A positive value indicates that the wheel was rotated forward, away from the user; a negative value indicates that the wheel was rotated backward, toward the user. One wheel click is defined as WHEEL_DELTA, which is 120.

(source).

Crate manifest

No response

Crate code

No response

[kennykerr]

MOUSEINPUT is originally defined as follows:

typedef struct tagMOUSEINPUT {
    LONG    dx;
    LONG    dy;
    DWORD   mouseData;
    DWORD   dwFlags;
    DWORD   time;
    ULONG_PTR dwExtraInfo;
} MOUSEINPUT, *PMOUSEINPUT, FAR* LPMOUSEINPUT;

The metadata defines it as follows:

public struct MOUSEINPUT
{
	public int dx;
	public int dy;
	public uint mouseData;
	public MOUSE_EVENT_FLAGS dwFlags;
	public uint time;
	public UIntPtr dwExtraInfo;
}

So mouseData looks correct to me.

[pentamassiv]

It looks like there is a lot of going back and forth between the two types (https://github.com/microsoft/win32metadata/issues?q=is%3Aissue+mouseinput+is%3Aclosed). It was changed to i32 because you have to be able to use negative values to scroll up (#933) and a special case was added to the metadata. Later that seems to have been reverted.

[kennykerr]

Yes, it seems that way. I haven't been tracking that, but I believe the value is correct as of now. I'll transfer to the Win32 metadata repo for their input.

[mikebattista]

As long as we're ok deviating from the headers, we can change it back to int.

[kennykerr]

The type is clearly uint - I don't see a compelling reason to change it - the caller is clearly expected to cast a negative value.

[riverar]

I originally marked it as int long before we had ABI and codegen compatibility concerns/discussions (#889). I now agree this should remain uint and the user should cast.

Metadata currently has this listed as uint, but then we have this in the emitter rsp... 🤔 https://github.com/riverar/win32metadata/blob/master/generation/WinSDK/emitter.settings.rsp#L464. Looks like something we need to clean up.

[mikebattista]

Everything looks correct to me in the latest builds.

[pentamassiv]

Okay, thank you for the quick response. I will cast the values.