Request for ability to get display scaling properties from a window

[john01dav]

It would be very useful to be able to get a display scaling factor for a window, in order to render what is inside it at an appropriate size. Otherwise, when things are rendered on, for example, a 4K display, then they may be quite inappropriately small. I'd imagine that this is as simple as adding a function that wraps a different native function for each platform to return this factor, maybe with some trivial formula to scale it to a common format (e.g. 2.0 vs 200, to represent 200% scaling), although I don't know anything about the native APIs that minifb wraps, so this may be wrong.

[emoon]

What you can do is to use this

https://github.com/emoon/rust_minifb/blob/master/src/lib.rs#L18 and pass it as something like this

        WindowOptions {
            scale: Scale::FitScreen,
            ..WindowOptions::default()
        },

So now if you have a 320x200 window on a 4k screen it will resize the window to fit your desktop resolution (but the render area if your window will be still 320x200)

Is that what your are looking for?

[john01dav]

That feature isn't quite what I'm looking for, as I prefer the ability to render the contents of my window at a higher resolution as appropriate in my own code as it has access to a vector graphics representation of the contents, rather than letting minifb do a raster resize, which can only produce inferior results. I just need an API hook like this function from winit (+ a way to know when it changes, ideally but not necessarily, whether that means polling it each frame, or an event). I can't use winit however, as its architecture is not what I need — I need something that enables 100% software rendering, without the involvement of a GPU. I've looked into the Windows and Linux APIs a bit (this winit documentation has a surprising amount of detail on this), and it looks like it wouldn't be too hard to implement this, so I may be able to do a pull request for the Linux and Windows implementations, but I don't have access to an Apple computer to test on.

[emoon]

I see. Doing the PR for Linux and Windows would be fine. I can add the macOS support as I don't think it would be much work as you say.

[emoon]

Being tracked in #236