Problems in RMX-OS, Input.update about mkxp HOT 14 CLOSED

And this is on Windows, yes?

from mkxp.

This is on Windows yes.

I've managed a temporary fix after some tinkering, by including another call to Win32API:

GetKeyState=Win32API.new("user32", "GetKeyState", "I", "I")
GetKeyState.call(0)

This makes input work just fine it has just one very big side-effect.

It continues grabbing and updating keystrokes even when the application isn't active/minimized.

Do you have any plans of adding full-keyboard input to mkxp, perhaps?

from mkxp.

Do you have any plans of adding full-keyboard input to mkxp, perhaps?

Yes, see #70. I'm still thinking about the best way to write something generic that can cover as many use cases as possible. I will probably open a separate issue for it soon.

from mkxp.

I created an issue for general keyboard input: #73. As for your specific problem khkramer, it could be related to how SDL sets up its window and processes messages on Windows. Why do you need it to work so badly btw., I thought you were trying to port the code over to mkxp anyway?

from mkxp.

Well I had to have the chat functional to test a lot of things, and I didn't know if you were going to add full keyboard input or something that allows us to use those full keyboard input scripts, so my plan was originally to first fix it on windows and then look up if there are similar api calls on Linux.

I'm pretty sure it's how SDL set up its windows then, since I'm checking the thread id of the caller (so the game) and then checking the thread id of the foreground window, and then comparing them to see if the game has focus and is in the foreground. However even when it is the foreground window it returns another thread id for the window than the 'normal' thread id if that makes sense

from mkxp.

Added full keyboard support to mkxp, you can check my win32 wrapper for how to use it.

from mkxp.

Cool.

I'll try it out later today, and let you know how it worked.

from mkxp.

Closing this for now, reopen if you need to.

from mkxp.

Very very late response from me :)

It works like a charm, however it breaks mouseclicks for me.
I've found the cause as well, in Input.h, MouseLeft has scancode 0x38, while in the RMX-OS input script and your win32-wrapper the 'Up' key has scancode 0x38. So whenever I press the up key it thinks the left mouse is being clicked as well, while clicking the actual left mouse does nothing.

from mkxp.

Oh, I see, it's because you're overriding Input functions so my wrapper ends up calling your modified code. The proper fix on the mkxp side is to just completely remove the mouse extensions from Input and stash them into MKXP (which I should have done from the start), but until that change has landed, the quick fix is to save an alias to the original update function and call that in the win32 wrapper instead.

from mkxp.

Input.update isn't being called in the win32-wrapper??

from mkxp.

Oh sorry, brainfart. I meant press? of course.

from mkxp.

I saved an alias like this

module Input
  class << self
    alias_method(:win32wrap_press?, :press?)
    def press?(num)
      win32wrap_press?(num)
    end
  end
end

And called that instead and it's not detecting arrow keys as mouse buttons anymore, however actual mouse clicks still aren't detected.

Edit: Tried another input script with which it does work. So this problem is pretty much fixed on my end.

from mkxp.

I'd still be interested into why Mouse input didn't work the first time, but oh well.

from mkxp.