Fix keyboard on windows issue 12937

[codec-abc]

This is an attempt to fix the behavior of the input elements on Windows to fix #12937. Now, the input fields mostly work for the most part but the backspace key has a weird behavior and there is probably a couple of funkyness elsewhere By the way, I tried to align the keyboard events so we got the same in MacOs & Linux and Windows. A couple of things to notes:

1. There are 2 types of events associated with keyboard: Key event (up or down) and Character Input.
2. A character is not necessary created one a key is pressed (like the home key).
3. A character may be created only using a combination of several key. For example, using a Qwerty International layout you have to type ` then e to get é. The first key press on ` will create no character.
4. In servo, we currently merge the 2 events together, meaning that we store a Character Input event in a field to fire a single event later on when we get a Key event.
5. The order of events seems to be system dependent. For example, let's say that we type the A key on a Qwerty Layout. In Linux and MacOs we will get this:

• Character Input for character A
• Key down event with virtual key code of the key A
• Key up event with virtual key code of the key A
  while in Windows we get:
• Key down event with virtual key code of the key A
• Character Input for character A
• Key up event with virtual key code of the key A

6. It seems that glutin make no attempt to reorder the event to make the order independent of the Operating System. I think it would be easier for Servo if it were handled by the library.
7. The Js event generated by Servo seems to be correct when testing on this page but not on Windows (before or after my changes). They may be no more correct in MacOs and Linux after my changes since I tried to align the event to a kind of middle ground. By the way this page may be serve a basic for a good test for input event order.

---

• ./mach build -d does not report any errors
• ./mach test-tidy does not report any errors
• These changes fix (partially) #12937.

• There are NO tests for these changes because I don't know how to write them and what to write.

---

This change is 

EDIT: I do not like the code I have written because there is quite a lot of platform dependent code in it. It would probably be proper create a file per platform so the logic behind may be encapsulated in those files.

[jonathandturner]

Just gave this a spin. While yeah, looks like we're getting the keys when we press them, I'm seeing things like backspace no longer work. It's not immediately obvious to me what's going wrong looking at the code, but I suspect non-printable characters aren't getting the same way they were before and we're missing a step, potentially.

[codec-abc]

Yeah, it seems weird since if I remember correctly when I pressed backspace the character was erased but then a square appear right away. Now while I think of it I remember a warning about not using the filter_nonprintable function and looking at the code it seems it is not used on Windows.

EDIT: After more looking at the code, I think the problem come from not using the filter_nonprintable function. I think the way to use it on Windows is to check if the value of the last last_pressed_key field is non printable when receiving a CharacterInput event. If the virtual keycode is non printable then the event should not be fired.

[jonathandturner]

Yeah, I'd believe that. When I was looking around before, I didn't see a way to check if the key was printable in the keydown event. Do you know of one?

[codec-abc]

I am not sure If I follow you since in the keydown event you should get the virtual key code (except if I missed something) and could use (almost as it is) the filter_nonprintable function.

[jonathandturner]

@codec-abc - Ahh, I misread what approach you were trying to take. I'm curious if your way works. If it does, it seems like the right way to do it.

[codec-abc]

Go ahead if you want to test it because I will not have time tomorrow.

[jonathandturner]

@codec-abc - done: #16198

I agree, still not a perfect solution, but this one appears to work decently in practice and may be enough to get us going until we come up with a better solution.

[codec-abc]

Thanks! I will close this one since your PR goes a step further and fix more bugs.