1015028 - Should compute scancode from virtual keycode at computing KeyboardEvent.code value if the key message doesn't have proper scancode

Status: RESOLVED FIXED

(Core :: Widget: Win32, defect)

Comment 1

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Patch

https://tbpl.mozilla.org/?tree=Try&usebuildbot=1&rev=29ac934b3fdf

Comment 2

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

D3E defines the behavior in this situation like this:
https://dvcs.w3.org/hg/dom3events/raw-file/tip/html/DOM3-Events.html#code-virtual-keyboards

If scancode is properly set, we should compute .code value with it. Otherwise, we should map code value from virtual keycode as far as possible for web apps which expect that .code value is set in all cases.

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Patch

I realized that the multimedia keys which are mapped with virtual keycode can be mapped only with scancode.

This is not completely emulating physical keyboard behavior. MapVirtualKeyEx() returns non-extended key's scancode as far as possible. Therefore, e.g., VK_DOWN causes "Numpad2" rather than "ArrowDown" even if NumLock is unlocked. I feel that this is not better behavior. However, I also don't have better and smarter implementation for some extended keys since any virtual keycodes might be mapped to different physical keys.

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/510b1b15f384

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/510b1b15f384