Switch KeyboardEvents over to NewEventHandler

Switch KeyboardEvents over to NewEventHandler

This CL moves KeyboardEvents from the old event model to NewEventHandler. This
CL keeps the basic structure of keydown, keypress, keyup events even though
that's a bit wacky. As with pointer and gesture events, this CL removes
PlatformKeyboardEvent in favor of just using WebKeyboardEvent. I've also made
WebKeyboardEvent align more closely with Mojo's keyboard event.

The CL does change one important aspect of key event handling: on the web the
"keyCode" property of KeyboardEvent changes its meaning depending on whether
the event is a keydown or a keypress event. For the former events, keyCode is
the "virtual" (i.e., windows) key code where for the latter events, keyCode is
the character code. To be more precise, I've renamed keyCode to virtualKeyCode
and I've given it a zero (unknown key code) value during keypress events.

R=ojan@chromium.org, eseidel@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/f5143561cc7cce29082a60238922124e4d8c0b25

[ojan, 2015-01-26 06:18:50]

ojan@chromium.org changed reviewers:
+ ojan@chromium.org

[ojan, 2015-01-26 06:18:52]

lgtm

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
File sky/engine/core/events/KeyboardEvent.h (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
sky/engine/core/events/KeyboardEvent.h:45: unsigned virtualKeyCode() const {
return m_virtualKeyCode; }
FWIW, virtual keycode means something to windows developers I think, but it
means nothing to me. I'm not sure what's "virtual" about it.

Also, being more future-looking, the new standard key events in DOM have "key"
and "char" that are the actual character being pressed. Seems more developer
friendly than codes.

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
File sky/engine/core/frame/NewEventHandler.cpp (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
sky/engine/core/frame/NewEventHandler.cpp:152: // Do we really need to suppress
keypress events for these keys anymore?
I think probably we don't need to. Although, really, I think we probably want to
get rid of keypress events entirely and just have keydown/keyup.

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
sky/engine/core/frame/NewEventHandler.cpp:176: if (!target)
Presumably this is the fix for https://github.com/domokit/mojo/issues/40?

https://codereview.chromium.org/872233002/diff/20001/sky/engine/web/WebViewIm...
File sky/engine/web/WebViewImpl.cpp (left):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/web/WebViewIm...
sky/engine/web/WebViewImpl.cpp:311: bool WebViewImpl::mapKeyCodeForScroll(
This is another place where if we were to plug in a mouse or keyboard, things
wouldn't work as we expect. It's easy to add back in later though. Although, it
will get harder and harder (e.g. well delete a bunch of the ScrollGranularity
code next since we'll find that it's dead code because we keep deleting all the
callers).

[abarth-chromium, 2015-01-26 06:28:16]

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
File sky/engine/core/events/KeyboardEvent.h (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
sky/engine/core/events/KeyboardEvent.h:45: unsigned virtualKeyCode() const {
return m_virtualKeyCode; }
On 2015/01/26 at 06:18:51, ojan wrote:
> FWIW, virtual keycode means something to windows developers I think, but it
means nothing to me. I'm not sure what's "virtual" about it.
> 
> Also, being more future-looking, the new standard key events in DOM have "key"
and "char" that are the actual character being pressed. Seems more developer
friendly than codes.

You need to use codes for some keys because not all of the keys have characters.
 For example, what is the character for F4?

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
File sky/engine/core/frame/NewEventHandler.cpp (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
sky/engine/core/frame/NewEventHandler.cpp:152: // Do we really need to suppress
keypress events for these keys anymore?
On 2015/01/26 at 06:18:51, ojan wrote:
> I think probably we don't need to. Although, really, I think we probably want
to get rid of keypress events entirely and just have keydown/keyup.

Yeah.  I wasn't sure how much of the nuttiness to remove in this CL.  I've tried
to keep the model basically the same but simplify the logic behind it.  Having
keydown/keyup + some sort of "text input" event make sense.

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
sky/engine/core/frame/NewEventHandler.cpp:176: if (!target)
On 2015/01/26 at 06:18:51, ojan wrote:
> Presumably this is the fix for https://github.com/domokit/mojo/issues/40?

Yeah.  I'm checking now.  If it is, I'll land separately.

https://codereview.chromium.org/872233002/diff/20001/sky/engine/web/WebViewIm...
File sky/engine/web/WebViewImpl.cpp (left):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/web/WebViewIm...
sky/engine/web/WebViewImpl.cpp:311: bool WebViewImpl::mapKeyCodeForScroll(
On 2015/01/26 at 06:18:51, ojan wrote:
> This is another place where if we were to plug in a mouse or keyboard, things
wouldn't work as we expect. It's easy to add back in later though. Although, it
will get harder and harder (e.g. well delete a bunch of the ScrollGranularity
code next since we'll find that it's dead code because we keep deleting all the
callers).

I'd imagine we'd want to do this work in the framework: translating unhandled
keydowns into scroll gestures.

[abarth-chromium, 2015-01-26 06:38:52]

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
File sky/engine/core/events/KeyboardEvent.h (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/events/K...
sky/engine/core/events/KeyboardEvent.h:45: unsigned virtualKeyCode() const {
return m_virtualKeyCode; }
On 2015/01/26 at 06:28:16, abarth wrote:
> On 2015/01/26 at 06:18:51, ojan wrote:
> > FWIW, virtual keycode means something to windows developers I think, but it
means nothing to me. I'm not sure what's "virtual" about it.
> > 
> > Also, being more future-looking, the new standard key events in DOM have
"key" and "char" that are the actual character being pressed. Seems more
developer friendly than codes.
> 
> You need to use codes for some keys because not all of the keys have
characters.  For example, what is the character for F4?

I renamed this to "key"

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
File sky/engine/core/frame/NewEventHandler.cpp (right):

https://codereview.chromium.org/872233002/diff/20001/sky/engine/core/frame/Ne...
sky/engine/core/frame/NewEventHandler.cpp:176: if (!target)
On 2015/01/26 at 06:28:16, abarth wrote:
> On 2015/01/26 at 06:18:51, ojan wrote:
> > Presumably this is the fix for https://github.com/domokit/mojo/issues/40?
> 
> Yeah.  I'm checking now.  If it is, I'll land separately.

I wasn't able to reproduce the crash...

[abarth-chromium, 2015-01-26 06:46:55]

Committed patchset #3 (id:40001) manually as
f5143561cc7cce29082a60238922124e4d8c0b25 (presubmit successful).