Import GTK WebInputEvent factory to content

Import GTK WebInputEvent factory to content
(Migrated from crrev.com/21133002)

This imports the GTK WebInputEvent factory into content as
a set of WebInputEventBuilders. The functionality should be exacly the
same except for the handling of the following keycodes which changed due
to moving from the WebCore GDK -> windows keycode mapping to ui/base:

0xD7 (XK_multiply) is mapped to 0xA6 (VKEY_MULTIPLY) instead of 0
(VKEY_UKNOWN).
0xA6 (XK_brokenbar), 0xAB (XK_guillemotleft), 0xB0 (XK_degree),
0xBB (XK_guillemotright), 0xD9 (XK_Ugrave) and 0xF9 (XK_ugrave)
map to 0xE2 (VKEY_OEM_102) instead of 0 (VKEY_UNKNOWN).

I think these are progressions.

BUG=263189
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=223533

[garykac, 2013-09-09 15:29:46]

(Taking over crrev.com/2113302 from jamesr)

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:89: //
Handle the AltGr (ISO Level3 Shift) key as Control + Alt.
nona: I added AltGr support here. I couldn't find docs that say GDK_MOD5_MASK is
AltGr, but determined it empirically.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:445: case
GDK_2BUTTON_PRESS:
nona: double- and triple-clicks are handled below (in result.clickCount), rather
than relying on these extra GDK events.

[Seigo Nonaka, 2013-09-10 04:14:32]

lgtm

lgtm with ntis.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:37: gint
double_click_time = 250;
nit: How about adding "const"?

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:89: //
Handle the AltGr (ISO Level3 Shift) key as Control + Alt.
On 2013/09/09 15:29:46, garykac wrote:
> nona: I added AltGr support here. I couldn't find docs that say GDK_MOD5_MASK
is
> AltGr, but determined it empirically.
Yes, Mod5Mask is for AltGr.
Thank you.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:445: case
GDK_2BUTTON_PRESS:
On 2013/09/09 15:29:46, garykac wrote:
> nona: double- and triple-clicks are handled below (in result.clickCount),
rather
> than relying on these extra GDK events.

Sure, make sense.

[garykac, 2013-09-10 18:54:35]

jam, can you OWNER review?

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:37: gint
double_click_time = 250;
On 2013/09/10 04:14:32, Seigo Nonaka wrote:
> nit: How about adding "const"?

This is passed by reference to g_object_get() below where the value can be
overwritten.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:89: //
Handle the AltGr (ISO Level3 Shift) key as Control + Alt.
On 2013/09/10 04:14:32, Seigo Nonaka wrote:
> On 2013/09/09 15:29:46, garykac wrote:
> > nona: I added AltGr support here. I couldn't find docs that say
GDK_MOD5_MASK
> is
> > AltGr, but determined it empirically.
> Yes, Mod5Mask is for AltGr.
> Thank you.

Thanks for the confirmation.

[jam, 2013-09-11 02:50:00]

On 2013/09/10 18:54:35, garykac wrote:
> jam, can you OWNER review?

lgtm for files outside content/browser/renderer_host/input. since that directory
has more specific owners, please get them to review these files as I'm not
familiar with them.

[garykac, 2013-09-11 19:59:24]

+nduca as OWNER for content/browser/renderer_host/input

[garykac, 2013-09-14 17:19:05]

ping for OWNER review

(cc'ing aelias, jdduke as alternate OWNERS)

[jdduke (slow), 2013-09-16 15:31:55]

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:25: static
GdkWindow* lack_click_event_window = 0;
nit: last_click_event_window

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:50: void
ResetClickCountState() {
I realize this is a direct port from WebKit, but I wonder how necessary the
global state tracking is? Does it have to be process-wide, or would it be
sufficient to have some state/context stored per-RWHVGTK?

Anyhow, if that is appropriate (it may not be), it should probably happen in a
separate patch as this is more or less just moving code; perhaps we could file a
bug?

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_gtk.cc (right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_gtk.cc:74: const float
kDefaultScrollPixelsPerTick = 160.0f / 3.0f;
Could we store this constant in a single location? Perhaps a static method on
WebMouseWheelEventBuilder?

[garykac, 2013-09-16 22:45:05]

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:25: static
GdkWindow* lack_click_event_window = 0;
On 2013/09/16 15:31:55, jdduke wrote:
> nit: last_click_event_window

Done.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:50: void
ResetClickCountState() {
On 2013/09/16 15:31:55, jdduke wrote:
> I realize this is a direct port from WebKit, but I wonder how necessary the
> global state tracking is? Does it have to be process-wide, or would it be
> sufficient to have some state/context stored per-RWHVGTK?
> 
> Anyhow, if that is appropriate (it may not be), it should probably happen in a
> separate patch as this is more or less just moving code; perhaps we could file
a
> bug?

I added crbug.com/292733 to track this since I don't know the answer to it
offhand.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_gtk.cc (right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_gtk.cc:74: const float
kDefaultScrollPixelsPerTick = 160.0f / 3.0f;
On 2013/09/16 15:31:55, jdduke wrote:
> Could we store this constant in a single location? Perhaps a static method on
> WebMouseWheelEventBuilder?

WebMouseWheelEventBuilder seems like a reasonable place, but it's in Blink, so
it would need to be a separate cl.
Bug filed to track: crbug.com/292920

[jdduke (slow), 2013-09-16 23:46:57]

lgtm

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/input/web_input_event_builders_gtk.cc
(right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/input/web_input_event_builders_gtk.cc:579: static
const float scrollbarPixelsPerTick = 160.0f / 3.0f;
Here's the other use I was thinking of.

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_gtk.cc (right):

https://codereview.chromium.org/23815005/diff/3001/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_gtk.cc:74: const float
kDefaultScrollPixelsPerTick = 160.0f / 3.0f;
I see, I was thinking more so of the use here, and in WebMouseWheelEventBuilder
(web_input_event_builders_gtk.cpp:579, static const float scrollbarPixelsPerTick
= 160.0f / 3.0f;).  As both lines fall in this patch, could we consolidate just
the two for now, and leave the Blink/remoting duplications for another day?

[commit-bot: I haz the power, 2013-09-17 01:25:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/garykac@chromium.org/23815005/36001

[commit-bot: I haz the power, 2013-09-17 04:06:05]

Change committed as 223533