Add Windows nativeKeyCode to USB keycode conversion table

[Chromoting] Add USB keycode handling to Windows client.

Update code that converts from native Windows scancodes into USB keycodes.
The code now uses a table that is shared across platforms.
Pepper uses this table when populating the |usb_key_code| field in the InputEvent.

This cl also bring the webkit/plugins/ppapi/usb_keycode_map.h file in sync with the one in remoting/host.
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=127157

[Wez, 2012-03-07 23:35:50]

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
File webkit/plugins/ppapi/usb_key_code_conversion_win.cc (right):

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:3: // found in the LICENSE
file.
We probably need a comment indicating that the table is derived from the RealVNC
version.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:17: uint8_t
win_scancode_to_usb[128] = {
This table should be laid out in blocks of e.g. 4 codes to a line, to make it
easier to read.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:17: uint8_t
win_scancode_to_usb[128] = {
Your table has 128 entries, but the scancode is an 8-bit value, not including
the extended bit.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:22: // 02-0b: 1!  2@  3#  4$
 5%  6^  7&  8*  9(  0)
Please put commas between the key names, to make this easier to read.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:155: //    repeat-count: The
number of times the key is auto-repeated.
This is more detail than we need here; we only care about the extended bit and
the scan-code, which are actually contiguous, allowing you to treat them as a
single 9-bit value.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:162: if (extended) {
Merge the two tables into a single 512-entry table to simplify this.

http://codereview.chromium.org/9619014/diff/1/webkit/plugins/ppapi/usb_key_co...
webkit/plugins/ppapi/usb_key_code_conversion_win.cc:172: uint32_t usbkeycode =
0x070000 | usb_id;
Why are you hard-coding all keys to the keyboard page now?  The original table
mapped some keys to the multimedia page.

[garykac, 2012-03-15 22:14:39]

CL modified to use newly-added cross-platform table.

+tony for webkit/glue

[garykac, 2012-03-15 22:15:30]

On 2012/03/15 22:14:39, garykac wrote:
> CL modified to use newly-added cross-platform table.
> 
> +tony for webkit/glue

+tony for real this time

[tony, 2012-03-15 22:32:21]

http://codereview.chromium.org/9619014/diff/5001/webkit/plugins/ppapi/usb_key...
File webkit/plugins/ppapi/usb_key_code_conversion.cc (right):

http://codereview.chromium.org/9619014/diff/5001/webkit/plugins/ppapi/usb_key...
webkit/plugins/ppapi/usb_key_code_conversion.cc:14: #if !defined(OS_LINUX) &&
!defined(OS_MACOSX) && !defined(OS_WIN)
Can we just delete this file?  What's left?

http://codereview.chromium.org/9619014/diff/5001/webkit/plugins/ppapi/usb_key...
File webkit/plugins/ppapi/usb_keycode_map.h (right):

http://codereview.chromium.org/9619014/diff/5001/webkit/plugins/ppapi/usb_key...
webkit/plugins/ppapi/usb_keycode_map.h:8: // Changes to this file must be made
in both locations.
Should you be updating the file in remoting/host as well?

[garykac, 2012-03-15 22:53:51]

On 2012/03/15 22:32:21, tony wrote:
>
http://codereview.chromium.org/9619014/diff/5001/webkit/plugins/ppapi/usb_key...
> webkit/plugins/ppapi/usb_key_code_conversion.cc:14: #if !defined(OS_LINUX) &&
> !defined(OS_MACOSX) && !defined(OS_WIN)
> Can we just delete this file?  What's left?

According to build_config.h, at least OS_ANDROID.

I have no idea if other OSs listed there (OS_FREEBSD, OS_OPENBSD, OS_SOLARIS)
are currently used, but I didn't want to break them.

> Should you be updating the file in remoting/host as well?

There was already a cl in flight to update the Win part of the table before the
webkit/plugins/ppapi one was added. This cl brings them in sync.

[tony, 2012-03-15 23:22:31]

LGTM!