[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] Fixes key mapping so all keys work
From: |
Peter Maydell |
Subject: |
Re: [Qemu-devel] [PATCH] Fixes key mapping so all keys work |
Date: |
Wed, 14 Jan 2015 00:29:38 +0000 |
On 14 January 2015 at 00:12, Programmingkid <address@hidden> wrote:
> This patch allows for all the keys on an Apple extended keyboard to work in
> QEMU.
>
> Signed-off-by: John Arbuckle <address@hidden>
>
> ---
> ui/cocoa.m | 29 ++++++++++++++++++++---------
> 1 files changed, 20 insertions(+), 9 deletions(-)
>
> diff --git a/ui/cocoa.m b/ui/cocoa.m
> index c8535a3..afac987 100644
> --- a/ui/cocoa.m
> +++ b/ui/cocoa.m
> @@ -128,18 +128,26 @@ int keymap[] =
> 14, // 51 0x33 0x0e BKSP QZ_BACKSPACE
> 0, // 52 0x34 Undefined
> 1, // 53 0x35 0x01 ESC QZ_ESCAPE
> - 220, // 54 0x36 0xdc E0,5C R GUI QZ_RMETA
> - 219, // 55 0x37 0xdb E0,5B L GUI QZ_LMETA
> + 219,// 54 0x36 QZ_RMETA
> + 219,// 55 0x37 QZ_LMETA
This looks wrong: we shouldn't be mapping right meta and
left meta to the same thing, they're different keys.
> 42, // 56 0x38 0x2a L SHFT QZ_LSHIFT
> 58, // 57 0x39 0x3a CAPS QZ_CAPSLOCK
> 56, // 58 0x3A 0x38 L ALT QZ_LALT
> 29, // 59 0x3B 0x1d L CTRL QZ_LCTRL
> - 54, // 60 0x3C 0x36 R SHFT QZ_RSHIFT
> - 184,// 61 0x3D 0xb8 E0,38 R ALT QZ_RALT
> - 157,// 62 0x3E 0x9d E0,1D R CTRL QZ_RCTRL
> +
> + /*
> + Map right shift, control, option, and command keys to left counterpart
> + to improve compatibility with Mac OS.
> + */
> +
> + 42, // 60 0x3C 0x36 R SHFT QZ_RSHIFT
> + 56, // 61 0x3D 0xb8 E0,38 R ALT QZ_RALT
> + 29, // 62 0x3E 0x9d E0,1D R CTRL QZ_RCTRL
Compatibility with Mac OS as a host or as a guest?
These are different keys, and if we do this then
we won't be able to work correctly with guests that
really do treat them as different.
> +
> +
> 0, // 63 0x3F Undefined
> 0, // 64 0x40 Undefined
> - 0, // 65 0x41 Undefined
> + 0x53,// 65 0x41 Keypad .
Can we be consistent with the other entries, please? SDLi entry
in decimal, and appropriate items in the other columns including
whatever the QZ_ constant is if we can find it.
> 0, // 66 0x42 Undefined
> 55, // 67 0x43 0x37 KP * QZ_KP_MULTIPLY
> 0, // 68 0x44 Undefined
> @@ -150,12 +158,12 @@ int keymap[] =
> 0, // 73 0x49 Undefined
> 0, // 74 0x4A Undefined
> 181,// 75 0x4B 0xb5 E0,35 KP / QZ_KP_DIVIDE
> - 152,// 76 0x4C 0x9c E0,1C KP EN QZ_KP_ENTER
> + 0xe01c, //152,// 76 0x4C 0x9c E0,1C KP EN QZ_KP_ENTER
Only entry in the entire table that's not a one-byte value?
> 0, // 77 0x4D undefined
> 74, // 78 0x4E 0x4a KP - QZ_KP_MINUS
> 0, // 79 0x4F Undefined
> 0, // 80 0x50 Undefined
> - 0, // 81 0x51 QZ_KP_EQUALS
> + 13, // 81 0x51 QZ_KP_EQUALS
The keypad = key and the normal = key ought to have different
values here. Otherwise the guest can't tell them apart.
> 82, // 82 0x52 0x52 KP 0 QZ_KP0
> 79, // 83 0x53 0x4f KP 1 QZ_KP1
> 80, // 84 0x54 0x50 KP 2 QZ_KP2
> @@ -201,10 +209,13 @@ int keymap[] =
> 205,// 124 0x7C 0xcd e0,4D R ARROW QZ_RIGHT
> 208,// 125 0x7D 0xd0 E0,50 D ARROW QZ_DOWN
> 200,// 126 0x7E 0xc8 E0,48 U ARROW QZ_UP
> +
> /* completed according to
> http://www.libsdl.org/cgi/cvsweb.cgi/SDL12/src/video/quartz/SDL_QuartzKeys.h?rev=1.6&content-type=text/x-cvsweb-markup
> */
>
> -/* Additional 104 Key XP-Keyboard Scancodes from
> http://www.computer-engineering.org/ps2keyboard/scancodes1.html */
> +/* Aditional 104 Key XP-Keyboard Scancodes from
> http://www.computer-engineering.org/ps2keyboard/scancodes1.html */
This is reintroducing a typo that was fixed back in 2013
(in commit 49b9bd4dc).
> /*
> + 219 // 0xdb e0,5b L GUI
> + 220 // 0xdc e0,5c R GUI
> 221 // 0xdd e0,5d APPS
> // E0,2A,E0,37 PRNT SCRN
> // E1,1D,45,E1,9D,C5 PAUSE
-- PMM