[PATCH v2 xserver] xkb: fix releasing overlay while keydown
walter harms
wharms at bfs.de
Sat Nov 5 21:06:10 UTC 2016
Am 05.11.2016 20:03, schrieb Mihail Konev:
> The overlay_keys array could be avoided.
>
> Testcase:
>
> // In the exact order:
> // - press Insert
> // - press Tilde
> // - release Insert
> // - wait
> // - release Tilde
> //
> // Keyboard input would be locked.
>
> xkb_keymap {
> xkb_keycodes { include "evdev" };
> xkb_types { include "complete" };
>
> xkb_compat { include "complete"
> interpret Overlay1_Enable+AnyOfOrNone(all) {
> action= SetControls(controls=Overlay1);
> };
> };
>
> xkb_symbols { include "pc+inet(evdev)+us"
> key <INS> { [ Overlay1_Enable ] };
> key <AE01> { overlay1 = <AE02> }; // Insert+1 => 2
> key <TLDE> { overlay1 = <I128> }; // Insert+~ => XF86LaunchA
> };
>
> xkb_geometry { include "pc(pc104)" };
> };
>
> Signed-off-by: Mihail Konev <k.mvc at ya.ru>
> ---
>
> v2: remove debug print
>
> xkb/xkbPrKeyEv.c | 49 ++++++++++++++++++++++++++++++++++++++-----------
> 1 file changed, 38 insertions(+), 11 deletions(-)
>
> diff --git a/xkb/xkbPrKeyEv.c b/xkb/xkbPrKeyEv.c
> index f7a6b4b14306..92a10bdc6fd2 100644
> --- a/xkb/xkbPrKeyEv.c
> +++ b/xkb/xkbPrKeyEv.c
> @@ -43,6 +43,9 @@ THE USE OR PERFORMANCE OF THIS SOFTWARE.
>
> /***====================================================================***/
>
> +static unsigned char overlay_in_effect[256];
> +static unsigned int overlay_keys[2][256];
> +
> void
> XkbProcessKeyboardEvent(DeviceEvent *event, DeviceIntPtr keybd)
> {
> @@ -51,6 +54,7 @@ XkbProcessKeyboardEvent(DeviceEvent *event, DeviceIntPtr keybd)
> int key;
> XkbBehavior behavior;
> unsigned ndx;
> + unsigned effective_overlay = 0;
>
> xkbi = keyc->xkbInfo;
> key = event->detail.key;
> @@ -68,6 +72,16 @@ XkbProcessKeyboardEvent(DeviceEvent *event, DeviceIntPtr keybd)
> /* do anything to implement the behavior, but it *does* report that */
> /* key is hardwired */
>
> + {
> + if (key >= 0 && key < 256) {
> + effective_overlay = overlay_in_effect[key];
> + switch (effective_overlay) {
> + case 1: behavior.type = XkbKB_Overlay1; break;
> + case 2: behavior.type = XkbKB_Overlay1; break;
this looks strange if fall truh is intended then do it.
> + }
> + }
> + }
> +
> if (!(behavior.type & XkbKB_Permanent)) {
> switch (behavior.type) {
> case XkbKB_Default:
> @@ -121,20 +135,33 @@ XkbProcessKeyboardEvent(DeviceEvent *event, DeviceIntPtr keybd)
> case XkbKB_Overlay2:
> {
> unsigned which;
> + unsigned which_overlay;
> + unsigned overlay_enabled;
> + unsigned behavior_data = behavior.data;
>
> - if (behavior.type == XkbKB_Overlay1)
> + if (behavior.type == XkbKB_Overlay1) {
> which = XkbOverlay1Mask;
> - else
> + which_overlay = 1;
> + } else {
> which = XkbOverlay2Mask;
> - if ((xkbi->desc->ctrls->enabled_ctrls & which) == 0)
> - break;
> - if ((behavior.data >= xkbi->desc->min_key_code) &&
> - (behavior.data <= xkbi->desc->max_key_code)) {
> - event->detail.key = behavior.data;
> - /* 9/11/94 (ef) -- XXX! need to match release with */
> - /* press even if the state of the */
> - /* corresponding overlay control */
> - /* changes while the key is down */
> + which_overlay = 2;
> + }
> +
> + overlay_enabled = (xkbi->desc->ctrls->enabled_ctrls & which);
> +
> + if (!overlay_enabled && event->type == ET_KeyRelease)
> + overlay_in_effect[(unsigned char)key] = 0;
> + else
> + overlay_in_effect[(unsigned char)key] = which_overlay;
> +
> + if (overlay_enabled)
> + overlay_keys[which_overlay][(unsigned char)key] = behavior.data;
> + else
> + behavior_data = overlay_keys[which_overlay][(unsigned char)key];
> +
> + if ((behavior_data >= xkbi->desc->min_key_code) &&
> + (behavior_data <= xkbi->desc->max_key_code)) {
> + event->detail.key = behavior_data;
> }
> }
the whole block is huge, is it possible to make it a function ?
just my 2 cents,
re,
wh
> break;
More information about the xorg-devel
mailing list