Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2010 The Android Open Source Project
|
|
|
|
*
|
|
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
* you may not use this file except in compliance with the License.
|
|
|
|
* You may obtain a copy of the License at
|
|
|
|
*
|
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
*
|
|
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
* See the License for the specific language governing permissions and
|
|
|
|
* limitations under the License.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef _UI_INPUT_H
|
|
|
|
#define _UI_INPUT_H
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Native input event structures.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <android/input.h>
|
|
|
|
#include <utils/Vector.h>
|
2010-07-24 04:28:06 +00:00
|
|
|
#include <utils/KeyedVector.h>
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
#include <utils/Timers.h>
|
2010-07-24 04:28:06 +00:00
|
|
|
#include <utils/RefBase.h>
|
|
|
|
#include <utils/String8.h>
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Additional private constants not defined in ndk/ui/input.h.
|
|
|
|
*/
|
|
|
|
enum {
|
|
|
|
/*
|
|
|
|
* Private control to determine when an app is tracking a key sequence.
|
|
|
|
*/
|
2010-07-15 01:48:53 +00:00
|
|
|
AKEY_EVENT_FLAG_START_TRACKING = 0x40000000
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Maximum number of pointers supported per motion event.
|
|
|
|
*/
|
|
|
|
#define MAX_POINTERS 10
|
|
|
|
|
2010-06-19 01:09:33 +00:00
|
|
|
/*
|
|
|
|
* Declare a concrete type for the NDK's input event forward declaration.
|
|
|
|
*/
|
2010-07-14 00:48:30 +00:00
|
|
|
struct AInputEvent {
|
|
|
|
virtual ~AInputEvent() { }
|
|
|
|
};
|
2010-06-19 01:09:33 +00:00
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
/*
|
2010-07-24 04:28:06 +00:00
|
|
|
* Declare a concrete type for the NDK's input device forward declaration.
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
*/
|
2010-07-24 04:28:06 +00:00
|
|
|
struct AInputDevice {
|
|
|
|
virtual ~AInputDevice() { }
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
};
|
|
|
|
|
2010-07-24 04:28:06 +00:00
|
|
|
|
|
|
|
namespace android {
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
/*
|
|
|
|
* Flags that flow alongside events in the input dispatch system to help with certain
|
|
|
|
* policy decisions such as waking from device sleep.
|
|
|
|
*/
|
|
|
|
enum {
|
|
|
|
/* These flags originate in RawEvents and are generally set in the key map. */
|
|
|
|
|
|
|
|
POLICY_FLAG_WAKE = 0x00000001,
|
|
|
|
POLICY_FLAG_WAKE_DROPPED = 0x00000002,
|
|
|
|
POLICY_FLAG_SHIFT = 0x00000004,
|
|
|
|
POLICY_FLAG_CAPS_LOCK = 0x00000008,
|
|
|
|
POLICY_FLAG_ALT = 0x00000010,
|
|
|
|
POLICY_FLAG_ALT_GR = 0x00000020,
|
|
|
|
POLICY_FLAG_MENU = 0x00000040,
|
|
|
|
POLICY_FLAG_LAUNCHER = 0x00000080,
|
|
|
|
|
2010-06-18 03:52:56 +00:00
|
|
|
POLICY_FLAG_RAW_MASK = 0x0000ffff,
|
|
|
|
|
2010-06-15 08:31:58 +00:00
|
|
|
/* These flags are set by the input reader policy as it intercepts each event. */
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
// Indicates that the screen was off when the event was received and the event
|
|
|
|
// should wake the device.
|
|
|
|
POLICY_FLAG_WOKE_HERE = 0x10000000,
|
|
|
|
|
|
|
|
// Indicates that the screen was dim when the event was received and the event
|
|
|
|
// should brighten the device.
|
|
|
|
POLICY_FLAG_BRIGHT_HERE = 0x20000000,
|
|
|
|
};
|
|
|
|
|
2010-06-15 08:31:58 +00:00
|
|
|
/*
|
|
|
|
* Describes the basic configuration of input devices that are present.
|
|
|
|
*/
|
|
|
|
struct InputConfiguration {
|
|
|
|
enum {
|
|
|
|
TOUCHSCREEN_UNDEFINED = 0,
|
|
|
|
TOUCHSCREEN_NOTOUCH = 1,
|
|
|
|
TOUCHSCREEN_STYLUS = 2,
|
|
|
|
TOUCHSCREEN_FINGER = 3
|
|
|
|
};
|
|
|
|
|
|
|
|
enum {
|
|
|
|
KEYBOARD_UNDEFINED = 0,
|
|
|
|
KEYBOARD_NOKEYS = 1,
|
|
|
|
KEYBOARD_QWERTY = 2,
|
|
|
|
KEYBOARD_12KEY = 3
|
|
|
|
};
|
|
|
|
|
|
|
|
enum {
|
|
|
|
NAVIGATION_UNDEFINED = 0,
|
|
|
|
NAVIGATION_NONAV = 1,
|
|
|
|
NAVIGATION_DPAD = 2,
|
|
|
|
NAVIGATION_TRACKBALL = 3,
|
|
|
|
NAVIGATION_WHEEL = 4
|
|
|
|
};
|
|
|
|
|
|
|
|
int32_t touchScreen;
|
|
|
|
int32_t keyboard;
|
|
|
|
int32_t navigation;
|
|
|
|
};
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
/*
|
|
|
|
* Pointer coordinate data.
|
|
|
|
*/
|
|
|
|
struct PointerCoords {
|
|
|
|
float x;
|
|
|
|
float y;
|
|
|
|
float pressure;
|
|
|
|
float size;
|
2010-07-15 01:48:53 +00:00
|
|
|
float touchMajor;
|
|
|
|
float touchMinor;
|
|
|
|
float toolMajor;
|
|
|
|
float toolMinor;
|
|
|
|
float orientation;
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Input events.
|
|
|
|
*/
|
2010-06-28 22:27:30 +00:00
|
|
|
class InputEvent : public AInputEvent {
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
public:
|
|
|
|
virtual ~InputEvent() { }
|
|
|
|
|
|
|
|
virtual int32_t getType() const = 0;
|
|
|
|
|
|
|
|
inline int32_t getDeviceId() const { return mDeviceId; }
|
|
|
|
|
2010-07-15 01:48:53 +00:00
|
|
|
inline int32_t getSource() const { return mSource; }
|
2010-06-30 02:20:40 +00:00
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
protected:
|
2010-07-15 01:48:53 +00:00
|
|
|
void initialize(int32_t deviceId, int32_t source);
|
2010-07-16 00:44:53 +00:00
|
|
|
void initialize(const InputEvent& from);
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
private:
|
|
|
|
int32_t mDeviceId;
|
2010-07-15 01:48:53 +00:00
|
|
|
int32_t mSource;
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
};
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
/*
|
|
|
|
* Key events.
|
|
|
|
*/
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
class KeyEvent : public InputEvent {
|
|
|
|
public:
|
|
|
|
virtual ~KeyEvent() { }
|
|
|
|
|
2010-07-15 01:48:53 +00:00
|
|
|
virtual int32_t getType() const { return AINPUT_EVENT_TYPE_KEY; }
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
inline int32_t getAction() const { return mAction; }
|
|
|
|
|
|
|
|
inline int32_t getFlags() const { return mFlags; }
|
|
|
|
|
|
|
|
inline int32_t getKeyCode() const { return mKeyCode; }
|
|
|
|
|
|
|
|
inline int32_t getScanCode() const { return mScanCode; }
|
|
|
|
|
|
|
|
inline int32_t getMetaState() const { return mMetaState; }
|
|
|
|
|
|
|
|
inline int32_t getRepeatCount() const { return mRepeatCount; }
|
|
|
|
|
|
|
|
inline nsecs_t getDownTime() const { return mDownTime; }
|
|
|
|
|
|
|
|
inline nsecs_t getEventTime() const { return mEventTime; }
|
|
|
|
|
2010-06-30 02:20:40 +00:00
|
|
|
// Return true if this event may have a default action implementation.
|
|
|
|
static bool hasDefaultAction(int32_t keyCode);
|
|
|
|
bool hasDefaultAction() const;
|
|
|
|
|
|
|
|
// Return true if this event represents a system key.
|
|
|
|
static bool isSystemKey(int32_t keyCode);
|
|
|
|
bool isSystemKey() const;
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
void initialize(
|
|
|
|
int32_t deviceId,
|
2010-07-15 01:48:53 +00:00
|
|
|
int32_t source,
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
int32_t action,
|
|
|
|
int32_t flags,
|
|
|
|
int32_t keyCode,
|
|
|
|
int32_t scanCode,
|
|
|
|
int32_t metaState,
|
|
|
|
int32_t repeatCount,
|
|
|
|
nsecs_t downTime,
|
|
|
|
nsecs_t eventTime);
|
2010-07-16 00:44:53 +00:00
|
|
|
void initialize(const KeyEvent& from);
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
private:
|
|
|
|
int32_t mAction;
|
|
|
|
int32_t mFlags;
|
|
|
|
int32_t mKeyCode;
|
|
|
|
int32_t mScanCode;
|
|
|
|
int32_t mMetaState;
|
|
|
|
int32_t mRepeatCount;
|
|
|
|
nsecs_t mDownTime;
|
|
|
|
nsecs_t mEventTime;
|
|
|
|
};
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
/*
|
|
|
|
* Motion events.
|
|
|
|
*/
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
class MotionEvent : public InputEvent {
|
|
|
|
public:
|
|
|
|
virtual ~MotionEvent() { }
|
|
|
|
|
2010-07-15 01:48:53 +00:00
|
|
|
virtual int32_t getType() const { return AINPUT_EVENT_TYPE_MOTION; }
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
inline int32_t getAction() const { return mAction; }
|
|
|
|
|
|
|
|
inline int32_t getEdgeFlags() const { return mEdgeFlags; }
|
|
|
|
|
|
|
|
inline int32_t getMetaState() const { return mMetaState; }
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getXOffset() const { return mXOffset; }
|
|
|
|
|
|
|
|
inline float getYOffset() const { return mYOffset; }
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
inline float getXPrecision() const { return mXPrecision; }
|
|
|
|
|
|
|
|
inline float getYPrecision() const { return mYPrecision; }
|
|
|
|
|
|
|
|
inline nsecs_t getDownTime() const { return mDownTime; }
|
|
|
|
|
|
|
|
inline size_t getPointerCount() const { return mPointerIds.size(); }
|
|
|
|
|
|
|
|
inline int32_t getPointerId(size_t pointerIndex) const { return mPointerIds[pointerIndex]; }
|
|
|
|
|
|
|
|
inline nsecs_t getEventTime() const { return mSampleEventTimes[getHistorySize()]; }
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getRawX(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).x;
|
|
|
|
}
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getRawY(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).y;
|
|
|
|
}
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
inline float getX(size_t pointerIndex) const {
|
2010-06-16 08:53:36 +00:00
|
|
|
return getRawX(pointerIndex) + mXOffset;
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
inline float getY(size_t pointerIndex) const {
|
2010-06-16 08:53:36 +00:00
|
|
|
return getRawY(pointerIndex) + mYOffset;
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
inline float getPressure(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).pressure;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getSize(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).size;
|
|
|
|
}
|
|
|
|
|
2010-07-15 01:48:53 +00:00
|
|
|
inline float getTouchMajor(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).touchMajor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getTouchMinor(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).touchMinor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getToolMajor(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).toolMajor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getToolMinor(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).toolMinor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getOrientation(size_t pointerIndex) const {
|
|
|
|
return getCurrentPointerCoords(pointerIndex).orientation;
|
|
|
|
}
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
inline size_t getHistorySize() const { return mSampleEventTimes.size() - 1; }
|
|
|
|
|
|
|
|
inline nsecs_t getHistoricalEventTime(size_t historicalIndex) const {
|
|
|
|
return mSampleEventTimes[historicalIndex];
|
|
|
|
}
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getHistoricalRawX(size_t pointerIndex, size_t historicalIndex) const {
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).x;
|
|
|
|
}
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getHistoricalRawY(size_t pointerIndex, size_t historicalIndex) const {
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).y;
|
|
|
|
}
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
inline float getHistoricalX(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalRawX(pointerIndex, historicalIndex) + mXOffset;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalY(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalRawY(pointerIndex, historicalIndex) + mYOffset;
|
|
|
|
}
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
inline float getHistoricalPressure(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).pressure;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalSize(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).size;
|
|
|
|
}
|
|
|
|
|
2010-07-15 01:48:53 +00:00
|
|
|
inline float getHistoricalTouchMajor(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).touchMajor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalTouchMinor(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).touchMinor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalToolMajor(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).toolMajor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalToolMinor(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).toolMinor;
|
|
|
|
}
|
|
|
|
|
|
|
|
inline float getHistoricalOrientation(size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return getHistoricalPointerCoords(pointerIndex, historicalIndex).orientation;
|
|
|
|
}
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
void initialize(
|
|
|
|
int32_t deviceId,
|
2010-07-15 01:48:53 +00:00
|
|
|
int32_t source,
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
int32_t action,
|
|
|
|
int32_t edgeFlags,
|
|
|
|
int32_t metaState,
|
2010-06-16 08:53:36 +00:00
|
|
|
float xOffset,
|
|
|
|
float yOffset,
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
float xPrecision,
|
|
|
|
float yPrecision,
|
|
|
|
nsecs_t downTime,
|
|
|
|
nsecs_t eventTime,
|
|
|
|
size_t pointerCount,
|
|
|
|
const int32_t* pointerIds,
|
|
|
|
const PointerCoords* pointerCoords);
|
|
|
|
|
|
|
|
void addSample(
|
|
|
|
nsecs_t eventTime,
|
|
|
|
const PointerCoords* pointerCoords);
|
|
|
|
|
|
|
|
void offsetLocation(float xOffset, float yOffset);
|
|
|
|
|
2010-06-16 08:53:36 +00:00
|
|
|
// Low-level accessors.
|
|
|
|
inline const int32_t* getPointerIds() const { return mPointerIds.array(); }
|
|
|
|
inline const nsecs_t* getSampleEventTimes() const { return mSampleEventTimes.array(); }
|
|
|
|
inline const PointerCoords* getSamplePointerCoords() const {
|
|
|
|
return mSamplePointerCoords.array();
|
|
|
|
}
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
private:
|
|
|
|
int32_t mAction;
|
|
|
|
int32_t mEdgeFlags;
|
|
|
|
int32_t mMetaState;
|
2010-06-16 08:53:36 +00:00
|
|
|
float mXOffset;
|
|
|
|
float mYOffset;
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
float mXPrecision;
|
|
|
|
float mYPrecision;
|
|
|
|
nsecs_t mDownTime;
|
|
|
|
Vector<int32_t> mPointerIds;
|
|
|
|
Vector<nsecs_t> mSampleEventTimes;
|
|
|
|
Vector<PointerCoords> mSamplePointerCoords;
|
|
|
|
|
|
|
|
inline const PointerCoords& getCurrentPointerCoords(size_t pointerIndex) const {
|
|
|
|
return mSamplePointerCoords[getHistorySize() * getPointerCount() + pointerIndex];
|
|
|
|
}
|
|
|
|
|
|
|
|
inline const PointerCoords& getHistoricalPointerCoords(
|
|
|
|
size_t pointerIndex, size_t historicalIndex) const {
|
|
|
|
return mSamplePointerCoords[historicalIndex * getPointerCount() + pointerIndex];
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Input event factory.
|
|
|
|
*/
|
|
|
|
class InputEventFactoryInterface {
|
|
|
|
protected:
|
|
|
|
virtual ~InputEventFactoryInterface() { }
|
|
|
|
|
|
|
|
public:
|
|
|
|
InputEventFactoryInterface() { }
|
|
|
|
|
|
|
|
virtual KeyEvent* createKeyEvent() = 0;
|
|
|
|
virtual MotionEvent* createMotionEvent() = 0;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A simple input event factory implementation that uses a single preallocated instance
|
|
|
|
* of each type of input event that are reused for each request.
|
|
|
|
*/
|
|
|
|
class PreallocatedInputEventFactory : public InputEventFactoryInterface {
|
|
|
|
public:
|
|
|
|
PreallocatedInputEventFactory() { }
|
|
|
|
virtual ~PreallocatedInputEventFactory() { }
|
|
|
|
|
|
|
|
virtual KeyEvent* createKeyEvent() { return & mKeyEvent; }
|
|
|
|
virtual MotionEvent* createMotionEvent() { return & mMotionEvent; }
|
|
|
|
|
|
|
|
private:
|
|
|
|
KeyEvent mKeyEvent;
|
|
|
|
MotionEvent mMotionEvent;
|
|
|
|
};
|
|
|
|
|
2010-07-24 04:28:06 +00:00
|
|
|
/*
|
|
|
|
* Describes the characteristics and capabilities of an input device.
|
|
|
|
*/
|
|
|
|
class InputDeviceInfo {
|
|
|
|
public:
|
|
|
|
InputDeviceInfo();
|
|
|
|
InputDeviceInfo(const InputDeviceInfo& other);
|
|
|
|
~InputDeviceInfo();
|
|
|
|
|
|
|
|
struct MotionRange {
|
|
|
|
float min;
|
|
|
|
float max;
|
|
|
|
float flat;
|
|
|
|
float fuzz;
|
|
|
|
};
|
|
|
|
|
|
|
|
void initialize(int32_t id, const String8& name);
|
|
|
|
|
|
|
|
inline int32_t getId() const { return mId; }
|
|
|
|
inline const String8 getName() const { return mName; }
|
|
|
|
inline uint32_t getSources() const { return mSources; }
|
|
|
|
|
|
|
|
const MotionRange* getMotionRange(int32_t rangeType) const;
|
|
|
|
|
|
|
|
void addSource(uint32_t source);
|
|
|
|
void addMotionRange(int32_t rangeType, float min, float max, float flat, float fuzz);
|
|
|
|
void addMotionRange(int32_t rangeType, const MotionRange& range);
|
|
|
|
|
|
|
|
inline void setKeyboardType(int32_t keyboardType) { mKeyboardType = keyboardType; }
|
|
|
|
inline int32_t getKeyboardType() const { return mKeyboardType; }
|
|
|
|
|
2010-08-30 10:02:23 +00:00
|
|
|
inline const KeyedVector<int32_t, MotionRange> getMotionRanges() const {
|
|
|
|
return mMotionRanges;
|
|
|
|
}
|
|
|
|
|
2010-07-24 04:28:06 +00:00
|
|
|
private:
|
|
|
|
int32_t mId;
|
|
|
|
String8 mName;
|
|
|
|
uint32_t mSources;
|
|
|
|
int32_t mKeyboardType;
|
|
|
|
|
|
|
|
KeyedVector<int32_t, MotionRange> mMotionRanges;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Provides remote access to information about an input device.
|
|
|
|
*
|
|
|
|
* Note: This is essentially a wrapper for Binder calls into the Window Manager Service.
|
|
|
|
*/
|
|
|
|
class InputDeviceProxy : public RefBase, public AInputDevice {
|
|
|
|
protected:
|
|
|
|
InputDeviceProxy();
|
|
|
|
virtual ~InputDeviceProxy();
|
|
|
|
|
|
|
|
public:
|
|
|
|
static void getDeviceIds(Vector<int32_t>& outIds);
|
|
|
|
|
|
|
|
static sp<InputDeviceProxy> getDevice(int32_t id);
|
|
|
|
|
|
|
|
inline const InputDeviceInfo* getInfo() { return & mInfo; }
|
|
|
|
|
|
|
|
// TODO add hasKeys, keymap, etc...
|
|
|
|
|
|
|
|
private:
|
|
|
|
InputDeviceInfo mInfo;
|
|
|
|
};
|
|
|
|
|
Native input dispatch rewrite work in progress.
The old dispatch mechanism has been left in place and continues to
be used by default for now. To enable native input dispatch,
edit the ENABLE_NATIVE_DISPATCH constant in WindowManagerPolicy.
Includes part of the new input event NDK API. Some details TBD.
To wire up input dispatch, as the ViewRoot adds a window to the
window session it receives an InputChannel object as an output
argument. The InputChannel encapsulates the file descriptors for a
shared memory region and two pipe end-points. The ViewRoot then
provides the InputChannel to the InputQueue. Behind the
scenes, InputQueue simply attaches handlers to the native PollLoop object
that underlies the MessageQueue. This way MessageQueue doesn't need
to know anything about input dispatch per-se, it just exposes (in native
code) a PollLoop that other components can use to monitor file descriptor
state changes.
There can be zero or more targets for any given input event. Each
input target is specified by its input channel and some parameters
including flags, an X/Y coordinate offset, and the dispatch timeout.
An input target can request either synchronous dispatch (for foreground apps)
or asynchronous dispatch (fire-and-forget for wallpapers and "outside"
targets). Currently, finding the appropriate input targets for an event
requires a call back into the WindowManagerServer from native code.
In the future this will be refactored to avoid most of these callbacks
except as required to handle pending focus transitions.
End-to-end event dispatch mostly works!
To do: event injection, rate limiting, ANRs, testing, optimization, etc.
Change-Id: I8c36b2b9e0a2d27392040ecda0f51b636456de25
2010-04-23 01:58:52 +00:00
|
|
|
|
|
|
|
} // namespace android
|
|
|
|
|
|
|
|
#endif // _UI_INPUT_H
|