gecko-dev/dom/animation/Animation.h

381 строка
15 KiB
C
Исходник Обычный вид История

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#ifndef mozilla_dom_Animation_h
#define mozilla_dom_Animation_h
#include "nsWrapperCache.h"
#include "nsCycleCollectionParticipant.h"
#include "mozilla/Attributes.h"
#include "mozilla/TimeStamp.h" // for TimeStamp, TimeDuration
#include "mozilla/dom/AnimationBinding.h" // for AnimationPlayState
#include "mozilla/dom/DocumentTimeline.h" // for DocumentTimeline
#include "mozilla/dom/KeyframeEffect.h" // for KeyframeEffectReadOnly
#include "mozilla/dom/Promise.h" // for Promise
#include "nsCSSProperty.h" // for nsCSSProperty
// X11 has a #define for CurrentTime.
#ifdef CurrentTime
#undef CurrentTime
#endif
// GetCurrentTime is defined in winbase.h as zero argument macro forwarding to
// GetTickCount().
#ifdef GetCurrentTime
#undef GetCurrentTime
#endif
struct JSContext;
class nsCSSPropertySet;
class nsIDocument;
class nsPresContext;
namespace mozilla {
struct AnimationCollection;
namespace css {
class AnimValuesStyleRule;
class CommonAnimationManager;
} // namespace css
namespace dom {
class CSSAnimation;
class CSSTransition;
class Animation
: public nsISupports
, public nsWrapperCache
{
protected:
virtual ~Animation() {}
public:
explicit Animation(DocumentTimeline* aTimeline)
: mTimeline(aTimeline)
, mPlaybackRate(1.0)
, mPendingState(PendingState::NotPending)
Bug 1171817 part 5 - Add a sequence number member to Animations; r=dbaron Web Animations defines Animations as having a globally unique sequence number for the purpose of prioritization: http://w3c.github.io/web-animations/#animation-sequence-number As of the writing of this patch, the spec says the sequence number is updated when the Animation is created. This is problematic and I have proposed that actually this should be updated from each transition from idle: https://lists.w3.org/Archives/Public/public-fx/2015AprJun/0054.html This doesn't seem to have met any opposition so I will update the spec to reflect this soon. This patch implements the behavior of updating the sequence number on each transition from idle. To make sure we perform this on each change to timing this patch removes a couple of instances of early returns to ensure that UpdateTiming is called. The current maximum sequence number is simply a class static and we make no attempt to deal with wraparound. This is because we only update this number when an animation transitions from idle which only happens when an animation is created or script calls cancel() followed by play() on the animation. Supposing that across all content this happenned an unlikely 1 billion times a second we still wouldn't exhaust the range of the unsigned 64-bit int for about 585 years. We'd like to make kUnsequenced be zero and make the static represent the current maximum. This would probably be easier to understand and recognize in a debugger. However, later in this patch series we will make CSS animations and CSS transitions override this sequencing behavior. If we define kUnsequenced to be zero and they accidentally assign zero as an actual sequence number then they'll run into trouble. To avoid that we set kUnsequenced to UINT64_MAX. --HG-- extra : commitid : DMw8uKjg4Hz extra : rebase_source : 9e98b3346f0297efce3ecfa0b2dd8a9c13075dca
2015-06-09 05:13:53 +03:00
, mSequenceNum(kUnsequenced)
, mIsRunningOnCompositor(false)
Bug 1078122 part 6 - Store the previous finished state; r=dholbert AnimationPlayer::CanThrottle determines if an animation player has just finished by inspecting the value of mLastNotification. This is problematic for two reasons: 1. mLastNotification is intended to be used for events (as the XXX comment notes) 2. mLastNotification is specific to CSS Animations and should be moved to CSSAnimationPlayer. To address this, this patch adds an extra member mIsPreviousStateFinished. The Web Animations spec already defines animation players as having such a member: http://w3c.github.io/web-animations/#previous-finished-state We set it to true when we calculate the style for an animation that has finished. This differs slightly from the code it is replacing as explained below. In the case of CSS Animations we perform the following sequence of steps on each sample. 1. EnsureStyleRuleFor (calls CanThrottle, and maybe ComposeStyle) 2. GetEventsForCurrentTime In the existing code, we update mLastNotification in (2) which happens on every sample, even throttled samples. In this patch, however, we update mIsPreviousStateFinished in (1) during the ComposeStyle step which only happens for unthrottled samples. So, as of this patch, in CanThrottle, we ask "have we newly entered the finished state since the last *unthrottled* sample?", whereas previously we simply looked for a change since the last sample, throttled or not. However, if the answer to the question is "yes", then we'll run an unthrottled sample and update mIsPreviousStateFinished so these should be functionally equivalent. Another subtle difference is that this patch looks at the player's finished state rather than the animation phase of its source content, and these will produce different results in the case where the player is paused. However, since paused animations are not run on the compositor, this should not matter. In the case of CSS Transitions, AnimationPlayer::CanThrottle() is not currently used and so mIsPreviousStateFinished is irrelevant. Ultimately, both the existing and the new code is somewhat fragile but hopefully this will be addressed by: * Replacing mIsPreviousStateFinished with inspecting whether the finished promise is settled (bug 1074630), * Merging more of the code in nsAnimationManager and nsTransitionManager and applying a unified approach to sampling that better accommodates these considerations.
2014-10-20 08:55:47 +04:00
, mIsPreviousStateFinished(false)
, mFinishedAtLastComposeStyle(false)
, mIsRelevant(false)
{
}
NS_DECL_CYCLE_COLLECTING_ISUPPORTS
NS_DECL_CYCLE_COLLECTION_SCRIPT_HOLDER_CLASS(Animation)
DocumentTimeline* GetParentObject() const { return mTimeline; }
virtual JSObject* WrapObject(JSContext* aCx,
JS::Handle<JSObject*> aGivenProto) override;
virtual CSSAnimation* AsCSSAnimation() { return nullptr; }
virtual const CSSAnimation* AsCSSAnimation() const { return nullptr; }
virtual CSSTransition* AsCSSTransition() { return nullptr; }
virtual const CSSTransition* AsCSSTransition() const { return nullptr; }
/**
* Flag to pass to Play to indicate whether or not it should automatically
* rewind the current time to the start point if the animation is finished.
* For regular calls to play() from script we should do this, but when a CSS
* animation's animation-play-state changes we shouldn't rewind the animation.
*/
enum class LimitBehavior {
AutoRewind,
Continue
};
// Animation interface methods
KeyframeEffectReadOnly* GetEffect() const { return mEffect; }
void SetEffect(KeyframeEffectReadOnly* aEffect);
DocumentTimeline* Timeline() const { return mTimeline; }
Nullable<TimeDuration> GetStartTime() const { return mStartTime; }
void SetStartTime(const Nullable<TimeDuration>& aNewStartTime);
Nullable<TimeDuration> GetCurrentTime() const;
void SetCurrentTime(const TimeDuration& aNewCurrentTime);
double PlaybackRate() const { return mPlaybackRate; }
void SetPlaybackRate(double aPlaybackRate);
AnimationPlayState PlayState() const;
virtual Promise* GetReady(ErrorResult& aRv);
virtual Promise* GetFinished(ErrorResult& aRv);
void Cancel();
virtual void Finish(ErrorResult& aRv);
virtual void Play(ErrorResult& aRv, LimitBehavior aLimitBehavior);
virtual void Pause(ErrorResult& aRv);
bool IsRunningOnCompositor() const { return mIsRunningOnCompositor; }
// Wrapper functions for Animation DOM methods when called
// from script.
//
// We often use the same methods internally and from script but when called
// from script we (or one of our subclasses) perform extra steps such as
// flushing style or converting the return type.
Nullable<double> GetStartTimeAsDouble() const;
void SetStartTimeAsDouble(const Nullable<double>& aStartTime);
Nullable<double> GetCurrentTimeAsDouble() const;
void SetCurrentTimeAsDouble(const Nullable<double>& aCurrentTime,
ErrorResult& aRv);
virtual AnimationPlayState PlayStateFromJS() const { return PlayState(); }
virtual void PlayFromJS(ErrorResult& aRv)
{
Play(aRv, LimitBehavior::AutoRewind);
}
/**
* PauseFromJS is currently only here for symmetry with PlayFromJS but
* in future we will likely have to flush style in
* CSSAnimation::PauseFromJS so we leave it for now.
*/
void PauseFromJS(ErrorResult& aRv) { Pause(aRv); }
Bug 1171817 part 2 - Add CSSAnimation::GetOwningElement; r=dbaron In order to sort CSS animation objects correctly, we need to know which element's animation-name property they appear in, if any. Normally that's simply the target element of the animation's keyframe effect but it can differ in the following cases: 1) When script modifies a CSSAnimation's effect to target a different element (or simply removes the effect altogether). In this case we use the *owning* element to determine the priority of the animation, not the target element. This scenario does not yet occur (bug 1049975). 2) When script creates a CSSAnimation object using the CSSAnimation constructor. In this case, the owning element should be empty (null) and we should determine the priority of the animation in the same way as any other Animation object. Again, this is not yet supported (or even specced) but will be eventually. 3) When script holds a reference to a CSSAnimation object but then updates the animation-name property such that the animation object is cancelled. In this case the owning element should be cleared (null) so we know to not to try and sort this with regard to any animation-name property. This is possible using code such as the following: elem.style.animation = 'a 5s'; var a = elem.getAnimations()[0]; elem.style.animation = 'b 5s'; a.play(); // Bring a back to life document.timeline.getAnimations(); // ^ At this point we need to know how to sort 'a' and 'b' which depends // on recognizing that a is no longer part of an animation-name list. Until we implement bug 1049975, we could support sorting animations without adding the reference to the owning element by setting a flag on the CSSAnimation object but (having tried this) it turns out to be cleaner to just introduce this reference now, particularly since we know we will need it later. Note that we will also need this information in future to dispatch events to the correct element in circumstances such as (1) once we separate updating timing information (including events) from applying animation values. --HG-- extra : commitid : 8o9bf6l7kj7 extra : rebase_source : 391a4e8769cc96584ebd625d4b1d0e873373fd41
2015-06-09 05:13:53 +03:00
// Wrapper functions for Animation DOM methods when called from style.
Bug 1171817 part 2 - Add CSSAnimation::GetOwningElement; r=dbaron In order to sort CSS animation objects correctly, we need to know which element's animation-name property they appear in, if any. Normally that's simply the target element of the animation's keyframe effect but it can differ in the following cases: 1) When script modifies a CSSAnimation's effect to target a different element (or simply removes the effect altogether). In this case we use the *owning* element to determine the priority of the animation, not the target element. This scenario does not yet occur (bug 1049975). 2) When script creates a CSSAnimation object using the CSSAnimation constructor. In this case, the owning element should be empty (null) and we should determine the priority of the animation in the same way as any other Animation object. Again, this is not yet supported (or even specced) but will be eventually. 3) When script holds a reference to a CSSAnimation object but then updates the animation-name property such that the animation object is cancelled. In this case the owning element should be cleared (null) so we know to not to try and sort this with regard to any animation-name property. This is possible using code such as the following: elem.style.animation = 'a 5s'; var a = elem.getAnimations()[0]; elem.style.animation = 'b 5s'; a.play(); // Bring a back to life document.timeline.getAnimations(); // ^ At this point we need to know how to sort 'a' and 'b' which depends // on recognizing that a is no longer part of an animation-name list. Until we implement bug 1049975, we could support sorting animations without adding the reference to the owning element by setting a flag on the CSSAnimation object but (having tried this) it turns out to be cleaner to just introduce this reference now, particularly since we know we will need it later. Note that we will also need this information in future to dispatch events to the correct element in circumstances such as (1) once we separate updating timing information (including events) from applying animation values. --HG-- extra : commitid : 8o9bf6l7kj7 extra : rebase_source : 391a4e8769cc96584ebd625d4b1d0e873373fd41
2015-06-09 05:13:53 +03:00
virtual void CancelFromStyle() { DoCancel(); }
void Tick();
/**
* Set the time to use for starting or pausing a pending animation.
*
* Typically, when an animation is played, it does not start immediately but
* is added to a table of pending animations on the document of its effect.
* In the meantime it sets its hold time to the time from which playback
* should begin.
*
* When the document finishes painting, any pending animations in its table
* are marked as being ready to start by calling StartOnNextTick.
* The moment when the paint completed is also recorded, converted to a
* timeline time, and passed to StartOnTick. This is so that when these
* animations do start, they can be timed from the point when painting
* completed.
*
* After calling TriggerOnNextTick, animations remain in the pending state
* until the next refresh driver tick. At that time they transition out of
* the pending state using the time passed to TriggerOnNextTick as the
* effective time at which they resumed.
*
* This approach means that any setup time required for performing the
* initial paint of an animation such as layerization is not deducted from
* the running time of the animation. Without this we can easily drop the
* first few frames of an animation, or, on slower devices, the whole
* animation.
*
* Furthermore:
*
* - Starting the animation immediately when painting finishes is problematic
* because the start time of the animation will be ahead of its timeline
* (since the timeline time is based on the refresh driver time).
* That's a problem because the animation is playing but its timing
* suggests it starts in the future. We could update the timeline to match
* the start time of the animation but then we'd also have to update the
* timing and style of all animations connected to that timeline or else be
* stuck in an inconsistent state until the next refresh driver tick.
*
* - If we simply use the refresh driver time on its next tick, the lag
* between triggering an animation and its effective start is unacceptably
* long.
*
* For pausing, we apply the same asynchronous approach. This is so that we
* synchronize with animations that are running on the compositor. Otherwise
* if the main thread lags behind the compositor there will be a noticeable
* jump backwards when the main thread takes over. Even though main thread
* animations could be paused immediately, we do it asynchronously for
* consistency and so that animations paused together end up in step.
*
* Note that the caller of this method is responsible for removing the
* animation from any PendingAnimationTracker it may have been added to.
*/
void TriggerOnNextTick(const Nullable<TimeDuration>& aReadyTime);
/**
* Testing only: Start or pause a pending animation using the current
* timeline time. This is used to support existing tests that expect
* animations to begin immediately. Ideally we would rewrite the those tests
* and get rid of this method, but there are a lot of them.
*
* As with TriggerOnNextTick, the caller of this method is responsible for
* removing the animation from any PendingAnimationTracker it may have been
* added to.
*/
void TriggerNow();
/**
* When StartOnNextTick is called, we store the ready time but we don't apply
* it until the next tick. In the meantime, GetStartTime() will return null.
*
* However, if we build layer animations again before the next tick, we
* should initialize them with the start time that GetStartTime() will return
* on the next tick.
*
* If we were to simply set the start time of layer animations to null, their
* start time would be updated to the current wallclock time when rendering
* finishes, thus making them out of sync with the start time stored here.
* This, in turn, will make the animation jump backwards when we build
* animations on the next tick and apply the start time stored here.
*
* This method returns the start time, if resolved. Otherwise, if we have
* a pending ready time, it returns the corresponding start time. If neither
* of those are available, it returns null.
*/
Nullable<TimeDuration> GetCurrentOrPendingStartTime() const;
bool IsPausedOrPausing() const
{
return PlayState() == AnimationPlayState::Paused ||
mPendingState == PendingState::PausePending;
}
bool HasInPlayEffect() const
{
return GetEffect() && GetEffect()->IsInPlay(*this);
}
bool HasCurrentEffect() const
{
return GetEffect() && GetEffect()->IsCurrent(*this);
}
bool IsInEffect() const
{
return GetEffect() && GetEffect()->IsInEffect();
}
/**
* "Playing" is different to "running". An animation in its delay phase is
* still running but we only consider it playing when it is in its active
* interval. This definition is used for fetching the animations that are
* candidates for running on the compositor (since we don't ship animations
* to the compositor when they are in their delay phase or paused).
*/
bool IsPlaying() const
{
// We need to have an effect in its active interval, and
// be either running or waiting to run.
return HasInPlayEffect() &&
(PlayState() == AnimationPlayState::Running ||
mPendingState == PendingState::PlayPending);
}
bool IsRelevant() const { return mIsRelevant; }
void UpdateRelevance();
void SetIsRunningOnCompositor() { mIsRunningOnCompositor = true; }
void ClearIsRunningOnCompositor() { mIsRunningOnCompositor = false; }
/**
* Returns true if this animation does not currently need to update
* style on the main thread (e.g. because it is empty, or is
* running on the compositor).
*/
bool CanThrottle() const;
/**
* Updates |aStyleRule| with the animation values of this animation's effect,
* if any.
* Any properties already contained in |aSetProperties| are not changed. Any
* properties that are changed are added to |aSetProperties|.
* |aNeedsRefreshes| will be set to true if this animation expects to update
* the style rule on the next refresh driver tick as well (because it
* is running and has an effect to sample).
*/
void ComposeStyle(nsRefPtr<css::AnimValuesStyleRule>& aStyleRule,
nsCSSPropertySet& aSetProperties,
bool& aNeedsRefreshes);
protected:
void SilentlySetCurrentTime(const TimeDuration& aNewCurrentTime);
void SilentlySetPlaybackRate(double aPlaybackRate);
void DoCancel();
void DoPlay(ErrorResult& aRv, LimitBehavior aLimitBehavior);
void DoPause(ErrorResult& aRv);
void ResumeAt(const TimeDuration& aReadyTime);
void PauseAt(const TimeDuration& aReadyTime);
void FinishPendingAt(const TimeDuration& aReadyTime)
{
if (mPendingState == PendingState::PlayPending) {
ResumeAt(aReadyTime);
} else if (mPendingState == PendingState::PausePending) {
PauseAt(aReadyTime);
} else {
NS_NOTREACHED("Can't finish pending if we're not in a pending state");
}
}
/**
* Finishing behavior depends on if changes to timing occurred due
* to a seek or regular playback.
*/
enum class SeekFlag {
NoSeek,
DidSeek
};
void UpdateTiming(SeekFlag aSeekFlag);
void UpdateFinishedState(SeekFlag aSeekFlag);
void UpdateEffect();
void FlushStyle() const;
void PostUpdate();
/**
* Remove this animation from the pending animation tracker and reset
* mPendingState as necessary. The caller is responsible for resolving or
* aborting the mReady promise as necessary.
*/
void CancelPendingTasks();
bool IsPossiblyOrphanedPendingAnimation() const;
StickyTimeDuration EffectEnd() const;
nsIDocument* GetRenderedDocument() const;
nsPresContext* GetPresContext() const;
virtual css::CommonAnimationManager* GetAnimationManager() const = 0;
AnimationCollection* GetCollection() const;
nsRefPtr<DocumentTimeline> mTimeline;
nsRefPtr<KeyframeEffectReadOnly> mEffect;
// The beginning of the delay period.
Nullable<TimeDuration> mStartTime; // Timeline timescale
Nullable<TimeDuration> mHoldTime; // Animation timescale
Nullable<TimeDuration> mPendingReadyTime; // Timeline timescale
Nullable<TimeDuration> mPreviousCurrentTime; // Animation timescale
double mPlaybackRate;
// A Promise that is replaced on each call to Play()
// and fulfilled when Play() is successfully completed.
// This object is lazily created by GetReady.
// See http://w3c.github.io/web-animations/#current-ready-promise
nsRefPtr<Promise> mReady;
// A Promise that is resolved when we reach the end of the effect, or
// 0 when playing backwards. The Promise is replaced if the animation is
// finished but then a state change makes it not finished.
// This object is lazily created by GetFinished.
// See http://w3c.github.io/web-animations/#current-finished-promise
nsRefPtr<Promise> mFinished;
// Indicates if the animation is in the pending state (and what state it is
// waiting to enter when it finished pending). We use this rather than
// checking if this animation is tracked by a PendingAnimationTracker because
// the animation will continue to be pending even after it has been removed
// from the PendingAnimationTracker while it is waiting for the next tick
// (see TriggerOnNextTick for details).
enum class PendingState { NotPending, PlayPending, PausePending };
PendingState mPendingState;
Bug 1171817 part 5 - Add a sequence number member to Animations; r=dbaron Web Animations defines Animations as having a globally unique sequence number for the purpose of prioritization: http://w3c.github.io/web-animations/#animation-sequence-number As of the writing of this patch, the spec says the sequence number is updated when the Animation is created. This is problematic and I have proposed that actually this should be updated from each transition from idle: https://lists.w3.org/Archives/Public/public-fx/2015AprJun/0054.html This doesn't seem to have met any opposition so I will update the spec to reflect this soon. This patch implements the behavior of updating the sequence number on each transition from idle. To make sure we perform this on each change to timing this patch removes a couple of instances of early returns to ensure that UpdateTiming is called. The current maximum sequence number is simply a class static and we make no attempt to deal with wraparound. This is because we only update this number when an animation transitions from idle which only happens when an animation is created or script calls cancel() followed by play() on the animation. Supposing that across all content this happenned an unlikely 1 billion times a second we still wouldn't exhaust the range of the unsigned 64-bit int for about 585 years. We'd like to make kUnsequenced be zero and make the static represent the current maximum. This would probably be easier to understand and recognize in a debugger. However, later in this patch series we will make CSS animations and CSS transitions override this sequencing behavior. If we define kUnsequenced to be zero and they accidentally assign zero as an actual sequence number then they'll run into trouble. To avoid that we set kUnsequenced to UINT64_MAX. --HG-- extra : commitid : DMw8uKjg4Hz extra : rebase_source : 9e98b3346f0297efce3ecfa0b2dd8a9c13075dca
2015-06-09 05:13:53 +03:00
static uint64_t sNextSequenceNum;
static const uint64_t kUnsequenced = UINT64_MAX;
// The sequence number assigned to this animation. This is kUnsequenced
// while the animation is in the idle state and is updated each time
// the animation transitions out of the idle state.
uint64_t mSequenceNum;
bool mIsRunningOnCompositor;
Bug 1078122 part 6 - Store the previous finished state; r=dholbert AnimationPlayer::CanThrottle determines if an animation player has just finished by inspecting the value of mLastNotification. This is problematic for two reasons: 1. mLastNotification is intended to be used for events (as the XXX comment notes) 2. mLastNotification is specific to CSS Animations and should be moved to CSSAnimationPlayer. To address this, this patch adds an extra member mIsPreviousStateFinished. The Web Animations spec already defines animation players as having such a member: http://w3c.github.io/web-animations/#previous-finished-state We set it to true when we calculate the style for an animation that has finished. This differs slightly from the code it is replacing as explained below. In the case of CSS Animations we perform the following sequence of steps on each sample. 1. EnsureStyleRuleFor (calls CanThrottle, and maybe ComposeStyle) 2. GetEventsForCurrentTime In the existing code, we update mLastNotification in (2) which happens on every sample, even throttled samples. In this patch, however, we update mIsPreviousStateFinished in (1) during the ComposeStyle step which only happens for unthrottled samples. So, as of this patch, in CanThrottle, we ask "have we newly entered the finished state since the last *unthrottled* sample?", whereas previously we simply looked for a change since the last sample, throttled or not. However, if the answer to the question is "yes", then we'll run an unthrottled sample and update mIsPreviousStateFinished so these should be functionally equivalent. Another subtle difference is that this patch looks at the player's finished state rather than the animation phase of its source content, and these will produce different results in the case where the player is paused. However, since paused animations are not run on the compositor, this should not matter. In the case of CSS Transitions, AnimationPlayer::CanThrottle() is not currently used and so mIsPreviousStateFinished is irrelevant. Ultimately, both the existing and the new code is somewhat fragile but hopefully this will be addressed by: * Replacing mIsPreviousStateFinished with inspecting whether the finished promise is settled (bug 1074630), * Merging more of the code in nsAnimationManager and nsTransitionManager and applying a unified approach to sampling that better accommodates these considerations.
2014-10-20 08:55:47 +04:00
// Indicates whether we were in the finished state during our
// most recent unthrottled sample (our last ComposeStyle call).
bool mIsPreviousStateFinished; // Spec calls this "previous finished state"
bool mFinishedAtLastComposeStyle;
// Indicates that the animation should be exposed in an element's
// getAnimations() list.
bool mIsRelevant;
};
} // namespace dom
} // namespace mozilla
#endif // mozilla_dom_Animation_h