There are a fair amount of Signal objects created. In the current
implementation these are somewhat "fat" objects for two reasons:
- A closure / ad-hoc `dispatch` is created for each new Signal - this
increases the retained size by 138+ bytes/Signal in Chrome.
- There are a number of instance variables that never change from their
default value, and the bindings array is always created, even if never
used.
This change "lightens" the Signals such that there is significantly less
penalty for having many (unusued) signals.
As an example of how much this _does_ play a role, in the "Random Sprite"
demo ~2000 Signals are created, with only 12 of these signals being
subscribed to. This results in a shallow size of ~300K and a retained size
of ~600K (which is almost as much memory as required by the
Sprites/Sprites themselves!) for .. nothing.
With these changes the shallow and retained sizes are less than 50K each -
and this is with only ~200 sprites!
This change addresses these issues by
- Changing it so there is _no_ `dispatch` closure created. This is a
_breaking change_ (although there is no usage of such in core where it
breaks); the code referenced "#24", but no such issue could be found on
github.
In the rare case that code needs to obtain a dispatch-closure, the
`boundDispatch` property can be used to trivially obtain a (cached)
closure.
- The properties and default values are moved into the prototype; and the
`_bindings` array creation is deferred. This change, coupled with the
removal of the automatic closure, results in a very lightweight
~24bytes/object (in Chrome) for unbound signals.
This is minor breaking change, as per the removal of the automatic
closure; but no such need/usage was found in the core. (There may be cases
in the examples, which were not checked.)
If merely opting for the array creation delay and keeping the default
properties in the prototype the shallow size is still halved; less
significant but still a consideration.
This de-optimization occurred between 2.0.7 and 2.1.0 and is currently
present through dev.
`Group.forEach`, which is used by QuadTree, had an extreme de-optimization
in assigning to `arguments` - _CPU profiling showed as much as 50% of the
time was used by Group.forEach_ (after the correction it is not
registered) due to this de-optimization making the "When Particles
Collide" demo run with an unsatisfactory performance, even on a Desktop.
The fix uses a separate array and push (which is optimizable; the previous
implementation was not optimizable in Chrome, FF, or IE!).
This also fixes usages of `slice(arguments,..); ushift` elsewhere in
Group, using the same convention. It applies the same update for `iterate`
as does https://github.com/photonstorm/phaser/pull/1353 so it can also
accept null/undefined for `args` from the invoking functions.
- Cleaned up abnormal case handling
- Disabled parallel loading by default; it can be enabled with
`enableParallel`.
- Minor quibbles
- Removed unused `dataLoadError`
Loaded/loading assets are skipped if they have been superceded. This makes
the behavior consistent with respect to `addToFileList`, if the queue is
inspected or modified while loading.
- Added `withSyncPoint` and `addSyncPoint` methods to allow explicit
adding of synchronization points (synchronization points are explained
in `withSyncPoint`.
- Changed the file/asset `sync` attribute to `syncPoint` to reflect
terminology.
Update iterate documentation to cover usage of `args` and added a guard so
that the callback can be used without requiring that `args` is specified.
Ref. https://github.com/photonstorm/phaser/issues/1352
Employee the use of image.complete and a width/height check to detect when
it is available "from cache" and skip having to run though the
onload/onerror cases.
Parallel loading is now supported, configured by
`loader.concurrentRequestCount`. Each file (a pack is now considered a
type of file) asset can be marked with `sync`, which is done for both pack
files and script files.
When a `sync` asset is encountered it must be loaded before any previous
resource any following resource is loaded (but it doesn't have to wait for
previous resources). Pack files are an exception in that they can download
(but are not processed) and they can fetch-around other `sync` assets.
Because of the concurrent nature there is no guarantee of the order in
which the individual events will file in relation to eachother, but local
ordering (e.g. onFileError always before onFileComplete) and overall
ordering (e.g. onLoadStart .. onFile? .. onLoadComplete) is preserved.
There is also increased error hardening and a few previous edge-cases
fixed (and likely a few bugs added).
Keyboard.justReleased has bee renamed to Keyboard.upDuration which is a much clearer name for what the method actually does.
Keyboard.downDuration, Keyboard.upDuration and Keyboard.isDown now all return `null` if the Key wasn't found in the local keys array.
Calling justPressed or justReleased on Phaser.Keyboard throws an exception. Changed to reflect new method names in Phaser.Key
I imagine you'd want these methods renamed as well, but it appears to be called by a few other classes and I didn't want a huge pull-request.
When specifying the ease in `Tween.to` or `Tween.from` you can now use a string instead of the Function. This makes your code less verbose. For example instead of `Phaser.Easing.Sinusoidal.Out` and you can now just use the string "Sine".The string names match those used by TweenMax and includes: "Linear", "Quad", "Cubic", "Quart", "Quint", "Sine", "Expo", "Circ", "Elastic", "Back", "Bounce", "Power0", "Power1", "Power2", "Power3" and "Power4". You can append ".easeIn", ".easeOut" and "easeInOut" variants. All are supported for each ease types.
Tweens now create a TweenData object. The Tween object itself acts like more of a timeline, managing multiple TweenData objects. You can now call `Tween.to` and each call will create a new child tween that is added to the timeline, which are played through in sequence.
Tweens are now bound to the new Time.desiredFps value and update based on the new Game core loop, rather than being bound to time calculations. This means that tweens are now running with the same update logic as physics and the core loop.
Tween.timeScale allows you to scale the duration of a tween (and any child tweens it may have). A value of 1.0 means it should play at the desiredFps rate. A value of 0.5 will run at half the frame rate, 2 at double and so on. You can even tween the timeScale value for interesting effects!
Tween.reverse allows you to instantly reverse an active tween. If the Tween has children then it will smoothly reverse through all child tweens as well.
Tween.repeatAll allows you to control how many times all child tweens will repeat before firing the Tween.onComplete event. You can set the value to -1 to repeat forever.
Tween.loop now controls the looping of all child tweens.
Tween.onRepeat is a new signal that is dispatched whenever a Tween repeats. If a Tween has many child tweens its dispatched once the sequence has repeated.
Tween.onChildComplete is a new signal that is dispatched whenever any child tweens have completed. If a Tween consists of 4 sections you will get 3 onChildComplete events followed by 1 onComplete event as the final tween finishes.
Chained tweens are now more intelligently handled. Because you can easily create child tweens (by simply calling Tween.to multiple times) chained tweens are now used to kick-off longer sequences. You can pass as many Tween objects to `Tween.chain` as you like as they'll all be played in sequence. As one Tween completes it passes on to the next until the entire chain is finished.
Tween.stop has a new `complete` parameter that if set will still fire the onComplete event and start the next chained tween, if there is one.
Tween.delay, Tween.repeat, Tween.yoyo, Tween.easing and Tween.interpolation all have a new `index` parameter. This allows you to target specific child tweens, or if set to -1 it will update all children at once.
Tween.totalDuration reports the total duration of all child tweens in ms.
There are new easing aliases:
* Phaser.Easing.Power0 = Phaser.Easing.Linear.None
* Phaser.Easing.Power1 = Phaser.Easing.Quadratic.Out
* Phaser.Easing.Power2 = Phaser.Easing.Cubic.Out
* Phaser.Easing.Power3 = Phaser.Easing.Quartic.Out
* Phaser.Easing.Power4 = Phaser.Easing.Quintic.Out
ScaleManager.windowContraints now allows specifing 'visual' or 'layout' as
the constraint. Using the 'layout' constraint should prevent a mobile
device from trying to resize the game when zooming.
Including the the new changes the defaults have been changed to
windowContraints = { right: 'layout', bottom: '' }
This changes the current scaling behavior as seen in "Game Scaling" (as it
will only scale for the right edge) but also prevents such scaling from
going wonkers in some mobile environtments like the newer Android browser.
(Automatic scroll-to-top, albeit configurable, enabled for non-desktop by
default is not a fun situation here.)
To obtain the current semantics on a desktop the bottom should be changed
to 'layout'; although this will result in different behavior depending on
mobile device. To make the sizing also follow mobile zooming they should
be changed to 'visual'.
Also added temp Rectangle re-used for various internal calculations.
---
Phaser.DOM now also special-cases desktops to align the layout bounds
correctly (this may disagree with CSS breakpoints but it aligns the with
actual CSS width), without applying a window height/width expansion as
required on mobile browsers.
(And the jshint error isn't mine..)
