- FIX#1306, hopefully, where an orientation change did not correclty
cause a screen/layout update.
- FIX/CHANGE where Paused games would not update the scale
- The new behavior "runs" the ScaleManager in a paused state via
`pauseUpdate`; a User paused game will now correctly track scale
changes. This is closer to the 2.1.3 behavior in some cases, such as
window resizing, when the updates were done in the DOM event.
- This change also affects device orientation change monitoring and
events, which are also deferred to the update cycle
- The update cycle is set to the maximum and is still dependent on the
RAF / primary loop running, so it should not affect background
apps/tabs
- FIX/CHANGE New better backoff timing; ie. continuous window resizing is
limited to ~10 fps update calculations. This makes it much harder to
crash Chrome by rapidly and continously resizing the window. Also
increases the scaling from 0..10..20..40 to 0..25..50..100.
- FIX an issue where the incorrect orientation was "one frame behind" the
scaling.
- UPDATE The contract for when the change orientation events occurs is
better defined - it now always happens in the update context as with
game sizing.
- UPDATE Unifies orientation-change code / handling and duplicate.
- CHANGE Added DOM.getScreenOrientation which obtains the orientation via
the Device Orientation API (WD) and provides comprehensive fallbacks
- This should cover all modern browsers
- FIX: Orientation on desktops now computed as screen ratio by default
which fixesi the false-portrait chain/detection when the page is made
more narrow than it is tall.
- CHANGE/FIX: window.orientation is now only used as fallback, if
requested (due to device differences). It may be appropriate to enable
this (via `scale.compatibility` on boot, for instance) in some
environments.
Signed-off-by: Paul <pstickne@gmail.com>
No known breaking changes - as it's still dev/internal stuff.
- Added Phaser.DOM to house new DOM functions, moved stuff
over from ScaleManager as appropriate
- Fixed a fiew cases of missing functions
- Changed some of the new signatures to protected for the interim.
(Maybe a `beta` tag would fit better? Public is promises!)
- Moved generic support from Canvas to DOM and added proxy/notes
- Updated internal usages
- Updated some comments for consistency
- Access always on bottom for members/properties, public assumed
There are no known breaking changes.
- Timer
- Uses standard Math.min/Math.max (it's better 2, 3 items).
- Math
- Updated documentation
- Marked various Math functions as deprecated, proxying as appropriate
- Array-based functions -> ArrayUtils
- RNG-based functions -> Utils
- Updated core-usage
- floor/ceil should not be used (alternatives provided)
- Altered for some equivalencies
- Also fixes some assorted issues
- Marked a few internal functions as private
- Utils
- Moved polyfills to their own file for better visibility
- Moved array functions to ArrayUtils and marked proxies as deprecated
- Created Phaser.ArrayUtils for array-related functions
- polyfills moved to their own file
- Functions given function names
- Added Math.trunc
- Added `compatibility` settings
- CHANGE (2.1.2-4): moved `supportsFullScreen` and `noMargins` into it
- Added additional properties for greater control and up-front settings.
- `scrollTo`: where the browser will scrollTo, if anywhere
- `forceMinimumDocumentHeight`: apply document element style?
- `allowShowAllExpand`: allow SHOW_ALL to try to expand? (It already
could, this allows configuration.)
- Removed `windowConstraints.top/left`. This may be a feature in the
future, but scrubbed for now.
- Added `USER_SCALE` scale mode. This is like NO_SCALE but it scales off
of a user-specified scale factor, as set by `setUserScale`. This is
marked as "experimental" as the exactly semantics of non-adjusting modes
(e.g. NO_SCALE and USER_SCALE) wrt. Canvas and "maximum" size clamps
need to be re-examined.
- FIX: `onSizeChange` now works as documented, which means it is also
fired if the game size changes even though the game canvas size does
not.
- CHANGE (no known breaking): `margins` is now non-Point/non-Rectangle
that uses top/left/bottom/right properties (any quasi-updated x/y). This
is to get around the issue that Rectangle is only designed for positive
width/height cases.
- Cleaned up property access / quotes for consistency
- Various documentation cleanup and consistency
- Fixed issue with not clearing an unparented `_createdFullScreenTarget`
- Added Phaser.Rectangle.sameDimensions which does a strict equality check
over the `width` and `height` properties of two objects, perhaps
Rectangles.
We have separated the logic and render updates to permit slow motion and time slicing effects. We've fixed time calling to fix physics problems caused by variable time updates (i.e. collisions sometimes missing, objects tunneling, etc)
Once per frame calling for rendering and tweening to keep things as smooth as possible
Calculates a `suggestedFps` value (in multiples of 5 fps) based on a 2 second average of actual elapsed time values in the `Time.update` method. This is recalculated every 2 seconds so it could be used on a level-by-level basis if a game varies dramatically. I.e. if the fps rate consistently drops, you can adjust your game effects accordingly.
Game loop now tries to "catch up" frames if it is falling behind by iterating the logic update. This will help if the logic is occasionally causing things to run too slow, or if the renderer occasionally pushes the combined frame time over the FPS time. It's not a band-aid for a game that floods a low powered device however, so you still need to code accordingly. But it should help capture issues such as gc spikes or temporarily overloaded CPUs.
It now detects 'spiralling' which happens if a lot of frames are pushed out in succession meaning the CPU can never "catch up". It skips frames instead of trying to catch them up in this case. Note: the time value passed to the logic update functions is always constant regardless of these shenanigans.
Signals to the game program if there is a problem which might be fixed by lowering the desiredFps
Time.desiredFps is the new desired frame rate for your game.
Time.suggestedFps is the suggested frame rate for the game based on system load.
Time.slowMotion allows you to push the game into a slow motion mode. The default value is 1.0. 2.0 would be half speed, and so on.
Time.timeCap is no longer used and now deprecated. All timing is now handled by the fixed time-step code we've introduced.
ScaleManager.calibrate is a private method that calibrates element coordinates for viewport checks.
ScaleManager.aspect gets the viewport aspect ratio (or the aspect ratio of an object or element)
ScaleManager.inViewport tests if the given DOM element is within the viewport, with an optional cushion parameter that allows you to specify a distance.
ScaleManager.scaleSprite takes a Sprite or Image object and scales it to fit the given dimensions. Scaling happens proportionally without distortion to the sprites texture. The letterBox parameter controls if scaling will produce a letter-box effect or zoom the sprite until it fills the given values.
ScaleManager.viewportWidth returns the viewport width in pixels.
ScaleManager.viewportHeight returns the viewport height in pixels.
ScaleManager.documentWidth returns the document width in pixels.
ScaleManager.documentHeight returns the document height in pixels.
- Adds `ScaleManager#windowContraints`
- In 2.1.3 and prior the scale modes (EXACT_FIT, SHOW_ALL) were actually
based off the window dimensions, even though the parent element did
not correctly reflect this nature.
- When set (the default now is that right and bottom are set) the
behavior will mostly correctly mimic the 2.1.3 (minus bugs) and
before.
- CHANGE (from 2.1.3): The window constraints also affect the RESIZE
mode, arguably this is more consistent.
- To disable this "constrain to window" behavior, simply set the
appropriate property to false, as in:
`game.scale.windowConstraints.bottom = false`
- Sizing events:
- CHANGE: The `onResize` callback is called only from `preUpdate` (which
may be triggered from a window resize) and it will be called on
refreshes even if the parent size has not actually changed.
- A new `onSizeChange` Signal has been added. It is called _only_ when
the Game size or Game canvas size has changed and is generally more
applicable for performing layout updates.
- Game documentation now links to ScaleManager#setGameSize (which was
renamed from #setGameDimensions)
- Removed extra/legacy full-screen restore code
- Margins:
- Added `noMargins` flag; if set to true the Canvas margins will never
be altered. This also means that
- Margins are now set/cleared individually to avoid conflict with
'margins' style compound property
- Code consistency updates
- NOTE: Changing `game.width/game.height` via user code was always
problematic. This commit updates the documentation for such members as
read-only. The only supported way to change the GAME SIZE after it is
created is to use `ScaleManager#setGameDimensions`, which has been
added.
- The GAME SIZE will be reset to the initial (or as set by
`setGameDimensions`) values upon changing the scale mode or
entering/leaving full screen. This may be a breaking from 2.1.2 (but
many mode changes acted oddly prior).
- SHOW_ALL will now EXPAND it's parent container if it can. As per
@tjkopena 's notes, this should more closely represented the expected
behavior.
- SHOW_ALL will first try to expand by the OVERFLOW AXIS and then
attempt to resize to fit into the possibly larger area; use the
parent's max-height/max-width properties to limit how far SHOW_ALL can
expand.
- RE-BREAKING: This changes the behavior from 2.1.4 and makes it more like
2.1.3, with fixes.
- As per previous commit the ScaleManager _owns_ the margins and size of
the GAME CANVAS. To control the dimensions of the GAME CANVAS, use the min/max
height/width of the parent. Setting padding on the parent is _NOT_
supported.
- Fixes various issues with switching between Scale Modes
This includes some minor breaking changes.
- Unifies SHOW_ALL and NO_SCALE being stretched in Firefox and IE
- As suggested by MDN: "..to emulate WebKit's behavior on Gecko, you
need to place the element you want to present inside another
element.."
- This done via an (overwritable) `createFullScreenTarget` function.
The (new) DOM element returned from this is placed into the DOM and
the canvas is added to (and later removed) as the full screen mode
changes.
- MINOR BREAK: may affect code that assumes the Phaser canvas has a
fixed DOM/CSS path (which should hopefully be nobody). To use to the
original behavior, where the canvas is not moved, simply set
`this.fullScreenTarget = game.canvas` manually.
- Updates the refresh/queue to be unified and uses a smarter back-off to
detect and react to parent dimension changes
- Cleans up some odd browser issues; not tried on mobile
- Fixes an issue were update might be called too much and spend time
doing nothing useful.
- `maxIterations` is no longer user and marked as deprecated
- MINOR BREAK: previous approach would occasionally (but not always)
back off updates the entire iteration/setTimeout sequence; under the
new approach "onResize" may be called more frequently.
- Fixes a number various transition issues, mostly around RESIZE
- MINOR BREAK, but correct: leaving RESIZE restores the original game
size possible
- Fixes assorted quirks with scales not being updated
- Layout
- MINOR BREAK: All Canvas margins are "OWNED" by the ScaleManager. They
will be reset in all modes as appropriate. This is for consistency
fixes as well as coping with the updated full screen.
- MINOR BREAK: Canvas right/bottom margins are set to negative margins
to counter left/top margins. This prevents Canvas margin adjustments
from affecting the flow .. much.
- `getParentBounds` rounds to the nearest pixel to avoid "close to"
value propagation from CSS.
- Fixes page-align center pushing canvas out of parent
- Misc.
- MINOR BREAK: `setScreenSize` will update the game size if the mode is
RESIZE. User-code shoulde use `refresh` instead to ensure that any
relevant changes are propagated.
- Corrected incorrect documentation
- Added prefixed event handlers for IE; it still doesn't work in IE, but that is for other reasons
- Added monitoring of the fullscreenerror event, exposed as a fullScreenFailed signal
- Added `supportsFullScreen` a read-only flag that indicates that this ScaleManager even attempts to support such
- Specific bug fixes (example)
- Scale modes can now be set independently
- Switching between fullscreen and normal correctly restores modes
- Alignment does not incorrectly offset in fullscreen mode
- Changing scale/alignment promptly refreshes layout
- `isFullScreen` returns a boolean, as it should
- Faster parent checks (if required)
- NO_SCALE should not not scale (vs previous behavior of having no behavior)
- Correct usage of scaleMode depending on mode
- Removed / unified code-paths, which helped address several issues
- fullScreenTarget adjustment/restoration is less brutal
- Updated documentation
- pageAlign* works as alignment on the containing element. It should still work were it worked before (so this is not a breaking change) as well as being more universal.
- Added @protected and @private attributes
- Some methods were (implicitly) @public even though using them out of context is invalid
- API Breaking:
- Renamed some INTERNAL/@private methods
- Should only affect methods that were not valid/sane to use publically
- Event callbacks (ie. checkResize changed to resizeWindow) renamed for meaning
- Changed some INTERNAL semantics to better align with usage
- eg. `check*` methods separated from response-to-check
- backgroundColor now uses valueToColor which supports hex strings, web/rgba strings, and hex numbers.
- valueToColor normalizes across hexToColor/webToColor/getRGB; see comments
- hexToColor now documents the prefix being optional and allows the `0x` prefix.
- webToColor now exctracts an alpha channel if present - as a slight misfeature it will also accept `rgb(..,a)`.
takes a predicate function and passes child, index, and the entire child array to it.
return an ArrayList containing all children that the predicate returns true for.
The width and height given to the Phaser.Game constructor can now be numbers or strings in which case the value is treated as a percentage. For example a value of "100%" for the width and height will tell Phaser to size the game to match the parent container dimensions exactly (or the browser window if no parent is given). Equally a size of "50%" would tell it to be half the size of the parent. The values are retained even through resize events, allowing it to maintain a percentage size based on the parent even as it updates.
Stage.offset has been moved to ScaleManager.offset
Stage.bounds has been removed, you can access it via Stage.getBounds.
Stage.checkOffsetInterval has been moved to ScaleManager.trackParentInterval
ScaleManager.hasResized signal has been removed. Use ScaleManager.setResizeCallback instead.
Moved the DOM offset from Stage to ScaleManager (a more logical location for it) and updated Pointer to use that.
Moved the Pointer offset check to look at the ScaleManager.
Used getBoundingClientRect(), will see if that works better than Phaser.Canvas.getOffset.
This PR depends on [#1150] (https://github.com/photonstorm/phaser/pull/1150)!
(I apologize for all the different pull requests in a short time period. I finally got some time to do some development tonight and have been making my way through many of my TODO items.)
This adds support for CocoonJS.App's 'onSuspended' and 'onActivated' events, making it so that the timers and sounds are stopped/started and muted/unmuted when the user swaps an app from the background to the fore or the reverse.
Because neither ['onActivated'] (http://doc.ludei.com/2.0.2/CocoonJS_App/symbols/CocoonJS.App.html#.event:onActivated) nor ['onSuspended'] (http://doc.ludei.com/2.0.2/CocoonJS_App/symbols/CocoonJS.App.html#.event:onSuspended) send an Event object themselves, this patch fakes sending an object by creating one during the function call and giving it a 'type' property for visibilityChange() to check against.
StateManager.clearCurrentState now handles the process of clearing down the current state and is now called if the Game is destroyed.
Game.destroy now clears the current state, activating its shutdown callback if it had one. It also now destroys the SoundManager, stopping any currently running sounds (#1092)
Input.Gamepad.destroy now destroys all connected SinglePads and clears event listeners.
SinglePad.destroy now clears all associated GamepadButton objects and signals.
Group.onDestroy is a new signal that is dispatched whenever the Group is being destroyed. It's dispatched at the start of the destroy process, allowing you to perform any additional house cleaning needed (thanks @jonkelling #1084)
A slightly obnoxious but necessary hack to prevent a race condition between the loading of Apache Cordova and Phaser itself.
Without waiting for the 'deviceready' event, Phaser can often load first, preventing any console messages from appearing to the user. Because Cordova writes to the platform's console (via CordovaLog), it must first be loaded and signal its own 'deviceready' event before console or plugin usage can occur. Otherwise, all messages and functionality is ignored.
Group.checkProperty allows you to check if the property exists on the given child of the Group and is set to the value specified (thanks @codevinsky #1013)
Phaser.Utils.setProperty will set an Objects property regardless of depth (thanks @codevinsky #1013)
Phaser.Utils.setProperty will set an Objects property regardless of depth (thanks @codevinsky #1013)
Phaser.Utils.getProperty will get an Objects property regardless of depth (thanks @codevinsky #1013)
Rebuilt the way items are polled for Pointer events (drag, click, move). Now faster and more efficient, especially when some items in the stack require pixel perfect checks.
If pause is called manually, codePaused should be set regardless of whether the game is currently paused or not. This would fix issues where a developer might not want the game to automatically resume when the screen regains focus.
Added deep-property getting and setting via strings:
Phaser.Util.getProperty(someObj, 'foo.bar.baz');
Phaser.Util.setProperty(someObj, 'foo.bar.baz', 'lol');
Added a "checkAll" method to Phaser.Group that returns true/false if all of the children's given properties match the value passed in.
this.someGroup.checkAll('foo.bar.baz', 'lol'); // will return true if child[n].foo.bar.baz === 'lol'
Comes with standard 'force' ability.
Group.sendToBack (and consequently Sprite.sendToBack) no longer removes the child from the InputManager if enabled.
Group.add has a new optional boolean parameter: `silent`. If set to `true` the child will not dispatch its `onAddedToGroup` event.
Group.addAt has a new optional boolean parameter: `silent`. If set to `true` the child will not dispatch its `onAddedToGroup` event.
Group.remove has a new optional boolean parameter: `silent`. If set to `true` the child will not dispatch its `onRemovedFromGroup` event.
Group.removeBetween has a new optional boolean parameter: `silent`. If set to `true` the children will not dispatch their `onRemovedFromGroup` events.
Group.removeAll has a new optional boolean parameter: `silent`. If set to `true` the children will not dispatch their `onRemovedFromGroup` events.
Internal child movements in Group (such as bringToTop) now uses the new `silent` parameter to avoid the child emitting incorrect Group addition and deletion events.
Pointer.withinGame is now accurate based on game scale and updated as the Pointer moves.
Stage.bounds is now updated if the game canvas offset changes position. Note that it gives the un-scaled game dimensions.
P2.Body.applyForce should have used pxmi instead of pxm (thanks @Trufi, fix#776)
P2 fixed creation of RevoluteConstraint by passing maxForce in the options (thanks @woutercommandeur, fix#783)
Input and Pointer now use the new ArrayList instead of a LinkedList, which resolve list item removable during callback issues.
Input.reset no longer resets every interactive item it knows of, because they are removed during the destroy phase and can now persist between States if needed.
Game checks if window.console exists before using it (should fix IE9 issues when dev tools are closed), however it is still used deeper in Pixi.
Body.loadData flagged as deprecated.
InputHandler._setHandCursor private var wasn't properly set, meaning the hand cursor could sometimes remain (during destroy sequence for example)
All Game Objects have a new property: destroyPhase (boolean) which is true if the object is in the process of being destroyed, otherwise false.
The PIXI.AbstractFilter is now included in the Phaser Pixi build by default, allowing for easier use of external Pixi Filters.
World.shutdown now removes all children iteratively, calling destroy on each one, ultimately performing a soft reset of the World.
Objects with a scale.x or y of 0 are no longer considered valid for input (fix#602)
InputHandler will set the browser pointer back to default if destroyed while over (fix#602)
Group.destroy has a new parameter: `soft`. A soft destruction won't remove the Group from its parent or null game references. Default is `false`.
InputHandler.validForInput is a new method that checks if the handler and its owner should be considered for Pointer input handling or not.
Group.replace will now return the old child, the one that was replaced in the Group.
All the Debug methods have had the word 'render' removed from the start. So where you did `debug.renderSpriteInfo` before, it's now just `debug.spriteInfo`.
Debug methods that rendered geometry (Rectangle, Circle, Line, Point) have been merged into the single method: `Debug.geom`.
Fixed issue where renderOrderID wasn't being assigned correctly, causing the Input Handler to be unable to select the "top" item on a display list (would all default to zero)
Fixed issue where Stage would assign renderOrderIDs in reverse, should be in sequence.
Fixed issue where objects where checking World for the currentRenderOrderID by mistake instead of Stage.
Basically, input handling works a lot better now for Groups and nested objects :)
Group.sendToBottom(child) is the handy opposite of Group.bringToTop()
Group.moveUp(child) will move a child up the display list, swapping with the child above it.
Group.moveDown(child) will move a child down the display list, swapping with the child below it.
If the game was set to NO_SCALE and you swapped orientation, it would pause and resize, then fail to resize when you swapped back (thanks starnut, fixes#258)
You can now safely destroy a Group and the 'destroyChildren' boolean will propogate fully down the display list.
Calling destroy on an already destroyed object would throw a run-time error. Now checked for and aborted.
Calling destroy while in an Input Event callback now works for either the parent Group or the calling object itself.
In Group.destroy the default for 'destroyChildren' was false. It's now `true` as this is a far more likely requirement when destroying a Group.
All GameObjects now have a 'destroyChildren' boolean as a parameter to their destroy method. It's default is true and the value propogates down its children.
AnimationManager.add documentation said that 'frames' could be null, but the code couldn't handle this so it defaults to an empty array if none given (thanks yig)
Also updated TypeScript definitions and StateManager.add docs.
If your game references the old Phaser.StageScaleMode consts like SHOW_ALL you need to update them to Phaser.ScaleManager, i.e. Phaser.ScaleManager.SHOW_ALL.
All of the Project Templates have been updated to reflect the above change.
Sprite.smoothed and Image.smoothed allows you to set per-Sprite smoothing, perfect if you just want to keep a few sprites smoothed (or not).
Fixes#381.
Loader.setPreloadSprite() will now set sprite.visible = true once the crop has been applied. Should help avoid issues (#430) on super-slow connections.
After defining tiles that collide on a Tilemap, you need to call Tilemap.generateCollisionData(layer) to populate the physics world with the data required.
Debug.renderPhysicsBody updated to take camera location and body rotation into account.
Body movement functions put back to velocity :)
Updated to latest dev version of pixi and latest p2.js
Updated docs
fixedToCamrea now works for Groups as well :) You can fix a Group to the camera and it will influence its children.
Also fixed the issue with World.preUpdate/postUpdate not being called and various small documentation issues.
