To preserve dynamic color capabilities for layer-backed
nodes, we need to track backgroundColor as a UIColor. In layer backed nodes we lose the dynamic color logic since we
extract a CGColor from the UIColor applied to ASDisplayNode. With this
change we can respond to changes in UITraitCollection.userInterfaceStyle
for backgroundColor.
* Add support for tinting layer-backed ASImageNode's
* Move tint color support ASDisplayNode
* Address comments, unlock before calling "tintColorDidChange" on
subclasses
* Address comments
- Add the ability to specify a tintColor for images / buttons that behave similarly to how they would under UIKit.
- Add flag to control if tint color is applied to ASTextNode to match how
UIKit controls if it is .applied to UILabel.
- Update tests for not-tinting automatic render mode for UIImage.
- Add property on ASTextNode2.
- Add tint color support to ASTextNode2.
Given that UIColor is now a dynamic provider we can no longer rely on the fact that there is a 1:1 mapping between UIColor <-> CGColorRef.
I've updated the UIViewBridge and ASPendingState accordingly to use UIColor and translate to CGColorRef only when interacting with layers. This should preserve any dynamic capabilities we might need in the future when trying to re-render layer backed nodes.
* Set background color directly to view or layer depending on if the node
is layerBacked.
Summary:
UIColor can now be a dynamic color provider to support capabilities like Dark Mode in applications. UIKit is able to re-render views when a user enters dark mode but this dynamic capabilities is not handled at the layer level.
Currently this takes the recommendation in facebookarchive/AsyncDisplayKit#1537 (comment) to set background color on the view directly or layer depending on if the node is layer backed. This change allows UIKit to track these colors and update the UI when the UITraitCollection.userInterfaceStyle changes.
* Updates to UIViewBridge to fix tests
* Remove opaque
* Test fixes
* Revert change to test checker method
* Fix strange issue where setting UIView backgroundColor sets the layer to
opaque to false
* Address Huy comments
* Address last comment
* Improve Accessibility Implementation
- Fixes calculations for accessibilityFrame of ASAccessibilityCustomAction and ASAccessibilityElement
- Fix crash for accessibility elements within layer backed nodes
- Fix accessibility action label updating after other actions change the accessibility label of the corresponding node
* Address comments
* Add support for basic ASTextNode2 link highlighting
* Add approach to fix line highlighting
* Enable proper highlighting for first row
* Fix snapshot highlighting test for ASTextNode2
* Update screenshot again
* Address comments
* ASSignpost: Add support for the os_signpost API
- Removes support for colored intervals. These weren't really useful and they aren't part of the signpost api.
- Fixes an issue with range controller interval logging.
- Adds an interval for interface orientation changes.
* Rename
* Support Xcode 9
* Commentary
Revert "Ensure ASControlMode properties lock before accessing their ivars (#1476)"
This reverts commit ce1e1956f4.
Revert "Make sure all ASDisplayNode properties have backing ivars for consistency. (#1475)"
This reverts commit d6061f4390.
- Except the unused property `interfaceStateSuspended` is still removed (this is a memory saver).
* Shrink ASImageNode by .6% and ASNetworkImageNode by 2.2%
ASImageNode goes from 1384 to 1376. ASNetworkImageNode goes from 1496 to 1464.
These objects accumulate in the heap, so reducing their size will allow more to accumulate before memory warnings.
Group the `BOOL`s into a struct. Shrink the various stored `enum`s to fit the size of their contents. Move the ivars around so that the smaller `enum` are near eachother and the bitfield struct.
* add comments as requested in garrett's review.
* access ivar directly since already locked as suggested in review.
* Shrink ASDisplayNode from 1088 to 976 bytes, reduction of 11.48%
These objects accumulate in the heap, so reducing their size will allow more to accumulate before memory warnings.
Group the `BOOL`s into a struct. Shrink the various stored `enum`s to fit the size of their contents. Move the ivars around so that the smaller `enum` are near eachother and the bitfield struct.
* Forgot to bit-field-ify placeholderEnable; new small size is 968, old size is 1072 (measured consistentlyl on iPhoneSE simulator). 10.74% reduction.
* Shrink _ASPendingState from 824 to 808 bytes, reduction of ~2%
These objects accumulate in the heap, so reducing their size will allow more to accumulate before memory warnings.
Shrink ASPendingState BOOLs. Use unsigned int for these bitfields for clarity. Separate from `_flags` as that struct is compared to `0` for testing. The `_boolFlags` struct is just storing other `BOOL`s.
* rename struct fields as suggested in garret's review.
* Make sure all ASDisplayNodes have backing ivars for consistency.
Found this by enabling #pragma clang diagnostic error "-Wobjc-missing-property-synthesis" for ASDisplayNode. One property is unused, saving 8 bytes of heap space per instance on 64-bit builds. Implement setter/getters for these properties, and add appropriate locking. add the warning as error to the build for this file.
* Remove ASEventLog. We'll focus on ASLog for the future, possibly making it triggerable on a per-node basis
* See ya!
* Hello semicolon
* Remove refs to gone tree methods
* Fix
* Remove irrelevant change, kick CI
* Clean up
* Add layer-action support to nodes, unify hierarchy notifications on it
* Better pending state
* Fix bool
* Skip extra copy
* Never run default actions
* Continue the search
~/Texture/Source/Private/ASMutableElementMap.mm:32:24: warning: Conversion from value of type 'NSMutableArray<NSMutableArray *> *' to incompatible type 'ASMutableCollectionElementTwoDimensionalArray *'
_sectionsOfItems = (id)ASTwoDimensionalArrayDeepMutableCopy(items);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
* Rename ASDN C++ namespace to "AS." Referring to the framework as ASDisplayNode is pretty out-dated and verbose. See CoreAnimation which usees CA for their internal namespace.
More using
* More cases
Right now when an image node enters preload state, we kick off an image request with the default priority. Then when it enters display state, we change the priority to "imminent" which is mapped to the default priority as well. This means that requests from preload and display nodes have the same priority and are put to the same pool. The right behavior would be that preload requests should have a lower priority from the beginning.
Another problem is that, due to the execution order of -didEnter(Preload|Display|Visible)State calls, a node may kick off a low priority request when it enters preload state even though it knows that it's also visible. By the time -didEnterVisibleState is called, the low priority request may have already been consumed and the download/data task won't pick up the new higher priority, or some work needs to be done to move it to another queue. A better behavior would be to always use the current interface state to determine the priority. This means that visible nodes will kick off high priority requests as soon as -didEnterPreloadState is called.
The last (and smaller) issue is that a node marks its request as preload/low priority as soon as it exits visible state. I'd argue that this is too agressive. It may be reasonble for nodes in the trailing direction. Even so, we already handle this case by (almost always) have smaller trailing buffers. So this diff makes sure that nodes that exited visible state will have imminent/default priority if they remain in the display range.
All of these new behaviors are wrapped in an experiment and will be tested carefully before being rolled out.
* Add imports
* Fix build failure
* Encapsulate common logics into methods
* Address comments
* Optimize ASTwoDimensionalArrayUtils
These methods are called on the main thread during range controller updates (i.e. every frame) and so they should be as fast as possible.
* Rename
* Use vector instead of stack array to handle really big cases (e.g. photos)
* Add an experimental flag to use native dispatch_apply instead of our core count * 2 approach. This has shown performance wins in some profiling.
* Add in other places
* Improve separation of code for layout method types
* Address PR comments
- Delegate to layout spec engine if the node is a layout spec node but yoga engine was asked for calculate the layout
- Change ASLayoutType to ASLayoutEngineType
- Improve layout engine fall through code
Initializing the LUT arrays at file level scope creates a large chunk of code retaining and releasing all of the NSStrings in the tables. Moving them to function level moves the initialization to being lazy.
Kevin Smith
Rahul Malik
Hanton Yang
Huy Nguyen
Adlai Holler
Michael Schneider
Greg Bolsinga
Adlai Holler
Wu Zhong
Huy Nguyen
Harry Shamansky
Ballad
Jacob Farkas
Tim Norman
Greg Bolsinga
dmaclach