* [Yoga] Implement ASYogaLayoutSpec, an experimental alternative to full-tree integration.
This approach allows us to avoid any ASDisplayNode.mm integration points.
However, it is not yet proven to be possible to achieve correctness with this approach.
The entry point (to start calculating), and the measurement function inputs, lack
the full expressiveness of ASSizeRange; we need to make sure that workarounds like
using style.minSize are successful in simulating the behavior of a full Yoga tree.
* [Yoga] Fix file comments, move towards <ASLayoutElement> support.
* [Yoga] Important fix for simplified, non-contiguous Yoga integration.
* [Yoga] Complete implementation of manual memory management (__bridge_transfer, YGNodeFree)
* Lock released between add to pend controller and modifying pend state
The existing design is pretty fraught with error. We should probably
rethink this but in the meantime, this fixes a bug where calling
setNeedsLayout can start failing for nodes.
Essentially the method ASDisplayNodeShouldApplyBridgedWriteToView has
a side effect of registering a node to apply it's pending state *if*
it doesn't currently need the pending state applied. My guess is this
was to avoid continually registering the node and this behavior actually
helped expose this bug.
The bug: after the node is registered for flushing it's state, several
code paths released the lock before applying that state to the pending
state object. Before it could re-obtain the lock to apply it to the pending
state, the pending state controller flushed it on the main thread.
On subsequent calls to setNeedsLayout, the pending state had pending state
already (from previous calls which missed the flush) and thus wasn't
registered for future flushing.
* Add changelog
* Fix alignment of ASImageNodeContentsKey struct to fix hashing
* Change the change log by logging a change
* Add the world's stupidest explicit cast
* Actually its simpler
* Add ASBatchFetchingDelegate
- In addition to checking remaining leading screens, ASBatchFetching now also calculates a remaining time and consults its delegate if needed.
- The delegate can override the decision of ASBatchFetching, for example based on remaining time and average time of past batch requests.
* Fix up tests
* Update CHANGELOG
* Add experimental text node implementation, based on YYText
* Fix warnings and alert when unimplemented experimental features are used.
* Address feedback from review
* Extend the cthulog
* Update license headers
* Implement ASPageTable
- It is a screen page table that can be used to quickly filter out objects in a certain rect without checking each and every one of them.
- ASCollectionLayoutState generates and keeps a table that maps page to layout attributes within that page.
- ASCollectionLayout (and later, ASCollectionGalleryLayoutDelegate) consults its layout state for `layoutAttributesForElementsInRect:`. This ensures the method can return as quickly as possible, especially on a large data set (I heard some people have galleries with thousands of photos!).
* Address comments
* Handle items that span multiple pages
* Make danger happy
* Make ASCellNode indexPath and supplementaryElementKind atomic
* Update the change log
* Fix licenses
* Be explicit with atomic
* Rename the protocol
* And the file
* [ASDisplayNode] Convert isSynchronous to an Objective-C atomic BOOL.
This reduces lock contention, and should also fix a very rarely seen deadlock.
* [ASDisplayNode] Implement a std::atomic-based flag system for superb performance
Although Objective-C atomics are equally fast, or better that std::atomic when
access through method calls, for the most intense use cases it is best to avoid
method calls entirely.
In ASDisplayNode, we could benefit significantly from avoiding both method calls
(already true today) but also avoid locking the mutex, for both CPU and contention
gains.
There will still be many methods that need locking for transactional
consistency - however, there are currently dozens of accessor methods that could
avoid frequent lock / unlock cycles with use of atomics, and other usages of the
ivars directly where locking could be delayed until after early-return conditions
are checked and passed.
* [Yoga Beta] Improvements to the experimental support for Yoga layout.
Yoga remains an unsupported / speculative feature, but these improvements are important for
the functionality of clients that are experimenting with it.
For example, without these changes, ASButtonNode is not able to lay out correctly. These
changes allow certain subtrees that use layout specs to coexist properly in a Yoga heirarchy.
The most significant change here is moving ASEdgeInsets into the #if YOGA gating. Although
this is technically an API change, this type was added with no known use cases and is
really only useful for flexbox layout specification. So, before usages of it are created,
it makes sense to constrain the Texture API surface until that time.
* [RTL] Bridge the UISemanticContentAttribute property for more convenient RTL support.
Although apps could handle this before by setting the view's property in didLoad, it's
useful to bridge this property for setting during off-main initialization.
This change also makes RTL fully functional / automatic for Yoga layout users.
* Remove RTL property addition and depend on PR #60 landing first.
* Fix warnings
* Add line to changelog
Although apps could handle this before by setting the view's property in didLoad, it's
useful to bridge this property for setting during off-main initialization.
This change also makes RTL fully functional / automatic for Yoga layout users.
* Add a thread-safe layoutIfNeeded implementation to ASDisplayNode
* Trigger a layout pass when a display node enters preload state
- This ensures that all the subnodes have the correct size to preload their content.
* ASCollectionNode to trigger its initial data load when it enters preload state
* Minor change in _ASCollectionViewCell
* Layout sublayouts before dispatch to main for subclass hooks
* Update comments
* Don't wait until updates are committed when the collection node enters display state
* Same deal for table node
* Explain the layout trigger in ASDisplayNode
* Introduce ASCollectionViewLayout
- `ASCollectionViewLayout` is an async `UICollectionViewLayout` that encapsulates its layout calculation logic into a separate thread-safe object which can be used ahead of time and/or on multiple threads.
- `ASDataController` now can prepare for a new layout resulted from a change set before `ASCollectionView` even knows about it. By the time the change set it ready to be consumed by `ASCollectionView`, its new layout is also ready.
- New `ASCollectionViewLayoutCalculating` protocol that is simple and generic enough that many types of calculators can be built on top. `ASCollectionViewLayoutSpecCalculator` conforms to `ASCollectionViewLayoutCalculating` protocol and can be backed by any layout spec (e.g `ASStackLayoutSpec`, `PIMasonryLayoutSpec`, etc). We can even build a `ASCollectionViewLayoutYogaCalculator` that uses Yoga internally.
- A built-in `ASCollectionViewFlowLayoutCalculator` that is a subclass of `ASCollectionViewLayoutSpecCalculator` and uses a multi-threaded multi-line `ASStackLayoutSpec` internally. The result is a performant and thread-safe flow layout calculator.
- Finally, `ASCollectionViewLayout` can be subclassed to handle a specific type of calculator with optimizations implemented based on the knowledge of such calculator. For example, `ASCollectionViewFlowLayout` can have a highly optimized implementation of `-layoutAttributesForElementsInRect:`.
Protocolize layout calculator providing and consuming
Add flex wrap documentation
Add a `multithreaded` flag to ASStackLayoutSpec that forces it to dispatch even if it's off main
- Update ASCollectionViewFlowLayoutSpecCalculator to use that flag.
Minor change in ASCollectionViewLayout
Implement Mosaic layout calculator
Minor change
Fix project file
Rename and fix project file
Skip fetching constrained size only if a layout calculator is available
Update examples/ASCollectionView
Remove unnecessary change in ASTableView
Address comments
Rename collection view calculator protocols
Minor changes after rebasing with master
Add ASLegacyCollectionLayoutCalculator for backward compatibility
Remove ASCollectionLayoutSpecCalculator
Remove ASLegacyCollectionLayoutCalculator
Introduce ASCollectionLayout
- A wrapper object that contains content size and an element to rect table.
- Collection layout calculators to return this new object instead of an ASLayout.
Before adding a content cache
Finishing hooking up ASCollectionLayoutDataSource to ASCollectionNode
Stash
Finish ASCollectionLayout
Rough impl of ASCollectionFlowLayout
Revert changes in CustomCollectionView example
Move ASRectTable back to Private
* Rename ASCollectionContentAttributes to ASCollectionLayoutState
* Address other comments
* Introduce ASCollectionLayoutDelegate and make ASCollectionLayout private
* Address comments
* API tweaks:
- Replace `-layoutContextWithElementMap:` in ASCollectionLayoutDelegate with `-additionalInfoForLayoutWithElements:`. The returned object is then stored in ASCollectionLayoutContext for later lookups.
- ASCollectionLayoutContext has no public initializer.
- ASDataControllerLayoutDelegate no longer requires a context of type ASCollectionLayoutContext but simply an `id`. This helps decouple ASDataController and ASCollectionLayout.
- Rename `elementMap` to `elements`.
- Rename `visibleMap` to `visibleElements`.
- Other minor changes.
* Rename ASCGSizeHash to ASHashFromCGSize
* Make sure to call super in -[ASCollectionLayout prepareLayout]
* Update example/ASCollectionView to use ASCollectionFlowLayoutDelegate
* Remove unnecessary change
* Need to check availability before use of macros.
* Get rid of PCH and enforce macro definition.
* Remove prefix
* Switch to global warning instead, much better.
* no message
* Go further
* Make the symbols public so that apps actually build
* Move ASAvailability into the umbrella header
* Remove duplicate define
* Put the file back where it was in the list
* Revert "Put the file back where it was in the list"
This reverts commit 6a80c15b5b5efe5ff39812a018114e8bdc1dc0cf.
We have to check in this case if the supernode of the subnode is indeed the _node that executes a layout transition. It can happen that a node already did a layout transition and added this subnode, in this case the subnode would would be removed from the new node instead of _node.
- Avoid returning a NaN flexShrinkAdjustment.
- Bail early (instead of assert) if a line has a positive `flexFactorSum` but then no flexible item can be found via flex adjustments, for example because the flexible items have zero size.
- Continue flexing other lines when one line is inflexible.
- Add unit test for the first 2 fixes.
appleguy
Garrett Moon
Huy Nguyen
Adlai Holler
Michael Schneider
Adlai Holler