Using the IsEnabled property disabled touch in addition to scroll. We'll need to keep some state about the current horizontal scroll status, but we can easily work around this by just disabling horizontal and vertical scroll.
Fixes#764
Currently, when a bundle has not been created properly using the `react-native bundle` command, an exception along the lines of 'Argument can not be null' is thrown. Now we have a more specific error message along the lines of "Cannot resolve '__fbBatchedBridge' ...".
Fixes#719
Getting null reference exceptions due to a condition where the ImageSource is not always set on the ImageBrush. Instead, gathering the image metadata directly from the OnLoad event.
Previously, failures on the CompositionTarget.Rendering event could crash the app. With these changes, each frame callback has an exception handling wrapper.
Fixes#750
Prior to this change, Images with one <Text> ancestor were supported:
```
<Text>
<Image />
</Text>
```
But Images with multiple <Text> ancestors resulted in an exception:
```
<Text>
<Text>
<Image />
</Text>
</Text>
```
This change fixes the latter scenario.
The problem was that some code assumed that it would only receive Inlines when actually it could also receive UIElements.
A number of users were seeing AccessViolationExceptions, especially in Release builds, where JS strings were being GC'd prior to use. I suspect this is because we need to increment the ref count of the marshalled arguments so they are not collected before use as call arguments.
Fixes#751
A recent change to the touch handler to support cases where overlays that did not support any pointer events were used (so the underlying React view that did support pointer events could be reached) ignored the fact that some clicks simply don't belong to the React hierarchy (e.g., native app overlays, as is the case for SplitView).
This change first checks that the original pointer target actually belongs to the React before calculating the target using pointer events logic.
Fixes#746
JavaScript may call native behaviors prior to the `_bridge` variable being set in `ReactInstance`, which would be invoked in the initial call to `FlushedQueue` following the call to run the bundle.
This change set ensures the `_bridge` is set before the bundle is loaded.
Fixes#744
* initial .Net46 project
* attempted merge of matt's net46 tests branch
* completed tracer migration
* tweaks per @rozele feedback
* re-add Shared reference to .Net46.Tests project
* removed NUnit from Universal Tests (again)
* restructured and documented Net46 null objects and added debug flag to Net46 project so tests will pass
* add documentation requirement to all build configurations and removed unnecessary documentation
Previously, we were only considering the hierarchy from the pointer "OriginalSource" in the pointer events calculations. Unfortunately, this meant that if a transparent overlay with "box-none" or "none" pointer events was set, as is the case for "react-native-drawer", then no alternative targets would be considered. This change uses VisualTreeHelper.FindElementsInHostCoordinates to consider all possible targets.
Fixes#715
* initial building Shared project
* Base working shared project and tests
* remove Any CPU references
* Modify react-native CLI solution template to include shared project
* remove duplicated files in ReactNative.Shared
* fixed Any CPU configuration agaain
* removed extraneous copy of file
* Fix proj to match filesystem
* fixed build configurations in VS and added NUnit3.TestAdapter
* initial migration of bridge and reflection to new shared project
* fix(tests): Quick fix to ensure all tests are run for Debug configurations.
* cleanup three unneded EoF newlines
* initial migration of chakra, common, and collections - brings us up to par with matthargett/shared-project
* migrated collections test
* removed unnecessary Facebook.CSSLayout reference
* initial building Shared project
* Base working shared project and tests
* remove Any CPU references
* Modify react-native CLI solution template to include shared project
* remove duplicated files in ReactNative.Shared
* fixed Any CPU configuration agaain
* removed extraneous copy of file
* Fix proj to match filesystem
* fixed build configurations in VS and added NUnit3.TestAdapter
* initial migration of bridge and reflection to new shared project
* fix(tests): Quick fix to ensure all tests are run for Debug configurations.
* cleanup three unneded EoF newlines
* initial migration of chakra, common, and collections - brings us up to par with matthargett/shared-project
* removed unnecessary Facebook.CSSLayout reference
* initial building Shared project
* Base working shared project and tests
* remove Any CPU references
* Modify react-native CLI solution template to include shared project
* remove duplicated files in ReactNative.Shared
* fixed Any CPU configuration agaain
* removed extraneous copy of file
* Fix proj to match filesystem
* fixed build configurations in VS and added NUnit3.TestAdapter
The mapping used from key to filename was not used to convert from filename back to key. Additionally, adds some optimizations to reduce allocations for AsyncStorage.
Fixes#724
Basic overview of the deadlock:
1) CSS layout operation in progress
2) Text node measurement started on native module thread
3) ReactContext disposed on dispatcher thread
4) ReactContext disposal blocks on completion of native module thread
5) Native module thread blocks on dispatcher operation for text measurement
6) We have deadlock
The best way to get around this is to not block from the native module thread, but this is a limitation we haven't figured out how to work around yet. In the meantime, it actually makes sense to make the teardown process for the React context asynchronous, which is what this changeset does.
Fixes#702
Getting a lot of issues filed because of exceptions thrown from WebSocket module. This tend to come from the devtools setup which no one (currently) seems to be using for Windows. We can delete the added file at some point when devtools should be enabled for Windows.
Fixes#701
We don't want to disable dev support for all Release builds (the 'Release' configuration still expects to communicate with the bundle server). Instead, we'll just disable it for the ReleaseBundle configuration.
Towards #692
There seems to be a problem with using BitmapImage.SetSourceAsync. Occasionally, when the many images are being loaded simultaneously, the app will freeze/crash without throwing an exception. The culprit seems to be the UI thread hanging in the BitmapImage.SetSourceAsync method. To get around this issue, I've reverted to using the UriSource property for BitmapImage. We still have open issues to implement prefetch, which is what @phongcao is working on and should be done with in the coming weeks.
Fixes#672
Kevin VanGelder
Eric Rozell
Adam Comella
Gant Laborde