WebXR Viewer

Version 2.0 of the WebXR Viewer is a complete rewrite of the application, based on the Firefox for iOS source code. The application now implements the official WebXR API, including the accepted and proposed features.

This update is focused on fixing stability issues and bug.

- The XRViewer keeps ARKit running for a short while between XR session, to allow users to switch between 2D and XR pages without incurring complete ARKit startup times each time they enter an XR session. We have seen issues where this causes problems, so we have changed the XRViewer to reset ARKit more often.
- Starting a new session clears out relocalization attempts to prevent immediately getting thrown into "relocalizing" loops
- Allow ARKit to automatically relocalize when returning from the background to prevent objects from disappearing after loading the worldMap
- Fix bug that caused the permissionLevelButton to become inactive after granting and subsequently denying "Device Motion" access on a WebXR site

Bug fixes, including
- the page URL should no longer be erased if a page causes an out-of-memory condition
- obscure issue resolved that sometimes caused the Permissions page to have a long delay before appearing

UI change to hide tracking status on non-XR pages, and when the URL Bar is hidden.

This version updates the userAgent string to included "WebXRViewer" and the version number. It also fixes a bug that allowed hit testing to return hits against more than the single plane selected in Lite mode.

This version has some performance improvements and bug fixes.
Update the splash page to point at the new set of examples.

Bug fixes.

This version represents a major internal overhaul of the application, in preparation for changes needed to support the official WebXR API when it is finalized. These changes also address some performance issues. While the external behavior has not changed substantially, we have reworked the UI somewhat. The most noticeable change is that starting a WebXR session now requires permission from you before a web page can access the WebXR APIs, offering you an opportunity to provide reduced information about the world as part of the permission request.

Detailed Changes:
- many bug fixes
- removed video and photo capture capabilities, in lieu of the newer iOS built in capabilities
- updated the built-in WebXR API to implement most of the older API functionality (no camera image access yet).
- Image targets can be tracked now (ARKit 2.0 functionality)
- accessing WebXR functionality now asks for user permission, as the official WebXR requires. In particular, entering AR mode now requires permission, as does access information about the world (such as ARKit Planes and Face Meshes), and accessing Camera data (when this is implemented).
- permissions (for basic WebXR and world knowledge give the user the option for an "AR Lite" Mode. When selected, the user is asked to pick a plane, and only that plane is made available to the web page. This allows interesting AR application functionality, without exposing the entire structure of the space to the app. The web page does not know if the user is using "AR Lite" mode.
- permission can be changed while a web app is running in XR mode
- added icon in menu bar to show current permission, and clicking it brings up a permission menu
- most of the application has been rewritten in Swift

- bug fixes

- bug fixes, updated for iOS 12 and ARKit 2.0
- cleaned up preferences so in-app and iOS preferences match
- added APIs for using ARKit ARWorldMap
- added option to "always allow world sensing access" for a site. Permissions are revoked by resetting all saved sites.
-

Updated wording on "Access to World Sensing" dialog to make it clearer what is being asked.

This version adds support for additional ARKit 1.5 features using experimental API extensions to WebXR.
- ARKit 1.5 features added: plane geometry, Image Recognition/Image Anchors, and Face Anchors.
- On phones with TrueDepth cameras (iPhoneX), user can switch to front-facing camera to detect and track a face with ARKit Face Anchors
- Expose API to create images for ARKit Image Detection
- Using the Image Recognition and Face Tracking first requires user to give permission for the app to "sense and track world knowledge"

Updated to support ARKit 1.5

- ARKit state is more carefully managed, so each load of a page doesn't require the user to scan the room and help ARKit restart. Moving from page to page is more seamless, and reloading pages is faster.
- a new WebXR video camera API is enabled, allowing web apps to do computer vision. The user is always prompted to give permission to any web page before it can access video image data.
- Support for AR-capable web pages that can switch between a 2D page and full screen immersive AR views.
- new permissions UI on first load: explains why we are asking for GPS and camera, before user is asked to select or decline each
- "Swipe up to show URL bar" message only shown once per run of the app
- various bug fixes

This version contains bug fixes, plus a few changes:
- We have started using Mozilla's telemetry service to collect anonymous usage information (you may opt-out in the app settings)
- Added in-app settings page with link to privacy notice, opt-out for data collection, and default URL

In addition to some small bug fixes, the app now supports viewing in all orientations properly.
To reduce screen clutter, we've replaced the "eye" icon in lower right (that was used to show / hide native UI) with swipe gesture at the top of screen. Swipe up to hide, down to reveal ... it's a wonderful thing.
We've also added a bit more detail to the error messages when ARKit doesn't initialize properly.

More UI and bug fixes. iPhoneX support.

This release fixes a number of issues we discovered since our initial submission, and updates the UI to make things clearer and easier to use.