v3.4.0

01.02.2022 3.4.0

Fixed

• UI - Fixed an issue where the UI scaled incorrectly when shifting between orientations and resizing the window
• UI - We've reduced the "page jank" when resizing the UI, when the window resizes change or the phone rotates

Changes

• UI - We've smoothed out the "move closer" prompt to avoid toggling between two UI states.
• Timeouts - We now handle timeout logic in a much clearer way:
  • The SDK emits an error event with feedback code error_token_timeout if the user starts the iProov session too late.
  • As before, if the user starts iProoving but doesn't stream imagery in time, the SDK emits a failure event with the
    feedback code failure_user_timeout.
• Error messaging and language - We've increased the level of detail, see iproov-en.json for the new strings:
  • When the SDK runs any of the following callbacks: error, cancelled, interrupted, permission_denied, no_camera or unsupported:
    • The language string mapping to the feedback event property is now displayed as the heading.
    • A new language string has been added for each feedback code, suffixed with _message.
    • Example: error_no_camera is the heading term, defaulting to "We couldn't find a camera connected to your device"
    • Example: error_no_camera_message is the text term displayed beneath the heading, defaulting to "A camera must be available to use iProov"
  • If your integration specifies language strings but doesn't implement the new _message suffix, then we fall back to the old behaviour:
    • Heading using the "error" language term, and text using the corresponding feedback code's language term.

Added

• Genuine Presence Assurance - added reflection strength estimation capabilities
• Language strings:
  • error_token_timeout and error_token_timeout_message when sessions don't start within 10 minutes of token create
  • iproov_ready_title to allow internationalisation of "Ready to iProov" on the ready callback
  • iproov_ready_button to allow internationalisation of "Scan face" for the ready callback button

v3.3.8

14.01.2022 3.3.8

Fixed

• Support - We've improved support logic where our partners embed the Web SDK into webpages which might lack the required feature policies to run properly.
• client_error - We've fixed a large proportion, if not all, of video NotAllowedError and fullscreen related TypeError in browsers that use different user activation paradigms.
• server_error - Some cases where tokens expired after 10 minutes were being classified as server_error. These are now correctly thrown as error_expired_token.

Internal changes

• Start button - After the user clicks the start button, we now briefly set the start button innerHTML to the prompt_loading language term while the video starts.
• Support checker - We've made a few changes that help give users a better experience:
  • Where possible in Liveness, we now check document.featurePolicy for accelerometer, gyroscope and magnetometer. If these aren't allowed, instead of starting the SDK and erroring, we now run the unsupported callback.
  • For both GPA and Liveness, we now require fullscreen to be enabled on the webpage, unless in iOS. This ensures correct behaviour of the SDK.
  • We no longer use the Permissions API to establish camera support due to its poor support and unreliable behaviour.

v3.3.6

09.12.2021 3.3.6

Fixed

• Client errors - We've fixed some, if not all, instances of NotAllowedError occurring when we get a video stream.
• WebGL - We now gracefully handle webglcontextlost events which might happen when devices are under high load.
• Iframe Bridge + Native Bridge - We fixed a regression introduced in 3.3.5. Sorry about that.

v3.3.5

26.11.2021 3.3.5

Fixed

• Supported Devices - Block custom webview user agents where we can't test if supported.
  • More info here.
• Interrupted Event - If exiting fullscreen, handle when there's no video feed.
• Unsupported video feeds - Further improve handling of very low resolutions.
• Permission button - the grant button delay has been fixed

Internal Additions

• Telemetry - Collects state of the SDK during various events, and in the case of incomplete transactions.
• Socket.io dependency - We've upgraded our socket.io client dependency to version 4 to match the server.

v3.3.4

Fixed

• a11y - Improve text to speech order while using some screen readers.
• No Face Timeouts - Only start detecting faces when ready is clicked.
• Error Handling - Ensure error messages are always present.

v3.3.3

Fixed

• Safari fullscreen issues - Handle Safari 13-14.1 not reporting fullscreen enter/exit events correctly.
• Safari permission issues - We fixed an issue where permission could be blocked on certain devices.
• Progress event - Progress event returned messages weren't being localised.

Changes

• Unsupported video feeds - Detect very low resolutions and attempt to use a different video device if possible.
• Safari 15.0 desktop - Regrettably, this browser is now blocked until Apple releases a version that fixes various regressions.
• Native Bridge - Added support for Native Bridge inside iframes which will be supported on iOS 14+ with iProov SDK v9.1 (coming soon). Iframes on iOS 13 and under cannot be supported and will raise an unsupported error.
• WebGL - THREE.js upgraded to 0.133.1.

v3.3.2

Fixed

• Liveness - We have worked around instances where devices do not send orientation events. These previously triggered device_motion_unsupported errors.

Changes

• Performance analysis - We have further tweaked our logging to enhance performance analysis in the long term.

v3.3.1

Fixed

• Safari on iOS - Worked around circumstances when Safari provides corrupt imagery.
• Safari on iOS - Liveness is now supported in an iframe on iOS 15 without the use of Iframe Bridge.

Changes

• Face detection - Blazeface + TFJS upgraded to 0.0.7 and 3.8.0 respectively, enabling SIMD support, making things faster.
• UI - Minor optimisations and simplified our shaders to reduce the WebGL load where possible.
• UI - Low FPS Mode no longer presents an overpixellated UI. This should improve completion rates.
• Crash handling - Improved detection of face detector crashes.
• Performance Analysis - Improved distinction between the user clicking the exit button or exiting fullscreen.

v3.3.0

New

• Camera selection - in multi camera environments, where browser support allows it, we have improved camera selection logic to prioritise user facing and higher resolution devices.

Fixed

• Camera selection - multiple camera selection no longer blocks fullscreen.
• Camera selection - we have fixed a regression where programmatically setting the camera device in slotted content failed.
• Safari on macOS Catalina - we have worked around Safari's lack of fullscreen event support for this OS.

Please note our Browserslist database has been updated to 1.0.30001248.

v3.2.2

Fixed

• Fixed crashing issue found in some JavaScript libraries.
• Fixed missing connecting event.
• Fixed missing feedback property from interrupted and cancelled events.

Deprecation Notice

The code property which is returned within event.details has been deprecated and will be removed in future versions. This was unintentionally added. If you are using code please switch to the feedback property. For more details on the correct way to use event data please see our docs.