IssueTracker

If this is a bug in the library, we would appreciate if you could attach:

• No bug is visible in the logcat but a log is visible and is triggered by this line Log.e(TAG, "Failure in canAuthenticate(). FingerprintManager was null.");
• Can be found in https://github.com/infinum/Android-Goldfinger/tree/release/2.1.0.
• Not a UI issue

BUILD INFO

• Device type: Tested on Huawei Mate 20 Pro (LYA-L29) and Xiaomi Mi 9T
• OS version: Android 10 (both devices)
• Biometric library version: Tested with 1.1.0-rc01 and 1.2.0-alpha01

STEPS TO REPRODUCE [Be as specific as possible please]

1. Run the sample project on an Android 10 device.
2. The initial screen will display the available authenticators where the Strong authenticator will look disabled which means it is not available for that specific device

EXPECTED RESULTS

I expect available hardware for these devices because they do have a Fingerprint sensor and they do classify as BIOMETRIC_STRONG because if run Biometric authentication without any set authenticators the internal lib will set the authenticator to BIOMETRIC_STRONG since I have a crypto object initialised and everything works as expected. The method that is responsible for setting the authenticator to strong is getConsolidatedAuthenticators(@NonNull BiometricPrompt.PromptInfo info, @Nullable BiometricPrompt.CryptoObject crypto) inside the first else block:

authenticators = crypto != null ? Authenticators.BIOMETRIC_STRONG : Authenticators.BIOMETRIC_WEAK;

OBSERVED RESULTS

I get a BIOMETRIC_ERROR_HW_UNAVAILABLE inside the BiometricManager class due to the fingerprintManager being null:

@VisibleForTesting
    BiometricManager(@NonNull Injector injector) {
        mInjector = injector;
        if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.Q) {
            mBiometricManager = injector.getBiometricManager();
            mFingerprintManager = null;
        } else {
            mBiometricManager = null;
            mFingerprintManager = injector.getFingerprintManager();
        }
    }

the result does not make sense to me because I do have a device which have the appropriate strong classified HW based on the behaviour in the EXPECTED RESULTS section. The fingerprint implementation with the CryptoObject worked fine on 1.0.1 version of the library.

NUMBER OF TIMES YOU WERE ABLE TO REPRODUCE (e.g. 3/10)

10/10