Firefox bug: https://bugzilla.mozilla.org/show_bug.cgi?id=2063492
#276 added FidoDevice::downgrade_to_ctap1() as an infallible operation to work-around #269 (AppID extension pre-flighting).
However:
In practice:
FidoDevice::downgrade_to_ctap1() should be a fallible operation, and there should be a FidoDevice method which indicates whether a authenticator supports CTAP1 (ie: supports_ctap1()).
When use_ctap2_fallback = true, CTAP2-only authenticators should be excluded as a candidate device.
This issue is reproducible with Firefox Nightly on Linux.
Firefox bug: https://bugzilla.mozilla.org/show_bug.cgi?id=2063492
#276 added
FidoDevice::downgrade_to_ctap1()as an infallible operation to work-around #269 (AppID extension pre-flighting).However:
CAPABILITY_NMSGbit set will not acceptCTAPHID_MSG(U2F) commands. It doesn't look like this library checks that flag.FIDO_2_0will not accept U2F commands.fidoServiceRevisionBitfield = 0x20does not support U2F commands.In practice:
FidoDevice::downgrade_to_ctap1()should be a fallible operation, and there should be aFidoDevicemethod which indicates whether a authenticator supports CTAP1 (ie:supports_ctap1()).When
use_ctap2_fallback = true, CTAP2-only authenticators should be excluded as a candidate device.This issue is reproducible with Firefox Nightly on Linux.