Fix recovery device in CLI

[FirelightFlagboy]

parsec-cli appear to use the recovery device mechanism the intended way.

Currently way creating a recovery device, it use the current device without creating a new device.
And when importing the recovery device, we import the device as is without using it to create a new device.

The shamir (wip) system behave like intended and we could reuse some part to fix the issues:

• When creating a shamir device:

  parsec-cloud/libparsec/crates/client/src/certif/shamir_recovery_setup.rs

  Lines 241 to 258 in 96eb831

  	let recovery_device = LocalDevice::generate_new_device(
  	certificate_ops.device.organization_addr.clone(),
  	certificate_ops.device.initial_profile,
  	certificate_ops.device.human_handle.clone(),
  	DeviceLabel::try_from(
  	format!(
  	"recovery-{timestamp}-{}",
  	u32::from_be_bytes(brief[..4].try_into().expect("u32 from 4 u8 should be ok"))
  	)
  	.as_str(),
  	)
  	.expect("Invalid device label"),
  	Some(certificate_ops.device.user_id),
  	None,
  	None,
  	None,
  	None,
  	);

• When importing a shamir device:

  parsec-cloud/libparsec/crates/client/src/certif/shamir_recovery_setup.rs

  Lines 310 to 318 in 96eb831

  	let (device_certificate, redacted_device_certificate) =
  	create_new_device(recovery_device, certificate_ops.device.clone(), timestamp);
  	match certificate_ops
  	.cmds
  	.send(device_create::Req {
  	device_certificate,
  	redacted_device_certificate,
  	})
  	.await?

TODOs

• [CLI] Create a dedicated device when creating a recovery device
• [CLI] Create a dedicated device when importing a recovery device

[NicoTuxx]

related to #8585