support/testing: do not print non-utf-8 chars on cbor test

support/testing: do not print non-utf-8 chars on cbor test

The python-cbor sample script to be run on the target dumps a binary
variable to the target stdout. This is done for debug purposes-only, in
the case the test on test infra fails.
This non-utf-8 is currently silently ignored by the Emulator class /
pexpect.spawn from the test infra because the infra uses Python 2, that
in turn do not differentiate between string and byte data.

Make the code Python3-friendly (Python 3 in the host) by doing the right
thing and encoding the data before printing it.

Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>