subplot acceptance tests are sometimes superficial
Consider the following test:
given an installed sq
when I run sq key generate --userid Alice --export alice.pgp
when I run sq key generate --userid Bob --export bob.pgp
when I run sq keyring join alice.pgp bob.pgp
then stdout contains "-----BEGIN PGP PUBLIC KEY BLOCK-----"
then stdout contains "-----END PGP PUBLIC KEY BLOCK-----"
This isn't really checking that the output contains two keys. It would be better if it checked the output of sq inspect
or sq packet dump
for two keys. Of course, the output of those two commands should not be grepped (at least not until there is stable, machine readable output). Alternatively, I think we could have an additional subplot step function that can check a few properties of a key or keyring.
@Lars: What do you think?