IPNS format keys in b36cid

[petar]

No description provided.

[petar]

To run sharness locally, during dev, run this from the test/sharness dir:

TEST_NO_FUSE=1 GOIPFSTEST_SKIP_LOCAL_DEVTREE_DEPS_CHECK=1 make

[aschmahmann]

We need more sharness tests here. In particular we should

1. Like with the previous PR, have a function that will let us test that Ed25519 + RSA keys both work (it's not quite in line with this PR but we need to do it anyway
2. Make sure that we can publish using a b58mh as well as a b36cid. The results can all come out as Publish to ${b36cidKey} though.

[petar]

We need more sharness tests here. In particular we should

1. Like with the previous PR, have a function that will let us test that Ed25519 + RSA keys both work (it's not quite in line with this PR but we need to do it anyway
2. Make sure that we can publish using a b58mh as well as a b36cid. The results can all come out as Publish to ${b36cidKey} though.

I've added a test for (2). For (1), it seems one good strategy is to wrap all tests in a function parameterized by ed25519/rsa and use this to initialize the self ID, because most tests are based on the self id. Would you prefer a different approach?

[aschmahmann]

(1), it seems one good strategy is to wrap all tests in a function parameterized by ed25519/rsa and use this to initialize the self ID, because most tests are based on the self id. Would you prefer a different approach?

We need to test publishing with the --key flag anyway. So maybe something like this:

• Have a function that tests with ed25519, rsa, and self
• Test that publishing with no --key flag means we published with self
• Test that we can publish with --key=peerIDRepresentationOfKey, or --key=friendlyName

[petar]

(1), it seems one good strategy is to wrap all tests in a function parameterized by ed25519/rsa and use this to initialize the self ID, because most tests are based on the self id. Would you prefer a different approach?

We need to test publishing with the --key flag anyway. So maybe something like this:

• Have a function that tests with ed25519, rsa, and self
• Test that publishing with no --key flag means we published with self
• Test that we can publish with --key=peerIDRepresentationOfKey, or --key=friendlyName

I think now all these cases are covered.

[aschmahmann]

This code tests publishing with the self key which what is tested in the other function test_name_with_self . Did you mean to use the generated key here?

[petar]

No: This test is focused on using an explicit key. Moved it to test_name_with_self.