WIP: Upgrade versions of RustCrypto to pre

[wiktor-k]

@baloo sorry to interrupt but maybe you'll have a clue on why these tests are failing.

This is a one-commit extension to #85 which only upgrades RustCrypto crates to latest versions (to pull changes such as RustCrypto/SSH#233) (and also it removes some example code that I didn't want to update).

The tests fail with error: ssh_agent_lib::proto::error::ProtoError - SSH signature error: signature error: unsupported algorithm: ssh-ed25519 but there are no functional changes, only the bump in versions. Is this a regression in the RustCrypto crates or am I holding it wrong?

Thanks in advance for your help! 🙇

[jcspencer]

@wiktor-k potentially related to this commit? RustCrypto/SSH@0782a5c

[baloo]

interestingly ... the tests would fail when running on my machine, but they work in CI?

    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.14s
     Running unittests src/lib.rs (target/debug/deps/ssh_agent_lib-7312c3c9364a567c)

running 2 tests
test proto::extension::constraint::tests::parse_destination_constraint ... FAILED
test proto::extension::message::tests::parse_bind ... ok

failures:

---- proto::extension::constraint::tests::parse_destination_constraint stdout ----
Destination constraint: RestrictDestination { constraints: [DestinationConstraint { from: HostTuple { username: "", hostname: "", keys: [] }, to: HostTuple { username: "", hostname: "github.com", keys: [KeySpec { keyblob: Ed25519(Ed25519PublicKey([227, 42, 170, 121, 21, 206, 185, 180, 73, 209, 186, 80, 234, 42, 40, 187, 26, 110, 1, 249, 11, 218, 36, 90, 45, 29, 135, 105, 125, 24, 162, 101])), is_ca: false }, KeySpec { keyblob: Rsa(RsaPublicKey { e: Mpint(010001), n: Mpint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}), is_ca: false }, KeySpec { keyblob: Ecdsa(NistP256(EncodedPoint(Uncompressed { x: Array([73, 138, 72, 67, 99, 64, 71, 179, 58, 108, 100, 100, 204, 187, 162, 146, 160, 192, 80, 125, 158, 75, 121, 97, 26, 216, 50, 51, 110, 27, 147, 124]), y: Array([238, 228, 96, 131, 160, 139, 173, 186, 57, 192, 7, 83, 255, 46, 175, 210, 98, 149, 209, 77, 176, 209, 102, 118, 96, 31, 254, 249, 58, 104, 114, 72]) }))), is_ca: false }] } }] }
thread 'proto::extension::constraint::tests::parse_destination_constraint' panicked at src/proto/extension/constraint.rs:313:38:
error: ssh_agent_lib::proto::error::ProtoError - SSH encoding error: length invalid
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    proto::extension::constraint::tests::parse_destination_constraint

test result: FAILED. 1 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.01s

EDIT: because the CI did not run.

[baloo]

(rebased)

[wiktor-k]

potentially related to this commit? RustCrypto/SSH@0782a5c

Excellent detective work, folks! 🙇

Sadly this means we still can't build the thing without patched sources 😢

What would be the way to proceed here @baloo? Wait until the RustCrypto crates stabilize a bit?

I'm wondering if it makes sense to merge #85 as is (I think @jcspencer is leaning on this approach), release 0.5 (with -pre in between). If RustCrypto stabilizes we can fix that in 0.5.1... or maybe I'm missing something? 🤔

[baloo]

This will require at least a pre-release of ed25519-dalek (and others in the repo, but I don't have access there) and a subsequent release of ssh-* after that (that I can do).

I have no idea what the timeline is in regard to ed25519.
We can go and use the non-pre-release, non-patched sources, and release a 0.5 here then bump and release a 0.6 (nice thing about numbers is we have a fairly unlimited supply of those).