Files
jaap-jan e3fd3e1728 Sync and authenticate with SSH keys on the client
Completes the client half of SSH keys: they sync alongside hosts, appear in
their own list, and can be selected to authenticate a connection instead of
typing a password.

The reconciler and the repository were Host-typed throughout, so the choice was
to generalise them or to keep a second copy per item type. Generalised, because
ItemReconciler's whole premise is that the pull and the push paths must answer
the same collision the same way — two copies would drift the first time one of
them was fixed. What is genuinely per-type now arrives through
IItemKind<TSecret>: the cipher, the merge, the plaintext columns, and the noun
to use when telling a person what happened to their item. Generic where the
server's IItemKind is not, and for the reason that reverses there — the client
needs the concrete type, because it merges field by field.

The pull filter is derived from the same registry that builds the reconcilers.
That is the specific failure being designed out: an item type that encrypts,
merges and lists perfectly and is never once requested from the server, so it
works on the machine that made it and exists nowhere else.

No client cache migration. The item table's primary key and the outbox's unique
index already carry the entity type, and AadResourceTypes already mapped SshKey
— so a host and a key may share an id and never see each other's rows, which
SshKeySyncTests now arranges deliberately.

A key hands the server nothing in plaintext. There is a public_key_fingerprint
column and it would be accepted; leaving it null is deliberate. A fingerprint is
not secret but it is a stable identifier for a key pair, so filling it would let
an operator tell which of their users hold the same key and correlate one across
vaults, for a column nothing reads. The design allows itself one plaintext
concession — the relay address, which the relay cannot work without — and this
is not that.

A key is chosen per connection rather than bound to a host, which works the way
ssh -i does. Binding one needs a field on HostSecret and therefore a payload
schema bump, which makes every host written afterwards read-only on an older
build; worth doing deliberately rather than as a side effect of adding keys.

Three things this found, all of them by being falsified rather than by review:

- Making the reconciler generic silently turned a record comparison into
  reference equality, because == on a type parameter is not value equality. The
  effect would have been a conflict recorded on every pass for an unacknowledged
  create that had in fact landed. Sabotaging the fix left all 73 tests passing —
  nothing covered that branch — so ConflictMatrixTests now has
  AnUnacknowledgedCreateThatDidLand_IsDroppedQuietly, which fails without it.

- A test asserting that a blank passphrase reaches SSH.NET as null was vacuous:
  it exercised the editor, not the credential path, and passed with the guard
  deleted. Resolved by making SshKeySecret.Passphrase normalise an empty string
  to null, so there is one spelling of one state — which also keeps two clients
  from producing different payload bytes for an identical key. That exposed a
  wider gap: SshKeySecret, its codec and its merge had no direct unit tests at
  all. They have 25 now.

- The reason first given for that normalisation was false. It claimed SSH.NET
  rejects a passphrase supplied for an unprotected key; measured against a real
  sshd it ignores it and authenticates anyway. Corrected everywhere it was
  stated and recorded in docs/platform-flags.md. The same test file also closes
  a real hole: SshPrivateKeyCredential had never been exercised against a
  server, because the existing key test builds SSH.NET's auth method directly
  and bypasses the path a vault-held key actually takes.

Only one editor may be open at a time. Both sit in the same 340-pixel column as
Auto rows and their heights together exceed it at the window's minimum size, so
two open editors put the lower one's Save and Cancel past the bottom edge — the
same failure this window already shipped once with the setup screens. Expressed
as a state rule because that is the only form of it this repository can check:
nothing here loads a .axaml. The refusal keeps what was typed, since in the key
editor that is a pasted private key the user may have nowhere else.

The end-to-end slice now carries a key as well as a host, so both item types go
through the real API, the real PostgreSQL and the real crypto in one pass — the
three hand-kept mappings between enums that do not line up are the reason that
is worth doing rather than trusting the unit suites.

735 tests green, including the container-backed SSH and end-to-end suites. Zero
warnings, dotnet format clean.
2026-07-29 20:27:23 +02:00

312 lines
12 KiB
C#

using DodoSSH.Client.Storage;
using DodoSSH.Contracts;
using static DodoSSH.Client.Sync.Tests.SyncHarness;
namespace DodoSSH.Client.Sync.Tests;
/// <summary>
/// SSH keys through the same two-machine harness as hosts.
/// </summary>
/// <remarks>
/// <para>
/// Deliberately not a copy of <see cref="ConflictMatrixTests"/> with the nouns changed. The six collision
/// outcomes are decided by <c>ItemReconciler&lt;TSecret&gt;</c>, which is one implementation shared by both
/// item types, so re-asserting all of them per type would test the same code twice and grow with every type
/// added. What is tested here is what is genuinely different about a key: its cipher, its merge, the fact
/// that it hands the server nothing in plaintext, that the reconciler's messages call it a key, and that its
/// items cannot be confused with a host's.
/// </para>
/// <para>
/// The two collision cases that <em>are</em> repeated — resurrection and an abandoned delete — are here
/// because they are the two that touch key material: one re-seals it under a new id, the other decides
/// whether a private key survives a deletion.
/// </para>
/// </remarks>
public sealed class SshKeySyncTests : IAsyncLifetime
{
private SyncHarness harness = null!;
private static CancellationToken Token => TestContext.Current.CancellationToken;
/// <inheritdoc />
public async ValueTask InitializeAsync() => harness = await CreateAsync();
/// <inheritdoc />
public ValueTask DisposeAsync()
{
harness.Dispose();
return ValueTask.CompletedTask;
}
// ---- The uncontested paths ----
[Fact]
public async Task AKeyCreatedOnOneMachine_ReachesTheOther()
{
var entityId = await harness.First.CreateKeyAsync(
Key("deploy", material: "LAPTOP-MATERIAL", passphrase: "hunter2", notes: "rotate in June"));
await harness.SettleAsync();
var seen = await harness.Second.FindKeyAsync(entityId);
seen.Secret.Label.ShouldBe("deploy");
seen.Secret.PrivateKeyPem.ShouldContain("LAPTOP-MATERIAL");
seen.Secret.Passphrase.ShouldBe("hunter2");
seen.Secret.Notes.ShouldBe("rotate in June");
seen.HasUnsyncedChanges.ShouldBeFalse();
}
[Fact]
public async Task ThePull_AsksForKeysAsWellAsHosts()
{
await harness.First.SyncAsync();
var asked = harness.Server.LastPullTypes.ShouldNotBeNull();
asked.ShouldContain(SyncEntityType.Host);
asked.ShouldContain(
SyncEntityType.SshKey,
"a type the engine can reconcile but never requests would work in every unit test and never "
+ "sync");
}
[Fact]
public async Task AHostAndAKeyQueuedTogether_BothGoInOnePush()
{
var hostId = await harness.First.CreateAsync(Host("prod-db"));
var keyId = await harness.First.CreateKeyAsync(Key("deploy"));
await harness.First.SyncAsync();
harness.Server.PushCount.ShouldBe(1, "one outbox, one batch, whatever the item types in it");
await harness.Second.SyncAsync();
(await harness.Second.FindAsync(hostId)).Secret.Label.ShouldBe("prod-db");
(await harness.Second.FindKeyAsync(keyId)).Secret.Label.ShouldBe("deploy");
}
[Fact]
public async Task AKeyList_DoesNotShowHosts()
{
await harness.First.CreateAsync(Host("prod-db"));
await harness.First.CreateKeyAsync(Key("deploy"));
await harness.SettleAsync();
(await harness.Second.ListKeysAsync()).Items.ShouldHaveSingleItem()
.Secret.Label.ShouldBe("deploy");
(await harness.Second.ListAsync()).Items.ShouldHaveSingleItem()
.Secret.Label.ShouldBe("prod-db");
}
// ---- What the server is told ----
[Fact]
public async Task AKeyHandsTheServerNothingInPlaintext()
{
// The public half is supplied, which is the case where a fingerprint could have been derived and
// sent. The server has a column for one and would accept it; this client does not fill it, because a
// fingerprint is a stable identifier for a key pair and nothing in the product reads the column.
var entityId = await harness.First.CreateKeyAsync(
Key("deploy", publicKey: "ssh-ed25519 AAAAC3Nz deploy@laptop"));
var queued = await harness.First.Outbox
.FindAsync(VaultId, SyncEntityType.SshKey, entityId, Token);
queued.ShouldNotBeNull();
queued.Fields.ShouldBeNull("a key sends no plaintext fields at all, not an empty set of them");
await harness.SettleAsync();
var row = harness.Server.Find(entityId, SyncEntityType.SshKey).ShouldNotBeNull();
row.Fields.PublicKeyFingerprint.ShouldBeNull();
row.Fields.RelayEnabled.ShouldBeFalse();
row.Fields.Hostname.ShouldBeNull();
}
[Fact]
public async Task AHostAndAKeyWithTheSameId_AreDifferentItems()
{
// Not reachable through the repositories, which mint UUIDv7s, so it is arranged on the server. The
// point is that two things defend the separation independently: the item table is keyed on the type
// as well as the id, and the payload's AAD binds a resource type — so neither payload can be opened
// as the other even if a lookup did confuse them.
var sharedId = Guid.CreateVersion7();
harness.First.Keyring.TryGet(VaultId, out var vaultKey, out var generation).ShouldBeTrue();
harness.Server.ExternalUpsert(
sharedId,
HostCipher.Seal(Host("prod-db"), vaultKey.Span, sharedId, generation, itemVersion: 1),
new SyncPlaintextFields(),
SyncEntityType.Host);
harness.Server.ExternalUpsert(
sharedId,
SshKeyCipher.Seal(Key("deploy"), vaultKey.Span, sharedId, generation, itemVersion: 1),
null,
SyncEntityType.SshKey);
await harness.Second.SyncAsync();
var hosts = await harness.Second.ListAsync();
var keys = await harness.Second.ListKeysAsync();
hosts.Items.ShouldHaveSingleItem().Secret.Label.ShouldBe("prod-db");
keys.Items.ShouldHaveSingleItem().Secret.Label.ShouldBe("deploy");
hosts.Unreadable.ShouldBe(0);
keys.Unreadable.ShouldBe(0);
}
// ---- Merging ----
[Fact]
public async Task TwoMachinesEditingDifferentFieldsOfAKey_BothSurvive()
{
var entityId = await harness.First.CreateKeyAsync(Key("deploy", material: "SHARED"));
await harness.SettleAsync();
await harness.First.UpdateKeyAsync(entityId, Key("deploy-laptop", material: "SHARED"));
await harness.Second.UpdateKeyAsync(
entityId, Key("deploy", material: "SHARED", notes: "from the desktop"));
await harness.SettleAsync();
var first = (await harness.First.FindKeyAsync(entityId)).Secret;
var second = (await harness.Second.FindKeyAsync(entityId)).Secret;
first.ShouldBe(second);
first.Label.ShouldBe("deploy-laptop");
first.Notes.ShouldBe("from the desktop");
first.PrivateKeyPem.ShouldContain("SHARED");
(await ConflictKindsAsync()).ShouldBeEmpty();
}
[Fact]
public async Task BothReplacedTheKeyMaterial_NeitherKeyIsWrittenToTheConflictLog()
{
// The reason SshKeySecretMerge redacts. A host conflict records the value that lost so the user can
// put it back; doing that with a private key would copy a secret into a log that is designed to be
// read and is deliberately kept after acknowledgement.
var entityId = await harness.First.CreateKeyAsync(Key("deploy", material: "ORIGINAL"));
await harness.SettleAsync();
await harness.First.UpdateKeyAsync(entityId, Key("deploy", material: "LAPTOP-SECRET"));
await harness.Second.UpdateKeyAsync(entityId, Key("deploy", material: "DESKTOP-SECRET"));
await harness.SettleAsync();
(await ConflictKindsAsync()).ShouldContain(kind => kind == ConflictKind.FieldOverridden);
var details = await ConflictDetailsAsync();
details.ShouldContain(
detail => detail.Contains("PrivateKeyPem", StringComparison.Ordinal),
"the user still has to be told which field clashed");
foreach (var detail in details)
{
detail.ShouldNotContain("LAPTOP-SECRET");
detail.ShouldNotContain("DESKTOP-SECRET");
}
}
[Fact]
public async Task BothChangedThePassphrase_ThePassphraseIsNotInTheLogEither()
{
var entityId = await harness.First.CreateKeyAsync(Key("deploy", passphrase: "original"));
await harness.SettleAsync();
await harness.First.UpdateKeyAsync(entityId, Key("deploy", passphrase: "laptop-passphrase"));
await harness.Second.UpdateKeyAsync(entityId, Key("deploy", passphrase: "desktop-passphrase"));
await harness.SettleAsync();
foreach (var detail in await ConflictDetailsAsync())
{
detail.ShouldNotContain("laptop-passphrase");
detail.ShouldNotContain("desktop-passphrase");
}
}
// ---- Deletes, where key material can be lost ----
[Fact]
public async Task AKeyDeletedElsewhereWhileEditedHere_KeepsTheMaterialUnderANewName()
{
var entityId = await harness.First.CreateKeyAsync(Key("deploy", material: "IRREPLACEABLE"));
await harness.SettleAsync();
await harness.First.DeleteKeyAsync(entityId);
await harness.Second.UpdateKeyAsync(
entityId, Key("deploy", material: "IRREPLACEABLE", notes: "still in use"));
await harness.SettleAsync();
var restored = (await harness.First.ListKeysAsync()).Items.ShouldHaveSingleItem();
restored.EntityId.ShouldNotBe(entityId);
restored.Secret.Label.ShouldBe("deploy (restored)");
restored.Secret.Notes.ShouldBe("still in use");
restored.Secret.PrivateKeyPem.ShouldContain(
"IRREPLACEABLE", Case.Sensitive, "a resurrection that lost the key would rescue nothing");
(await ConflictKindsAsync()).ShouldContain(kind => kind == ConflictKind.RemoteDeleteResurrected);
}
[Fact]
public async Task AKeyEditedElsewhereAfterBeingDeletedHere_SurvivesAndIsCalledAKey()
{
var entityId = await harness.First.CreateKeyAsync(Key("deploy"));
await harness.SettleAsync();
await harness.First.UpdateKeyAsync(entityId, Key("deploy", notes: "still in use"));
await harness.Second.DeleteKeyAsync(entityId);
await harness.SettleAsync();
(await harness.First.FindKeyAsync(entityId)).Secret.Notes.ShouldBe("still in use");
(await ConflictKindsAsync()).ShouldContain(kind => kind == ConflictKind.LocalDeleteOverridden);
// The noun matters: someone told a host was edited elsewhere goes looking in the host list.
var details = await ConflictDetailsAsync();
details.ShouldContain(detail => detail.Contains("This SSH key was edited", StringComparison.Ordinal));
details.ShouldNotContain(detail => detail.Contains("This host was edited", StringComparison.Ordinal));
}
// ---- Helpers ----
private async Task<IReadOnlyList<ConflictKind>> ConflictKindsAsync()
{
var first = await harness.First.ConflictsAsync();
var second = await harness.Second.ConflictsAsync();
return [.. first.Concat(second).Select(conflict => conflict.Kind)];
}
/// <remarks>
/// The raw stored detail, decoded as text rather than parsed. The redaction claims are claims about what
/// is <em>absent</em> from the bytes, and reading through the JSON model would only prove the material
/// is absent from the fields the model happens to name.
/// </remarks>
private async Task<IReadOnlyList<string>> ConflictDetailsAsync()
{
var first = await harness.First.ConflictsAsync();
var second = await harness.Second.ConflictsAsync();
return
[
.. first.Concat(second)
.Select(conflict => System.Text.Encoding.UTF8.GetString(conflict.Detail)),
];
}
}