[PM-37886] refactor: Update Key Connector vault unlock to use SDK's keyConnectorUrl method

[matt-livefront]

🎟️ Tracking

PM-37886

📔 Objective

Migrates the Key Connector vault unlock to let the SDK handle Key Connector communication, removing the app's direct GET /user-keys network call.

Changes

• AuthRepository.unlockVaultWithKeyConnectorKey: switches from InitUserCryptoMethod.keyConnector(masterKey:userKey:) (which required a prior getMasterKeyFromKeyConnector API call) to InitUserCryptoMethod.keyConnectorUrl(url:keyConnectorKeyWrappedUserKey:), which delegates the Key Connector fetch to the SDK.
• Removes getMasterKeyFromKeyConnector from KeyConnectorService and KeyConnectorAPIService protocols and implementations.
• Deletes KeyConnectorUserKeyRequest and KeyConnectorUserKeyResponseModel.

[github-actions]

🤖 Bitwarden Claude Code Review

Overall Assessment: APPROVE

Reviewed the Key Connector vault-unlock refactor that delegates the Key Connector fetch to the SDK via InitUserCryptoMethod.keyConnectorUrl, removing the app's direct GET /user-keys call along with the now-unused request/response models and service methods. Verified the SDK method is wired correctly, all removed symbols have no remaining references (the retained Post* types are the separate migration path), and the SingleSignOnProcessor loading-overlay handling correctly hides the overlay on every branch including the cancel path. Test updates are consistent with the code changes.

Code Review Details

No findings. This is a clean, well-scoped tech-debt refactor:

• The .keyConnectorUrl(url:keyConnectorKeyWrappedUserKey:) migration preserves unlock semantics; keyConnectorURL.absoluteString is passed correctly.
• Dead code removal is complete — no orphaned references to getMasterKeyFromKeyConnector, KeyConnectorUserKeyRequest, KeyConnectorUserKeyResponseModel, or the keyConnectorUserKey fixture.
• The loading-overlay stuck-on-cancel issue from earlier review threads is resolved; the remaining open thread is tracked via PM-38748.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.82%. Comparing base (f5067ec) to head (22b8a42).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2748      +/-   ##
==========================================
+ Coverage   80.80%   80.82%   +0.01%     
==========================================
  Files        1019     1018       -1     
  Lines       64849    64840       -9     
==========================================
+ Hits        52402    52407       +5     
+ Misses      12447    12433      -14     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[KatherineInCode]

🎨 To one of the things Claude generated for this, is there a way we could test to make sure the overlay is hidden here?

🤔 Would the hiding of the overlay be something we could defer?

[matt-livefront]

🎨 To one of the things Claude generated for this, is there a way we could test to make sure the overlay is hidden here?

Yep, I updated the tests for this.

🤔 Would the hiding of the overlay be something we could defer?

That would be ideal. At one point we ran into a problem where defer was causing the loading overlay to not hide when transitioning from the auth flow to the vault, which is why they are manually hidden in some places like this before completing auth. I'd like to clean this up, but I'd prefer to do it separately. I created a ticket: https://bitwarden.atlassian.net/browse/PM-38748.

[github-actions]

Warning

@matt-livefront Uploading code coverage report failed. Please check the "Upload to codecov.io" step of Process Test Reports job for more details.