Acuant SDK Test Plan

General Information

Automated testing of the Acuant SDK, used for capturing pictures of a user’s ID and face, is not currently available. Prior to each bi-weekly deploy, Team Timnit wants to manually verify that the SDK is functioning as expected.

Cadence

When to test: ✅

Manual testing should be done prior to the regular full deploy of IDP to prod. Typically we do a full deploy twice weekly, on Tuesdays and Thursdays - see Deploying new IDP and PKI code for details and up-to-date cadence.

Test Procedure

  • Visit the Environments status page of the Dashboard.
    • Under Staging, click on either the oidc-sinatra or the saml-sinatra links to go to a sample app.
    • Note the git sha of the idp in staging.
  • From the sample app, under Options > Level of Service, select Biometric Comparison, then click Sign in.
  • From the Login.gov sign in page, choose Create an account, enter your email with a modifier to make it unique, e.g., first.last+yymmdd@gsa.gov, and create your new account.
  • Using a mobile device or the hybrid handoff flow, go through identity verification up through document/selfie capture and note any issues with the SDK. Things to look for may include:
    • Does the SDK start?
    • Does the SDK use the full screen?
    • Does capture happen automatically?
    • Does voice over work as expected?
    • Is your ID successfully verified?
    • Is the new feature we’re deploying working properly?

Communications

Slack: @login-oncall-timnit receives a reminder, in #login-team-timnit on Tuesday and Thursday mornings, to test the release prior to the day’s deploy (others are welcome to test as well!). When conducting the test, react to the reminder with 👀 and reply in thread with:

  • The mobile platform used (OS, browser, version)
  • The path to document capture
    • hybrid (start on desktop, switch to phone for capture)
    • standard (start on phone)
  • The git sha of the idp tested (available from the Environment status page)
  • Any issues encountered!

If an issue is found, we need to work quickly to triage and determine if the problem is being introduced in the current deploy or if it is an existing bug. If new, work with @login-deployer to try and revert the breaking changes prior to deployment. If old, file a bug in Jira and share on the team’s channel for visibility.