liveBook · Manning

End-to-end testing

As explained in Claude at Work

An end-to-end test verifies an application flow across its user interface, services, and persistent data.

End-to-end tests drive a browser as a user would, then query the database to verify writes. They wipe the database to establish a known state, so run them against a separate stack and database to protect exploratory data. section 6.8

Listing 6.5 Part of the Gherkin file for the channels feature
@bdd
Feature: Channels

  Conversation in CollabHub happens in channels. A channel belongs to
  exactly one workspace and a public channel is visible to everyone in
  that workspace whether or not they have joined it. Joining is what
  will gate reading messages, from a later slice; it does not gate
  seeing that a channel exists.

  ...

  Background:
    Given Ada is signed in

  Scenario: Ada creates a public channel and lands in it
    When Ada creates a public channel named "general"
    Then Ada is looking at the "general" channel
    And "general" is in Ada's channel list

A Gherkin scenario describes creating a public channel and checking that the user lands in it and sees it in the channel list. section 6.8

How Claude at Work covers it

How deep it goes
Explains browser interaction, database checks, clean-state requirements, and separate test containers.
Examples
Tests Ada creating a public channel, landing in it, and seeing it in her channel list.
What you can do after
Operate tests against a separate stack and database that can be reset without losing exploratory data.
How it connects
Connects BDD tests with Gherkin scenarios, browser interaction, Docker Compose, and database verification.

Also explained in

Compare how 17 books explain end-to-end testing →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Claude at Work. The section links point to the parts of the book it draws on.