The speaker highlights the "new tests" command as a favorite tool for ensuring code quality. After making changes, particularly in the context of fixing a font size issue in the Solo desktop application's terminal, the speaker runs "new tests" followed by "finalize" (covered in another video).
The "new tests" command focuses on recent work (git diffs, unstaged changes) and prioritizes the current conversation, identifying potential gaps in test coverage. In the example, while fixing the font size bug, the speaker also addressed similar issues with line height, letter spacing, font family, and font weight, but initially only wrote tests for font size. The "new tests" command recognized this discrepancy and suggested adding tests for the other affected properties, preventing future regressions.
The speaker emphasizes the importance of iterative development: building, manually verifying, and then adding tests to lock in working functionality. This contrasts with an autonomous loop that might generate tests without human validation. The speaker believes in "keeping hands on the wheel," building and refining iteratively before committing, using "new tests" to ensure thorough coverage before the "finalize" command.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.