New procedure defines rules for freezing flaky tests based on run disagreements

·1 views

DEV Community outlines a technical procedure for handling flaky software tests. The process requires documenting a two-run disagreement on a specific fixture snapshot to justify skipping a test. Six mandatory fields must be validated, including test identifiers and file hashes, before a freeze record is approved. The system is designed to prevent stable bugs from being quarantined while allowing genuinely flaky tests to be temporarily skipped. The record must be stored in a specific file format for consistent review by both automated systems and human reviewers.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Tutorial demonstrates API testing using GitHub's public user data

A technical tutorial outlines how to test APIs using GitHub's public user information endpoint. The guide provides a practical example using curl to fetch data without authentication requirements. It explains how to retrieve specific JSON fields including username, ID, and profile links from the API response. The article emphasizes that API testing is a fundamental skill for software developers working with web services.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

New app Luna uses AI and astronomy to guide stargazers, then tells them to put phone away

Luna is a new stargazing application designed to minimize screen time by encouraging users to look at the actual night sky. The app, built for mobile devices, uses the on-device Gemma language model to answer questions about celestial objects. Every answer is cross-checked for factual accuracy against a real-time astronomy engine that calculates positions based on the user's location and time. Luna provides features like dark sky site navigation and personalized alerts, but its core function ends with prompting the user to put their phone down and observe. It was recently tested in Bangalore, where it successfully identified stars and planets.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Software development guide advises freezing test outputs before refactoring code

A software development article outlines a method for safely refactoring legacy code. The approach involves first creating a set of nine synthetic test cases that define expected outputs. These test cases must pass before any code changes are made to ensure behavior is preserved. The article uses a specific function that maps payment webhook payloads as a concrete example. This technique aims to prevent accidental introduction of bugs during code maintenance.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

OpenRouter's Free Tier Allows Testing of AI Model Endpoints

A recent tutorial explains how to test a free AI model endpoint using OpenRouter's platform. The process requires a free API key and the use of curl commands to send requests to the chat completion endpoint. The guide demonstrates sending a query to a specific model and parsing the JSON response to extract the AI's answer. This method is presented as a cost-effective way for developers to perform demos or quick checks, with reminders to adhere to rate limits.

0 comments Read more at DEV Community