What-Happens-When-a-Job-Is-Reproduced.md

What Happens When a Job Is Reproduced

Below is a high-level overview of what happens when a job is reproduced. The code for these steps are located in {travis,github}-reproducer/pipeline/. Read the source comments in that directory for details that are not described here, including checking for skipping, storing temporary files for reuse, etc.

Step 1 - Generate Files

1. Copy and reset the repository: The repository is copied to a temporary directory and reset to the commit hash of the job to be reproduced.
2. Generate the build script: Based on the job's configuration, a shell script is generated that will be used to build the project.
3. Generate the Dockerfile: A Dockerfile is generated that will be used to build the Docker image. The base image is determined by the job's configuration.

Step 2 - Build the Docker Image and Spawn the Docker Container

The Docker image is built using the generated Dockerfile.

Step 3 - Examine the Reproducibility

The reproduced log is analyzed by Analyzer and compared with the original log.