This script guesses the numbers for the merged pull requests from the last tag, then fetches the corresponding information from GitHub and formats it into markdown format suitable for use in e.g. Discourse.

You should invoke this script from the toplevel folder of your repo, and redirect the output to a file, or copy/paste the output as required:

$ python fetch_changelog > changelog.md

What it does

This script relies on git to detect your currently checked-out branch, its upstream remote, the URL of that remote, and verifies that it matches 'github.com' (it is designed to fetch the issue list from GitHub using its API, and will not work with other git hosting providers).

It will then find the most recent signed tag on that branch, and find all merge commits that start with 'Merge pull request #'. It will extract the pull request number from these commit messages, and then use the GitHub API to fetch the JSON entries for each of these pull requests, from which it will extract the title, user and body of the pull request, which will be printed on stdout in markdown.

Command-line options

• --tag=XX: Override the automatic detection of the base tag, and specify it manually. The script will then fetch all pull requests that have been merged after that tag.

• --no-fetch: Report all pull request IDs detected, but do not use the GitHub API to fetch any further information. This is useful for testing, to avoid hitting GitHub API rate limits.

GitHub Authentication

If necessary (for example, if your repo is private, or you encounter issues with rate limiting), you can provide GitHub authentication information in your own fetch_changelog_settings file, which should be created in the toplevel of your repo folder, with the correct information:

user = 'jdtournier'
password = 'gngkasgndfgknldsafnfknxgkjd'

where password above should contain your personal authentication token, as generated from your GitHub acccount.