Skip to content

Latest commit

 

History

History
159 lines (122 loc) · 6.46 KB

CONTRIBUTING.md

File metadata and controls

159 lines (122 loc) · 6.46 KB
Raw

Table of Contents

Pull request workflow

  • assuming you are working with git from a command line
  • assuming your GitHub username is username
  • github.com/sappelhoff/pyprep is upstream
  • github.com/username/pyprep is origin (your fork)

Syncing your fork's main with upstream main

  • first, you start with forking upstream
  • then, you continue by cloning your fork: git clone https://github.com/username/pyprep
    • you'll have your own main branch there
  • you always want to make sure that your fork's main and upstream main are aligned
    • to do this, you work with git remotes
    • Note: this also means that you NEVER work on main (unless you know what you are doing) ... because you want to always be able to SYNC your fork with upstream, which would mean losing your own work on main
  • use git remote -v to list your configured remotes
    • initially this will only list origin ... a bit like this maybe:
origin	https://github.com/username/pyprep (fetch)
origin	https://github.com/username/pyprep (push)
  • Now you want to add upstream as a remote. Use git remote add upstream https://github.com/sappelhoff/pyprep
  • again, do git remote -v, it should look like this:
origin	https://github.com/username/pyprep (fetch)
origin	https://github.com/username/pyprep (push)
upstream	https://github.com/sappelhoff/pyprep (fetch)
upstream	https://github.com/sappelhoff/pyprep (push)
  • Now you can use your upstream remote to make sure your fork's main is up to date.
    1. git checkout main to make sure you are on your main branch
    2. Make sure you do not have any changes on your main, because we will discard them!
    3. git pull upstream main SYNC your fork and upstream
    4. sometimes there are issues, so to be safe, do: git reset --hard upstream/main ... this makes sure that both branches are really synced.
    5. ensure with another git pull upstream main ... this should say "already up to date"

Working on a feature (and rebasing)

Working on a feature

  • before working on any feature: always do git checkout main and git pull upstream main
  • then make your new branch to work on and check it out, for example git checkout -b my_feature
    • do your work
    • submit a pull request
    • hope you are lucky and nobody did work in between
  • however IF somebody did work in between, we need to rebase. Just follow the steps below

Rebasing without conflicts

  1. sync main through: git checkout main and git pull upstream main
  2. go back to your branch and rebase it: git checkout my_feature and then git rebase main

Now it could be that you are lucky and there no conflicts ... in that case, the rebase just works and you can then finish up by force pushing your rebased branch: git push -f my_feature ... you need to force it, because rebasing changed the history of your branch. But don't worry, if rebasing "just worked" without any conflicts, this should be very safe.

Rebasing WITH conflicts

In case you are unlucky, there are conflicts and you'll have to resolve them step by step ... git will be in rebase mode and try to rebase one commit after another ... for each commit where conflicts are detected, it'll stop.

Then you have to do: git status to see conflicting files ... then edit these files to resolve conflicts ... then git add <filename> ... and then git rebase --continue to go on to the next commit, rinse and repeat.

NOTE: the conflict resolution part is the dangerous part that can get very messy and where you can actually lose stuff ... so make backups of your branch before.

After everything is resolved, you can again do git push -f my_feature.

If you screw up during rebasing and you panic, you can do git rebase --abort and start again.

Rebasing ... panic mode

If nothing helps and you just don't know how to resolve the issues and conflicts that arise during rebasing, just make a new branch: 1. git checkout main 1. git pull upstream main 1. git checkout -b my_feature_2nd_attempt

... and apply your changes manually.

This method is not really a git workflow, ... but in cases where there are only few changes, this is often a practical solution.

Info about versioning

We follow a semantic versioning scheme. This is implemented via hatch-vcs.

Info about docs

The documentation is build and hosted by https://readthedocs.org/.

Admin credentials are needed to access the setup.

How to make a release

Follow this workflow:

  1. go to your python environment for pyprep
  2. make sure all tests pass and the docs are built cleanly
  3. update docs/changelog.rst, renaming the "current" headline to the new version and updating the "Authors" section. "Authors" are all people who committed code or in other ways contributed to pyprep (e.g., by reviewing PRs, moderating discussions).
  4. commit the change and git push to main (or make a pull request). Start your commit message with [REL].
  5. make an annotated tag git tag -a -m "1.2.3" 1.2.3 upstream/main (This assumes that you have a git remote configured with the name "upstream" and pointing to https://github.com/sappelhoff/pyprep).
  6. git push --follow-tags upstream
  7. make a release on GitHub, using the git tag from the previous step (e.g., 1.2.3). Fill the tag name into all fields of the release.

Then the release is done and main has to be prepared for development of the next release:

  1. add a "current" headline to docs/changelog.rst
  2. commit the changes and git push to main (or make a pull request)