Releasing PyInstaller

How to put PyInstaller on PyPI

Prerequisites

• It's assumed that you are running in a UNIX like environment with tools like bash and perl available.

• It's also assumed that your Python environment is suitably de-xkcd-1987-ified. i.e which python points to a sensible, non-EOL version of Python 3 with user-writable site-packages directory and which pip points to that same environment.

• You need to install and be running fish. (Just entering fish into a terminal is enough, no need to chsh).

• You need docker with qemu (cross architecture emulation). These are typically just called docker and qemu on Linux repositories. Windows and macOS can get them via docker desktop. To test your setup run:

  $ docker run --rm --privileged multiarch/qemu-user-static --reset -p yes
  ... lots of noise ...
  $ docker run --rm --platform=ppc64le alpine uname -m
  ... some more noise ...
  ppc64le

To release PyInstaller to PyPI

The steps to do a release are below. Note that only explicitly typing the commands pyi_upload_to_pypi or pyi_github_release or a git push will make persistent online changes. All other steps are harmless in that they apply changes locally only and can be undone by running git reset --hard origin/develop. It should be impossible to accidentally create a release or other irreversible change just by monkeying around.

1. Ensure that you are on the develop branch, your working directory is clean and you have pulled the latest upstream changes.

2. cd to the root of this repo and enter fish:

   $ fish

3. Source the release.fish script.

   > source release/release.fish

   That script should prompt you for everything you need to do from there on.