Controlling package uploads

The command flit publish will upload your package to a package index server. The default settings let you upload to PyPI, the default Python Package Index, with a single user account.

If you want to upload to other servers, or with more than one user account, or upload packages from a continuous integration job, you can configure Flit in two main ways:

Using .pypirc

You can create or edit a config file in your home directory, ~/.pypirc that will be used by default or you can specify a custom location. This is also used by other Python tools such as twine.

For instance, to upload a package to the Test PyPI server instead of the normal PyPI, use a config file looking like this:

[distutils]
index-servers =
   pypi
   testpypi

[pypi]
repository = https://upload.pypi.org/legacy/
username = sirrobin  # Replace with your PyPI username

[testpypi]
repository = https://test.pypi.org/legacy/
username = sirrobin  # Replace with your TestPyPI username

You can select an index server from this config file with the --repository option:

flit publish --repository testpypi

If you don’t use this option, Flit will use the server called pypi in the config file. If that doesn’t exist, it uploads to PyPI at https://upload.pypi.org/legacy/ by default.

If you publish a package and you don’t have a .pypirc file, Flit will create it to store your username.

Flit tries to store your password securely using the keyring library. If keyring is not installed, Flit will ask for your password for each upload. Alternatively, you can also manually add your password to the .pypirc file (password = ...)

Using environment variables

You can specify a server to upload to with FLIT_INDEX_URL, and pass credentials with FLIT_USERNAME and FLIT_PASSWORD. Environment variables take precedence over the config file, except if you use the --repository option to explicitly pick a server from the config file.

This can make it easier to automate uploads, for example to release packages from a continuous integration job.

警告

Storing a password in an environment variable is convenient, but it’s easy to accidentally leak it. Look out for scripts that helpfully print all environment variables for debugging, and remember that other scripts and libraries you run in that environment have access to your password.