Pillow currently has a Pyroma test that is implemented as part of the runtime test suite.
This test needed some small tweaks to work on iOS (See #9116). I have tried to get those changes incorporated upstream to make the test less fragile for future PyRoma changes, but the maintainer has determined that using Pyroma as a library is out of scope: see regebro/pyroma#116.
So - Pillow should restructure the Pyroma test so that it isn't implemented as a runtime test, but as a validation of the actual generated wheel using PyRoma as an "external tool", rather than a library.
What are your OS, Python and Pillow versions?
The issue to be resolved is with the Pillow test suite; the relevant test has been in the current form since at least Pillow 11.3.
Pillow currently has a Pyroma test that is implemented as part of the runtime test suite.
This test needed some small tweaks to work on iOS (See #9116). I have tried to get those changes incorporated upstream to make the test less fragile for future PyRoma changes, but the maintainer has determined that using Pyroma as a library is out of scope: see regebro/pyroma#116.
So - Pillow should restructure the Pyroma test so that it isn't implemented as a runtime test, but as a validation of the actual generated wheel using PyRoma as an "external tool", rather than a library.
What are your OS, Python and Pillow versions?
The issue to be resolved is with the Pillow test suite; the relevant test has been in the current form since at least Pillow 11.3.