Is is a harbinger of finally devendored libffi? Currently, CPython ships 1.20 (1998) while Debian LTS supplies 3.2.1 (2014).

libffi was de-vendored in 3.7 (Linux, BSD) and 3.8 (Windows). We only keep a copy for macos around.

🤖 New build scheduled with the buildbot fleet by @tiran for commit cc5cb24 🤖

If you want to schedule another build, you need to add the "🔨 test-with-buildbots" label again.

ctypes tests are failing on ARM64 macOS

I have no clue how the SDK stuff works on macOS. In order to land the PR I would need some assistance from @ned-deily or @ronaldoussoren . Or can we drop the SDK and use system-ffi on all system?

🤖 New build scheduled with the buildbot fleet by @tiran for commit 54e6b11 🤖

If you want to schedule another build, you need to add the "🔨 test-with-buildbots" label again.

I'll try to look at the PR over the weekend.

As a general statement I'd prefer to use the system install of libffi on macOS because Apple includes some magic in its version to ensure libffi closures work properly with their code signature system (in particular some magic to ensure that you don't need the "JIT" entitlement to use closures).

🤖 New build scheduled with the buildbot fleet by @tiran for commit 53607e3 🤖

If you want to schedule another build, you need to add the "🔨 test-with-buildbots" label again.

🤖 New build scheduled with the buildbot fleet by @tiran for commit e4ab8cf 🤖

If you want to schedule another build, you need to add the "🔨 test-with-buildbots" label again.

Users without pkg-config can fall back to LIBFFI_CFLAGS and LIBFFI_LIBS env vars, too.

🤖 New build scheduled with the buildbot fleet by @tiran for commit d854d7f 🤖

If you want to schedule another build, you need to add the "🔨 test-with-buildbots" label again.

@ned-deily @ronaldoussoren I would like to merge this PR rather sooner than later into main. It is required to test ctypes on wasm32-emscripten and will allow us to remove setup.py. Tests are passing on macOS.

Status check is done, and it's a success ✅ .