Recipe conflict between pip and conda

[amgcc]

There is a mismatch on the cffi requirement between the conda and pip recipes. Is this intentional and if not, can they be updated to match? As a result, my rattler packaging fails on a pip check as the dependencies can't resolve. I have other packages I'm using that require cffi < 2. I would prefer having to manually pin pynacl < 1.6.1 and if the current version was marked to require cffi >= 2 that would fix the issue for me..

The pip package requirements are as follows:

[build-system]
# Must be kept in sync with `dependencies` below
requires = [
    "setuptools>=61.0.0,!=74.0.0",
    "wheel",
    "cffi>=1.4.1; platform_python_implementation != 'PyPy' and python_version < '3.9'",
    "cffi>=2.0.0; platform_python_implementation != 'PyPy' and python_version >= '3.9'",
]

Conda requirements are as follows:

host:
    - _libgcc_mutex 0.1 conda_forge
    - _openmp_mutex 4.5 2_gnu
    - bzip2 1.0.8 hda65f42_8
    - ca-certificates 2025.11.12 hbd8a1cb_0
    - cffi 2.0.0 py311h03d9500_1
...
  run:
    - __glibc >=2.17,<3.0.a0
    - cffi >=1.4.1
    - libgcc >=14
    - libsodium >=1.0.20,<1.0.21.0a0
...

[alex]

We have no control of the conda requirements, if they're a problem, you'll need to complain to the conda upstream.