No longer able to build from source with `pip install -v .`

🐛 Describe the bug

After PR https://github.com/pytorch/vision/pull/4384, I'm no longer able to build torchvision with

pip install -v .

, though

python setup.py develop

still works.

Is this the intended behavior now?

The pip install -v . now reports error message

xxx@yyy:/opt/pytorch/vision# pip install -v .
Using pip 21.2.2 from /opt/conda/lib/python3.8/site-packages/pip (python 3.8)
Processing /opt/pytorch/vision
  DEPRECATION: A future pip version will change local packages to be built in-place without first copying to a temporary directory. We recommend you use --use-feature=in-tree-build to test your packages with this new behavior before it becomes the default.
   pip 21.3 will remove support for this functionality. You can find discussion regarding this at https://github.com/pypa/pip/issues/7555.
  Running command /opt/conda/bin/python /tmp/pip-standalone-pip-qzzditq9/__env_pip__.zip/pip install --ignore-installed --no-user --prefix /tmp/pip-build-env-klifg95l/overlay --no-warn-script-location --no-binary :none: --only-binary :none: -i https://pypi.org/simple -- 'setuptools>=40.8.0' wheel
  Collecting setuptools>=40.8.0
    Using cached setuptools-58.2.0-py3-none-any.whl (946 kB)
  Collecting wheel
    Using cached wheel-0.37.0-py2.py3-none-any.whl (35 kB)
  Installing collected packages: wheel, setuptools
  Successfully installed setuptools-58.2.0 wheel-0.37.0
  WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
  Installing build dependencies ... done
  Running command /opt/conda/bin/python /opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py get_requires_for_build_wheel /tmp/tmptw3tr8z6
  Traceback (most recent call last):
    File "/opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 349, in <module>
      main()
    File "/opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 331, in main
      json_out['return_val'] = hook(**hook_input['kwargs'])
    File "/opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py", line 117, in get_requires_for_build_wheel
      return hook(config_settings)
    File "/tmp/pip-build-env-klifg95l/overlay/lib/python3.8/site-packages/setuptools/build_meta.py", line 154, in get_requires_for_build_wheel
      return self._get_build_requires(
    File "/tmp/pip-build-env-klifg95l/overlay/lib/python3.8/site-packages/setuptools/build_meta.py", line 135, in _get_build_requires
      self.run_setup()
    File "/tmp/pip-build-env-klifg95l/overlay/lib/python3.8/site-packages/setuptools/build_meta.py", line 258, in run_setup
      super(_BuildMetaLegacyBackend,
    File "/tmp/pip-build-env-klifg95l/overlay/lib/python3.8/site-packages/setuptools/build_meta.py", line 150, in run_setup
      exec(compile(code, __file__, 'exec'), locals())
    File "setup.py", line 12, in <module>
      import torch
  ModuleNotFoundError: No module named 'torch'
  Getting requirements to build wheel ... error
WARNING: Discarding file:///opt/pytorch/vision. Command errored out with exit status 1: /opt/conda/bin/python /opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py get_requires_for_build_wheel /tmp/tmptw3tr8z6 Check the logs for full command output.
ERROR: Command errored out with exit status 1: /opt/conda/bin/python /opt/conda/lib/python3.8/site-packages/pip/_vendor/pep517/in_process/_in_process.py get_requires_for_build_wheel /tmp/tmptw3tr8z6 Check the logs for full command output.

Versions

I'm using the latest pytorch version 1.11.0a0+git458a00b https://github.com/pytorch/pytorch/commit/458a00bacb68998a75a87af153e6ae3690698bc2,

OS: Ubuntu 20.04.3 LTS (x86_64)
GCC version: (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
Clang version: Could not collect
CMake version: version 3.19.6
Libc version: glibc-2.31

Python version: 3.8.11 (default, Aug  3 2021, 15:09:35)  [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-4.15.0-123-generic-x86_64-with-glibc2.17
Is CUDA available: True

cc @ptrblck @pmeier @NicolasHug

Reproduce the failure with pip install -v . and compare it with python setup.py develop. Start in setup.py, especially the top-level import torch, and review the build changes from PR 4384. Done means the source installation path no longer fails with ModuleNotFoundError: No module named 'torch', while the existing development install remains functional.