Switch from print_exported_headers to CMake's installed headers in scripts/build_apple_frameworks.sh

[swolchok]

This removes the need to use Buck in this script. The previous PR (#13485) adds enforcement that we don't install non-public headers from CMake.

I built the frameworks before and after. Here is a diff of all headers
present in the frameworks:
https://gist.github.com/swolchok/70647b551da7827a9cbdb083fad1e209

Generally, we now include more headers than before. The only header we
used to include that we no longer include is
schema/extended_header.h. It looks like we are not including the rest
of the headers for the PTE schema anyway, so I propose that we fix
installation of the schema headers separately.

[swolchok]

Stack from ghstack (oldest at bottom):

• Delete tools/cmake/{buck_util,resolve_buck}.py #13588
• -> Switch from print_exported_headers to CMake's installed headers in scripts/build_apple_frameworks.sh #13559
• install headers from runtime/backend #13558
• Add check_for_installed_private_headers_in_cmake_out #13485

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13559

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit 7f0e6fc with merge base 335de46 ():

NEW FAILURES - The following jobs have failed:

• Build documentation / build (buck2) / Build doc (gh)
  At least one of the pre-conditions you specified did not hold
• pull / test-llava-runner-linux / linux-job (gh)
  RuntimeError: Command docker exec -t e25d35fbc2d09e519dbadbd889bc3ea1385651a6b19ff342fe46fc94f3e5ebf1 /exec failed with exit code 139

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[shoumikhin]

There are a lot of headers for the code not shipped with the executorch runtime framework, so anyone using them by mistake will run into missing symbols error.
E.g. extension/llm headers technically belong to a different framework executorch_llm (although we currently don't expose them in favor of the native ObjC/Swift APIs).
Let's try to keep the set of headers we ship to the minimum and make sure we do have the corresponding implementation in .a we ship. Ideally, nothing should change.