Unconditionally register schema even for manual registration.

[ezyang]

Stack from ghstack:

• Functionalize qconcat and register with new registration API #36536 Functionalize qconcat and register with new registration API
• Convert qclamp to new operator registration API. #36535 Convert qclamp to new operator registration API.
• Convert qbatch_norm to new operator registration API. #36534 Convert qbatch_norm to new operator registration API.
• Update quantized README for registering operators with new API. #36531 Update quantized README for registering operators with new API.
• Functionalize qadd and register with new registration API. #36527 Functionalize qadd and register with new registration API.
• Add a centralized TORCH_LIBRARY declaration for quantized and xnnpack ops. #36520 Add a centralized TORCH_LIBRARY declaration for quantized and xnnpack ops.
• Add explicit schema for quantized conv/conv_prepack #36513 Add explicit schema for quantized conv/conv_prepack
• Port detach/detach to new registration API #36512 Port detach/detach to new registration API
• Make it an error to TORCH_LIBRARY the same namespace twice. #36505 Make it an error to TORCH_LIBRARY the same namespace twice.
• Port Resize to use new registration API. #36390 Port Resize to use new registration API.
• Port to new registration API (part 1) #36389 Update reference to RegisterOperators in error message in Convolution
• Propagate and report the file/line number of registrations on errors. #36264 Propagate and report the file/line number of registrations on errors.
• Switch to pybind11 style registration function API. #36258 Switch to pybind11 style registration function API.
• Unconditionally register schema even for manual registration. #36250 Unconditionally register schema even for manual registration.

The general concept is that I want a centralized location where you
can find all of the registrations for a library. I cannot do this
if I don't codegen all of the schemas in one spot--right now,
most schemas get generated, but not manually registered ones. Let us
assume that manual registration has to do with the actual
implementations; nothing strange is going on with the schema
definition itself. Make it so.

Signed-off-by: Edward Z. Yang ezyang@fb.com

Differential Revision: D20929258

[dr-ci]

💊 Build failures summary and remediations

As of commit fc81e9b (more details on the Dr. CI page):

---

✅ None of the build failures appear to be your fault 💚

---

• 1/1 broken upstream at merge base fb70b4f since Apr 14

  Please rebase on the viable/strict branch (expand for instructions)

  If your commit is newer than viable/strict, you can try basing on an older, stable commit:

  git fetch https://github.com/pytorch/pytorch viable/strict
  git rebase --onto FETCH_HEAD $(git merge-base origin/master HEAD)
  

  If your commit is older than viable/strict:

  git fetch https://github.com/pytorch/pytorch viable/strict
  git rebase FETCH_HEAD
  

  Check out the recency history of this "viable master" tracking branch.

---

🚧 1 upstream failure:

These were probably caused by upstream breakages:

• pytorch_linux_backward_compatibility_check_test since Apr 14
  • 🔁 rerun

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker.

See how this bot performed.

This comment has been revised 32 times.

[facebook-github-bot]

@ezyang merged this pull request in 5150334.