Separate clang lint? (#119575)

25 min -> 17 + 13 min, which is still not as fast as I want it to be but I'll take it
Lintrunner provides some parallelism by default, but it's not perfect

Reducing fetch-depth from all to 1 further reduces time by ~2-3 minutes

From non clang's logs:
```
2024-02-09T22:05:39.5297616Z Requirement already satisfied: PyYAML==6.0 in /opt/conda/lib/python3.11/site-packages (6.0)
2024-02-09T22:12:23.6164708Z Collecting black==23.12.1
```
I don't know why this part takes so long, maybe it's just buffering?  Clang version doesn't show this issue

See 5a750c8035
Pull Request resolved: https://github.com/pytorch/pytorch/pull/119575
Approved by: https://github.com/huydhn, https://github.com/malfet

.github/scripts/lintrunner.sh

@@ -0,0 +1,43 @@
+#!/usr/bin/env bash
+set -x
+
+# The generic Linux job chooses to use base env, not the one setup by the image
+CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
+conda activate "${CONDA_ENV}"
+
+CACHE_DIRECTORY="/tmp/.lintbin"
+# Try to recover the cached binaries
+if [[ -d "${CACHE_DIRECTORY}" ]]; then
+    # It's ok to fail this as lintrunner init would download these binaries
+    # again if they do not exist
+    cp -r "${CACHE_DIRECTORY}" . || true
+fi
+
+# This has already been cached in the docker image
+lintrunner init 2> /dev/null
+
+# Do build steps necessary for linters
+if [[ "${CLANG}" == "1" ]]; then
+    python3 -m tools.linter.clang_tidy.generate_build_files
+fi
+python3 -m tools.generate_torch_version --is_debug=false
+python3 -m tools.pyi.gen_pyi \
+    --native-functions-path aten/src/ATen/native/native_functions.yaml \
+    --tags-path aten/src/ATen/native/tags.yaml \
+    --deprecated-functions-path "tools/autograd/deprecated.yaml"
+
+RC=0
+# Run lintrunner on all files
+if ! lintrunner --force-color --all-files --tee-json=lint.json ${ADDITIONAL_LINTRUNNER_ARGS} 2> /dev/null; then
+    echo ""
+    echo -e "\e[1m\e[36mYou can reproduce these results locally by using \`lintrunner -m origin/main\`. (If you don't get the same results, run \'lintrunner init\' to update your local linter)\e[0m"
+    echo -e "\e[1m\e[36mSee https://github.com/pytorch/pytorch/wiki/lintrunner for setup instructions.\e[0m"
+    RC=1
+fi
+
+# Use jq to massage the JSON lint output into GitHub Actions workflow commands.
+jq --raw-output \
+    '"::\(if .severity == "advice" or .severity == "disabled" then "warning" else .severity end) file=\(.path),line=\(.line),col=\(.char),title=\(.code) \(.name)::" + (.description | gsub("\\n"; "%0A"))' \
+    lint.json || true
+
+exit $RC

.github/workflows/lint.yml

@@ -15,53 +15,32 @@ permissions: read-all
 # The names of steps that actually test the code should be suffixed with `(nonretryable)`.
 # When any other step fails, it's job will be retried once by retryBot.
 jobs:
-  lintrunner:
+  lintrunner-clang:
     uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
     with:
       timeout: 120
       runner: linux.2xlarge
       docker-image: pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter
-      fetch-depth: 0
+      fetch-depth: 1
+      submodules: true
       ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
       script: |
-        # The generic Linux job chooses to use base env, not the one setup by the image
-        CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
-        conda activate "${CONDA_ENV}"
+        export ADDITIONAL_LINTRUNNER_ARGS="--take CLANGTIDY,CLANGFORMAT"
+        export CLANG=1
+        .github/scripts/lintrunner.sh
 
-        CACHE_DIRECTORY="/tmp/.lintbin"
-        # Try to recover the cached binaries
-        if [[ -d "${CACHE_DIRECTORY}" ]]; then
-          # It's ok to fail this as lintrunner init would download these binaries
-          # again if they do not exist
-          cp -r "${CACHE_DIRECTORY}" . || true
-        fi
-
-        # This has already been cached in the docker image
-        lintrunner init 2> /dev/null
-
-        # Do build steps necessary for linters
-        python3 -m tools.linter.clang_tidy.generate_build_files
-        python3 -m tools.generate_torch_version --is_debug=false
-        python3 -m tools.pyi.gen_pyi \
-          --native-functions-path aten/src/ATen/native/native_functions.yaml \
-          --tags-path aten/src/ATen/native/tags.yaml \
-          --deprecated-functions-path "tools/autograd/deprecated.yaml"
-
-        RC=0
-        # Run lintrunner on all files
-        if ! lintrunner --force-color --all-files --tee-json=lint.json 2> /dev/null; then
-          echo ""
-          echo -e "\e[1m\e[36mYou can reproduce these results locally by using \`lintrunner -m origin/main\`. (If you don't get the same results, run \'lintrunner init\' to update your local linter)\e[0m"
-          echo -e "\e[1m\e[36mSee https://github.com/pytorch/pytorch/wiki/lintrunner for setup instructions.\e[0m"
-          RC=1
-        fi
-
-        # Use jq to massage the JSON lint output into GitHub Actions workflow commands.
-        jq --raw-output \
-          '"::\(if .severity == "advice" or .severity == "disabled" then "warning" else .severity end) file=\(.path),line=\(.line),col=\(.char),title=\(.code) \(.name)::" + (.description | gsub("\\n"; "%0A"))' \
-          lint.json || true
-
-        exit $RC
+  lintrunner-noclang:
+    uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
+    with:
+      timeout: 120
+      runner: linux.2xlarge
+      docker-image: pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter
+      fetch-depth: 1
+      submodules: true
+      ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
+      script: |
+        export ADDITIONAL_LINTRUNNER_ARGS="--skip CLANGTIDY,CLANGFORMAT"
+        .github/scripts/lintrunner.sh
 
   quick-checks:
     uses: pytorch/test-infra/.github/workflows/linux_job.yml@main