CI: use a minimal local clone to get changes in patch

This aims at circumventing limit of 300 changed files returned by GH API
0penBrain · May 29, 2023 · 587ade5 · 587ade5
1 parent 413f9fe
commit 587ade5
Showing 1 changed file with 19 additions and 34 deletions.
diff --git a/.github/workflows/sub_prepare.yml b/.github/workflows/sub_prepare.yml
@@ -80,7 +80,6 @@ jobs:
         run: |
           mkdir -p ${{ env.logdir }}
           mkdir -p ${{ env.reportdir }}
-          commitCnt=0
           touch ${{ env.logdir }}changedFiles.lst ${{ env.logdir }}changedCppFiles.lst ${{ env.logdir }}changedPythonFiles.lst
           echo "reportFile=${{ env.reportfilename }}" >> $GITHUB_OUTPUT
       - name: Determine base and head SHA in case of PR
@@ -102,7 +101,7 @@ jobs:
           # Exit the step with appropriate code
           if [ $dateDiff -gt ${{ inputs.maxRebaseHours }} ]
           then
-            echo -n ":warning: Pull request should be rebased" | tee -a ${{env.reportdir}}${{ env.reportfilename }}
+            echo ":warning: Pull request should be rebased" | tee -a ${{env.reportdir}}${{ env.reportfilename }}
             exit 1
           fi
       - name: Determine base and head SHA in case of push
@@ -121,20 +120,23 @@ jobs:
             echo "Base SHA is $baseSha, Head SHA is $headSha" | tee -a ${{env.reportdir}}${{ env.reportfilename }}
             echo "baseSha=$baseSha" >> $GITHUB_ENV
           fi
-      - name: Get compare between head and base
-        if: env.isPR == 'true' || env.isPush == 'true'
+      - name: Prepare minimal local clone for subsequent operations
         run: |
-          echo "$GITHUB_API_URL/repos/$GITHUB_REPOSITORY/compare/$baseSha...$headSha"
-          curl -H "Accept: application/vnd.github+json" -H "authorization: Bearer ${{ secrets.GITHUB_TOKEN }}" $GITHUB_API_URL/repos/$GITHUB_REPOSITORY/compare/$baseSha...$headSha > ${{ env.logdir }}compare.json
+          set -x
+          git clone --filter=blob:none --no-checkout $GITHUB_SERVER_URL/$GITHUB_REPOSITORY.git
       - name: Get number of commits in the changeset
         if: env.isPR == 'true' || env.isPush == 'true'
         run: |
-          commitCnt=$(jq -re '.ahead_by' ${{ env.logdir }}compare.json)
+          set -x
+          cd $(basename "$GITHUB_REPOSITORY")
+          commitCnt=$(git log $baseSha..$headSha --oneline | wc -l)
           echo "Changeset is composed of $commitCnt commit(s)" | tee -a ${{env.reportdir}}${{ env.reportfilename }}
       - name: Get files modified in changeset
         if: env.isPR == 'true' || env.isPush == 'true'
         run: |
-          jq -r '.files[] | if .status != "removed" then .filename else empty end' ${{ env.logdir }}compare.json > ${{ env.logdir }}changedFiles.lst
+          set -x
+          cd $(basename "$GITHUB_REPOSITORY")
+          git diff $baseSha...$headSha --name-only --diff-filter=d | cut -f 1 > ${{ env.logdir }}changedFiles.lst
           grep -E '\.(py|py3)"' ${{ env.logdir }}changedFiles.lst > ${{ env.logdir }}changedPythonFiles.lst || true
           grep -E '\.(c|c\+\+|cc|cpp|cu|cuh|cxx|h|h\+\+|hh|hpp|hxx)"' ${{ env.logdir }}changedFiles.lst > ${{ env.logdir }}changedCppFiles.lst || true
           # Write the report
@@ -148,35 +150,18 @@ jobs:
         run: |
           set -x
           sudo apt-get install -y --no-install-recommends pre-commit
-          BASEADDR="$GITHUB_SERVER_URL/$GITHUB_REPOSITORY/raw/$headSha"
-          mkdir precommit
-          cd precommit
-          git init
-          wget $BASEADDR/.pre-commit-config.yaml
-          mkdir files
-          pushd files
-          for file in $(cat ${{ env.logdir }}changedFiles.lst)
-          do
-            if [ $(dirname "$file") != '.' ]
-            then
-              mkdir -p $(dirname $file)
-            fi
-            wget -O $file $BASEADDR/$file
-          done
-          popd
-          pre-commit run --files $(find files/ -type f)
+          cd $(basename "$GITHUB_REPOSITORY")
+          IFS=$'\n' files=($(cat ${{ env.logdir }}changedFiles.lst))
+          git config gc.auto 0
+          git fetch --depth=1 origin $GITHUB_REF
+          git checkout $headSha -- .pre-commit-config.yaml ${files[@]}
+          pre-commit run -a
       - name: Transmitting outputs
         id: Output
         run: |
-          echo "changedFiles<<EOD" >> $GITHUB_OUTPUT
-          cat ${{ env.logdir }}changedFiles.lst >> $GITHUB_OUTPUT
-          echo "EOD" >> $GITHUB_OUTPUT
-          echo "changedPythonFiles<<EOD" >> $GITHUB_OUTPUT
-          cat ${{ env.logdir }}changedPythonFiles.lst >> $GITHUB_OUTPUT
-          echo "EOD" >> $GITHUB_OUTPUT
-          echo "changedCppFiles<<EOD" >> $GITHUB_OUTPUT
-          cat ${{ env.logdir }}changedCppFiles.lst >> $GITHUB_OUTPUT
-          echo "EOD" >> $GITHUB_OUTPUT
+          echo "changedFiles=$(cat ${{ env.logdir }}changedFiles.lst | xargs -I{file} echo -n "\"{file}\" ")" >> $GITHUB_OUTPUT
+          echo "changedPythonFiles=$(cat ${{ env.logdir }}changedPythonFiles.lst | xargs -I{file} echo -n "\"{file}\" ")" >> $GITHUB_OUTPUT
+          echo "changedCppFiles=$(cat ${{ env.logdir }}changedCppFiles.lst | xargs -I{file} echo -n "\"{file}\" ")" >> $GITHUB_OUTPUT
       - name: Upload logs
         if: always()
         uses: actions/upload-artifact@v3