Improve format check

.github/workflows/format_check.yml

@@ -26,6 +26,7 @@ jobs:
       - name: Fetch latest changes
         run: git fetch origin
 
+      # Step 4: Find changed Java files
       - name: Get changed Java files
         id: changed_files
         run: |
@@ -37,24 +38,61 @@ jobs:
           # Write the multiline content to a file
           echo "$CHANGED_FILES" > changed_files.txt
 
-      # Step 4: Get a list of changed Java files in the PR
-      - name: Check Java file format
+      # Step 5: Get formatting issues for source branch 
+      - name: Get formatting issues for source branch 
+        id: number_of_formatting_issues_source
         run: |
-          # Check if the changed_files.txt exists
+          # Ensure there are changed files
           if [ ! -f changed_files.txt ]; then
-            echo "No changed files found."
-            exit 0
-          fi
+            echo "No Java files changed."
+          else
+            echo "Processing the following changed Java files:"
+
+            cat changed_files.txt
+            # Read the multiline content from the file
+            CHANGED_FILES=$(cat changed_files.txt)
 
-          # Read the multiline content from the file
-          CHANGED_FILES=$(cat changed_files.txt)
+            # # Checkout source branch and store the unformatted file
+            # git checkout ${{ github.event.pull_request.head.ref }}
 
+            echo "::group::Iterating changed files in source branch"
+            # Iterate over the CHANGED_FILES variable, assuming files are separated by newlines
+            while IFS= read -r FILE; do
+              # Skip empty lines if any
+              if [ -n "$FILE" ]; then
+                FILE_NAME=$(basename "$FILE")
+                echo "Checking for $FILE_NAME"
+
+                # Run your formatter validation for each file
+                mvn formatter:format -f formatter-pom.xml "-Dformatter.includes=**/$FILE_NAME"
+              fi
+            done <<< "$CHANGED_FILES"
+            echo "::endgroup::"
+
+            git diff --stat
+            # Write diff into a file
+            git diff --stat > formatting_diff_source.txt
+          fi
+          git reset --hard
+
+      # Step 6: Get formatting issues for base branch 
+      - name: Get formatting issues for base branch 
+        id: number_of_formatting_issues_base
+        run: |
           # Ensure there are changed files
-          if [ -z "$CHANGED_FILES" ]; then
+          if [ ! -f changed_files.txt ]; then
             echo "No Java files changed."
           else
             echo "Processing the following changed Java files:"
+
+            cat changed_files.txt
+            # Read the multiline content from the file
+            CHANGED_FILES=$(cat changed_files.txt)
+
+            # Checkout source branch and store the unformatted file
+            git checkout ${{ github.event.pull_request.base.ref }}
 
+            echo "::group::Iterating changed files in base branch"
             # Iterate over the CHANGED_FILES variable, assuming files are separated by newlines
             while IFS= read -r FILE; do
               # Skip empty lines if any
@@ -63,7 +101,51 @@ jobs:
                 echo "Checking for $FILE_NAME"
 
                 # Run your formatter validation for each file
-                mvn formatter:validate -f formatter-pom.xml "-Dformatter.includes=**/$FILE_NAME"
+                mvn formatter:format -f formatter-pom.xml "-Dformatter.includes=**/$FILE_NAME"
               fi
             done <<< "$CHANGED_FILES"
+            echo "::endgroup::"
+
+            git diff --stat
+            # Write diff into a file
+            git diff --stat > formatting_diff_base.txt
           fi
+          git reset --hard
+
+      # Step 7: Compare format diffs
+      - name: Compare format diffs
+        run: |
+          format_source=$(head -n -1 formatting_diff_source.txt | awk '{print $1 " " $3}')
+          echo -e "Formatting diff in source branch: \n $format_source"
+          format_base=$(head -n -1 formatting_diff_base.txt | awk '{print $1 " " $3}')
+          echo -e "Formatting diff in base branch: \n $format_base"
+          comm -13 <(echo -e "$format_base") <(echo -e "$format_source") >diff_result.txt
+
+          if [ -s diff_result.txt ]; then
+
+            while read -r line; do
+              source_path=$(echo "$line" | awk '{print $1}')
+              source_index=$(echo "$line" | awk '{print $2}')
+
+              base_line=$(grep "^$source_path " formatting_diff_base.txt)
+              # If the line does not exist, it means it is new in the source branch, and thus, any formatting changes there are treated as new issues.
+              if [ -z "$base_line" ]; then
+                echo "$source_path $source_index" >> new_formatting_issues.txt
+                continue
+              fi
+
+              base_index=$(echo "$base_line" | awk '{print $2}')
+              if [ "$source_index" -gt "$base_index" ]; then
+                echo "$source_path $source_index" >> new_formatting_issues.txt
+              fi
+            done < diff_result.txt
+
+          fi
+
+          if [ -s new_formatting_issues.txt ]; then
+            echo "There are formatting issues introduced with this PR."
+            cat new_formatting_issues.txt
+            exit 1  # Fail the workflow
+          else
+            echo "No new formatting issue detected."
+          fi