fix: ci/cd and windows line endings

Author: acefsm

.github/CODECOV_SETUP.md

@@ -0,0 +1,33 @@
+# Codecov Setup Instructions
+
+To enable code coverage reporting for this repository:
+
+## 1. Sign up for Codecov
+1. Go to [codecov.io](https://about.codecov.io/)
+2. Sign in with your GitHub account
+3. Grant access to this repository
+
+## 2. Get the Upload Token
+1. Navigate to your repository in Codecov
+2. Go to Settings → General
+3. Copy the Repository Upload Token
+
+## 3. Add the Token to GitHub Secrets
+1. Go to your GitHub repository
+2. Navigate to Settings → Secrets and variables → Actions
+3. Click "New repository secret"
+4. Name: `CODECOV_TOKEN`
+5. Value: (paste the token from Codecov)
+
+## 4. Update the CI Workflow (Optional)
+Once the token is configured, you can update `.github/workflows/ci.yml`:
+- Change `fail_ci_if_error: false` to `fail_ci_if_error: true`
+- Remove `continue-on-error: true`
+
+## Current Configuration
+The codecov integration is currently configured to:
+- Not fail the CI build if coverage upload fails
+- Run in informational mode (won't block PRs)
+- Ignore test files and the binary entry point
+
+This allows the CI to pass while Codecov is not fully configured.

.github/workflows/ci.yml

@@ -89,7 +89,7 @@ jobs:
       shell: bash
       run: |
         # Create test data
-        echo -e "3\n1\n2" > test.txt
+        printf "3\n1\n2\n" > test.txt
         
         # Test basic sorting
         if [[ "${{ matrix.os }}" == "windows-latest" ]]; then
@@ -98,22 +98,20 @@ jobs:
           ./target/release/sort test.txt > output.txt
         fi
         
-        # Check output
-        expected="1\n2\n3"
-        if [[ "${{ matrix.os }}" == "windows-latest" ]]; then
-          # Windows line endings
-          if ! echo -e "1\r\n2\r\n3" | cmp -s - output.txt; then
-            echo "Basic sort test failed"
-            cat output.txt
-            exit 1
-          fi
-        else
-          if ! echo -e "$expected" | cmp -s - output.txt; then
-            echo "Basic sort test failed"
-            cat output.txt
-            exit 1
-          fi
+        # Check output - simplified cross-platform test
+        # Extract just the numbers and check they're in order
+        sorted_output=$(cat output.txt | tr -d '\r' | tr '\n' ' ')
+        expected="1 2 3 "
+        
+        if [[ "$sorted_output" != "$expected" ]]; then
+          echo "Basic sort test failed"
+          echo "Expected: $expected"
+          echo "Got: $sorted_output"
+          echo "Raw output:"
+          cat output.txt | od -c
+          exit 1
         fi
+        
         echo "✅ Basic functionality test passed"
 
   benchmark:
@@ -199,7 +197,9 @@ jobs:
       uses: codecov/codecov-action@v3
       with:
         file: lcov.info
-        fail_ci_if_error: true
+        fail_ci_if_error: false
+        verbose: true
+      continue-on-error: true
 
   docs:
     name: Documentation
@@ -217,12 +217,13 @@ jobs:
     - name: Build documentation
       run: cargo doc --no-deps --all-features
 
-    - name: Deploy documentation (main branch only)
-      if: github.ref == 'refs/heads/main' && github.event_name == 'push'
-      uses: peaceiris/actions-gh-pages@v3
-      with:
-        github_token: ${{ secrets.GITHUB_TOKEN }}
-        publish_dir: ./target/doc
+    # Documentation deployment disabled until gh-pages branch is configured
+    # - name: Deploy documentation (main branch only)
+    #   if: github.ref == 'refs/heads/main' && github.event_name == 'push'
+    #   uses: peaceiris/actions-gh-pages@v3
+    #   with:
+    #     github_token: ${{ secrets.GITHUB_TOKEN }}
+    #     publish_dir: ./target/doc
 
   release:
     name: Create Release

codecov.yml

@@ -0,0 +1,29 @@
+codecov:
+  require_ci_to_pass: false
+  notify:
+    wait_for_ci: false
+
+coverage:
+  status:
+    project:
+      default:
+        # Basic settings for code coverage
+        target: auto
+        threshold: 1%
+        informational: true  # Won't fail the build
+    patch:
+      default:
+        target: auto
+        threshold: 1%
+        informational: true  # Won't fail the build
+
+comment:
+  layout: "reach,diff,flags,files,footer"
+  behavior: default
+  require_changes: false
+
+ignore:
+  - "tests/**"
+  - "benches/**"
+  - "examples/**"
+  - "src/main.rs"  # Ignore the binary entry point

src/zero_copy.rs

@@ -601,15 +601,26 @@ fn parse_lines(data: &[u8]) -> Vec<Line> {
 
     for (i, &byte) in data.iter().enumerate() {
         if byte == b'\n' {
-            let line_data = &data[start..i];
+            // Handle both Unix (\n) and Windows (\r\n) line endings
+            let end = if i > 0 && data[i - 1] == b'\r' {
+                i - 1
+            } else {
+                i
+            };
+            let line_data = &data[start..end];
             lines.push(Line::new(line_data));
             start = i + 1;
         }
     }
 
     // Handle last line if it doesn't end with newline
     if start < data.len() {
-        let line_data = &data[start..];
+        let mut end = data.len();
+        // Strip trailing \r if present
+        if end > start && data[end - 1] == b'\r' {
+            end -= 1;
+        }
+        let line_data = &data[start..end];
         lines.push(Line::new(line_data));
     }
 
@@ -899,4 +910,45 @@ mod tests {
         assert_eq!(parse_int(b"12.34"), None); // Not simple
         assert_eq!(parse_int(b"abc"), None); // Not numeric
     }
+
+    #[test]
+    fn test_parse_lines_with_different_endings() {
+        // Test Unix line endings
+        let unix_data = b"line1\nline2\nline3";
+        let unix_lines = parse_lines(unix_data);
+        assert_eq!(unix_lines.len(), 3);
+        unsafe {
+            assert_eq!(unix_lines[0].as_bytes(), b"line1");
+            assert_eq!(unix_lines[1].as_bytes(), b"line2");
+            assert_eq!(unix_lines[2].as_bytes(), b"line3");
+        }
+
+        // Test Windows line endings
+        let windows_data = b"line1\r\nline2\r\nline3\r\n";
+        let windows_lines = parse_lines(windows_data);
+        assert_eq!(windows_lines.len(), 3);
+        unsafe {
+            assert_eq!(windows_lines[0].as_bytes(), b"line1");
+            assert_eq!(windows_lines[1].as_bytes(), b"line2");
+            assert_eq!(windows_lines[2].as_bytes(), b"line3");
+        }
+
+        // Test mixed line endings
+        let mixed_data = b"line1\r\nline2\nline3\r";
+        let mixed_lines = parse_lines(mixed_data);
+        assert_eq!(mixed_lines.len(), 3);
+        unsafe {
+            assert_eq!(mixed_lines[0].as_bytes(), b"line1");
+            assert_eq!(mixed_lines[1].as_bytes(), b"line2");
+            assert_eq!(mixed_lines[2].as_bytes(), b"line3");
+        }
+
+        // Test single line without ending
+        let single_data = b"single_line";
+        let single_lines = parse_lines(single_data);
+        assert_eq!(single_lines.len(), 1);
+        unsafe {
+            assert_eq!(single_lines[0].as_bytes(), b"single_line");
+        }
+    }
 }