refactor(merge-diff): rewrite bash into go to better handle complexity of multi-parent pull requests

apps/ci-os/packages/default.nix

@@ -1,12 +1,12 @@
 { pkgs, lib, gitSvPkg, ... }:
 
 let 
-  fluxLocal = import ./flux-local { inherit pkgs lib; };
+  fluxLocal = import ./flux-local { inherit lib pkgs; };
 in
 {
   flux-local = fluxLocal;
-  flux-diff = import ./flux-diff { inherit pkgs lib fluxLocal; };
+  flux-diff = import ./flux-diff { inherit lib pkgs fluxLocal; };
-  merge-diff = import ./merge-diff { inherit pkgs lib gitSvPkg; };
+  merge-diff = import ./merge-diff { inherit lib pkgs gitSvPkg; };
   forgejo-label = import ./forgejo-label { inherit lib pkgs; };
   forgejo-comment = import ./forgejo-comment { inherit lib pkgs; };
   forgejo-release = import ./forgejo-release { inherit lib pkgs; };

apps/ci-os/packages/merge-diff/default.nix

@@ -1,131 +1,15 @@
-{ pkgs, lib, gitSvPkg, ... }:
+{ lib, pkgs, gitSvPkg }:
 
-with pkgs;
+pkgs.buildGoModule rec {
+  pname = "merge-diff";
+  version = "2.0.0";
 
-writeShellApplication rec {
+  buildInputs = with pkgs; [ gitFull jq coreutils gnugrep gawk gitSvPkg ];
-  name = "merge-diff";
 
-  runtimeInputs = [ gitFull jq coreutils gnugrep gawk gitSvPkg ];
+  src = ./src;
 
-  # Define the shell script
+  # Vendor dependencies for reproducibility
-  text = ''
+  vendorHash = null;
-    #!/usr/bin/env bash
-    set -euo pipefail
 
-    # Function to display usage
+  subPackages = [ "." ];
-    usage() {
-      echo "Usage: $0 -p <path_pattern> -d <directory depth>"
-      echo "  -p, --pattern    Path pattern to filter changed files (can be specified multiple times)"
-      echo "  -d, --depth      Depth of changeset (default is 2)"
-      echo "  -h, --help       Display this help message"
-      exit 1
-    }
-
-    # Parse command-line arguments
-    PATTERNS=()
-    DEPTH=2
-    while [[ $# -gt 0 ]]; do
-      case "$1" in
-        -p|--pattern)
-          PATTERNS+=("$2")
-          shift 2
-          ;;
-        -d|--depth)
-          DEPTH="$2"
-          shift 2
-          ;;
-        -h|--help)
-          usage
-          ;;
-        *)
-          echo "Unknown option: $1"
-          usage
-          ;;
-      esac
-    done
-
-    if [ ''${#PATTERNS[@]} -eq 0 ]; then
-      echo "Error: At least one path pattern must be specified."
-      usage
-    fi
-
-    # Get the current commit SHA
-    CURRENT_SHA=$(git rev-parse HEAD)
-
-    # Get parent commits
-    PARENT_COUNT=$(git rev-list --parents -n 1 "$CURRENT_SHA" | wc -w)
-
-    if [ "$PARENT_COUNT" -gt 2 ]; then
-      echo "This is a merge commit with $((PARENT_COUNT - 1)) parents."
-      # For simplicity, comparing with the first parent
-      PARENT_SHA=$(git rev-parse HEAD^1)
-    elif [ "$PARENT_COUNT" -eq 2 ]; then
-      # Regular commit with one parent
-      PARENT_SHA=$(git rev-parse HEAD^1)
-    else
-      echo "Single commit with no parents. Listing all changes."
-      PARENT_SHA=HEAD~1
-    fi
-
-    # List changed files between parent and current commit
-    CHANGED_FILES=$(git diff --name-only "$PARENT_SHA" "$CURRENT_SHA")
-
-    #echo "Changed files:"
-    #echo "$CHANGED_FILES"
-
-    # Initialize an empty array for filtered changes
-    FILTERED_CHANGES=()
-
-    # Iterate over each pattern and filter changed files
-    for pattern in "''${PATTERNS[@]}"; do
-      while IFS= read -r file; do
-        if [[ "$file" =~ $pattern ]]; then
-          file=$(echo "$file"| cut -d/ -f1-"$DEPTH")
-          FILTERED_CHANGES+=("$file")
-        fi
-      done <<< "$CHANGED_FILES"
-    done
-
-    # Remove duplicates
-    UNIQUE_FILTERED_CHANGES=$(printf "%s\n" "''${FILTERED_CHANGES[@]}" | sort -u)
-
-    #echo "Filtered changed files:"
-    #echo "$UNIQUE_FILTERED_CHANGES"
-
-
-    # Determine semantic version bump using git-sv
-    # Ensure that git-sv is available
-    if ! command -v git-sv &> /dev/null; then
-      echo "Error: git-sv is not installed or not in PATH."
-      exit 1
-    fi
-
-    # Get the semantic version bump
-    VERSION_BUMP=$(git-sv next-version)
-
-    # Validate the version bump
-    #if [[ "$VERSION_BUMP" =~ ^(patch|minor|major)$ ]]; then
-    #  echo "Version bump: $VERSION_BUMP"
-    #else
-    #  echo "Error: Unexpected version bump value: $VERSION_BUMP"
-    #  exit 1
-    #fi
-
-    # Convert to JSON
-    RESULTS=$(jq -n \
-      --argjson filteredChanges "$(printf '%s\n' "$UNIQUE_FILTERED_CHANGES" | jq -R -s -c 'split("\n") | map(select(length > 0))')" \
-      --arg versionBump "$VERSION_BUMP" \
-      '{filteredChanges: $filteredChanges, versionBump: $versionBump}')
-
-    # Output results
-    echo "$RESULTS"
-  '';
-
-  # Metadata for the package
-  meta = with lib; {
-    homepage = "https://code.252.no/tommy/containers";
-    description = "Identify Merge Commit and List Changed Files with Flexible Filtering and Semantic Versioning";
-    license = licenses.mit;
-    maintainers = with maintainers; [ "tommy-skaug" ];
-  };
 }

apps/ci-os/packages/merge-diff/src/go.mod

@@ -0,0 +1,3 @@
+module code.252.no/tommy/merge-diff
+
+go 1.22.6

apps/ci-os/packages/merge-diff/src/main.go

@@ -0,0 +1,179 @@
+package main
+
+import (
+	"encoding/json"
+	"flag"
+	"fmt"
+	"log"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"regexp"
+	"strings"
+)
+
+type Result struct {
+	FilteredChanges []string `json:"filteredChanges"`
+	NextVersion     string   `json:"nextVersion"`
+}
+
+func main() {
+	// Parse command-line arguments
+	var patterns patternList
+	var depth int
+
+	flag.Var(&patterns, "p", "Path pattern to filter changed files (can be specified multiple times)")
+	flag.Var(&patterns, "pattern", "Path pattern to filter changed files (can be specified multiple times)")
+	flag.IntVar(&depth, "d", 2, "Depth of changeset (default is 2)")
+	flag.IntVar(&depth, "depth", 2, "Depth of changeset (default is 2)")
+	help := flag.Bool("h", false, "Display this help message")
+	helpLong := flag.Bool("help", false, "Display this help message")
+	flag.Parse()
+
+	if *help || *helpLong {
+		flag.Usage()
+		os.Exit(1)
+	}
+
+	if len(patterns) == 0 {
+		fmt.Println("Error: At least one path pattern must be specified.")
+		flag.Usage()
+		os.Exit(1)
+	}
+
+	// Get the current commit SHA
+	currentSHA := getCurrentSHA()
+
+	// Get parent commits
+	parentSHAs := getParentSHAs(currentSHA)
+
+	if len(parentSHAs) == 0 {
+		fmt.Println("Single commit with no parents. Listing all changes.")
+		parentSHAs = append(parentSHAs, currentSHA+"~1")
+	}
+
+	// Collect all changed files from all parents
+	changedFiles := make(map[string]struct{})
+	for _, parentSHA := range parentSHAs {
+		files := getChangedFiles(parentSHA, currentSHA)
+		for _, file := range files {
+			changedFiles[file] = struct{}{}
+		}
+	}
+
+	// Filter and process changed files
+	filteredChanges := filterAndProcessFiles(changedFiles, patterns, depth)
+
+	// Get the next semantic version using git-sv
+	nextVersion := getNextVersion()
+
+	// Prepare the result
+	result := Result{
+		FilteredChanges: uniqueStrings(filteredChanges),
+		NextVersion:     nextVersion,
+	}
+
+	// Output results as JSON
+	output, err := json.Marshal(result)
+	if err != nil {
+		log.Fatalf("Error marshaling JSON: %v", err)
+	}
+
+	fmt.Println(string(output))
+}
+
+// Custom flag type for multiple patterns
+type patternList []string
+
+func (p *patternList) String() string {
+	return strings.Join(*p, ", ")
+}
+
+func (p *patternList) Set(value string) error {
+	*p = append(*p, value)
+	return nil
+}
+
+// Helper functions
+func getCurrentSHA() string {
+	output, err := exec.Command("git", "rev-parse", "HEAD").Output()
+	if err != nil {
+		log.Fatalf("Error getting current SHA: %v", err)
+	}
+	return strings.TrimSpace(string(output))
+}
+
+func getParentSHAs(commitSHA string) []string {
+	output, err := exec.Command("git", "rev-list", "--parents", "-n", "1", commitSHA).Output()
+	if err != nil {
+		log.Fatalf("Error getting parent SHAs: %v", err)
+	}
+	parts := strings.Fields(string(output))
+	if len(parts) > 1 {
+		return parts[1:] // Exclude the current commit SHA
+	}
+	return []string{}
+}
+
+func getChangedFiles(parentSHA, currentSHA string) []string {
+	output, err := exec.Command("git", "diff", "--name-only", parentSHA, currentSHA).Output()
+	if err != nil {
+		log.Fatalf("Error getting changed files: %v", err)
+	}
+	lines := strings.Split(strings.TrimSpace(string(output)), "\n")
+	if len(lines) == 1 && lines[0] == "" {
+		return []string{}
+	}
+	return lines
+}
+
+func filterAndProcessFiles(changedFiles map[string]struct{}, patterns []string, depth int) []string {
+	var filtered []string
+	for file := range changedFiles {
+		for _, pattern := range patterns {
+			matched, err := regexp.MatchString(pattern, file)
+			if err != nil {
+				log.Fatalf("Invalid regex pattern: %v", err)
+			}
+			if matched {
+				processedFile := processFilePath(file, depth)
+				filtered = append(filtered, processedFile)
+				break // Stop checking other patterns for this file
+			}
+		}
+	}
+	return filtered
+}
+
+func processFilePath(filePath string, depth int) string {
+	parts := strings.Split(filePath, string(os.PathSeparator))
+	if len(parts) > depth {
+		parts = parts[:depth]
+	}
+	return filepath.Join(parts...)
+}
+
+func uniqueStrings(input []string) []string {
+	uniqueMap := make(map[string]struct{})
+	for _, item := range input {
+		uniqueMap[item] = struct{}{}
+	}
+	var result []string
+	for item := range uniqueMap {
+		result = append(result, item)
+	}
+	return result
+}
+
+func getNextVersion() string {
+	// Ensure that git-sv is available
+	if _, err := exec.LookPath("git-sv"); err != nil {
+		log.Fatalf("Error: git-sv is not installed or not in PATH.")
+	}
+	output, err := exec.Command("git-sv", "next-version").Output()
+	if err != nil {
+		log.Fatalf("Error getting next version: %v", err)
+	}
+	nextVersion := strings.TrimSpace(string(output))
+	return nextVersion
+}