WIP: feat: exclude file (ariary#3)

Specify an exclude file with -e or --exclude

cmd/tsfinder/main.go

@@ -25,7 +25,7 @@ func main() {
 		Args:  cobra.MinimumNArgs(1),
 		Run: func(cmd *cobra.Command, args []string) {
 			path := args[0]
-			cfg := &config.Config{Verbose: verbose, Color: color, Sibling: &sibling, OnlyText: onlyText, Exclude: exclude}
+			cfg := &config.Config{Verbose: verbose, Color: color, Sibling: &sibling, OnlyText: onlyText, ExcludelistFilename: exclude}
 
 			homoglyph.Scan(path, cfg)
 		},
@@ -41,7 +41,7 @@ func main() {
 		Run: func(cmd *cobra.Command, args []string) {
 
 			name := args[0]
-			cfg := &config.Config{Verbose: verbose, Color: color, OnlyText: onlyText, Exclude: exclude}
+			cfg := &config.Config{Verbose: verbose, Color: color, OnlyText: onlyText, ExcludelistFilename: exclude}
 
 			bidirectional.Scan(name, cfg)
 		},
@@ -51,7 +51,7 @@ func main() {
 	rootCmd.PersistentFlags().BoolVarP(&verbose, "verbose", "v", false, "make tsfinder more verbose")
 	rootCmd.PersistentFlags().BoolVarP(&color, "color", "c", false, "make tsfinder print with color")
 	rootCmd.PersistentFlags().BoolVarP(&onlyText, "text-file", "t", false, "make tsfinder scan only on 'human readable' file (ie  looks like correct UTF-8). Add verbosity (-v) to see which files has been skipped. This could help to rule out false positives")
-	rootCmd.PersistentFlags().StringVarP(&exclude, "exclude", "e", "", "specify an exclude file containing one path per line to be excluded from scanning")
+	rootCmd.PersistentFlags().StringVarP(&exclude, "exclude", "e", "", "specify an exclude file containing one path per line to be excluded from scanning. The paths are relative to the working directory.")
 
 	rootCmd.AddCommand(cmdSHomoglyph)
 	rootCmd.Execute()

pkg/bidirectional/bidirectionnal.go

@@ -12,6 +12,7 @@ import (
 	"golang.org/x/tools/godoc/vfs"
 
 	"github.com/ariary/TrojanSourceFinder/pkg/config"
+	"github.com/ariary/TrojanSourceFinder/pkg/excludelist"
 	"github.com/ariary/TrojanSourceFinder/pkg/utils"
 )
 
@@ -113,9 +114,19 @@ func Scan(path string, cfg *config.Config) {
 		log.Fatal(err)
 	}
 
+	excludedPaths, err := excludelist.GetExcludelist(*&cfg.ExcludelistFilename)
+	if err != nil {
+		log.Fatal(err)
+	}
+
+	// Skip the given path if it is contained in the exclude list.
+	if _, ignoreEntry := (*excludedPaths)[filepath.Clean(path)]; ignoreEntry {
+		os.Exit(0)
+	}
+
 	var detected int
 	if fileInfo.IsDir() {
-		detected = scanDirectory(path, cfg)
+		detected = scanDirectory(path, cfg, excludedPaths)
 	} else {
 		detected = scanFile(path, cfg)
 	}
@@ -216,13 +227,24 @@ func scanFile(filename string, cfg *config.Config) int {
 // Browse the directory using filepath.Walk package => does not follow symbolic link
 // and for very large directories Walk can be inefficient
 // return 0 if no trojan source was detected
-func scanDirectory(pathD string, cfg *config.Config) (result int) {
-	err := filepath.Walk(pathD, func(path string, info os.FileInfo, err error) error {
+func scanDirectory(pathD string, cfg *config.Config, excludedPaths *map[string]struct{}) (result int) {
+	err := filepath.WalkDir(pathD, func(path string, dirEntry os.DirEntry, err error) error {
 		if err != nil {
 			fmt.Println(err)
 			return err
 		}
-		if !info.IsDir() {
+
+		if dirEntry.IsDir() {
+			// Skip the whole directory if it is contained in the exclude list.
+			if _, ignoreEntry := (*excludedPaths)[path]; ignoreEntry {
+				return filepath.SkipDir
+			}
+		} else {
+			// Skip the file if it is contained in the exclude list.
+			if _, ignoreEntry := (*excludedPaths)[path]; ignoreEntry {
+				return nil
+			}
+
 			result += scanFile(path, cfg)
 		}
 		return nil

pkg/config/config.go

@@ -2,9 +2,9 @@ package config
 
 // Config holds the tsfinder configuration
 type Config struct {
-	Verbose  bool
-	Color    bool
-	Sibling  *[]string
-	OnlyText bool
-	Exclude  string
+	Verbose             bool
+	Color               bool
+	Sibling             *[]string
+	OnlyText            bool
+	ExcludelistFilename string
 }

excludelist/excludelist.go → pkg/excludelist/excludelist.go

@@ -4,6 +4,8 @@ import (
 	"bufio"
 	"fmt"
 	"os"
+	"path/filepath"
+	"strings"
 )
 
 func GetExcludelist(excludelistFilename string) (*map[string]struct{}, error) {
@@ -18,7 +20,7 @@ func GetExcludelist(excludelistFilename string) (*map[string]struct{}, error) {
 	}
 
 	if excludelistFile.IsDir() {
-		return nil, fmt.Errorf("'%s' must not be a directory", excludelistFilename)
+		return nil, fmt.Errorf("excludefile %s must not be a directory", excludelistFilename)
 	}
 
 	readfile, err := os.Open(excludelistFilename)
@@ -29,7 +31,12 @@ func GetExcludelist(excludelistFilename string) (*map[string]struct{}, error) {
 
 	scanner := bufio.NewScanner(readfile)
 	for scanner.Scan() {
-		excludelist[scanner.Text()] = struct{}{}
+		line := strings.TrimSpace(scanner.Text())
+		if line == "" || strings.HasPrefix(line, "#") {
+			continue
+		}
+
+		excludelist[filepath.Clean(line)] = struct{}{}
 	}
 
 	return &excludelist, scanner.Err()