fix(gitparse): use bytes.Cut instead of regexp

trufflesecurity · Feb 6, 2024 · 586a813 · 586a813
1 parent 2e2a928
commit 586a813
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 25 deletions.
diff --git a/pkg/gitparse/gitparse.go b/pkg/gitparse/gitparse.go
@@ -12,8 +12,6 @@ import (
 	"strings"
 	"time"
 
-	regexp "github.com/wasilibs/go-re2"
-
 	"github.com/go-logr/logr"
 
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
@@ -460,8 +458,9 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, diffChan chan
 		case isBinaryLine(latestState, line):
 			latestState = BinaryFileLine
 
-			path, err := pathFromBinaryLine(line)
-			if err != nil {
+			path, ok := pathFromBinaryLine(line)
+			if !ok {
+				err = fmt.Errorf(`expected line to match 'Binary files a/fileA and b/fileB differ', got "%s"`, line)
 				ctx.Logger().Error(err, "Failed to parse binary file line")
 				latestState = ParseFailure
 				continue
@@ -715,24 +714,28 @@ func isBinaryLine(latestState ParseState, line []byte) bool {
 	return false
 }
 
-var binaryPathPat = regexp.MustCompile(`Binary files .+ and (?:/dev/null|b/(.+)|"b/(.+)") differ`)
-
 // Get the b/ file path. Ignoring the edge case of files having `and /b` in the name for simplicity.
-func pathFromBinaryLine(line []byte) (string, error) {
-	matches := binaryPathPat.FindSubmatch(line)
-	if len(matches) == 0 {
-		err := fmt.Errorf(`expected line to match 'Binary files a/fileA and b/fileB differ', got "%s"`, line)
-		return "", err
+func pathFromBinaryLine(line []byte) (string, bool) {
+	if bytes.Index(line, []byte("and /dev/null")) == -1 {
+		return "", true
 	}
 
-	var path string
-	for _, match := range matches[1:] { // the first match is the entire input
-		if len(match) > 0 {
-			path = string(match)
-			break
-		}
+	_, after, ok := bytes.Cut(line, []byte(" and b/"))
+	if ok {
+		// drop the " differ\n"
+		return string(after[:len(after)-8]), true
+	}
+
+	// Edge case where the path is quoted.
+	// https://github.com/trufflesecurity/trufflehog/issues/2384
+	_, after, ok = bytes.Cut(line, []byte(` and "b/`))
+	if ok {
+		// drop the `" differ\n`
+		return string(after[:len(after)-9]), true
 	}
-	return path, nil
+
+	// Unknown format.
+	return "", false
 }
 
 // --- a/internal/addrs/move_endpoint_module.go

diff --git a/pkg/gitparse/gitparse_test.go b/pkg/gitparse/gitparse_test.go
@@ -591,16 +591,16 @@ func TestLineChecksNoStaged(t *testing.T) {
 
 func TestBinaryPathParse(t *testing.T) {
 	cases := map[string]string{
-		"Binary files a/trufflehog_3.42.0_linux_arm64.tar.gz and /dev/null differ\n":                                 "",
-		"Binary files /dev/null and b/plugin.sig differ\n":                                                           "plugin.sig",
-		"Binary files /dev/null and b/ Lunch and Learn - HCDiag.pdf differ\n":                                        " Lunch and Learn - HCDiag.pdf",
-		"Binary files /dev/null and \"b/assets/retailers/ON-ikony-Platforma-ecom \\342\\200\\224 kopia.png\" differ": "assets/retailers/ON-ikony-Platforma-ecom \\342\\200\\224 kopia.png",
+		"Binary files a/trufflehog_3.42.0_linux_arm64.tar.gz and /dev/null differ\n":                                   "",
+		"Binary files /dev/null and b/plugin.sig differ\n":                                                             "plugin.sig",
+		"Binary files /dev/null and b/ Lunch and Learn - HCDiag.pdf differ\n":                                          " Lunch and Learn - HCDiag.pdf",
+		"Binary files /dev/null and \"b/assets/retailers/ON-ikony-Platforma-ecom \\342\\200\\224 kopia.png\" differ\n": "assets/retailers/ON-ikony-Platforma-ecom \\342\\200\\224 kopia.png",
 	}
 
 	for name, expected := range cases {
-		filename, err := pathFromBinaryLine([]byte(name))
-		if err != nil {
-			t.Errorf("Got unexpected error: %s", err)
+		filename, ok := pathFromBinaryLine([]byte(name))
+		if !ok {
+			t.Errorf("Failed to get path: %s", name)
 		}
 		if filename != expected {
 			t.Errorf("Expected: %s, Got: %s", expected, filename)