From 2e2a928e75ef1c3e1439303def1f4acac763285f Mon Sep 17 00:00:00 2001
From: Richard Gomez <rmgomez368@gmail.com>
Date: Tue, 6 Feb 2024 09:50:36 -0500
Subject: [PATCH] fix(gitparse): quoted binary files

---
 pkg/gitparse/gitparse.go      | 38 +++++++++++++++++++++++++----------
 pkg/gitparse/gitparse_test.go | 11 +++++++---
 2 files changed, 35 insertions(+), 14 deletions(-)

diff --git a/pkg/gitparse/gitparse.go b/pkg/gitparse/gitparse.go
index 1462b5c9397a4..380617f276620 100644
--- a/pkg/gitparse/gitparse.go
+++ b/pkg/gitparse/gitparse.go
@@ -12,6 +12,8 @@ import (
 	"strings"
 	"time"
 
+	regexp "github.com/wasilibs/go-re2"
+
 	"github.com/go-logr/logr"
 
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
@@ -458,10 +460,16 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, diffChan chan
 		case isBinaryLine(latestState, line):
 			latestState = BinaryFileLine
 
-			currentDiff.PathB = pathFromBinaryLine(line)
+			path, err := pathFromBinaryLine(line)
+			if err != nil {
+				ctx.Logger().Error(err, "Failed to parse binary file line")
+				latestState = ParseFailure
+				continue
+			}
 
 			// Don't do anything if the file is deleted. (pathA has file path, pathB is /dev/null)
-			if currentDiff.PathB != "" {
+			if path != "" {
+				currentDiff.PathB = path
 				currentDiff.IsBinary = true
 			}
 		case isFromFileLine(latestState, line):
@@ -707,16 +715,24 @@ 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 {
-	logger := context.Background().Logger()
-	sbytes := bytes.Split(line, []byte(" and b/"))
-	if len(sbytes) != 2 {
-		logger.V(2).Info("Expected binary line to be in 'Binary files a/fileA and b/fileB differ' format.", "got", line)
-		return ""
-	}
-	bRaw := sbytes[1]
-	return string(bRaw[:len(bRaw)-8]) // drop the "b/" and " differ\n"
+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
+	}
+
+	var path string
+	for _, match := range matches[1:] { // the first match is the entire input
+		if len(match) > 0 {
+			path = string(match)
+			break
+		}
+	}
+	return path, nil
 }
 
 // --- a/internal/addrs/move_endpoint_module.go
diff --git a/pkg/gitparse/gitparse_test.go b/pkg/gitparse/gitparse_test.go
index 505d6db9d7f7e..3fee460889484 100644
--- a/pkg/gitparse/gitparse_test.go
+++ b/pkg/gitparse/gitparse_test.go
@@ -591,12 +591,17 @@ func TestLineChecksNoStaged(t *testing.T) {
 
 func TestBinaryPathParse(t *testing.T) {
 	cases := map[string]string{
-		"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 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",
 	}
 
 	for name, expected := range cases {
-		filename := pathFromBinaryLine([]byte(name))
+		filename, err := pathFromBinaryLine([]byte(name))
+		if err != nil {
+			t.Errorf("Got unexpected error: %s", err)
+		}
 		if filename != expected {
 			t.Errorf("Expected: %s, Got: %s", expected, filename)
 		}