From bc5041a381539522c2f615c80df37bc5281904ce Mon Sep 17 00:00:00 2001
From: frjcomp <107982661+frjcomp@users.noreply.github.com>
Date: Thu, 16 Jan 2025 15:02:45 +0000
Subject: [PATCH] base logic for line numbers, still wrong as the gitlab ui
 strangely counts newlines

---
 src/pipeleak/helper/helper.go |  9 +++++++++
 src/pipeleak/scanner/queue.go |  2 +-
 src/pipeleak/scanner/rules.go | 30 ++++++++++++++++++++++++++++--
 3 files changed, 38 insertions(+), 3 deletions(-)

diff --git a/src/pipeleak/helper/helper.go b/src/pipeleak/helper/helper.go
index 272a010..ff00b58 100644
--- a/src/pipeleak/helper/helper.go
+++ b/src/pipeleak/helper/helper.go
@@ -11,6 +11,7 @@ import (
 	"os/signal"
 	"path"
 	"regexp"
+	"strconv"
 	"strings"
 	"syscall"
 	"time"
@@ -220,3 +221,11 @@ func IsDirectory(path string) bool {
 
 	return fileInfo.IsDir()
 }
+
+func LineNRToWebURL(webUrl string, lineNumber int) string {
+	if (lineNumber < 0) {
+		return webUrl
+	}
+	
+	return webUrl + "#L" + strconv.Itoa(lineNumber)
+}
\ No newline at end of file
diff --git a/src/pipeleak/scanner/queue.go b/src/pipeleak/scanner/queue.go
index ced66b9..376ecc3 100644
--- a/src/pipeleak/scanner/queue.go
+++ b/src/pipeleak/scanner/queue.go
@@ -104,7 +104,7 @@ func analyzeJobTrace(git *gitlab.Client, item QueueItem, options *ScanOptions) {
 
 	findings := DetectHits(trace, options.MaxScanGoRoutines, options.TruffleHogVerification)
 	for _, finding := range findings {
-		log.Warn().Str("confidence", finding.Pattern.Pattern.Confidence).Str("ruleName", finding.Pattern.Pattern.Name).Str("value", finding.Text).Str("url", item.Meta.JobWebUrl).Str("jobName", item.Meta.JobName).Msg("HIT")
+		log.Warn().Str("confidence", finding.Pattern.Pattern.Confidence).Str("ruleName", finding.Pattern.Pattern.Name).Str("value", finding.Text).Str("url", helper.LineNRToWebURL(item.Meta.JobWebUrl, finding.LineNumber)).Str("jobName", item.Meta.JobName).Msg("HIT")
 	}
 }
 
diff --git a/src/pipeleak/scanner/rules.go b/src/pipeleak/scanner/rules.go
index 27ce204..fd3ac1f 100644
--- a/src/pipeleak/scanner/rules.go
+++ b/src/pipeleak/scanner/rules.go
@@ -41,6 +41,7 @@ type PatternPattern struct {
 type Finding struct {
 	Pattern PatternElement
 	Text    string
+	LineNumber int
 }
 
 // hold patterns in memory during runtime
@@ -152,6 +153,7 @@ func DetectHits(text []byte, maxThreads int, enableTruffleHogVerification bool)
 			hits := m.FindAllIndex(text, -1)
 
 			for _, hit := range hits {
+				lineNumber := countNewlinesUntilIndex(text, hit[0])
 				// truncate output to max 1024 chars for output readability
 				hitStr := extractHitWithSurroundingText(text, hit, 50)
 				hitStr = cleanHitLine(hitStr)
@@ -160,7 +162,7 @@ func DetectHits(text []byte, maxThreads int, enableTruffleHogVerification bool)
 				}
 
 				if hitStr != "" {
-					findingsYml = append(findingsYml, Finding{Pattern: pattern, Text: hitStr})
+					findingsYml = append(findingsYml, Finding{Pattern: pattern, Text: hitStr, LineNumber: lineNumber})
 				}
 			}
 
@@ -190,7 +192,9 @@ func DetectHits(text []byte, maxThreads int, enableTruffleHogVerification bool)
 				if len(result.RawV2) > 0 {
 					secret = result.RawV2
 				}
-				finding := Finding{Pattern: PatternElement{Pattern: PatternPattern{Name: result.DetectorType.String(), Confidence: "high-verified"}}, Text: string(secret)}
+				resultIndex := strings.Index(string(text), string(secret))
+				lineNumber := countNewlinesUntilIndex(text, resultIndex)
+				finding := Finding{Pattern: PatternElement{Pattern: PatternPattern{Name: result.DetectorType.String(), Confidence: "high-verified"}}, Text: string(secret), LineNumber: lineNumber}
 
 				// if trufflehog verification is enalbed ONLY verified rules are reported
 				if result.Verified {
@@ -277,3 +281,25 @@ func cleanHitLine(text string) string {
 	text = strings.ReplaceAll(text, "\n", " ")
 	return stripansi.Strip(text)
 }
+
+func countNewlinesUntilIndex(s []byte, index int) int {
+	if index > len(s) {
+		index = len(s)
+	}
+	if index == 0 {
+		return 0
+	}
+	if index < 0 {
+		return -1
+	}
+	substring := s[:index]
+
+	
+	count := 0
+	for i := 0; i < len(substring); i++ {
+		if substring[i] == '[' {
+			count++
+		}
+	}
+	return count
+}