Update test

Currently it's a unstructured map[string]string that gets copied into
a Result.ExtraData after detection, but we might want to revisit this
decision when we get more handlers.

go.mod

@@ -8,6 +8,7 @@ require (
 	cloud.google.com/go/secretmanager v1.11.1
 	cloud.google.com/go/storage v1.31.0
 	github.com/Azure/go-autorest/autorest/azure/auth v0.5.11
+	github.com/BobuSumisu/aho-corasick v1.0.3
 	github.com/TheZeroSlave/zapsentry v1.17.0
 	github.com/aws/aws-sdk-go v1.44.83
 	github.com/bill-rich/disk-buffer-reader v0.1.7
@@ -44,7 +45,6 @@ require (
 	github.com/mholt/archiver/v4 v4.0.0-alpha.8
 	github.com/patrickmn/go-cache v2.1.0+incompatible
 	github.com/paulbellamy/ratecounter v0.2.0
-	github.com/petar-dambovaliev/aho-corasick v0.0.0-20211021192214-5ab2d9280aa9
 	github.com/pkg/errors v0.9.1
 	github.com/prometheus/client_golang v1.16.0
 	github.com/rabbitmq/amqp091-go v1.8.1
@@ -94,7 +94,7 @@ require (
 	github.com/cloudflare/circl v1.3.3 // indirect
 	github.com/connesc/cipherio v0.2.1 // indirect
 	github.com/containerd/stargz-snapshotter/estargz v0.14.3 // indirect
-	github.com/couchbase/gocbcore/v10 v10.2.3 // indirect
+	github.com/couchbase/gocbcore/v10 v10.2.4 // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/dimchansky/utfbom v1.1.1 // indirect
 	github.com/docker/cli v23.0.5+incompatible // indirect

go.sum

@@ -51,6 +51,8 @@ github.com/Azure/go-autorest/tracing v0.6.0 h1:TYi4+3m5t6K48TGI9AUdb+IzbnSxvnvUM
 github.com/Azure/go-autorest/tracing v0.6.0/go.mod h1:+vhtPC754Xsa23ID7GlGsrdKBpUA79WCAKPPZVC2DeU=
 github.com/Azure/go-ntlmssp v0.0.0-20221128193559-754e69321358 h1:mFRzDkZVAjdal+s7s0MwaRv9igoPqLRdzOLzw/8Xvq8=
 github.com/Azure/go-ntlmssp v0.0.0-20221128193559-754e69321358/go.mod h1:chxPXzSsl7ZWRAuOIE23GDNzjWuZquvFlgA8xmpunjU=
+github.com/BobuSumisu/aho-corasick v1.0.3 h1:uuf+JHwU9CHP2Vx+wAy6jcksJThhJS9ehR8a+4nPE9g=
+github.com/BobuSumisu/aho-corasick v1.0.3/go.mod h1:hm4jLcvZKI2vRF2WDU1N4p/jpWtpOzp3nLmi9AzX/XE=
 github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
 github.com/BurntSushi/toml v1.2.1/go.mod h1:CxXYINrC8qIiEnFrOxCa7Jy5BFHlXnUU2pbicEuybxQ=
 github.com/BurntSushi/xgb v0.0.0-20160522181843-27f122750802/go.mod h1:IVnqGOEym/WlBOVXweHU+Q+/VP0lqqI8lqeDx9IjBqo=
@@ -119,8 +121,9 @@ github.com/containerd/stargz-snapshotter/estargz v0.14.3 h1:OqlDCK3ZVUO6C3B/5FSk
 github.com/containerd/stargz-snapshotter/estargz v0.14.3/go.mod h1:KY//uOCIkSuNAHhJogcZtrNHdKrA99/FCCRjE3HD36o=
 github.com/couchbase/gocb/v2 v2.6.3 h1:5RsMo+RRfK0mVxHLAfpBz3/tHlgXZb1WBNItLk9Ab+c=
 github.com/couchbase/gocb/v2 v2.6.3/go.mod h1:yF5F6BHTZ/ZowhEuZbySbXrlI4rHd1TIhm5azOaMbJU=
-github.com/couchbase/gocbcore/v10 v10.2.3 h1:PEkRSNSkKjUBXx82Ucr094+anoiCG5GleOOQZOHo6D4=
 github.com/couchbase/gocbcore/v10 v10.2.3/go.mod h1:lYQIIk+tzoMcwtwU5GzPbDdqEkwkH3isI2rkSpfL0oM=
+github.com/couchbase/gocbcore/v10 v10.2.4 h1:TDTQ1mSBUw9eajuV71ZDJcLRbmfGcsvZV9SMeUDAVGc=
+github.com/couchbase/gocbcore/v10 v10.2.4/go.mod h1:lYQIIk+tzoMcwtwU5GzPbDdqEkwkH3isI2rkSpfL0oM=
 github.com/couchbaselabs/gocaves/client v0.0.0-20230307083111-cc3960c624b1/go.mod h1:AVekAZwIY2stsJOMWLAS/0uA/+qdp7pjO8EHnl61QkY=
 github.com/couchbaselabs/gocaves/client v0.0.0-20230404095311-05e3ba4f0259 h1:2TXy68EGEzIMHOx9UvczR5ApVecwCfQZ0LjkmwMI6g4=
 github.com/couchbaselabs/gocaves/client v0.0.0-20230404095311-05e3ba4f0259/go.mod h1:AVekAZwIY2stsJOMWLAS/0uA/+qdp7pjO8EHnl61QkY=
@@ -382,8 +385,6 @@ github.com/patrickmn/go-cache v2.1.0+incompatible h1:HRMgzkcYKYpi3C8ajMPV8OFXaaR
 github.com/patrickmn/go-cache v2.1.0+incompatible/go.mod h1:3Qf8kWWT7OJRJbdiICTKqZju1ZixQ/KpMGzzAfe6+WQ=
 github.com/paulbellamy/ratecounter v0.2.0 h1:2L/RhJq+HA8gBQImDXtLPrDXK5qAj6ozWVK/zFXVJGs=
 github.com/paulbellamy/ratecounter v0.2.0/go.mod h1:Hfx1hDpSGoqxkVVpBi/IlYD7kChlfo5C6hzIHwPqfFE=
-github.com/petar-dambovaliev/aho-corasick v0.0.0-20211021192214-5ab2d9280aa9 h1:lL+y4Xv20pVlCGyLzNHRC0I0rIHhIL1lTvHizoS/dU8=
-github.com/petar-dambovaliev/aho-corasick v0.0.0-20211021192214-5ab2d9280aa9/go.mod h1:EHPiTAKtiFmrMldLUNswFwfZ2eJIYBHktdaUTZxYWRw=
 github.com/pierrec/lz4/v4 v4.1.15 h1:MO0/ucJhngq7299dKLwIMtgTfbkoSPF6AoMYDd8Q4q0=
 github.com/pierrec/lz4/v4 v4.1.15/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
 github.com/pingcap/errors v0.11.4 h1:lFuQV/oaUMGcD2tqt+01ROSmJs75VG1ToEOkZIZ4nE4=

pkg/context/context.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"runtime/debug"
+	"sync"
 	"time"
 
 	"github.com/go-logr/logr"
@@ -28,8 +29,9 @@ type CancelFunc = context.CancelFunc
 type logCtx struct {
 	// Embed context.Context to get all methods for free.
 	context.Context
-	log logr.Logger
-	err *error
+	log     logr.Logger
+	err     *error
+	errLock *sync.Mutex
 }
 
 // Logger returns a structured logger.
@@ -145,8 +147,11 @@ func captureCancelCallstack(ctx logCtx, f context.CancelFunc) (Context, context.
 	if ctx.err == nil {
 		var err error
 		ctx.err = &err
+		ctx.errLock = &sync.Mutex{}
 	}
 	return ctx, func() {
+		ctx.errLock.Lock()
+		defer ctx.errLock.Unlock()
 		// We must check Err() before calling f() since f() sets the error.
 		// If there's already an error, do nothing special.
 		if ctx.Err() != nil {

pkg/context/context_test.go

@@ -212,3 +212,9 @@ func TestErrCallstackTimeoutCancel(t *testing.T) {
 	cancel()
 	assert.Equal(t, err, ctx.Err())
 }
+
+func TestRace(t *testing.T) {
+	_, cancel := WithCancel(Background())
+	go cancel()
+	cancel()
+}

pkg/detectors/npmtoken/npmtoken.go

@@ -36,7 +36,6 @@ func (s Scanner) Keywords() []string {
 // FromData will find and optionally verify NpmToken secrets in a given set of bytes.
 func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (results []detectors.Result, err error) {
 	dataStr := string(data)
-
 	matches := keyPat.FindAllStringSubmatch(dataStr, -1)
 	for _, match := range matches {
 		if len(match) != 2 {

pkg/engine/engine.go

@@ -10,7 +10,7 @@ import (
 	"sync/atomic"
 	"time"
 
-	ahocorasick "github.com/petar-dambovaliev/aho-corasick"
+	ahocorasick "github.com/BobuSumisu/aho-corasick"
 	"golang.org/x/sync/errgroup"
 	"google.golang.org/protobuf/proto"
 
@@ -43,7 +43,7 @@ type Engine struct {
 
 	// prefilter is a ahocorasick struct used for doing efficient string
 	// matching given a set of words (keywords from the rules in the config)
-	prefilter ahocorasick.AhoCorasick
+	prefilter ahocorasick.Trie
 }
 
 type EngineOption func(*Engine)
@@ -145,18 +145,16 @@ func Start(ctx context.Context, options ...EngineOption) *Engine {
 	// on keywords
 	keywords := []string{}
 	for _, d := range e.detectors[false] {
-		keywords = append(keywords, d.Keywords()...)
+		for _, kw := range d.Keywords() {
+			keywords = append(keywords, strings.ToLower(kw))
+		}
 	}
 	for _, d := range e.detectors[true] {
-		keywords = append(keywords, d.Keywords()...)
+		for _, kw := range d.Keywords() {
+			keywords = append(keywords, strings.ToLower(kw))
+		}
 	}
-	builder := ahocorasick.NewAhoCorasickBuilder(ahocorasick.Opts{
-		AsciiCaseInsensitive: true,
-		MatchOnlyWholeWords:  false,
-		MatchKind:            ahocorasick.LeftMostLongestMatch,
-		DFA:                  true,
-	})
-	e.prefilter = builder.Build(keywords)
+	e.prefilter = *ahocorasick.NewTrieBuilder().AddStrings(keywords).Build()
 
 	ctx.Logger().Info("loaded decoders", "count", len(e.decoders))
 	ctx.Logger().Info("loaded detectors",
@@ -297,8 +295,8 @@ func (e *Engine) detectorWorker(ctx context.Context) {
 				}
 
 				// build a map of all keywords that were matched in the chunk
-				for _, m := range e.prefilter.FindAll(string(decoded.Data)) {
-					matchedKeywords[strings.ToLower(string(decoded.Data[m.Start():m.End()]))] = struct{}{}
+				for _, m := range e.prefilter.MatchString(strings.ToLower(string(decoded.Data))) {
+					matchedKeywords[strings.ToLower(m.MatchString())] = struct{}{}
 				}
 
 				for verify, detectorsSet := range e.detectors {
@@ -351,8 +349,15 @@ func (e *Engine) detectorWorker(ctx context.Context) {
 								continue
 							}
 							result.DecoderType = decoderType
+							if result.ExtraData == nil && chunk.HandleMetadata != nil {
+								result.ExtraData = chunk.HandleMetadata
+							} else {
+								for k, v := range chunk.HandleMetadata {
+									// TODO: Check key collisions.
+									result.ExtraData[k] = v
+								}
+							}
 							chunkResults = append(chunkResults, detectors.CopyMetadata(resultChunk, result))
-
 						}
 						if len(results) > 0 {
 							elapsed := time.Since(start)

pkg/gitparse/gitparse.go

@@ -51,6 +51,47 @@ type Parser struct {
 	dateFormat    string
 }
 
+type ParseState int
+
+const (
+	Initial ParseState = iota
+	CommitLine
+	AuthorLine
+	DateLine
+	MessageStartLine
+	MessageLine
+	MessageEndLine
+	DiffLine
+	ModeLine
+	IndexLine
+	FromFileLine
+	ToFileLine
+	BinaryFileLine
+	HunkLineNumberLine
+	HunkContentLine
+	ParseFailure
+)
+
+func (state ParseState) String() string {
+	return [...]string{
+		"Initial",
+		"CommitLine",
+		"AuthorLine",
+		"DateLine",
+		"MessageStartLine",
+		"MessageLine",
+		"MessageEndLine",
+		"DiffLine",
+		"ModeLine",
+		"IndexLine",
+		"FromFileLine",
+		"ToFileLine",
+		"BinaryFileLine",
+		"HunkLineNumberLine",
+		"HunkContentLine",
+	}[state]
+}
+
 // WithMaxDiffSize sets maxDiffSize option. Diffs larger than maxDiffSize will
 // be truncated.
 func WithMaxDiffSize(maxDiffSize int) Option {
@@ -137,7 +178,7 @@ func (c *Parser) RepoPath(ctx context.Context, source string, head string, abbre
 		cmd.Env = append(cmd.Env, fmt.Sprintf("GIT_DIR=%s", filepath.Join(absPath, ".git")))
 	}
 
-	return c.executeCommand(ctx, cmd)
+	return c.executeCommand(ctx, cmd, false)
 }
 
 // Staged parses the output of the `git diff` command for the `source` path.
@@ -152,11 +193,11 @@ func (c *Parser) Staged(ctx context.Context, source string) (chan Commit, error)
 		cmd.Env = append(cmd.Env, fmt.Sprintf("GIT_DIR=%s", filepath.Join(absPath, ".git")))
 	}
 
-	return c.executeCommand(ctx, cmd)
+	return c.executeCommand(ctx, cmd, true)
 }
 
 // executeCommand runs an exec.Cmd, reads stdout and stderr, and waits for the Cmd to complete.
-func (c *Parser) executeCommand(ctx context.Context, cmd *exec.Cmd) (chan Commit, error) {
+func (c *Parser) executeCommand(ctx context.Context, cmd *exec.Cmd, isStaged bool) (chan Commit, error) {
 	commitChan := make(chan Commit, 64)
 
 	stdOut, err := cmd.StdoutPipe()
@@ -181,7 +222,7 @@ func (c *Parser) executeCommand(ctx context.Context, cmd *exec.Cmd) (chan Commit
 	}()
 
 	go func() {
-		c.FromReader(ctx, stdOut, commitChan)
+		c.FromReader(ctx, stdOut, commitChan, isStaged)
 		if err := cmd.Wait(); err != nil {
 			ctx.Logger().V(2).Info("Error waiting for git command to complete.", "error", err)
 		}
@@ -190,14 +231,15 @@ func (c *Parser) executeCommand(ctx context.Context, cmd *exec.Cmd) (chan Commit
 	return commitChan, nil
 }
 
-func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan chan Commit) {
+func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan chan Commit, isStaged bool) {
 	outReader := bufio.NewReader(stdOut)
 	var (
-		currentCommit        *Commit
-		currentDiff          *Diff
-		recentlyPassedAuthor bool
-		totalLogSize         int
+		currentCommit *Commit
+		currentDiff   *Diff
+
+		totalLogSize int
 	)
+	var latestState = Initial
 
 	defer common.RecoverWithExit(ctx)
 	defer close(commitChan)
@@ -205,12 +247,16 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan ch
 		if common.IsDone(ctx) {
 			break
 		}
+
 		line, err := outReader.ReadBytes([]byte("\n")[0])
 		if err != nil && len(line) == 0 {
 			break
 		}
+
 		switch {
-		case isCommitLine(line):
+		case isCommitLine(isStaged, latestState, line):
+			latestState = CommitLine
+
 			// If there is a currentDiff, add it to currentCommit.
 			if currentDiff != nil && currentDiff.Content.Len() > 0 {
 				currentCommit.Diffs = append(currentCommit.Diffs, *currentDiff)
@@ -230,16 +276,31 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan ch
 			if len(line) >= 47 {
 				currentCommit.Hash = string(line[7:47])
 			}
-		case isAuthorLine(line):
+		case isAuthorLine(isStaged, latestState, line):
+			latestState = AuthorLine
+
 			currentCommit.Author = strings.TrimRight(string(line[8:]), "\n")
-			recentlyPassedAuthor = true
-		case isDateLine(line):
+		case isDateLine(isStaged, latestState, line):
+			latestState = DateLine
+
 			date, err := time.Parse(c.dateFormat, strings.TrimSpace(string(line[6:])))
 			if err != nil {
 				ctx.Logger().V(2).Info("Could not parse date from git stream.", "error", err)
 			}
 			currentCommit.Date = date
-		case isDiffLine(line):
+		case isMessageStartLine(isStaged, latestState, line):
+			latestState = MessageStartLine
+			// NoOp
+		case isMessageLine(isStaged, latestState, line):
+			latestState = MessageLine
+
+			currentCommit.Message.Write(line[4:])
+		case isMessageEndLine(isStaged, latestState, line):
+			latestState = MessageEndLine
+			// NoOp
+		case isDiffLine(isStaged, latestState, line):
+			latestState = DiffLine
+
 			// This should never be nil, but check in case the stdin stream is messed up.
 			if currentCommit == nil {
 				currentCommit = &Commit{}
@@ -267,27 +328,29 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan ch
 				}
 			}
 			currentDiff = &Diff{}
-		case isModeLine(line):
+		case isModeLine(isStaged, latestState, line):
+			latestState = ModeLine
 			// NoOp
-		case isIndexLine(line):
-			recentlyPassedAuthor = false
+		case isIndexLine(isStaged, latestState, line):
+			latestState = IndexLine
 			// NoOp
-		case isPlusFileLine(line):
-			currentDiff.PathB = strings.TrimRight(strings.TrimRight(string(line[6:]), "\n"), "\t") // Trim the newline and tab characters. https://github.com/trufflesecurity/trufflehog/issues/1060
-		case isMinusFileLine(line):
-			// NoOp
-		case isPlusDiffLine(line):
-			currentDiff.Content.Write(line[1:])
-		case isMinusDiffLine(line):
-			// NoOp. We only care about additions.
-		case (isMessageLine(line) && recentlyPassedAuthor):
-			currentCommit.Message.Write(line[4:])
-		case isContextDiffLine(line):
-			currentDiff.Content.Write([]byte("\n"))
-		case isBinaryLine(line):
+		case isBinaryLine(isStaged, latestState, line):
+			latestState = BinaryFileLine
+
 			currentDiff.IsBinary = true
 			currentDiff.PathB = pathFromBinaryLine(line)
-		case isLineNumberDiffLine(line):
+		case isFromFileLine(isStaged, latestState, line):
+			latestState = FromFileLine
+			// NoOp
+		case isToFileLine(isStaged, latestState, line):
+			latestState = ToFileLine
+
+			// TODO: Is this fix still required?
+			currentDiff.PathB = strings.TrimRight(strings.TrimRight(string(line[6:]), "\n"), "\t") // Trim the newline and tab characters. https://github.com/trufflesecurity/trufflehog/issues/1060
+		case isHunkLineNumberLine(isStaged, latestState, line):
+			latestState = HunkLineNumberLine
+
+			// TODO: Is it still necessary to check whether the currentDiff is nil?
 			if currentDiff != nil && currentDiff.Content.Len() > 0 {
 				currentCommit.Diffs = append(currentCommit.Diffs, *currentDiff)
 			}
@@ -305,7 +368,54 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan ch
 					currentDiff.LineStart = lineStart
 				}
 			}
+		case isHunkContextLine(isStaged, latestState, line):
+			if latestState != HunkContentLine {
+				latestState = HunkContentLine
+			}
+			// TODO: Why do we care about this? It creates empty lines in the diff. If there are no plusLines, it's just newlines.
+			currentDiff.Content.Write([]byte("\n"))
+		case isHunkPlusLine(isStaged, latestState, line):
+			if latestState != HunkContentLine {
+				latestState = HunkContentLine
+			}
+
+			currentDiff.Content.Write(line[1:])
+		case isHunkMinusLine(isStaged, latestState, line):
+			if latestState != HunkContentLine {
+				latestState = HunkContentLine
+			}
+			// NoOp. We only care about additions.
+		case isHunkNewlineWarningLine(isStaged, latestState, line):
+			if latestState != HunkContentLine {
+				latestState = HunkContentLine
+			}
+			// NoOp
+		case isHunkEmptyLine(isStaged, latestState, line):
+			if latestState != HunkContentLine {
+				latestState = HunkContentLine
+			}
+			// NoOp
+		case isCommitSeparatorLine(isStaged, latestState, line):
+			// NoOp
+		default:
+			// Skip ahead until we find the next diff or commit.
+			if latestState == ParseFailure {
+				continue
+			}
+
+			// Here be dragons...
+			// Build an informative error message.
+			var err error
+			if currentCommit != nil && currentCommit.Hash != "" {
+				err = fmt.Errorf(`failed to parse line "%s" after state "%s" (commit=%s)`, line, latestState, currentCommit.Hash)
+			} else {
+				err = fmt.Errorf(`failed to parse line "%s" after state "%s"`, line, latestState)
+			}
+			ctx.Logger().V(2).Error(err, "Recovering at the latest commit or diff...\n")
+
+			latestState = ParseFailure
 		}
+
 		if currentDiff.Content.Len() > c.maxDiffSize {
 			ctx.Logger().V(2).Info(fmt.Sprintf(
 				"Diff for %s exceeded MaxDiffSize(%d)", currentDiff.PathB, c.maxDiffSize,
@@ -318,125 +428,142 @@ func (c *Parser) FromReader(ctx context.Context, stdOut io.Reader, commitChan ch
 	ctx.Logger().V(2).Info("finished parsing git log.", "total_log_size", totalLogSize)
 }
 
-func cleanupParse(currentCommit *Commit, currentDiff *Diff, commitChan chan Commit, totalLogSize *int) {
-	if currentDiff != nil && currentDiff.Content.Len() > 0 {
-		currentCommit.Diffs = append(currentCommit.Diffs, *currentDiff)
+// commit 7a95bbf0199e280a0e42dbb1d1a3f56cdd0f6e05
+func isCommitLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || !(latestState == Initial ||
+		latestState == MessageStartLine ||
+		latestState == MessageEndLine ||
+		latestState == ModeLine ||
+		latestState == IndexLine ||
+		latestState == BinaryFileLine ||
+		latestState == ToFileLine ||
+		latestState == HunkContentLine ||
+		latestState == ParseFailure) {
+		return false
 	}
-	if currentCommit != nil {
-		commitChan <- *currentCommit
-		if totalLogSize != nil {
-			*totalLogSize += currentCommit.Size
-		}
-	}
-}
 
-// Date:   Tue Aug 10 15:20:40 2021 +0100
-func isDateLine(line []byte) bool {
-	if len(line) > 7 && bytes.Equal(line[:5], []byte("Date:")) {
+	if len(line) > 7 && bytes.Equal(line[:7], []byte("commit ")) {
 		return true
 	}
 	return false
 }
 
 // Author: Bill Rich <bill.rich@trufflesec.com>
-func isAuthorLine(line []byte) bool {
+func isAuthorLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || !(latestState == CommitLine) {
+		return false
+	}
 	if len(line) > 8 && bytes.Equal(line[:7], []byte("Author:")) {
 		return true
 	}
 	return false
 }
 
-// commit 7a95bbf0199e280a0e42dbb1d1a3f56cdd0f6e05
-func isCommitLine(line []byte) bool {
-	if len(line) > 7 && bytes.Equal(line[:6], []byte("commit")) {
+// Date:   Tue Aug 10 15:20:40 2021 +0100
+func isDateLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || latestState != AuthorLine {
+		return false
+	}
+	if len(line) > 7 && bytes.Equal(line[:5], []byte("Date:")) {
 		return true
 	}
 	return false
 }
 
-// diff --git a/internal/addrs/move_endpoint_module.go b/internal/addrs/move_endpoint_module.go
-func isDiffLine(line []byte) bool {
-	if len(line) > 5 && bytes.Equal(line[:4], []byte("diff")) {
-		return true
+// Line directly after Date with only a newline.
+func isMessageStartLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || latestState != DateLine {
+		return false
 	}
-	return false
-}
-
-// index 1ed6fbee1..aea1e643a 100644
-func isIndexLine(line []byte) bool {
-	if len(line) > 6 && bytes.Equal(line[:5], []byte("index")) {
-		return true
-	}
-	return false
-}
-
-// new file mode 100644
-func isModeLine(line []byte) bool {
-	if len(line) > 13 && bytes.Equal(line[:13], []byte("new file mode")) {
-		return true
-	}
-	return false
-}
-
-// --- a/internal/addrs/move_endpoint_module.go
-func isMinusFileLine(line []byte) bool {
-	if len(line) >= 6 && bytes.Equal(line[:3], []byte("---")) {
-		return true
-	}
-	return false
-}
-
-// +++ b/internal/addrs/move_endpoint_module.go
-func isPlusFileLine(line []byte) bool {
-	if len(line) >= 6 && bytes.Equal(line[:4], []byte("+++ ")) {
-		return true
-	}
-	return false
-}
-
-// +fmt.Println("ok")
-func isPlusDiffLine(line []byte) bool {
-	if len(line) >= 1 && bytes.Equal(line[:1], []byte("+")) {
-		return true
-	}
-	return false
-}
-
-// -fmt.Println("ok")
-func isMinusDiffLine(line []byte) bool {
-	if len(line) >= 1 && bytes.Equal(line[:1], []byte("-")) {
-		return true
-	}
-	return false
-}
-
-// fmt.Println("ok")
-func isContextDiffLine(line []byte) bool {
-	if len(line) >= 1 && bytes.Equal(line[:1], []byte(" ")) {
+	// TODO: Improve the implementation of this and isMessageEndLine
+	if len(strings.TrimRight(string(line[:]), "\r\n")) == 0 {
 		return true
 	}
 	return false
 }
 
 // Line that starts with 4 spaces
-func isMessageLine(line []byte) bool {
+func isMessageLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || !(latestState == MessageStartLine || latestState == MessageLine) {
+		return false
+	}
 	if len(line) > 4 && bytes.Equal(line[:4], []byte("    ")) {
 		return true
 	}
 	return false
 }
 
-// Binary files /dev/null and b/plugin.sig differ
-func isBinaryLine(line []byte) bool {
-	if len(line) > 7 && bytes.Equal(line[:6], []byte("Binary")) {
+// Line directly after MessageLine with only a newline.
+func isMessageEndLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if isStaged || latestState != MessageLine {
+		return false
+	}
+	if len(strings.TrimRight(string(line[:]), "\r\n")) == 0 {
 		return true
 	}
 	return false
 }
 
-// @@ -298 +298 @@ func maxRetryErrorHandler(resp *http.Response, err error, numTries int)
-func isLineNumberDiffLine(line []byte) bool {
-	if len(line) >= 8 && bytes.Equal(line[:2], []byte("@@")) {
+// diff --git a/internal/addrs/move_endpoint_module.go b/internal/addrs/move_endpoint_module.go
+func isDiffLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == MessageStartLine || // Empty commit messages can go from MessageStart->Diff
+		latestState == MessageEndLine ||
+		latestState == BinaryFileLine ||
+		latestState == IndexLine ||
+		latestState == HunkContentLine ||
+		latestState == ParseFailure) {
+		if latestState == Initial && !isStaged {
+			return false
+		}
+	}
+	if len(line) > 11 && bytes.Equal(line[:11], []byte("diff --git ")) {
+		return true
+	}
+	return false
+}
+
+// old mode 100644
+// new mode 100755
+// new file mode 100644
+// similarity index 100%
+// rename from old.txt
+// rename to new.txt
+// deleted file mode 100644
+func isModeLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == DiffLine || latestState == ModeLine) {
+		return false
+	}
+	// This could probably be better written.
+	if (len(line) > 17 && bytes.Equal(line[:17], []byte("deleted file mode"))) ||
+		(len(line) > 16 && bytes.Equal(line[:16], []byte("similarity index"))) ||
+		(len(line) > 13 && bytes.Equal(line[:13], []byte("new file mode"))) ||
+		(len(line) > 11 && bytes.Equal(line[:11], []byte("rename from"))) ||
+		(len(line) > 9 && bytes.Equal(line[:9], []byte("rename to"))) ||
+		(len(line) > 8 && bytes.Equal(line[:8], []byte("old mode"))) ||
+		(len(line) > 8 && bytes.Equal(line[:8], []byte("new mode"))) {
+		return true
+	}
+	return false
+}
+
+// index 1ed6fbee1..aea1e643a 100644
+// index 00000000..e69de29b
+func isIndexLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == DiffLine || latestState == ModeLine) {
+		return false
+	}
+	if len(line) > 6 && bytes.Equal(line[:6], []byte("index ")) {
+		return true
+	}
+	return false
+}
+
+// Binary files /dev/null and b/plugin.sig differ
+func isBinaryLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if latestState != IndexLine {
+		return false
+	}
+	if len(line) > 7 && bytes.Equal(line[:6], []byte("Binary")) {
 		return true
 	}
 	return false
@@ -453,3 +580,117 @@ func pathFromBinaryLine(line []byte) string {
 	bRaw := sbytes[1]
 	return string(bRaw[:len(bRaw)-7]) // drop the "b/" and " differ"
 }
+
+// --- a/internal/addrs/move_endpoint_module.go
+func isFromFileLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if latestState != IndexLine {
+		return false
+	}
+	if len(line) >= 6 && bytes.Equal(line[:4], []byte("--- ")) {
+		return true
+	}
+	return false
+}
+
+// +++ b/internal/addrs/move_endpoint_module.go
+func isToFileLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if latestState != FromFileLine {
+		return false
+	}
+	if len(line) >= 6 && bytes.Equal(line[:4], []byte("+++ ")) {
+		return true
+	}
+	return false
+}
+
+// @@ -298 +298 @@ func maxRetryErrorHandler(resp *http.Response, err error, numTries int)
+func isHunkLineNumberLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == ToFileLine || latestState == HunkContentLine) {
+		return false
+	}
+	if len(line) >= 8 && bytes.Equal(line[:2], []byte("@@")) {
+		return true
+	}
+	return false
+}
+
+// fmt.Println("ok")
+// (There's a space before `fmt` that gets removed by the formatter.)
+func isHunkContextLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == HunkLineNumberLine || latestState == HunkContentLine) {
+		return false
+	}
+	if len(line) >= 1 && bytes.Equal(line[:1], []byte(" ")) {
+		return true
+	}
+	return false
+}
+
+// +fmt.Println("ok")
+func isHunkPlusLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == HunkLineNumberLine || latestState == HunkContentLine) {
+		return false
+	}
+	if len(line) >= 1 && bytes.Equal(line[:1], []byte("+")) {
+		return true
+	}
+	return false
+}
+
+// -fmt.Println("ok")
+func isHunkMinusLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == HunkLineNumberLine || latestState == HunkContentLine) {
+		return false
+	}
+	if len(line) >= 1 && bytes.Equal(line[:1], []byte("-")) {
+		return true
+	}
+	return false
+}
+
+// \ No newline at end of file
+func isHunkNewlineWarningLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if latestState != HunkContentLine {
+		return false
+	}
+	if len(line) >= 27 && bytes.Equal(line[:27], []byte("\\ No newline at end of file")) {
+		return true
+	}
+	return false
+}
+
+// Newline after hunk, or an empty line, e.g.
+// +}
+//
+// commit 00920984e3435057f09cee5468850f7546dfa637 (tag: v3.42.0)
+func isHunkEmptyLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if !(latestState == HunkLineNumberLine || latestState == HunkContentLine) {
+		return false
+	}
+	// TODO: Can this also be `\n\r`?
+	if len(line) == 1 && bytes.Equal(line[:1], []byte("\n")) {
+		return true
+	}
+	return false
+}
+
+func isCommitSeparatorLine(isStaged bool, latestState ParseState, line []byte) bool {
+	if (latestState == ModeLine || latestState == IndexLine || latestState == BinaryFileLine || latestState == ToFileLine) &&
+		len(line) == 1 && bytes.Equal(line[:1], []byte("\n")) {
+		return true
+	}
+	return false
+}
+
+func cleanupParse(currentCommit *Commit, currentDiff *Diff, commitChan chan Commit, totalLogSize *int) {
+	// Ignore empty or binary diffs (this condition may be redundant).
+	if currentDiff != nil && currentDiff.Content.Len() > 0 {
+		currentCommit.Diffs = append(currentCommit.Diffs, *currentDiff)
+	}
+	if currentCommit != nil {
+		commitChan <- *currentCommit
+		if totalLogSize != nil {
+			*totalLogSize += currentCommit.Size
+		}
+	}
+}

pkg/handlers/archive.go

@@ -6,12 +6,14 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"strings"
 	"time"
 
 	"github.com/mholt/archiver/v4"
 
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
 	logContext "github.com/trufflesecurity/trufflehog/v3/pkg/context"
+	"github.com/trufflesecurity/trufflehog/v3/pkg/sources"
 )
 
 type ctxKey int
@@ -52,14 +54,14 @@ func SetArchiveMaxTimeout(timeout time.Duration) {
 }
 
 // FromFile extracts the files from an archive.
-func (d *Archive) FromFile(originalCtx context.Context, data io.Reader) chan ([]byte) {
-	archiveChan := make(chan ([]byte), 512)
+func (d *Archive) FromFile(originalCtx context.Context, data io.Reader) chan ChunkOpt {
+	archiveChan := make(chan ChunkOpt, 512)
 	go func() {
 		ctx, cancel := context.WithTimeout(originalCtx, maxTimeout)
 		logger := logContext.AddLogger(ctx).Logger()
 		defer cancel()
 		defer close(archiveChan)
-		err := d.openArchive(ctx, 0, data, archiveChan)
+		err := d.openArchive(ctx, nil, data, archiveChan)
 		if err != nil {
 			if errors.Is(err, archiver.ErrNoMatch) {
 				return
@@ -71,18 +73,21 @@ func (d *Archive) FromFile(originalCtx context.Context, data io.Reader) chan ([]
 }
 
 // openArchive takes a reader and extracts the contents up to the maximum depth.
-func (d *Archive) openArchive(ctx context.Context, depth int, reader io.Reader, archiveChan chan ([]byte)) error {
-	if depth >= maxDepth {
+func (d *Archive) openArchive(ctx context.Context, depth []string, reader io.Reader, archiveChan chan ChunkOpt) error {
+	if len(depth) >= maxDepth {
 		return fmt.Errorf("max archive depth reached")
 	}
 	format, reader, err := archiver.Identify("", reader)
 	if err != nil {
-		if errors.Is(err, archiver.ErrNoMatch) && depth > 0 {
+		if errors.Is(err, archiver.ErrNoMatch) && len(depth) > 0 {
 			chunkSize := 10 * 1024
 			for {
-				chunk := make([]byte, chunkSize)
-				n, _ := reader.Read(chunk)
-				archiveChan <- chunk
+				data := make([]byte, chunkSize)
+				n, _ := reader.Read(data)
+				archiveChan <- func(chunk *sources.Chunk) {
+					chunk.Data = data
+					chunk.HandleMetadata = map[string]string{"archive path": strings.Join(depth, " > ")}
+				}
 				if n < chunkSize {
 					break
 				}
@@ -102,9 +107,9 @@ func (d *Archive) openArchive(ctx context.Context, depth int, reader io.Reader,
 			return err
 		}
 		newReader := bytes.NewReader(fileBytes)
-		return d.openArchive(ctx, depth+1, newReader, archiveChan)
+		return d.openArchive(ctx, append(depth, "decompressed"), newReader, archiveChan)
 	case archiver.Extractor:
-		err := archive.Extract(context.WithValue(ctx, depthKey, depth+1), reader, nil, d.extractorHandler(archiveChan))
+		err := archive.Extract(context.WithValue(ctx, depthKey, depth), reader, nil, d.extractorHandler(archiveChan))
 		if err != nil {
 			return err
 		}
@@ -129,12 +134,12 @@ func (d *Archive) IsFiletype(ctx context.Context, reader io.Reader) (io.Reader,
 }
 
 // extractorHandler is applied to each file in an archiver.Extractor file.
-func (d *Archive) extractorHandler(archiveChan chan ([]byte)) func(context.Context, archiver.File) error {
+func (d *Archive) extractorHandler(archiveChan chan ChunkOpt) func(context.Context, archiver.File) error {
 	return func(ctx context.Context, f archiver.File) error {
 		logger := logContext.AddLogger(ctx).Logger()
 		logger.V(5).Info("Handling extracted file.", "filename", f.Name())
-		depth := 0
-		if ctxDepth, ok := ctx.Value(depthKey).(int); ok {
+		var depth []string
+		if ctxDepth, ok := ctx.Value(depthKey).([]string); ok {
 			depth = ctxDepth
 		}
 
@@ -148,7 +153,7 @@ func (d *Archive) extractorHandler(archiveChan chan ([]byte)) func(context.Conte
 		}
 		fileContent := bytes.NewReader(fileBytes)
 
-		err = d.openArchive(ctx, depth, fileContent, archiveChan)
+		err = d.openArchive(ctx, append(depth, f.Name()), fileContent, archiveChan)
 		if err != nil {
 			return err
 		}

pkg/handlers/archive_test.go

@@ -85,9 +85,11 @@ func TestArchiveHandler(t *testing.T) {
 		count := 0
 		re := regexp.MustCompile(testCase.matchString)
 		matched := false
-		for chunk := range archiveChan {
+		var chunk sources.Chunk
+		for opt := range archiveChan {
 			count++
-			if re.Match(chunk) {
+			opt(&chunk)
+			if re.Match(chunk.Data) {
 				matched = true
 			}
 		}

pkg/handlers/handlers.go

@@ -13,8 +13,10 @@ func DefaultHandlers() []Handler {
 	}
 }
 
+type ChunkOpt func(*sources.Chunk)
+
 type Handler interface {
-	FromFile(context.Context, io.Reader) chan ([]byte)
+	FromFile(context.Context, io.Reader) chan ChunkOpt
 	IsFiletype(context.Context, io.Reader) (io.Reader, bool)
 	New()
 }
@@ -38,13 +40,13 @@ func HandleFile(ctx context.Context, file io.Reader, chunkSkel *sources.Chunk, c
 	handlerChan := handler.FromFile(ctx, file)
 	for {
 		select {
-		case data, open := <-handlerChan:
+		case opt, open := <-handlerChan:
 			if !open {
 				// We finished reading everything from handlerChan.
 				return true
 			}
 			chunk := *chunkSkel
-			chunk.Data = data
+			opt(&chunk)
 			// Send data on chunksChan.
 			select {
 			case chunksChan <- &chunk:

pkg/output/plain.go

@@ -96,9 +96,9 @@ func structToMap(obj interface{}) (m map[string]map[string]interface{}, err erro
 }
 
 type outputFormat struct {
-	DetectorType,
-	DecoderType string
-	Verified bool
-	Raw      string
+	DetectorType string
+	DecoderType  string
+	Verified     bool
+	Raw          string
 	*source_metadatapb.MetaData
 }

pkg/sources/chunker.go

@@ -12,6 +12,8 @@ const (
 	ChunkSize = 10 * 1024
 	// PeekSize is the size of the peek into the previous chunk.
 	PeekSize = 3 * 1024
+	// TotalChunkSize is the total size of a chunk with peek data.
+	TotalChunkSize = ChunkSize + PeekSize
 )
 
 // Chunker takes a chunk and splits it into chunks of ChunkSize.
@@ -19,27 +21,31 @@ func Chunker(originalChunk *Chunk) chan *Chunk {
 	chunkChan := make(chan *Chunk)
 	go func() {
 		defer close(chunkChan)
-		if len(originalChunk.Data) <= ChunkSize+PeekSize {
+		if len(originalChunk.Data) <= TotalChunkSize {
 			chunkChan <- originalChunk
 			return
 		}
 		r := bytes.NewReader(originalChunk.Data)
 		reader := bufio.NewReaderSize(bufio.NewReader(r), ChunkSize)
 		for {
-			chunkBytes := make([]byte, ChunkSize)
+			chunkBytes := make([]byte, TotalChunkSize)
 			chunk := *originalChunk
+			chunkBytes = chunkBytes[:ChunkSize]
 			n, err := reader.Read(chunkBytes)
 			if err != nil && !errors.Is(err, io.EOF) {
 				break
 			}
+			if n == 0 {
+				if errors.Is(err, io.EOF) {
+					break
+				}
+				continue
+			}
 			peekData, _ := reader.Peek(PeekSize)
-			chunk.Data = append(chunkBytes[:n], peekData...)
-			if n > 0 {
-				chunkChan <- &chunk
-			}
-			if errors.Is(err, io.EOF) {
-				break
-			}
+			copy(chunkBytes[n:], peekData)
+			chunk.Data = chunkBytes[:n+len(peekData)]
+
+			chunkChan <- &chunk
 		}
 	}()
 	return chunkChan

pkg/sources/chunker_test.go

@@ -72,3 +72,15 @@ func TestChunker(t *testing.T) {
 	}
 
 }
+
+func BenchmarkChunker(b *testing.B) {
+	data := bytes.Repeat([]byte("a"), ChunkSize*100)
+	chunk := &Chunk{
+		Data: data,
+	}
+	b.ResetTimer()
+	for i := 0; i < b.N; i++ {
+		for range Chunker(chunk) {
+		}
+	}
+}

pkg/sources/git/git.go

@@ -51,9 +51,14 @@ type Git struct {
 	jobID              int64
 	sourceMetadataFunc func(file, email, commit, timestamp, repository string, line int64) *source_metadatapb.MetaData
 	verify             bool
+	metrics            metrics
 	concurrency        *semaphore.Weighted
 }
 
+type metrics struct {
+	commitsScanned int
+}
+
 func NewGit(sourceType sourcespb.SourceType, jobID, sourceID int64, sourceName string, verify bool, concurrency int,
 	sourceMetadataFunc func(file, email, commit, timestamp, repository string, line int64) *source_metadatapb.MetaData,
 ) *Git {
@@ -339,6 +344,10 @@ func CloneRepoUsingSSH(ctx context.Context, gitUrl string, args ...string) (stri
 	return CloneRepo(ctx, userInfo, gitUrl, args...)
 }
 
+func (s *Git) CommitsScanned() int {
+	return s.metrics.commitsScanned
+}
+
 func (s *Git) ScanCommits(ctx context.Context, repo *git.Repository, path string, scanOptions *ScanOptions, chunksChan chan *sources.Chunk) error {
 	if err := GitCmdCheck(); err != nil {
 		return err
@@ -371,6 +380,7 @@ func (s *Git) ScanCommits(ctx context.Context, repo *git.Repository, path string
 			break
 		}
 		depth++
+		s.metrics.commitsScanned++
 		logger.V(5).Info("scanning commit", "commit", commit.Hash)
 		for _, diff := range commit.Diffs {
 			if !scanOptions.Filter.Pass(diff.PathB) {

pkg/sources/sources.go

@@ -21,7 +21,8 @@ type Chunk struct {
 	SourceType sourcespb.SourceType
 	// SourceMetadata holds the context of where the Chunk was found.
 	SourceMetadata *source_metadatapb.MetaData
-
+	// HandleMetadata holds the metadata from a handler if one was used.
+	HandleMetadata map[string]string
 	// Data is the data to decode and scan.
 	Data []byte
 	// Verify specifies whether any secrets in the Chunk should be verified.