syz-cluster/series-tracker: collect Cc addresses

Extract Cc addresses for the processed patch series.

2 files changed, 81 insertions, 2 deletions

diff --git a/syz-cluster/series-tracker/main.go b/syz-cluster/series-tracker/main.go
index 6a0379fc2..a9de34a31 100644
--- a/syz-cluster/series-tracker/main.go
+++ b/syz-cluster/series-tracker/main.go
@@ -4,12 +4,17 @@
 package main
 
 import (
+	"bytes"
 	"context"
+	"errors"
 	"flag"
 	"fmt"
 	"log"
+	"maps"
 	"path/filepath"
 	"regexp"
+	"slices"
+	"sort"
 	"time"
 
 	"github.com/google/syzkaller/pkg/email"
@@ -141,11 +146,18 @@ func (sf *SeriesFetcher) handleSeries(ctx context.Context, series *lore.Series,
 		Link:        "https://lore.kernel.org/all/" + series.MessageID,
 		PublishedAt: date,
 	}
-	for _, patch := range series.Patches {
-		body, err := idToReader[patch.MessageID].Read()
+	sp := seriesProcessor{}
+	for i, patch := range series.Patches {
+		raw, err := idToReader[patch.MessageID].Read()
 		if err != nil {
 			return fmt.Errorf("failed to extract %q: %w", patch.MessageID, err)
 		}
+		body, err := sp.Process(raw)
+		if err != nil {
+			// Fall back to the raw message.
+			body = raw
+			log.Printf("failed to parse %d: %v", i, err)
+		}
 		apiSeries.Patches = append(apiSeries.Patches, api.SeriesPatch{
 			Seq:   patch.Seq,
 			Title: patch.Subject,
@@ -153,6 +165,7 @@ func (sf *SeriesFetcher) handleSeries(ctx context.Context, series *lore.Series,
 			Body:  body,
 		})
 	}
+	apiSeries.Cc = sp.Emails()
 	ret, err := sf.client.UploadSeries(ctx, apiSeries)
 	if err != nil {
 		return fmt.Errorf("failed to save series: %w", err)
@@ -170,6 +183,27 @@ func (sf *SeriesFetcher) handleSeries(ctx context.Context, series *lore.Series,
 	return nil
 }
 
+type seriesProcessor map[string]struct{}
+
+var errFailedToParse = errors.New("failed to parse the email")
+
+func (sp seriesProcessor) Process(raw []byte) ([]byte, error) {
+	msg, err := email.Parse(bytes.NewReader(raw), nil, nil, nil)
+	if err != nil {
+		return raw, fmt.Errorf("%w: %w", errFailedToParse, err)
+	}
+	for _, email := range msg.Cc {
+		sp[email] = struct{}{}
+	}
+	return []byte(msg.Body), nil
+}
+
+func (sp seriesProcessor) Emails() []string {
+	list := slices.Collect(maps.Keys(sp))
+	sort.Strings(list)
+	return list
+}
+
 func logSeries(series *lore.Series) {
 	log.Printf("series ID=%s Subject=%s Patches=%d Version=%d Corrupted=%q",
 		series.MessageID, series.Subject, len(series.Patches), series.Version,
diff --git a/syz-cluster/series-tracker/main_test.go b/syz-cluster/series-tracker/main_test.go
new file mode 100644
index 000000000..805335076
--- /dev/null
+++ b/syz-cluster/series-tracker/main_test.go
@@ -0,0 +1,45 @@
+// Copyright 2025 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package main
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestSeriesProcessor(t *testing.T) {
+	emails := []string{
+		`Date: Sun, 7 May 2017 19:54:00 -0700
+Message-ID: <123>
+Subject: test subject
+From: Bob <bob@example.com>
+To: A <a@a.com>
+Cc: B <b@b.com>, C <b@b.com>
+
+first body`,
+		`Date: Sun, 7 May 2017 19:55:00 -0700
+Message-ID: <234>
+Subject: test subject2
+From: Bob <bob@example.com>
+To: A <a@a.com>
+Cc: D <d@d.com>
+
+second body`,
+	}
+	bodies := []string{"first body", "second body"}
+
+	sp := seriesProcessor{}
+	for i, raw := range emails {
+		body, err := sp.Process([]byte(raw))
+		if err != nil {
+			t.Fatal(err)
+		}
+		assert.Equal(t, []byte(bodies[i]), body)
+	}
+	assert.Equal(t, []string{
+		"a@a.com", "b@b.com",
+		"bob@example.com", "d@d.com",
+	}, sp.Emails())
+}