fix: strip publisher URL from RPM CPE vendor (#5081)

Signed-off-by: Eljees <yurytumanov.r@yandex.ru>

syft/pkg/cataloger/internal/cpegenerate/rpm.go

@@ -1,26 +1,49 @@
 package cpegenerate
 
-import "github.com/anchore/syft/syft/pkg"
+import (
+	"net/url"
+	"strings"
+
+	"github.com/anchore/syft/syft/pkg"
+)
 
 func candidateVendorsForRPM(p pkg.Package) fieldCandidateSet {
 	vendors := newFieldCandidateSet()
+	var vendor string
 
 	switch m := p.Metadata.(type) {
 	case pkg.RpmDBEntry:
-		if m.Vendor != "" {
-			vendors.add(fieldCandidate{
-				value:                 normalizeName(m.Vendor),
-				disallowSubSelections: true,
-			})
-		}
+		vendor = m.Vendor
 	case pkg.RpmArchive:
-		if m.Vendor != "" {
-			vendors.add(fieldCandidate{
-				value:                 normalizeName(m.Vendor),
-				disallowSubSelections: true,
-			})
-		}
+		vendor = m.Vendor
+	}
+
+	vendor = stripTrailingURL(vendor)
+	if vendor != "" {
+		vendors.add(fieldCandidate{
+			value:                 normalizeName(vendor),
+			disallowSubSelections: true,
+		})
 	}
 
 	return vendors
 }
+
+func stripTrailingURL(value string) string {
+	trimmed := strings.TrimSpace(value)
+	if !strings.HasSuffix(trimmed, ">") {
+		return value
+	}
+
+	open := strings.LastIndex(trimmed, "<")
+	if open == -1 {
+		return value
+	}
+
+	parsed, err := url.Parse(trimmed[open+1 : len(trimmed)-1])
+	if err != nil || parsed.Host == "" || (parsed.Scheme != "http" && parsed.Scheme != "https") {
+		return value
+	}
+
+	return strings.TrimSpace(trimmed[:open])
+}

syft/pkg/cataloger/internal/cpegenerate/rpm_test.go

@@ -0,0 +1,60 @@
+package cpegenerate
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	"github.com/anchore/syft/syft/pkg"
+)
+
+func TestCandidateVendorsForRPM(t *testing.T) {
+	tests := []struct {
+		name     string
+		metadata any
+		expected []string
+	}{
+		{
+			name: "database vendor with publisher URL",
+			metadata: pkg.RpmDBEntry{
+				Vendor: "SUSE LLC <https://www.suse.com/>",
+			},
+			expected: []string{"susellc"},
+		},
+		{
+			name: "archive vendor with publisher URL",
+			metadata: pkg.RpmArchive{
+				Vendor: "SUSE LLC <https://www.suse.com/>",
+			},
+			expected: []string{"susellc"},
+		},
+		{
+			name: "plain vendor",
+			metadata: pkg.RpmDBEntry{
+				Vendor: "Red Hat, Inc.",
+			},
+			expected: []string{"redhat"},
+		},
+		{
+			name: "non-URL angle-bracket suffix",
+			metadata: pkg.RpmDBEntry{
+				Vendor: "Example <support@example.com>",
+			},
+			expected: []string{"example<support@example.com>"},
+		},
+		{
+			name: "non-HTTP URL",
+			metadata: pkg.RpmDBEntry{
+				Vendor: "Example <ftp://example.com>",
+			},
+			expected: []string{"example<ftp://example.com>"},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			p := pkg.Package{Metadata: test.metadata}
+			assert.ElementsMatch(t, test.expected, candidateVendorsForRPM(p).uniqueValues())
+		})
+	}
+}