fix(binary): preserve semver build metadata in ffmpeg version (#5452)

Signed-off-by: Adam Shaver <ashaver@nvidia.com>

syft/pkg/cataloger/binary/classifier_cataloger_test.go

@@ -1,6 +1,7 @@
 package binary
 
 import (
+	"bytes"
 	"context"
 	"errors"
 	"flag"
@@ -20,6 +21,7 @@ import (
 	"github.com/anchore/syft/internal/testutils"
 	"github.com/anchore/syft/syft/cpe"
 	"github.com/anchore/syft/syft/file"
+	"github.com/anchore/syft/syft/internal/unionreader"
 	"github.com/anchore/syft/syft/pkg"
 	"github.com/anchore/syft/syft/pkg/cataloger/binary/internal/manager/testutil"
 	"github.com/anchore/syft/syft/pkg/cataloger/internal/binutils"
@@ -3192,6 +3194,134 @@ func Test_Cataloger_DefaultClassifiers_PositiveCases_Image(t *testing.T) {
 // Test_Cataloger_PackageTypes covers what assertPackagesAreEqual deliberately ignores: the package type and CPEs.
 // The perl interpreter is reported as cpan so it lines up with the rest of the CPAN ecosystem, everything else
 // keeps the default binary type.
+// Test_Cataloger_ffmpeg_versionSuffix verifies the ffmpeg classifiers no longer
+// unconditionally truncate the embedded "FFmpeg version <ver>" string to X.Y[.Z].
+// Semver build metadata (a backported-patch stamp such as
+// "9.0.3+dev.9.0.2.p1.eb9cef3") must be preserved in the reported version, because
+// it compares equal to its core version and so does not disturb vulnerability
+// matching. A "-" pre-release suffix (a distro revision such as "6.1.1-3ubuntu5")
+// must stay truncated: it sorts below the release in semver and would corrupt
+// version-range matching. Suffix-free binaries match exactly as before, and
+// trailing non-version text (or NUL padding) is never absorbed into the version.
+func Test_Cataloger_ffmpeg_versionSuffix(t *testing.T) {
+	// collect the ffmpeg classifiers under test, keyed by their FileGlob
+	classifiers := map[string]binutils.Classifier{}
+	for _, c := range DefaultClassifiers() {
+		if c.Package == "ffmpeg" {
+			classifiers[c.FileGlob] = c
+		}
+	}
+	require.NotEmpty(t, classifiers, "expected to find ffmpeg classifiers")
+
+	tests := []struct {
+		name     string
+		fileGlob string // which classifier to run
+		data     string // simulated binary contents
+		want     string // expected reported version ("" = no match expected)
+	}{
+		{
+			name:     "library bare version unchanged",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 7.1.1",
+			want:     "7.1.1",
+		},
+		{
+			name:     "library two-component version unchanged",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 6.0",
+			want:     "6.0",
+		},
+		{
+			name:     "library build metadata with patch provenance preserved",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 9.0.3+dev.9.0.2.p1.eb9cef3",
+			want:     "9.0.3+dev.9.0.2.p1.eb9cef3",
+		},
+		{
+			name:     "library build metadata short sha preserved",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 9.0.2+eb9cef3",
+			want:     "9.0.2+eb9cef3",
+		},
+		{
+			// a "-" pre-release suffix is intentionally NOT captured: it sorts below
+			// the release in semver and would corrupt version-range vulnerability
+			// matching, so distro revisions stay truncated exactly as before
+			name:     "library distro hyphen revision truncated",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 6.1.1-3ubuntu5",
+			want:     "6.1.1",
+		},
+		{
+			name:     "library distro dotted hyphen revision truncated",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 4.4.2-0ubuntu0.22.04.1",
+			want:     "4.4.2",
+		},
+		{
+			name:     "library trailing text not captured",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 6.1.1 Copyright (c) 2000-2023 the FFmpeg developers",
+			want:     "6.1.1",
+		},
+		{
+			name:     "library NUL-padded version not over-captured",
+			fileGlob: "**/libav*",
+			data:     "FFmpeg version 9.0.3\x00libavformat.so",
+			want:     "9.0.3",
+		},
+		{
+			name:     "libswresample build metadata preserved",
+			fileGlob: "**/libswresample*",
+			data:     "FFmpeg version 9.0.3+dev.9.0.2.p1.eb9cef3",
+			want:     "9.0.3+dev.9.0.2.p1.eb9cef3",
+		},
+		{
+			// the compiled-in printf format string in the ffmpeg binary is the
+			// literal "%s version <ver>" (the program name is substituted at runtime)
+			name:     "binary build metadata preserved",
+			fileGlob: "**/ffmpeg",
+			data:     "%s version 9.0.3+dev.9.0.2.p1.eb9cef3",
+			want:     "9.0.3+dev.9.0.2.p1.eb9cef3",
+		},
+		{
+			name:     "binary bare version unchanged",
+			fileGlob: "**/ffmpeg",
+			data:     "%s version 7.1.1",
+			want:     "7.1.1",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			cls, ok := classifiers[tt.fileGlob]
+			require.Truef(t, ok, "no ffmpeg classifier for glob %q", tt.fileGlob)
+
+			ctx := binutils.MatcherContext{
+				GetReader: func(binutils.MatcherContext) (unionreader.UnionReader, error) {
+					return unionreader.GetUnionReader(io.NopCloser(bytes.NewBufferString(tt.data)))
+				},
+			}
+
+			pkgs, err := cls.EvidenceMatcher(cls, ctx)
+			require.NoError(t, err)
+
+			if tt.want == "" {
+				require.Empty(t, pkgs, "expected no package match")
+				return
+			}
+			require.Len(t, pkgs, 1, "expected exactly one package")
+			assert.Equal(t, tt.want, pkgs[0].Version)
+			// the fuller version must also flow into the PURL so the SBOM is not lossy;
+			// compare the decoded PURL version, since packageurl-go percent-encodes
+			// reserved characters (e.g. '+' -> %2B) in the serialized form
+			parsed, err := packageurl.FromString(pkgs[0].PURL)
+			require.NoError(t, err, "PURL should parse")
+			assert.Equal(t, tt.want, parsed.Version)
+		})
+	}
+}
+
 func Test_Cataloger_PackageTypes(t *testing.T) {
 	tests := []struct {
 		logicalFixture string

syft/pkg/cataloger/binary/classifiers.go

@@ -972,7 +972,10 @@ func DefaultClassifiers() []binutils.Classifier {
 			EvidenceMatcher: m.FileContentsVersionMatcher(
 				// Pattern found in the binary: "%s version 7.1.1" or "%s version 6.0"
 				// When executed outputs: "ffmpeg version 7.1.1" or "ffmpeg version 6.0"
-				`(?m)%s version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?)`,
+				// The optional "+" suffix preserves semver build metadata (e.g. a
+				// backported-patch stamp like "9.0.3+dev.9.0.2.p1.eb9cef3"); see the
+				// ffmpeg-library classifier for why "-" pre-release suffixes stay truncated.
+				`(?m)%s version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?(?:\+[0-9A-Za-z]+(?:\.[0-9A-Za-z]+)*)?)`,
 			),
 			Package: "ffmpeg",
 			PURL:    mustPURL("pkg:generic/ffmpeg@version"),
@@ -982,8 +985,15 @@ func DefaultClassifiers() []binutils.Classifier {
 			Class:    "ffmpeg-library",
 			FileGlob: "**/libav*",
 			EvidenceMatcher: binutils.MatchAny(
-				// Primary pattern: FFmpeg version found in most libraries
-				m.FileContentsVersionMatcher(`(?m)FFmpeg version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?)`),
+				// Primary pattern: FFmpeg version found in most libraries.
+				// The optional "+" suffix preserves semver build metadata (e.g. a
+				// backported-patch stamp like "9.0.3+dev.9.0.2.p1.eb9cef3") instead of
+				// truncating to X.Y[.Z]. Build metadata compares equal to its core
+				// version, so this does not change vulnerability matching. A "-"
+				// pre-release suffix (e.g. a distro revision "6.1.1-3ubuntu5") is
+				// deliberately NOT captured: it sorts below the release in semver and
+				// would corrupt version-range matching, so it stays truncated as before.
+				m.FileContentsVersionMatcher(`(?m)FFmpeg version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?(?:\+[0-9A-Za-z]+(?:\.[0-9A-Za-z]+)*)?)`),
 				// Fallback: library-specific version patterns for libavcodec and libavformat
 				m.FileContentsVersionMatcher(`(?m)Lavc(?P<version>[0-9]+\.[0-9]+\.[0-9]+)`),
 				m.FileContentsVersionMatcher(`(?m)Lavf(?P<version>[0-9]+\.[0-9]+\.[0-9]+)`),
@@ -996,8 +1006,9 @@ func DefaultClassifiers() []binutils.Classifier {
 			Class:    "ffmpeg-library",
 			FileGlob: "**/libswresample*",
 			EvidenceMatcher: m.FileContentsVersionMatcher(
-				// FFmpeg version pattern for libswresample
-				`(?m)FFmpeg version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?)`),
+				// FFmpeg version pattern for libswresample. Same optional "+" build
+				// metadata handling as the ffmpeg-library classifier above.
+				`(?m)FFmpeg version (?P<version>[0-9]+\.[0-9]+(\.[0-9]+)?(?:\+[0-9A-Za-z]+(?:\.[0-9A-Za-z]+)*)?)`),
 			Package: "ffmpeg",
 			PURL:    mustPURL("pkg:generic/ffmpeg@version"),
 			CPEs:    singleCPE("cpe:2.3:a:ffmpeg:ffmpeg:*:*:*:*:*:*:*:*", cpe.NVDDictionaryLookupSource),