read relative etc/apk/repositories for alpine version when no OS provided (#1615)

Signed-off-by: Avi Deitcher <avi@deitcher.net>

syft/pkg/cataloger/apkdb/cataloger_test.go

@@ -24,6 +24,7 @@ func TestCataloger_Globs(t *testing.T) {
 			pkgtest.NewCatalogTester().
 				FromDirectory(t, test.fixture).
 				ExpectsResolverContentQueries(test.expected).
+				IgnoreUnfulfilledPathResponses("etc/apk/repositories").
 				TestCataloger(t, NewApkdbCataloger())
 		})
 	}

syft/pkg/cataloger/apkdb/parse_apk_db.go

@@ -3,7 +3,9 @@ package apkdb
 import (
 	"bufio"
 	"fmt"
+	"io"
 	"path"
+	"regexp"
 	"strconv"
 	"strings"
 
@@ -20,11 +22,15 @@ import (
 // integrity check
 var _ generic.Parser = parseApkDB
 
+var (
+	repoRegex = regexp.MustCompile(`(?m)^https://.*\.alpinelinux\.org/alpine/v([^/]+)/([a-zA-Z0-9_]+)$`)
+)
+
 // parseApkDB parses packages from a given APK installed DB file. For more
 // information on specific fields, see https://wiki.alpinelinux.org/wiki/Apk_spec.
 //
-//nolint:funlen
-func parseApkDB(_ source.FileResolver, env *generic.Environment, reader source.LocationReadCloser) ([]pkg.Package, []artifact.Relationship, error) {
+//nolint:funlen,gocognit
+func parseApkDB(resolver source.FileResolver, env *generic.Environment, reader source.LocationReadCloser) ([]pkg.Package, []artifact.Relationship, error) {
 	scanner := bufio.NewScanner(reader)
 
 	var apks []pkg.ApkMetadata
@@ -101,6 +107,19 @@ func parseApkDB(_ source.FileResolver, env *generic.Environment, reader source.L
 	if env != nil {
 		r = env.LinuxRelease
 	}
+	// this is somewhat ugly, but better than completely failing when we can't find the release,
+	// e.g. embedded deeper in the tree, like containers or chroots.
+	// but we now have no way of handling different repository sources. On the other hand,
+	// we never could before this. At least now, we can handle some.
+	// This should get fixed with https://gitlab.alpinelinux.org/alpine/apk-tools/-/issues/10875
+	if r == nil {
+		// find the repositories file from the relative directory of the DB file
+		releases := findReleases(resolver, reader.Location.RealPath)
+
+		if len(releases) > 0 {
+			r = &releases[0]
+		}
+	}
 
 	pkgs := make([]pkg.Package, 0, len(apks))
 	for _, apk := range apks {
@@ -110,6 +129,58 @@ func parseApkDB(_ source.FileResolver, env *generic.Environment, reader source.L
 	return pkgs, discoverPackageDependencies(pkgs), nil
 }
 
+func findReleases(resolver source.FileResolver, dbPath string) []linux.Release {
+	if resolver == nil {
+		return nil
+	}
+
+	reposLocation := path.Clean(path.Join(path.Dir(dbPath), "../../../etc/apk/repositories"))
+	locations, err := resolver.FilesByPath(reposLocation)
+	if err != nil {
+		log.Tracef("unable to find APK repositories file %q: %+v", reposLocation, err)
+		return nil
+	}
+
+	if len(locations) == 0 {
+		return nil
+	}
+	location := locations[0]
+
+	reposReader, err := resolver.FileContentsByLocation(location)
+	if err != nil {
+		log.Tracef("unable to fetch contents for APK repositories file %q: %+v", reposLocation, err)
+		return nil
+	}
+
+	return parseReleasesFromAPKRepository(source.LocationReadCloser{
+		Location:   location,
+		ReadCloser: reposReader,
+	})
+}
+
+func parseReleasesFromAPKRepository(reader source.LocationReadCloser) []linux.Release {
+	var releases []linux.Release
+
+	reposB, err := io.ReadAll(reader)
+	if err != nil {
+		log.Tracef("unable to read APK repositories file %q: %+v", reader.Location.RealPath, err)
+		return nil
+	}
+
+	parts := repoRegex.FindAllStringSubmatch(string(reposB), -1)
+	for _, part := range parts {
+		if len(part) >= 3 {
+			releases = append(releases, linux.Release{
+				Name:      "Alpine Linux",
+				ID:        "alpine",
+				VersionID: part[1],
+			})
+		}
+	}
+
+	return releases
+}
+
 func parseApkField(line string) *apkField {
 	parts := strings.SplitN(line, ":", 2)
 	if len(parts) != 2 {

syft/pkg/cataloger/apkdb/parse_apk_db_test.go

@@ -1,8 +1,10 @@
 package apkdb
 
 import (
+	"io"
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"
 
 	"github.com/google/go-cmp/cmp"
@@ -1186,3 +1188,49 @@ func Test_stripVersionSpecifier(t *testing.T) {
 		})
 	}
 }
+
+func TestParseReleasesFromAPKRepository(t *testing.T) {
+	tests := []struct {
+		repos string
+		want  []linux.Release
+		desc  string
+	}{
+		{
+			"https://foo.alpinelinux.org/alpine/v3.14/main",
+			[]linux.Release{
+				{Name: "Alpine Linux", ID: "alpine", VersionID: "3.14"},
+			},
+			"single repo",
+		},
+		{
+			`https://foo.alpinelinux.org/alpine/v3.14/main
+https://foo.alpinelinux.org/alpine/v3.14/community`,
+			[]linux.Release{
+				{Name: "Alpine Linux", ID: "alpine", VersionID: "3.14"},
+				{Name: "Alpine Linux", ID: "alpine", VersionID: "3.14"},
+			},
+			"multiple repos",
+		},
+		{
+			``,
+			nil,
+			"empty",
+		},
+		{
+			`https://foo.bar.org/alpine/v3.14/main
+https://foo.them.org/alpine/v3.14/community`,
+			nil,
+			"invalid repos",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.desc, func(t *testing.T) {
+			reposReader := io.NopCloser(strings.NewReader(tt.repos))
+			got := parseReleasesFromAPKRepository(source.LocationReadCloser{
+				Location:   source.NewLocation("test"),
+				ReadCloser: reposReader,
+			})
+			assert.Equal(t, tt.want, got)
+		})
+	}
+}