ignore target link files based on path (#667)

* ignore target link files based on path

log when files are actually indexed

add test for sym link resolution

golang test nits

Signed-off-by: Jonas Galvão Xavier <jonas.agx@gmail.com>

* nil catalog should act like an empty catalog

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* remove dir path filtering in favor of file type filtering

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* split out addPathToIndex into specialized functions

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* add test for nul catalog enumeration

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* conditionally discover MIME types for file based on file resolver index

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* change logging around cataloging

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

* add tests to cover possible infinite symlink loop for resolver

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

Co-authored-by: Alex Goodman <alex.goodman@anchore.com>

syft/pkg/catalog.go

@@ -103,6 +103,10 @@ func (c *Catalog) Enumerate(types ...Type) <-chan Package {
 	channel := make(chan Package)
 	go func() {
 		defer close(channel)
+		if c == nil {
+			// we should allow enumerating from a catalog that was never created (which will result in no packages enumerated)
+			return
+		}
 		for ty, ids := range c.idsByType {
 			if len(types) != 0 {
 				found := false

syft/pkg/catalog_test.go

@@ -3,6 +3,8 @@ package pkg
 import (
 	"testing"
 
+	"github.com/stretchr/testify/assert"
+
 	"github.com/scylladb/go-set/strset"
 
 	"github.com/anchore/syft/syft/source"
@@ -144,3 +146,8 @@ func TestCatalog_PathIndexDeduplicatesRealVsVirtualPaths(t *testing.T) {
 	}
 
 }
+
+func TestCatalog_EnumerateNilCatalog(t *testing.T) {
+	var c *Catalog
+	assert.Empty(t, c.Enumerate())
+}

syft/pkg/cataloger/catalog.go

@@ -51,6 +51,7 @@ func Catalog(resolver source.FileResolver, theDistro *distro.Distro, catalogers
 	var errs error
 	for _, theCataloger := range catalogers {
 		// find packages from the underlying raw data
+		log.Debugf("cataloging with %q", theCataloger.Name())
 		packages, relationships, err := theCataloger.Catalog(resolver)
 		if err != nil {
 			errs = multierror.Append(errs, err)
@@ -59,7 +60,7 @@ func Catalog(resolver source.FileResolver, theDistro *distro.Distro, catalogers
 
 		catalogedPackages := len(packages)
 
-		log.Debugf("package cataloger %q discovered %d packages", theCataloger.Name(), catalogedPackages)
+		log.Debugf("discovered %d packages", catalogedPackages)
 		packagesDiscovered.N += int64(catalogedPackages)
 
 		for _, p := range packages {

syft/pkg/cataloger/javascript/parse_package_json.go

@@ -208,7 +208,14 @@ func (p PackageJSON) hasNameAndVersionValues() bool {
 	return p.Name != "" && p.Version != ""
 }
 
-func pathContainsNodeModulesDirectory(path string) bool {
-	isNodeModulesPath, _ := regexp.MatchString("[\\/]node_modules[\\/]", path)
-	return isNodeModulesPath
+// this supports both windows and unix paths
+var filepathSeparator = regexp.MustCompile(`[\\/]`)
+
+func pathContainsNodeModulesDirectory(p string) bool {
+	for _, subPath := range filepathSeparator.Split(p, -1) {
+		if subPath == "node_modules" {
+			return true
+		}
+	}
+	return false
 }

syft/pkg/cataloger/javascript/parse_package_json_test.go

@@ -6,6 +6,7 @@ import (
 
 	"github.com/anchore/syft/syft/pkg"
 	"github.com/go-test/deep"
+	"github.com/stretchr/testify/assert"
 )
 
 func TestParsePackageJSON(t *testing.T) {
@@ -160,3 +161,50 @@ func TestParsePackageJSON_Partial(t *testing.T) { // see https://github.com/anch
 		t.Errorf("no packages should've been returned (but got %d packages)", actualCount)
 	}
 }
+
+func Test_pathContainsNodeModulesDirectory(t *testing.T) {
+	tests := []struct {
+		path     string
+		expected bool
+	}{
+		// positive
+		{
+			path:     "something/node_modules/package",
+			expected: true,
+		},
+		{
+			path:     "node_modules/package",
+			expected: true,
+		},
+		{
+			path:     "something/node_modules",
+			expected: true,
+		},
+		{
+			path:     "\\something\\node_modules\\",
+			expected: true,
+		},
+		{
+			path:     "\\something\\node_modules",
+			expected: true,
+		},
+		// negative
+		{
+			path:     "something/node_bogus_modules",
+			expected: false,
+		},
+		{
+			path:     "something/node_modules_bogus",
+			expected: false,
+		},
+		{
+			path:     "something/node_bogus_modules/package",
+			expected: false,
+		},
+	}
+	for _, test := range tests {
+		t.Run(test.path, func(t *testing.T) {
+			assert.Equal(t, test.expected, pathContainsNodeModulesDirectory(test.path))
+		})
+	}
+}