finalize gemspec parser + update json schema to match

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

schema/json/schema.json

@@ -40,6 +40,9 @@
                             "architecture": {
                                 "type": "string"
                             },
+                            "authors": {
+                                "type": "null"
+                            },
                             "description": {
                                 "type": "string"
                             },
@@ -48,31 +51,38 @@
                             },
                             "files": {
                                 "items": {
-                                    "properties": {
+                                    "anyOf": [
-                                        "checksum": {
+                                        {
                                             "type": "string"
                                         },
-                                        "ownerGid": {
+                                        {
-                                            "type": "string"
+                                            "properties": {
-                                        },
+                                                "checksum": {
-                                        "ownerUid": {
+                                                    "type": "string"
-                                            "type": "string"
+                                                },
-                                        },
+                                                "ownerGid": {
-                                        "path": {
+                                                    "type": "string"
-                                            "type": "string"
+                                                },
-                                        },
+                                                "ownerUid": {
-                                        "permissions": {
+                                                    "type": "string"
-                                            "type": "string"
+                                                },
+                                                "path": {
+                                                    "type": "string"
+                                                },
+                                                "permissions": {
+                                                    "type": "string"
+                                                }
+                                            },
+                                            "required": [
+                                                "checksum",
+                                                "ownerGid",
+                                                "ownerUid",
+                                                "path",
+                                                "permissions"
+                                            ],
+                                            "type": "object"
                                         }
-                                    },
+                                    ]
-                                    "required": [
-                                        "checksum",
-                                        "ownerGid",
-                                        "ownerUid",
-                                        "path",
-                                        "permissions"
-                                    ],
-                                    "type": "object"
                                 },
                                 "type": "array"
                             },
@@ -85,6 +95,12 @@
                             "license": {
                                 "type": "string"
                             },
+                            "licenses": {
+                                "items": {
+                                    "type": "string"
+                                },
+                                "type": "array"
+                            },
                             "maintainer": {
                                 "type": "string"
                             },

syft/cataloger/cataloger.go

@@ -8,13 +8,13 @@ package cataloger
 import (
 	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/syft/syft/cataloger/apkdb"
-	"github.com/anchore/syft/syft/cataloger/bundler"
 	"github.com/anchore/syft/syft/cataloger/deb"
 	"github.com/anchore/syft/syft/cataloger/golang"
 	"github.com/anchore/syft/syft/cataloger/java"
 	"github.com/anchore/syft/syft/cataloger/javascript"
 	"github.com/anchore/syft/syft/cataloger/python"
 	"github.com/anchore/syft/syft/cataloger/rpmdb"
+	"github.com/anchore/syft/syft/cataloger/ruby"
 	"github.com/anchore/syft/syft/pkg"
 	"github.com/anchore/syft/syft/scope"
 )
@@ -36,7 +36,7 @@ type Cataloger interface {
 // ImageCatalogers returns a slice of locally implemented catalogers that are fit for detecting installations of packages.
 func ImageCatalogers() []Cataloger {
 	return []Cataloger{
-		bundler.NewGemSpecCataloger(),
+		ruby.NewGemSpecCataloger(),
 		python.NewPythonCataloger(),         // TODO: split and replace me
 		javascript.NewJavascriptCataloger(), // TODO: split and replace me
 		deb.NewDpkgdbCataloger(),
@@ -50,7 +50,7 @@ func ImageCatalogers() []Cataloger {
 // DirectoryCatalogers returns a slice of locally implemented catalogers that are fit for detecting packages from index files (and select installations)
 func DirectoryCatalogers() []Cataloger {
 	return []Cataloger{
-		bundler.NewGemFileLockCataloger(),
+		ruby.NewGemFileLockCataloger(),
 		python.NewPythonCataloger(),         // TODO: split and replace me
 		javascript.NewJavascriptCataloger(), // TODO: split and replace me
 		deb.NewDpkgdbCataloger(),

syft/cataloger/ruby/catalogers.go

@@ -1,7 +1,7 @@
 /*
 Package bundler provides a concrete Cataloger implementation for Ruby Gemfile.lock bundler files.
 */
-package bundler
+package ruby
 
 import (
 	"github.com/anchore/syft/syft/cataloger/common"

syft/cataloger/ruby/parse_gemfile_lock.go

@@ -1,4 +1,4 @@
-package bundler
+package ruby
 
 import (
 	"bufio"

syft/cataloger/ruby/parse_gemfile_lock_test.go

@@ -1,4 +1,4 @@
-package bundler
+package ruby
 
 import (
 	"os"

syft/cataloger/ruby/parse_gemspec.go

@@ -1,11 +1,11 @@
-package bundler
+package ruby
 
 import (
 	"bufio"
+	"encoding/json"
 	"fmt"
 	"io"
 	"regexp"
-	"strconv"
 	"strings"
 
 	"github.com/mitchellh/mapstructure"
@@ -48,6 +48,7 @@ var postProcessors = map[string]postProcessor{
 }
 
 func processList(s string) []string {
+	// nolint:prealloc
 	var results []string
 	for _, item := range strings.Split(s, ",") {
 		results = append(results, strings.Trim(item, "\" "))
@@ -106,47 +107,19 @@ func parseGemSpecEntries(_ string, reader io.Reader) ([]pkg.Package, error) {
 
 // renderUtf8 takes any string escaped string sub-sections from the ruby string and replaces those sections with the UTF8 runes.
 func renderUtf8(s string) string {
-	pattern := regexp.MustCompile(`\\u(?P<unicode>[0-9A-F]{4,8})`)
+	pattern := regexp.MustCompile(`\\u(?P<unicode>[0-9A-F]{4})`)
-	fullReplacement := replaceAllStringSubmatchFunc(pattern, s, func(unicodeSection []string) string {
+	fullReplacement := pattern.ReplaceAllStringFunc(s, func(unicodeSection string) string {
-		replacement := ""
+		var replacement string
-		if len(unicodeSection) == 1 {
+		// note: the json parser already has support for interpreting hex-representations of unicode escaped strings as unicode runes.
-			return unicodeSection[0]
+		// we can do this ourselves with strconv.Atoi, or leverage the existing json package.
-		}
+		if err := json.Unmarshal([]byte(`"`+unicodeSection+`"`), &replacement); err != nil {
-		for idx, m := range unicodeSection {
+			return unicodeSection
-			if idx == 0 {
-				continue
-			}
-			value, err := strconv.ParseInt(m, 16, 64)
-			if err != nil {
-				// TODO: log?
-				panic(err)
-				//return unicodeSection[0]
-			}
-			replacement = strings.ReplaceAll(unicodeSection[0], "\\u"+m, string(rune(value)))
 		}
 		return replacement
 	})
 	return fullReplacement
 }
 
-// replaceAllStringSubmatchFunc finds and replaces the given capture groups from the
-func replaceAllStringSubmatchFunc(re *regexp.Regexp, str string, repl func([]string) string) string {
-	result := ""
-	lastIndex := 0
-
-	for _, v := range re.FindAllSubmatchIndex([]byte(str), -1) {
-		var groups []string
-		for i := 0; i < len(v); i += 2 {
-			groups = append(groups, str[v[i]:v[i+1]])
-		}
-
-		result += str[lastIndex:v[0]] + repl(groups)
-		lastIndex = v[1]
-	}
-
-	return result + str[lastIndex:]
-}
-
 // matchCaptureGroups takes a regular expression and string and returns all of the named capture group results in a map.
 func matchCaptureGroups(regEx *regexp.Regexp, str string) map[string]string {
 	match := regEx.FindStringSubmatch(str)

syft/cataloger/ruby/parse_gemspec_test.go

@@ -1,4 +1,4 @@
-package bundler
+package ruby
 
 import (
 	"os"
@@ -19,7 +19,7 @@ func TestParseGemspec(t *testing.T) {
 			Name:     "bundler",
 			Version:  "2.1.4",
 			Files:    []string{"exe/bundle", "exe/bundler"},
-			Authors:  []string{"André Arko", "Samuel Giddins", "Colby Swandale", "Hiroshi Shibata", "David Rodréguez", "Grey Baker", "Stephanie Morillo", "Chris Morris", "James Wen", "Tim Moore", "André Medeiros", "Jessica Lynn Suttles", "Terence Lee", "Carl Lerche", "Yehuda Katz"},
+			Authors:  []string{"André Arko", "Samuel Giddins", "Colby Swandale", "Hiroshi Shibata", "David Rodríguez", "Grey Baker", "Stephanie Morillo", "Chris Morris", "James Wen", "Tim Moore", "André Medeiros", "Jessica Lynn Suttles", "Terence Lee", "Carl Lerche", "Yehuda Katz"},
 			Licenses: []string{"MIT"},
 		},
 	}

syft/pkg/gem_metadata.go

@@ -1,9 +1,9 @@
 package pkg
 
 type GemMetadata struct {
-	Name    string   `mapstructure:"name" json:"name"`
+	Name     string   `mapstructure:"name" json:"name"`
-	Version string   `mapstructure:"version" json:"version"`
+	Version  string   `mapstructure:"version" json:"version"`
-	Files   []string `mapstructure:"files" json:"files"`
+	Files    []string `mapstructure:"files" json:"files"`
 	Authors  []string `mapstructure:"authors" json:"authors"`
 	Licenses []string `mapstructure:"licenses" json:"licenses"`
 }