Guess go main module version based on binary contents (#2608)

* guess go main module version based on binary contents Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> * add configuration options for golang main module version heuristics Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> * fix test setup for go bin cataloger Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> * fix unit test Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> * fix incorrect test assert ordering Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> * handle error from seek Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> --------- Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
anchore · Feb 9, 2024 · 84576b9 · 84576b9
1 parent 737c4e4
commit 84576b9
Show file tree

Hide file tree

Showing 9 changed files with 300 additions and 85 deletions.
diff --git a/README.md b/README.md
@@ -691,6 +691,24 @@ golang:
    # SYFT_GOLANG_NOPROXY env var
    no-proxy: ""
 
+   # the go main module version discovered from binaries built with the go compiler will
+   # always show (devel) as the version. Use these options to control heuristics to guess
+   # a more accurate version from the binary.
+   main-module-version:
+
+      # look for LD flags that appear to be setting a version (e.g. -X main.version=1.0.0)
+      # SYFT_GOLANG_MAIN_MODULE_VERSION_FROM_LD_FLAGS env var
+      from-ld-flags: true
+
+      # use the build settings (e.g. vcs.version & vcs.time) to craft a v0 pseudo version 
+      # (e.g. v0.0.0-20220308212642-53e6d0aaf6fb) when a more accurate version cannot be found otherwise.
+      # SYFT_GOLANG_MAIN_MODULE_VERSION_FROM_BUILD_SETTINGS env var
+      from-build-settings: true
+
+      # search for semver-like strings in the binary contents.
+      # SYFT_GOLANG_MAIN_MODULE_VERSION_FROM_CONTENTS env var
+      from-contents: true
+
 java:
    maven-url: "https://repo1.maven.org/maven2"
    max-parent-recursive-depth: 5

diff --git a/cmd/syft/internal/options/catalog.go b/cmd/syft/internal/options/catalog.go
@@ -62,6 +62,7 @@ func DefaultCatalog() Catalog {
 		Scope:         source.SquashedScope.String(),
 		Package:       defaultPackageConfig(),
 		LinuxKernel:   defaultLinuxKernelConfig(),
+		Golang:        defaultGolangConfig(),
 		File:          defaultFileConfig(),
 		Relationships: defaultRelationshipsConfig(),
 		Source:        defaultSourceConfig(),
@@ -131,7 +132,13 @@ func (cfg Catalog) ToPackagesConfig() pkgcataloging.Config {
 			WithLocalModCacheDir(cfg.Golang.LocalModCacheDir).
 			WithSearchRemoteLicenses(cfg.Golang.SearchRemoteLicenses).
 			WithProxy(cfg.Golang.Proxy).
-			WithNoProxy(cfg.Golang.NoProxy),
+			WithNoProxy(cfg.Golang.NoProxy).
+			WithMainModuleVersion(
+				golang.DefaultMainModuleVersionConfig().
+					WithFromContents(cfg.Golang.MainModuleVersion.FromContents).
+					WithFromBuildSettings(cfg.Golang.MainModuleVersion.FromBuildSettings).
+					WithFromLDFlags(cfg.Golang.MainModuleVersion.FromLDFlags),
+			),
 		JavaScript: javascript.DefaultCatalogerConfig().
 			WithSearchRemoteLicenses(cfg.JavaScript.SearchRemoteLicenses).
 			WithNpmBaseURL(cfg.JavaScript.NpmBaseURL),

diff --git a/cmd/syft/internal/options/golang.go b/cmd/syft/internal/options/golang.go
@@ -1,9 +1,38 @@
 package options
 
+import (
+	"strings"
+
+	"github.com/anchore/syft/syft/pkg/cataloger/golang"
+)
+
 type golangConfig struct {
-	SearchLocalModCacheLicenses bool   `json:"search-local-mod-cache-licenses" yaml:"search-local-mod-cache-licenses" mapstructure:"search-local-mod-cache-licenses"`
-	LocalModCacheDir            string `json:"local-mod-cache-dir" yaml:"local-mod-cache-dir" mapstructure:"local-mod-cache-dir"`
-	SearchRemoteLicenses        bool   `json:"search-remote-licenses" yaml:"search-remote-licenses" mapstructure:"search-remote-licenses"`
-	Proxy                       string `json:"proxy" yaml:"proxy" mapstructure:"proxy"`
-	NoProxy                     string `json:"no-proxy" yaml:"no-proxy" mapstructure:"no-proxy"`
+	SearchLocalModCacheLicenses bool                          `json:"search-local-mod-cache-licenses" yaml:"search-local-mod-cache-licenses" mapstructure:"search-local-mod-cache-licenses"`
+	LocalModCacheDir            string                        `json:"local-mod-cache-dir" yaml:"local-mod-cache-dir" mapstructure:"local-mod-cache-dir"`
+	SearchRemoteLicenses        bool                          `json:"search-remote-licenses" yaml:"search-remote-licenses" mapstructure:"search-remote-licenses"`
+	Proxy                       string                        `json:"proxy" yaml:"proxy" mapstructure:"proxy"`
+	NoProxy                     string                        `json:"no-proxy" yaml:"no-proxy" mapstructure:"no-proxy"`
+	MainModuleVersion           golangMainModuleVersionConfig `json:"main-module-version" yaml:"main-module-version" mapstructure:"main-module-version"`
+}
+
+type golangMainModuleVersionConfig struct {
+	FromLDFlags       bool `json:"from-ld-flags" yaml:"from-ld-flags" mapstructure:"from-ld-flags"`
+	FromContents      bool `json:"from-contents" yaml:"from-contents" mapstructure:"from-contents"`
+	FromBuildSettings bool `json:"from-build-settings" yaml:"from-build-settings" mapstructure:"from-build-settings"`
+}
+
+func defaultGolangConfig() golangConfig {
+	def := golang.DefaultCatalogerConfig()
+	return golangConfig{
+		SearchLocalModCacheLicenses: def.SearchLocalModCacheLicenses,
+		LocalModCacheDir:            def.LocalModCacheDir,
+		SearchRemoteLicenses:        def.SearchRemoteLicenses,
+		Proxy:                       strings.Join(def.Proxies, ","),
+		NoProxy:                     strings.Join(def.NoProxy, ","),
+		MainModuleVersion: golangMainModuleVersionConfig{
+			FromLDFlags:       def.MainModuleVersion.FromLDFlags,
+			FromContents:      def.MainModuleVersion.FromContents,
+			FromBuildSettings: def.MainModuleVersion.FromBuildSettings,
+		},
+	}
 }
diff --git a/cmd/syft/internal/test/integration/package_cataloger_convention_test.go b/cmd/syft/internal/test/integration/package_cataloger_convention_test.go
@@ -31,7 +31,9 @@ func Test_packageCatalogerExports(t *testing.T) {
 	for pkg, expected := range expectAtLeast {
 		actual, ok := exports[pkg]
 		require.True(t, ok, pkg)
-		require.True(t, expected.IsSubset(actual.Names()), pkg)
+		if !assert.True(t, actual.Names().IsSubset(expected), pkg) {
+			t.Logf("missing: %s", strset.SymmetricDifference(expected, actual.Names()))
+		}
 	}
 
 }

diff --git a/syft/pkg/cataloger/golang/cataloger.go b/syft/pkg/cataloger/golang/cataloger.go
@@ -38,12 +38,12 @@ func NewGoModuleFileCataloger(opts CatalogerConfig) pkg.Cataloger {
 
 // NewGoModuleBinaryCataloger returns a new cataloger object that searches within binaries built by the go compiler.
 func NewGoModuleBinaryCataloger(opts CatalogerConfig) pkg.Cataloger {
-	c := goBinaryCataloger{
-		licenses: newGoLicenses(binaryCatalogerName, opts),
-	}
 	return &progressingCataloger{
 		cataloger: generic.NewCataloger(binaryCatalogerName).
-			WithParserByMimeTypes(c.parseGoBinary, mimetype.ExecutableMIMETypeSet.List()...),
+			WithParserByMimeTypes(
+				newGoBinaryCataloger(opts).parseGoBinary,
+				mimetype.ExecutableMIMETypeSet.List()...,
+			),
 	}
 }
 

diff --git a/syft/pkg/cataloger/golang/config.go b/syft/pkg/cataloger/golang/config.go
@@ -20,19 +20,28 @@ var (
 )
 
 type CatalogerConfig struct {
-	SearchLocalModCacheLicenses bool     `yaml:"search-local-mod-cache-licenses" json:"search-local-mod-cache-licenses" mapstructure:"search-local-mod-cache-licenses"`
-	LocalModCacheDir            string   `yaml:"local-mod-cache-dir" json:"local-mod-cache-dir" mapstructure:"local-mod-cache-dir"`
-	SearchRemoteLicenses        bool     `yaml:"search-remote-licenses" json:"search-remote-licenses" mapstructure:"search-remote-licenses"`
-	Proxies                     []string `yaml:"proxies,omitempty" json:"proxies,omitempty" mapstructure:"proxies"`
-	NoProxy                     []string `yaml:"no-proxy,omitempty" json:"no-proxy,omitempty" mapstructure:"no-proxy"`
+	SearchLocalModCacheLicenses bool                    `yaml:"search-local-mod-cache-licenses" json:"search-local-mod-cache-licenses" mapstructure:"search-local-mod-cache-licenses"`
+	LocalModCacheDir            string                  `yaml:"local-mod-cache-dir" json:"local-mod-cache-dir" mapstructure:"local-mod-cache-dir"`
+	SearchRemoteLicenses        bool                    `yaml:"search-remote-licenses" json:"search-remote-licenses" mapstructure:"search-remote-licenses"`
+	Proxies                     []string                `yaml:"proxies,omitempty" json:"proxies,omitempty" mapstructure:"proxies"`
+	NoProxy                     []string                `yaml:"no-proxy,omitempty" json:"no-proxy,omitempty" mapstructure:"no-proxy"`
+	MainModuleVersion           MainModuleVersionConfig `yaml:"main-module-version" json:"main-module-version" mapstructure:"main-module-version"`
+}
+
+type MainModuleVersionConfig struct {
+	FromLDFlags       bool `yaml:"from-ld-flags" json:"from-ld-flags" mapstructure:"from-ld-flags"`
+	FromContents      bool `yaml:"from-contents" json:"from-contents" mapstructure:"from-contents"`
+	FromBuildSettings bool `yaml:"from-build-settings" json:"from-build-settings" mapstructure:"from-build-settings"`
 }
 
 // DefaultCatalogerConfig create a CatalogerConfig with default options, which includes:
 // - setting the default remote proxy if none is provided
 // - setting the default no proxy if none is provided
 // - setting the default local module cache dir if none is provided
 func DefaultCatalogerConfig() CatalogerConfig {
-	g := CatalogerConfig{}
+	g := CatalogerConfig{
+		MainModuleVersion: DefaultMainModuleVersionConfig(),
+	}
 
 	// first process the proxy settings
 	if len(g.Proxies) == 0 {
@@ -76,6 +85,14 @@ func DefaultCatalogerConfig() CatalogerConfig {
 	return g
 }
 
+func DefaultMainModuleVersionConfig() MainModuleVersionConfig {
+	return MainModuleVersionConfig{
+		FromLDFlags:       true,
+		FromContents:      true,
+		FromBuildSettings: true,
+	}
+}
+
 func (g CatalogerConfig) WithSearchLocalModCacheLicenses(input bool) CatalogerConfig {
 	g.SearchLocalModCacheLicenses = input
 	return g
@@ -112,3 +129,23 @@ func (g CatalogerConfig) WithNoProxy(input string) CatalogerConfig {
 	g.NoProxy = strings.Split(input, ",")
 	return g
 }
+
+func (g CatalogerConfig) WithMainModuleVersion(input MainModuleVersionConfig) CatalogerConfig {
+	g.MainModuleVersion = input
+	return g
+}
+
+func (g MainModuleVersionConfig) WithFromLDFlags(input bool) MainModuleVersionConfig {
+	g.FromLDFlags = input
+	return g
+}
+
+func (g MainModuleVersionConfig) WithFromContents(input bool) MainModuleVersionConfig {
+	g.FromContents = input
+	return g
+}
+
+func (g MainModuleVersionConfig) WithFromBuildSettings(input bool) MainModuleVersionConfig {
+	g.FromBuildSettings = input
+	return g
+}
diff --git a/syft/pkg/cataloger/golang/config_test.go b/syft/pkg/cataloger/golang/config_test.go
@@ -7,7 +7,7 @@ import (
 	"github.com/stretchr/testify/assert"
 )
 
-func Test_Options(t *testing.T) {
+func Test_Config(t *testing.T) {
 	type opts struct {
 		local    bool
 		cacheDir string
@@ -51,6 +51,7 @@ func Test_Options(t *testing.T) {
 				SearchRemoteLicenses:        false,
 				Proxies:                     []string{"https://my.proxy"},
 				NoProxy:                     []string{"my.private", "no.proxy"},
+				MainModuleVersion:           DefaultMainModuleVersionConfig(),
 			},
 		},
 		{
@@ -74,6 +75,7 @@ func Test_Options(t *testing.T) {
 				SearchRemoteLicenses:        true,
 				Proxies:                     []string{"https://alt.proxy", "direct"},
 				NoProxy:                     []string{"alt.no.proxy"},
+				MainModuleVersion:           DefaultMainModuleVersionConfig(),
 			},
 		},
 	}