decode golang symbols (#5089)

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

syft/pkg/cataloger/golang/symbols.go

@@ -52,10 +52,7 @@ func getSymbols(r io.ReaderAt) (syms []binarySymbol, err error) {
 			continue
 		}
 		seen[fn.Name] = struct{}{}
-		syms = append(syms, binarySymbol{
+		syms = append(syms, makeBinarySymbol(fn.Name, fn.PackageName()))
-			packagePath: fn.PackageName(),
-			name:        fn.Name,
-		})
 	}
 
 	// debug/gosym only exposes top-level functions; functions that the compiler inlined into their
@@ -71,7 +68,7 @@ func getSymbols(r io.ReaderAt) (syms []binarySymbol, err error) {
 			continue
 		}
 		seen[name] = struct{}{}
-		syms = append(syms, binarySymbol{packagePath: pkgPath, name: name})
+		syms = append(syms, makeBinarySymbol(name, pkgPath))
 	}
 
 	return syms, nil
@@ -94,6 +91,55 @@ func packagePathFromSymbolName(name string) string {
 	return name[:slash+1+dot]
 }
 
+// makeBinarySymbol builds a binarySymbol, unescaping the %xx sequences the go linker introduces in the
+// import-path portion of a symbol name (see unescapePackagePath). The local-symbol suffix (method and
+// function names) is left untouched, so only the path prefix shared by packagePath and name is rewritten.
+func makeBinarySymbol(name, pkgPath string) binarySymbol {
+	unescaped := unescapePackagePath(pkgPath)
+	if unescaped != pkgPath && strings.HasPrefix(name, pkgPath) {
+		name = unescaped + name[len(pkgPath):]
+	}
+	return binarySymbol{packagePath: unescaped, name: name}
+}
+
+// unescapePackagePath reverses the escaping cmd/internal/objabi.PathToPrefix applies to import paths in
+// symbol names: bytes like '.' (at or after the final '/'), '%', '"', control bytes, and high bytes are
+// written as lowercase "%xx". For example "gopkg.in/yaml.v2" is stored as "gopkg.in/yaml%2ev2", so this
+// restores it before matching against the (unescaped) module paths from build info. A lone or malformed
+// '%' sequence is left as-is.
+func unescapePackagePath(path string) string {
+	if !strings.Contains(path, "%") {
+		return path
+	}
+	var b strings.Builder
+	b.Grow(len(path))
+	for i := 0; i < len(path); i++ {
+		if path[i] == '%' && i+2 < len(path) {
+			if hi, ok1 := unhex(path[i+1]); ok1 {
+				if lo, ok2 := unhex(path[i+2]); ok2 {
+					b.WriteByte(hi<<4 | lo)
+					i += 2
+					continue
+				}
+			}
+		}
+		b.WriteByte(path[i])
+	}
+	return b.String()
+}
+
+func unhex(c byte) (byte, bool) {
+	switch {
+	case c >= '0' && c <= '9':
+		return c - '0', true
+	case c >= 'a' && c <= 'f':
+		return c - 'a' + 10, true
+	case c >= 'A' && c <= 'F':
+		return c - 'A' + 10, true
+	}
+	return 0, false
+}
+
 // nameWithoutTypeArgs strips the type-argument portion from an instantiated generic symbol name, e.g.
 // "foo/bar.Do[net/url.Values]" -> "foo/bar.Do". The slashes and dots inside the brackets would otherwise
 // corrupt package-path derivation (yielding "foo/bar.Do[net" for the example above). Mirrors

syft/pkg/cataloger/golang/symbols_test.go

@@ -97,6 +97,58 @@ func Test_moduleSymbols(t *testing.T) {
 	}
 }
 
+func Test_escapedImportPathAttribution(t *testing.T) {
+	// the go linker escapes the '.' in the last path element, so "gopkg.in/yaml.v2" appears in the symbol
+	// table as "gopkg.in/yaml%2ev2". attribution must still land these under the unescaped module path.
+	mainModule := &debug.Module{Path: "github.com/someorg/somecli"}
+	deps := []*debug.Module{{Path: "gopkg.in/yaml.v2"}}
+
+	symbols := []binarySymbol{
+		makeBinarySymbol("gopkg.in/yaml%2ev2.(*decoder).alias", "gopkg.in/yaml%2ev2"),
+		makeBinarySymbol("gopkg.in/yaml%2ev2.(*TypeError).Error", "gopkg.in/yaml%2ev2"),
+	}
+
+	gotByModule, gotStdlib := moduleSymbols(symbols, mainModule, deps)
+	assert.Nil(t, gotStdlib)
+	assert.Equal(t, map[string]map[string][]string{
+		"gopkg.in/yaml.v2": {
+			"gopkg.in/yaml.v2": {"(*TypeError).Error", "(*decoder).alias"},
+		},
+	}, gotByModule)
+}
+
+func Test_unescapePackagePath(t *testing.T) {
+	tests := []struct {
+		input    string
+		expected string
+	}{
+		{"gopkg.in/yaml%2ev2", "gopkg.in/yaml.v2"},
+		{"foo%25bar", "foo%bar"},
+		{"foo%22bar", "foo\"bar"},
+		// no escapes: unchanged
+		{"github.com/foo/bar", "github.com/foo/bar"},
+		{"", ""},
+		// malformed sequences are left as-is
+		{"foo%2", "foo%2"},
+		{"foo%zz", "foo%zz"},
+		{"foo%", "foo%"},
+		// uppercase hex is accepted too
+		{"foo%2Ebar", "foo.bar"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.input, func(t *testing.T) {
+			assert.Equal(t, tt.expected, unescapePackagePath(tt.input))
+		})
+	}
+}
+
+func Test_makeBinarySymbol(t *testing.T) {
+	// only the import-path prefix is unescaped; the local-symbol suffix is preserved verbatim
+	sym := makeBinarySymbol("gopkg.in/yaml%2ev2.(*decoder).alias", "gopkg.in/yaml%2ev2")
+	assert.Equal(t, "gopkg.in/yaml.v2", sym.packagePath)
+	assert.Equal(t, "gopkg.in/yaml.v2.(*decoder).alias", sym.name)
+}
+
 func Test_localSymbolName(t *testing.T) {
 	tests := []struct {
 		name       string