sftp: Add SplitShellArgs

src/restic/backend/sftp/split.go

@@ -0,0 +1,73 @@
+package sftp
+
+import (
+	"errors"
+	"unicode"
+)
+
+const data = `"foo" "bar" baz "test argument" another 'test arg' "last \" argument" 'another \" last argument'`
+
+// shellSplitter splits a command string into separater arguments. It supports
+// single and double quoted strings.
+type shellSplitter struct {
+	quote    rune
+	lastChar rune
+}
+
+func (s *shellSplitter) isSplitChar(c rune) bool {
+	// only test for quotes if the last char was not a backslash
+	if s.lastChar != '\\' {
+
+		// quote ended
+		if s.quote != 0 && c == s.quote {
+			s.quote = 0
+			return true
+		}
+
+		// quote starts
+		if s.quote == 0 && (c == '"' || c == '\'') {
+			s.quote = c
+			return true
+		}
+	}
+
+	s.lastChar = c
+
+	// within quote
+	if s.quote != 0 {
+		return false
+	}
+
+	// outside quote
+	return c == '\\' || unicode.IsSpace(c)
+}
+
+// SplitShellArgs returns the list of arguments from a shell command string.
+func SplitShellArgs(data string) (list []string, err error) {
+	s := &shellSplitter{}
+
+	// derived from strings.SplitFunc
+	fieldStart := -1 // Set to -1 when looking for start of field.
+	for i, rune := range data {
+		if s.isSplitChar(rune) {
+			if fieldStart >= 0 {
+				list = append(list, data[fieldStart:i])
+				fieldStart = -1
+			}
+		} else if fieldStart == -1 {
+			fieldStart = i
+		}
+	}
+	if fieldStart >= 0 { // Last field might end at EOF.
+		list = append(list, data[fieldStart:])
+	}
+
+	switch s.quote {
+	case '\'':
+		return nil, errors.New("single-quoted string not terminated")
+	case '"':
+		return nil, errors.New("double-quoted string not terminated")
+	}
+
+	return list, nil
+}

src/restic/backend/sftp/split_test.go

@@ -0,0 +1,105 @@
+package sftp
+
+import (
+	"reflect"
+	"testing"
+)
+
+func TestShellSplitter(t *testing.T) {
+	var tests = []struct {
+		data string
+		want []string
+	}{
+		{
+			`foo`,
+			[]string{"foo"},
+		},
+		{
+			`'foo'`,
+			[]string{"foo"},
+		},
+		{
+			`foo bar baz`,
+			[]string{"foo", "bar", "baz"},
+		},
+		{
+			`foo 'bar' baz`,
+			[]string{"foo", "bar", "baz"},
+		},
+		{
+			`foo 'bar box' baz`,
+			[]string{"foo", "bar box", "baz"},
+		},
+		{
+			`"bar 'box'" baz`,
+			[]string{"bar 'box'", "baz"},
+		},
+		{
+			`'bar "box"' baz`,
+			[]string{`bar "box"`, "baz"},
+		},
+		{
+			`\"bar box baz`,
+			[]string{`"bar`, "box", "baz"},
+		},
+		{
+			`"bar/foo/x" "box baz"`,
+			[]string{"bar/foo/x", "box baz"},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run("", func(t *testing.T) {
+			res, err := SplitShellArgs(test.data)
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			if !reflect.DeepEqual(res, test.want) {
+				t.Fatalf("wrong data returned, want:\n  %#v\ngot:\n  %#v",
+					test.want, res)
+			}
+		})
+	}
+}
+
+func TestShellSplitterInvalid(t *testing.T) {
+	var tests = []struct {
+		data string
+		err  string
+	}{
+		{
+			"foo'",
+			"single-quoted string not terminated",
+		},
+		{
+			`foo"`,
+			"double-quoted string not terminated",
+		},
+		{
+			"foo 'bar",
+			"single-quoted string not terminated",
+		},
+		{
+			`foo "bar`,
+			"double-quoted string not terminated",
+		},
+	}
+
+	for _, test := range tests {
+		t.Run("", func(t *testing.T) {
+			res, err := SplitShellArgs(test.data)
+			if err == nil {
+				t.Fatalf("expected error not found: %v", test.err)
+			}
+
+			if err.Error() != test.err {
+				t.Fatalf("expected error not found, want:\n  %q\ngot:\n  %q", test.err, err.Error())
+			}
+
+			if len(res) > 0 {
+				t.Fatalf("splitter returned fields from invalid data: %v", res)
+			}
+		})
+	}
+}