Merge pull request #43674 from shiywang/fixjsonpath

Automatic merge from submit-queue (batch tested with PRs 48538, 43674)

fix JSONPath parser will not filter strings containing parentheses

Fixes https://github.com/kubernetes/client-go/issues/158
@mtaufen  @daizuozhuo @caesarxuchao

staging/src/k8s.io/client-go/util/jsonpath/parser.go

@@ -17,6 +17,7 @@ limitations under the License.
 package jsonpath
 
 import (
+	"errors"
 	"fmt"
 	"regexp"
 	"strconv"
@@ -42,6 +43,8 @@ type Parser struct {
 	width int
 }
 
+var ErrSyntax = errors.New("invalid syntax")
+
 // Parse parsed the given text and return a node Parser.
 // If an error is encountered, parsing stops and an empty
 // Parser is returned with the error
@@ -159,8 +162,8 @@ func (p *Parser) parseInsideAction(cur *ListNode) error {
 		p.consumeText()
 	case r == '[':
 		return p.parseArray(cur)
-	case r == '"':
+	case r == '"' || r == '\'':
-		return p.parseQuote(cur)
+		return p.parseQuote(cur, r)
 	case r == '.':
 		return p.parseField(cur)
 	case r == '+' || r == '-' || unicode.IsDigit(r):
@@ -334,15 +337,35 @@ Loop:
 func (p *Parser) parseFilter(cur *ListNode) error {
 	p.pos += len("[?(")
 	p.consumeText()
+	begin := false
+	end := false
+	var pair rune
+
 Loop:
 	for {
-		switch p.next() {
+		r := p.next()
+		switch r {
 		case eof, '\n':
 			return fmt.Errorf("unterminated filter")
+		case '"', '\'':
+			if begin == false {
+				//save the paired rune
+				begin = true
+				pair = r
+				continue
+			}
+			//only add when met paired rune
+			if p.input[p.pos-2] != '\\' && r == pair {
+				end = true
+			}
 		case ')':
+			//in rightParser below quotes only appear zero or once
+			//and must be paired at the beginning and end
+			if begin == end {
 				break Loop
 			}
 		}
+	}
 	if p.next() != ']' {
 		return fmt.Errorf("unclosed array expect ]")
 	}
@@ -370,19 +393,22 @@ Loop:
 	return p.parseInsideAction(cur)
 }
 
-// parseQuote unquotes string inside double quote
+// parseQuote unquotes string inside double or single quote
-func (p *Parser) parseQuote(cur *ListNode) error {
+func (p *Parser) parseQuote(cur *ListNode, end rune) error {
 Loop:
 	for {
 		switch p.next() {
 		case eof, '\n':
 			return fmt.Errorf("unterminated quoted string")
-		case '"':
+		case end:
+			//if it's not escape break the Loop
+			if p.input[p.pos-2] != '\\' {
 				break Loop
 			}
 		}
+	}
 	value := p.consumeText()
-	s, err := strconv.Unquote(value)
+	s, err := UnquoteExtend(value)
 	if err != nil {
 		return fmt.Errorf("unquote string %s error %v", value, err)
 	}
@@ -447,3 +473,51 @@ func isAlphaNumeric(r rune) bool {
 func isBool(s string) bool {
 	return s == "true" || s == "false"
 }
+
+//UnquoteExtend is almost same as strconv.Unquote(), but it support parse single quotes as a string
+func UnquoteExtend(s string) (string, error) {
+	n := len(s)
+	if n < 2 {
+		return "", ErrSyntax
+	}
+	quote := s[0]
+	if quote != s[n-1] {
+		return "", ErrSyntax
+	}
+	s = s[1 : n-1]
+
+	if quote != '"' && quote != '\'' {
+		return "", ErrSyntax
+	}
+
+	// Is it trivial?  Avoid allocation.
+	if !contains(s, '\\') && !contains(s, quote) {
+		return s, nil
+	}
+
+	var runeTmp [utf8.UTFMax]byte
+	buf := make([]byte, 0, 3*len(s)/2) // Try to avoid more allocations.
+	for len(s) > 0 {
+		c, multibyte, ss, err := strconv.UnquoteChar(s, quote)
+		if err != nil {
+			return "", err
+		}
+		s = ss
+		if c < utf8.RuneSelf || !multibyte {
+			buf = append(buf, byte(c))
+		} else {
+			n := utf8.EncodeRune(runeTmp[:], c)
+			buf = append(buf, runeTmp[:n]...)
+		}
+	}
+	return string(buf), nil
+}
+
+func contains(s string, c byte) bool {
+	for i := 0; i < len(s); i++ {
+		if s[i] == c {
+			return true
+		}
+	}
+	return false
+}

staging/src/k8s.io/client-go/util/jsonpath/parser_test.go

@@ -61,6 +61,22 @@ var parserTests = []parserTest{
 		newList(), newIdentifier("end"),
 	}, false},
 	{"malformat input", `{\\\}`, []Node{}, true},
+	{"paired parentheses in quotes", `{[?(@.status.nodeInfo.osImage == "()")]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("()")}, false},
+	{"paired parentheses in double quotes and with double quotes escape", `{[?(@.status.nodeInfo.osImage == "(\"\")")]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("(\"\")")}, false},
+	{"unregular parentheses in double quotes", `{[?(@.test == "())(")]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("test"), newList(), newText("())(")}, false},
+	{"plain text in single quotes", `{[?(@.status.nodeInfo.osImage == 'Linux')]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("Linux")}, false},
+	{"test filter suffix", `{[?(@.status.nodeInfo.osImage == "{[()]}")]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("{[()]}")}, false},
+	{"double inside single", `{[?(@.status.nodeInfo.osImage == "''")]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("''")}, false},
+	{"single inside double", `{[?(@.status.nodeInfo.osImage == '""')]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("\"\"")}, false},
+	{"single containing escaped single", `{[?(@.status.nodeInfo.osImage == '\\\'')]}`,
+		[]Node{newList(), newFilter(newList(), newList(), "=="), newList(), newField("status"), newField("nodeInfo"), newField("osImage"), newList(), newText("\\'")}, false},
 }
 
 func collectNode(nodes []Node, cur Node) []Node {