chore: give the tokenizer the ability to reject unclosed items

Author: ThomasRooney

pkg/jsonpath/parser_test.go

@@ -158,6 +158,18 @@ func TestParserPropertyNameExtension(t *testing.T) {
 			enabled: false,
 			valid:   false,
 		},
+		{
+			name:    "Missing closing a filter expression shouldn't crash",
+			input:   "$.paths.*.*[?([email protected])",
+			enabled: false,
+			valid:   false,
+		},
+		{
+			name:    "Missing closing a array crash",
+			input:   "$.paths.*[?@[\"x-my-ignore\"]",
+			enabled: false,
+			valid:   false,
+		},
 	}
 
 	for _, test := range tests {

pkg/jsonpath/token/token.go

@@ -533,6 +533,9 @@ func (t *Tokenizer) Tokenize() Tokens {
 		t.column++
 	}
 
+	if len(t.stack) > 0 {
+		t.addToken(ILLEGAL, 1, fmt.Sprintf("unmatched %s", t.stack[len(t.stack)-1].String()))
+	}
 	return t.tokens
 }
 

pkg/jsonpath/token/token_test.go

@@ -413,7 +413,7 @@ func TestTokenizer_categorize(t *testing.T) {
 		{name: "dot child with malformed array subscript", path: "$.child[1:2:3:4]"},
 		{name: "dot child with array subscript with zero step", path: "$.child[1:2:0]"},
 		{name: "dot child with non-integer array subscript", path: "$.child[1:2:a]"},
-		{name: "dot child with unclosed array subscript", path: "$.child[*"},
+		{name: "dot child with unclosed array subscript", path: "$.child[*", illegal: true},
 		{name: "dot child with missing array subscript", path: "$.child[]", simple: true},
 		{name: "dot child with embedded space", path: "$.child more", simple: true},
 		{name: "bracket child", path: "$['child']", simple: true},
@@ -429,7 +429,7 @@ func TestTokenizer_categorize(t *testing.T) {
 		{name: "bracket child with array subscript", path: "$['child'][*]"},
 		{name: "bracket child with malformed array subscript", path: "$['child'][1:2:3:4]"},
 		{name: "bracket child with non-integer array subscript", path: "$['child'][1:2:a]"},
-		{name: "bracket child with unclosed array subscript", path: "$['child'][*"},
+		{name: "bracket child with unclosed array subscript", path: "$['child'][*", illegal: true},
 		{name: "bracket child with missing array subscript", path: "$['child'][]", simple: true},
 		{name: "bracket child followed by space", path: "$['child'] ", illegal: true},
 		{name: "bracket dotted child", path: "$['child1.child2']", simple: true},
@@ -497,7 +497,7 @@ func TestTokenizer_categorize(t *testing.T) {
 		{name: "simple filter with bracket with extra whitespace", path: "$[?( ( @.child ) )]"},
 		{name: "simple filter with more complex subpath", path: "$[?((@.child[0]))]"},
 		{name: "missing filter ", path: "$[?()]"},
-		{name: "unclosed filter", path: "$[?("},
+		{name: "unclosed filter", path: "$[?(", illegal: true},
 		{name: "filter with missing operator", path: "$[?(@.child @.other)]"},
 		{name: "filter with malformed term", path: "$[?([)]", illegal: true},
 		{name: "filter with misplaced open bracket", path: "$[?(@.child ()]", illegal: true},
@@ -564,13 +564,13 @@ func TestTokenizer_categorize(t *testing.T) {
 		{name: "filter conjunction", path: "$[?(@.child&&@.other)]"},
 		{name: "filter conjunction with literals and whitespace", path: "$[?(@.child == 'x' && -9 == @.other)]"},
 		{name: "filter conjunction with bracket children", path: "$[?(@['child'][*]&&@['other'])]"},
-		{name: "filter invalid leading conjunction", path: "$[?(&&"},
+		{name: "filter invalid leading conjunction", path: "$[?(&&", illegal: true},
 		{name: "filter conjunction with extra whitespace", path: "$[?(@.child && @.other)]"},
 		{name: "filter disjunction", path: "$[?(@.child||@.other)]"},
-		{name: "filter invalid leading disjunction", path: "$[?(||"},
+		{name: "filter invalid leading disjunction", path: "$[?(||", illegal: true},
 		{name: "filter disjunction with extra whitespace", path: "$[?(@.child || @.other)]"},
 		{name: "simple filter of child", path: "$.child[?(@.child)]"},
-		{name: "filter with missing end", path: "$[?(@.child"},
+		{name: "filter with missing end", path: "$[?(@.child", illegal: true},
 		{name: "nested filter (edge case)", path: "$[?(@.y[?(@.z)])]"},
 		{name: "filter negation", path: "$[?([email protected])]"},
 		{name: "filter negation of comparison (edge case)", path: "$[?([email protected]>1)]"},