5f59a46 fix(expr): unescape every doubled quote in string literals, as the runner does · ermite-actions

M expr/PORTING.md +8 -5


@@ -64,11 +64,14 @@ Each item is implemented as the JavaScript does. Check the runner's tests
64	64		(`src/Test/L0/Expressions*`, `Sdk/Expressions2/...`) and decide for each one
65	65		whether to keep the JS behaviour or follow the runner.
66	66
67		-	- [ ] String escapes: only the first `''` is unescaped. The lexer does
68		-	`value.replace("''", "'")` with a string pattern, which replaces one
69		-	occurrence: `'a''b''c'` is `a'b''c`. The C# lexer should unescape every
70		-	pair (`StringBuilder`/`Replace`). Probable JS bug. (`lexer.go`,
71		-	`consumeString`; tests: `TestEvalLiteralsAndOperators`.)
	67	+	- [x] String escapes: every `''` is unescaped. Resolved: follows the C#.
	68	+	`LexicalAnalyzer.ReadStringToken` (`Tokens/LexicalAnalyzer.cs`, the
	69	+	"Escaped single quote" branch) skips the second quote of each pair and
	70	+	appends the first, so `'a''b''c'` is `a'b'c`. The JavaScript lexer does
	71	+	`value.replace("''", "'")` (first pair only) and gives `a'b''c`; that is
	72	+	a bug of the editor library. Tests: `TestEvalLiteralsAndOperators`; the
	73	+	oracle excuses the JS mismatch because `testdata/runner.json.gz` agrees
	74	+	with this package.
72	75		- [ ] `case()` availability. The JS gates it behind the experimental flag
73	76		`allowCaseFunction`, but only in completion; the parser and the
74	77		evaluator always accept it. Here it is always on. Confirm whether the

 (`src/Test/L0/Expressions*`, `Sdk/Expressions2/...`) and decide for each one
 whether to keep the JS behaviour or follow the runner.
-- [ ] **String escapes: only the first `''` is unescaped.** The lexer does
+- [x] **String escapes: every `''` is unescaped.** Resolved: follows the C#.
-      `value.replace("''", "'")` with a string pattern, which replaces one
+      `LexicalAnalyzer.ReadStringToken` (`Tokens/LexicalAnalyzer.cs`, the
-      occurrence: `'a''b''c'` is `a'b''c`. The C# lexer should unescape every
+      "Escaped single quote" branch) skips the second quote of each pair and
-      pair (`StringBuilder`/`Replace`). Probable JS bug. (`lexer.go`,
+      appends the first, so `'a''b''c'` is `a'b'c`. The JavaScript lexer does
-      `consumeString`; tests: `TestEvalLiteralsAndOperators`.)
+      `value.replace("''", "'")` (first pair only) and gives `a'b''c`; that is
+      a bug of the editor library. Tests: `TestEvalLiteralsAndOperators`; the
+      oracle excuses the JS mismatch because `testdata/runner.json.gz` agrees
+      with this package.
 - [ ] **`case()` availability.** The JS gates it behind the experimental flag
       `allowCaseFunction`, but only in completion; the parser and the
       evaluator always accept it. Here it is always on. Confirm whether the

M expr/eval_test.go +4 -4


@@ -88,11 +88,11 @@ func TestEvalLiteralsAndOperators(t *testing.T) {
88	88		{"'a'", "String:a"},
89	89		{"''", "String:"},
90	90		{"'it''s'", "String:it's"},
91		-	// Only the first doubled quote is unescaped: the reference's
92		-	// value.replace("''", "'") is not global.
93		-	{"'a''b''c'", "String:a'b''c"},
	91	+	// Every doubled quote is unescaped, as the C# runner does (the
	92	+	// JavaScript reference stops after the first).
	93	+	{"'a''b''c'", "String:a'b'c"},
94	94		{"''''", "String:'"},
95		-	{"'a''''b'", "String:a'''b"},
	95	+	{"'a''''b'", "String:a''b"},
96	96
97	97		{"1 == 1", "Boolean:true"},
98	98		{"1 == 2", "Boolean:false"},


@@ -88,11 +88,11 @@ func TestEvalLiteralsAndOperators(t *testing.T) {
88		{"'a'", "String:a"},	88		{"'a'", "String:a"},
89		{"''", "String:"},	89		{"''", "String:"},
90		{"'it''s'", "String:it's"},	90		{"'it''s'", "String:it's"},
91	-	// Only the first doubled quote is unescaped: the reference's	91	+	// Every doubled quote is unescaped, as the C# runner does (the
92	-	// value.replace("''", "'") is not global.	92	+	// JavaScript reference stops after the first).
93	-	{"'a''b''c'", "String:a'b''c"},	93	+	{"'a''b''c'", "String:a'b'c"},
94		{"''''", "String:'"},	94		{"''''", "String:'"},
95	-	{"'a''''b'", "String:a'''b"},	95	+	{"'a''''b'", "String:a''b"},
96			96
97		{"1 == 1", "Boolean:true"},	97		{"1 == 1", "Boolean:true"},
98		{"1 == 2", "Boolean:false"},	98		{"1 == 2", "Boolean:false"},

M expr/lexer.go +7 -5


@@ -295,11 +295,13 @@ func (l *lexer) consumeString() error {
295	295		}
296	296		l.next() // closing quote
297	297		value := l.substring(l.start+1, l.offset-1)
298		-	// The reference calls value.replace("''", "'") with a string pattern,
299		-	// which replaces only the FIRST doubled quote: 'a''b''c' is a'b''c.
300		-	// GitHub's C# lexer unescapes every pair. Followed as written; see
301		-	// PORTING.md.
302		-	value = strings.Replace(value, "''", "'", 1)
	298	+	// Every doubled quote is one quote. The JavaScript reference calls
	299	+	// value.replace("''", "'") with a string pattern, which replaces only the
	300	+	// first pair ('a''b''c' is a'b''c there); the C# runner, which is what
	301	+	// runs on github.com, appends one quote per pair
	302	+	// (Tokens/LexicalAnalyzer.cs ReadStringToken, "Escaped single quote"), and
	303	+	// this package follows it. See PORTING.md.
	304	+	value = strings.ReplaceAll(value, "''", "'")
303	305		l.addToken(TokenString, String(value))
304	306		return nil
305	307		}

 	}
 	l.next() // closing quote
 	value := l.substring(l.start+1, l.offset-1)
-	// The reference calls value.replace("''", "'") with a string pattern,
+	// Every doubled quote is one quote. The JavaScript reference calls
-	// which replaces only the FIRST doubled quote: 'a''b''c' is a'b''c.
+	// value.replace("''", "'") with a string pattern, which replaces only the
-	// GitHub's C# lexer unescapes every pair. Followed as written; see
+	// first pair ('a''b''c' is a'b''c there); the C# runner, which is what
-	// PORTING.md.
+	// runs on github.com, appends one quote per pair
-	value = strings.Replace(value, "''", "'", 1)
+	// (Tokens/LexicalAnalyzer.cs ReadStringToken, "Escaped single quote"), and
+	// this package follows it. See PORTING.md.
+	value = strings.ReplaceAll(value, "''", "'")
 	l.addToken(TokenString, String(value))
 	return nil
 }