vm: rework `in` operator semantics

- Ensure that testing for array membership does strict equality tests

- Ensure that `(NaN in [ NaN ]) == true`

- Do not perform implicit value conversion when testing for object keys,
  to avoid nonsensical results such as `([] in { "[ ]": true }) == true`

- Add test cases for the `in` operator

Fixes: #193
Signed-off-by: Jo-Philipp Wich <jo@mein.io>

2 files changed, 163 insertions, 15 deletions

diff --git a/tests/custom/00_syntax/28_in_operator b/tests/custom/00_syntax/28_in_operator
new file mode 100644
index 0000000..7d58a43
--- /dev/null
+++ b/tests/custom/00_syntax/28_in_operator
@@ -0,0 +1,129 @@
+The "in" operator allows testing whether a given value is an item of
+a specified array or whether a given key is present in a specified
+dictionary.
+
+
+1. The `in` operator returns true if the given element is an item of
+the specified array. Strict equality tests are performed.
+
+-- Expect stdout --
+[
+	true,
+	false,
+	true,
+	false,
+	true,
+	false,
+	true,
+	false
+]
+-- End --
+
+-- Testcase --
+{%
+	let o = {};
+	let a = [ o, {}, "", null, false ];
+
+	printf("%.J\n", [
+		o in a,
+		{} in a,
+		"" in a,
+		"test" in a,
+		null in a,
+		[] in a,
+		false in a,
+		true in a
+	]);
+%}
+-- End --
+
+2. Strict equality when testing array membership should rule out implict
+type coercion.
+
+-- Expect stdout --
+[
+	true,
+	false,
+	false,
+	false,
+	true,
+	false,
+	false
+]
+-- End --
+
+-- Testcase --
+{%
+	let a = [ "", true ];
+
+	printf("%.J\n", [
+		"" in a,
+		0 in a,
+		false in a,
+		null in a,
+		true in a,
+		1 in a,
+		1.0 in a
+	]);
+%}
+-- End --
+
+3. While there is the rule that `(NaN === NaN) == false`, testing for NaN
+in a given array containing NaN should yield `true`.
+
+-- Expect stdout --
+[
+	true
+]
+-- End --
+
+-- Testcase --
+{%
+	let a = [ NaN ];
+
+	printf("%.J\n", [
+		NaN in a
+	]);
+%}
+-- End --
+
+4. When the `in` operator is applied to an object, it tests whether the given
+string value is a key of the specified object. 
+
+-- Expect stdout --
+[
+	true,
+	true,
+	true,
+	false,
+	false,
+	false,
+	false,
+	false
+]
+-- End --
+
+-- Testcase --
+{%
+	let o = { 
+		"1": true,
+		"test": false,
+		"empty": null,
+		"false": 0,
+		"true": 1,
+		"[ ]": "array",
+		"{ }": "object"
+	};
+
+	printf("%.J\n", [
+		"1" in o,
+		"test" in o,
+		"empty" in o,
+		1 in o,        // not implicitly converted to "1"
+		false in o,    // not implicitly converted to "false"
+		true in o,     // not implicitly converted to "true"
+		[] in o,       // not implicitly converted to "[ ]"
+		{} in o        // not implicitly converted to "{ }"
+	]);
+%}
+-- End --
diff --git a/vm.c b/vm.c
index 4642380..bb6dc2f 100644
--- a/vm.c
+++ b/vm.c
@@ -993,6 +993,37 @@ uc_vm_raise_exception(uc_vm_t *vm, uc_exception_type_t type, const char *fmt, ..
 	vm->exception.stacktrace = uc_vm_get_error_context(vm);
 }
 
+static bool
+uc_vm_test_strict_equality(uc_value_t *v1, uc_value_t *v2, bool nan_equal)
+{
+	uc_type_t t1 = ucv_type(v1);
+	uc_type_t t2 = ucv_type(v2);
+	double d1, d2;
+
+	if (t1 != t2)
+		return false;
+
+	switch (t1) {
+	case UC_DOUBLE:
+		d1 = ((uc_double_t *)v1)->dbl;
+		d2 = ((uc_double_t *)v2)->dbl;
+
+		if (isnan(d1) && isnan(d2))
+			return nan_equal;
+
+		return (d1 == d2);
+
+	case UC_NULL:
+	case UC_BOOLEAN:
+	case UC_INTEGER:
+	case UC_STRING:
+		return ucv_is_equal(v1, v2);
+
+	default:
+		return (v1 == v2);
+	}
+}
+
 
 static void
 uc_vm_insn_load(uc_vm_t *vm, uc_vm_insn_t insn)
@@ -2066,7 +2097,6 @@ uc_vm_insn_in(uc_vm_t *vm, uc_vm_insn_t insn)
 	uc_value_t *item;
 	size_t arrlen, arridx;
 	bool found = false;
-	char *key;
 
 	switch (ucv_type(r2)) {
 	case UC_ARRAY:
@@ -2074,7 +2104,7 @@ uc_vm_insn_in(uc_vm_t *vm, uc_vm_insn_t insn)
 		     arridx < arrlen; arridx++) {
 			item = ucv_array_get(r2, arridx);
 
-			if (ucv_compare(I_EQ, r1, item, NULL)) {
+			if (uc_vm_test_strict_equality(r1, item, true)) {
 				found = true;
 				break;
 			}
@@ -2083,14 +2113,8 @@ uc_vm_insn_in(uc_vm_t *vm, uc_vm_insn_t insn)
 		break;
 
 	case UC_OBJECT:
-		if (ucv_type(r1) == UC_STRING) {
+		if (ucv_type(r1) == UC_STRING)
 			ucv_object_get(r2, ucv_string_get(r1), &found);
-		}
-		else {
-			key = ucv_to_string(vm, r1);
-			ucv_object_get(r2, key, &found);
-			free(key);
-		}
 
 		break;
 
@@ -2109,12 +2133,7 @@ uc_vm_insn_equality(uc_vm_t *vm, uc_vm_insn_t insn)
 {
 	uc_value_t *r2 = uc_vm_stack_pop(vm);
 	uc_value_t *r1 = uc_vm_stack_pop(vm);
-	bool equal;
-
-	if (ucv_is_scalar(r1) && ucv_is_scalar(r2))
-		equal = ucv_is_equal(r1, r2);
-	else
-		equal = (r1 == r2);
+	bool equal = uc_vm_test_strict_equality(r1, r2, false);
 
 	ucv_put(r1);
 	ucv_put(r2);