LibJS: Hook up Regex<ECMA262> to RegExpObject and implement `test()'

This makes RegExpObject compile and store a Regex<ECMA262>, adds
all flag-related properties, and implements `RegExpPrototype.test()`
(complete with 'lastIndex' support) :^)
It should be noted that this only implements `test()' using the builtin
`exec()'.

Libraries/LibJS/Runtime/CommonPropertyNames.h

@@ -86,6 +86,7 @@ namespace JS {
     P(deleteProperty)                        \
     P(description)                           \
     P(done)                                  \
+    P(dotAll)                                \
     P(entries)                               \
     P(enumerable)                            \
     P(error)                                 \
@@ -96,6 +97,7 @@ namespace JS {
     P(filter)                                \
     P(find)                                  \
     P(findIndex)                             \
+    P(flags)                                 \
     P(floor)                                 \
     P(forEach)                               \
     P(from)                                  \
@@ -122,9 +124,11 @@ namespace JS {
     P(getUTCMinutes)                         \
     P(getUTCMonth)                           \
     P(getUTCSeconds)                         \
+    P(global)                                \
     P(globalThis)                            \
     P(has)                                   \
     P(hasOwnProperty)                        \
+    P(ignoreCase)                            \
     P(includes)                              \
     P(indexOf)                               \
     P(info)                                  \
@@ -138,6 +142,7 @@ namespace JS {
     P(join)                                  \
     P(keyFor)                                \
     P(keys)                                  \
+    P(lastIndex)                             \
     P(lastIndexOf)                           \
     P(length)                                \
     P(log)                                   \
@@ -146,6 +151,7 @@ namespace JS {
     P(max)                                   \
     P(message)                               \
     P(min)                                   \
+    P(multiline)                             \
     P(name)                                  \
     P(next)                                  \
     P(now)                                   \
@@ -174,12 +180,15 @@ namespace JS {
     P(sin)                                   \
     P(slice)                                 \
     P(some)                                  \
+    P(source)                                \
     P(splice)                                \
     P(sqrt)                                  \
     P(startsWith)                            \
     P(stringify)                             \
+    P(sticky)                                \
     P(substring)                             \
     P(tan)                                   \
+    P(test)                                  \
     P(toDateString)                          \
     P(toISOString)                           \
     P(toJSON)                                \
@@ -196,6 +205,7 @@ namespace JS {
     P(trimStart)                             \
     P(trunc)                                 \
     P(undefined)                             \
+    P(unicode)                               \
     P(unshift)                               \
     P(value)                                 \
     P(valueOf)                               \

Libraries/LibJS/Runtime/ErrorTypes.h

@@ -146,6 +146,9 @@
     M(ReflectBadArgumentsList, "Arguments list must be an object")                                                                      \
     M(ReflectBadNewTarget, "Optional third argument of Reflect.construct() must be a constructor")                                      \
     M(ReflectBadDescriptorArgument, "Descriptor argument is not an object")                                                             \
+    M(RegExpCompileError, "RegExp compile error: '{}'")                                                                                 \
+    M(RegExpObjectBadFlag, "Invalid RegExp flag '{}'")                                                                                  \
+    M(RegExpObjectRepeatedFlag, "Repeated RegExp flag '{}'")                                                                            \
     M(StringRawCannotConvert, "Cannot convert property 'raw' to object from {}")                                                        \
     M(StringRepeatCountMustBe, "repeat count must be a {} number")                                                                      \
     M(ThisHasNotBeenInitialized, "|this| has not been initialized")                                                                     \

Libraries/LibJS/Runtime/RegExpObject.cpp

@@ -24,6 +24,7 @@
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
+#include <AK/Function.h>
 #include <AK/StringBuilder.h>
 #include <LibJS/Heap/Heap.h>
 #include <LibJS/Runtime/GlobalObject.h>
@@ -33,6 +34,73 @@
 
 namespace JS {
 
+static Flags options_from(const String& flags, VM& vm, GlobalObject& global_object)
+{
+    bool g = false, i = false, m = false, s = false, u = false, y = false;
+    Flags options {
+        { (regex::ECMAScriptFlags)regex::AllFlags::Global }, // JS regexps are all 'global' by default as per our definition, but the "global" flag enables "stateful".
+        {},
+    };
+
+    for (auto ch : flags) {
+        switch (ch) {
+        case 'g':
+            if (g)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            g = true;
+            options.effective_flags |= regex::ECMAScriptFlags::Global;
+            options.declared_flags |= regex::ECMAScriptFlags::Global;
+            break;
+        case 'i':
+            if (i)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            i = true;
+            options.effective_flags |= regex::ECMAScriptFlags::Insensitive;
+            options.declared_flags |= regex::ECMAScriptFlags::Insensitive;
+            break;
+        case 'm':
+            if (m)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            m = true;
+            options.effective_flags |= regex::ECMAScriptFlags::Multiline;
+            options.declared_flags |= regex::ECMAScriptFlags::Multiline;
+            break;
+        case 's':
+            if (s)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            s = true;
+            options.effective_flags |= regex::ECMAScriptFlags::SingleLine;
+            options.declared_flags |= regex::ECMAScriptFlags::SingleLine;
+            break;
+        case 'u':
+            if (u)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            u = true;
+            options.effective_flags |= regex::ECMAScriptFlags::Unicode;
+            options.declared_flags |= regex::ECMAScriptFlags::Unicode;
+            break;
+        case 'y':
+            if (y)
+                vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectRepeatedFlag, ch);
+            y = true;
+            // Now for the more interesting flag, 'sticky' actually unsets 'global', part of which is the default.
+            options.effective_flags.reset_flag(regex::ECMAScriptFlags::Global);
+            // "What's the difference between sticky and global, then", that's simple.
+            // all the other flags imply 'global', and the "global" flag implies 'stateful';
+            // however, the "sticky" flag does *not* imply 'global', only 'stateful'.
+            options.effective_flags |= (regex::ECMAScriptFlags)regex::AllFlags::Internal_Stateful;
+            options.effective_flags |= regex::ECMAScriptFlags::Sticky;
+            options.declared_flags |= regex::ECMAScriptFlags::Sticky;
+            break;
+        default:
+            vm.throw_exception<SyntaxError>(global_object, ErrorType::RegExpObjectBadFlag, ch);
+            return options;
+        }
+    }
+
+    return options;
+}
+
 RegExpObject* RegExpObject::create(GlobalObject& global_object, String pattern, String flags)
 {
     return global_object.heap().allocate<RegExpObject>(global_object, pattern, flags, *global_object.regexp_prototype());
@@ -42,11 +110,61 @@ RegExpObject::RegExpObject(String pattern, String flags, Object& prototype)
     : Object(prototype)
     , m_pattern(pattern)
     , m_flags(flags)
+    , m_active_flags(options_from(m_flags, this->vm(), this->global_object()))
+    , m_regex(pattern, m_active_flags.effective_flags)
 {
+    if (m_regex.parser_result.error != regex::Error::NoError) {
+        vm().throw_exception<SyntaxError>(global_object(), ErrorType::RegExpCompileError, m_regex.error_string());
+    }
+}
+
+void RegExpObject::initialize(GlobalObject& global_object)
+{
+    auto& vm = this->vm();
+    Object::initialize(global_object);
+
+    define_native_property(vm.names.lastIndex, last_index, set_last_index, Attribute::Writable);
 }
 
 RegExpObject::~RegExpObject()
 {
 }
 
+static RegExpObject* regexp_object_from(VM& vm, GlobalObject& global_object)
+{
+    auto* this_object = vm.this_value(global_object).to_object(global_object);
+    if (!this_object)
+        return nullptr;
+    if (!this_object->is_regexp_object()) {
+        vm.throw_exception<TypeError>(global_object, ErrorType::NotA, "RegExp");
+        return nullptr;
+    }
+    return static_cast<RegExpObject*>(this_object);
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpObject::last_index)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value((unsigned)regexp_object->regex().start_offset);
+}
+
+JS_DEFINE_NATIVE_SETTER(RegExpObject::set_last_index)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return;
+
+    auto index = value.to_i32(global_object);
+    if (vm.exception())
+        return;
+
+    if (index < 0)
+        index = 0;
+
+    regexp_object->regex().start_offset = index;
+}
+
 }

Libraries/LibJS/Runtime/RegExpObject.h

@@ -28,6 +28,12 @@
 
 #include <LibJS/AST.h>
 #include <LibJS/Runtime/Object.h>
+#include <LibRegex/Regex.h>
+
+struct Flags {
+    regex::RegexOptions<ECMAScriptFlags> effective_flags;
+    regex::RegexOptions<ECMAScriptFlags> declared_flags;
+};
 
 namespace JS {
 
@@ -38,16 +44,25 @@ public:
     static RegExpObject* create(GlobalObject&, String pattern, String flags);
 
     RegExpObject(String pattern, String flags, Object& prototype);
+    virtual void initialize(GlobalObject&) override;
     virtual ~RegExpObject() override;
 
     const String& pattern() const { return m_pattern; }
     const String& flags() const { return m_flags; }
+    const regex::RegexOptions<ECMAScriptFlags>& declared_options() { return m_active_flags.declared_flags; }
+    const Regex<ECMA262>& regex() { return m_regex; }
+    const Regex<ECMA262>& regex() const { return m_regex; }
 
 private:
     virtual bool is_regexp_object() const override { return true; }
 
+    JS_DECLARE_NATIVE_GETTER(last_index);
+    JS_DECLARE_NATIVE_SETTER(set_last_index);
+
     String m_pattern;
     String m_flags;
+    Flags m_active_flags;
+    Regex<ECMA262> m_regex;
 };
 
 }

Libraries/LibJS/Runtime/RegExpPrototype.cpp

@@ -43,6 +43,17 @@ void RegExpPrototype::initialize(GlobalObject& global_object)
     Object::initialize(global_object);
     u8 attr = Attribute::Writable | Attribute::Configurable;
     define_native_function(vm.names.toString, to_string, 0, attr);
+    define_native_function(vm.names.test, test, 1, attr);
+
+    u8 readable_attr = Attribute::Configurable;
+    define_native_property(vm.names.dotAll, dot_all, nullptr, readable_attr);
+    define_native_property(vm.names.flags, flags, nullptr, readable_attr);
+    define_native_property(vm.names.global, global, nullptr, readable_attr);
+    define_native_property(vm.names.ignoreCase, ignore_case, nullptr, readable_attr);
+    define_native_property(vm.names.multiline, multiline, nullptr, readable_attr);
+    define_native_property(vm.names.source, source, nullptr, readable_attr);
+    define_native_property(vm.names.sticky, sticky, nullptr, readable_attr);
+    define_native_property(vm.names.unicode, unicode, nullptr, readable_attr);
 }
 
 RegExpPrototype::~RegExpPrototype()
@@ -61,6 +72,124 @@ static RegExpObject* regexp_object_from(VM& vm, GlobalObject& global_object)
     return static_cast<RegExpObject*>(this_object);
 }
 
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::dot_all)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::SingleLine));
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::flags)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    auto flags = regexp_object->declared_options();
+    StringBuilder builder(8);
+
+    if (flags.has_flag_set(ECMAScriptFlags::Global))
+        builder.append('g');
+    if (flags.has_flag_set(ECMAScriptFlags::Insensitive))
+        builder.append('i');
+    if (flags.has_flag_set(ECMAScriptFlags::Multiline))
+        builder.append('m');
+    if (flags.has_flag_set(ECMAScriptFlags::SingleLine))
+        builder.append('s');
+    if (flags.has_flag_set(ECMAScriptFlags::Unicode))
+        builder.append('u');
+    if (flags.has_flag_set(ECMAScriptFlags::Sticky))
+        builder.append('y');
+
+    return js_string(vm, builder.to_string());
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::global)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::Global)); // Note that this "Global" is actually "Global | Stateful"
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::ignore_case)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::Insensitive));
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::multiline)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::Multiline));
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::source)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return js_string(vm, regexp_object->pattern());
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::sticky)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::Sticky));
+}
+
+JS_DEFINE_NATIVE_GETTER(RegExpPrototype::unicode)
+{
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    return Value(regexp_object->declared_options().has_flag_set(ECMAScriptFlags::Unicode));
+}
+
+RegexResult RegExpPrototype::do_match(const Regex<ECMA262>& re, const StringView& subject)
+{
+    auto result = re.match(subject);
+    // The 'lastIndex' property is reset on failing tests (if 'global')
+    if (!result.success && re.options().has_flag_set(ECMAScriptFlags::Global))
+        re.start_offset = 0;
+
+    return result;
+}
+
+JS_DEFINE_NATIVE_FUNCTION(RegExpPrototype::test)
+{
+    // FIXME: This should try using dynamic properties for 'exec' first,
+    //        before falling back to builtin_exec.
+    auto regexp_object = regexp_object_from(vm, global_object);
+    if (!regexp_object)
+        return {};
+
+    auto str = vm.argument(0).to_string(global_object);
+    if (vm.exception())
+        return {};
+
+    // RegExps without "global" and "sticky" always start at offset 0.
+    if (!regexp_object->regex().options().has_flag_set((ECMAScriptFlags)regex::AllFlags::Internal_Stateful))
+        regexp_object->regex().start_offset = 0;
+
+    auto result = do_match(regexp_object->regex(), str);
+    return Value(result.success);
+}
+
 JS_DEFINE_NATIVE_FUNCTION(RegExpPrototype::to_string)
 {
     auto* regexp_object = regexp_object_from(vm, global_object);

Libraries/LibJS/Runtime/RegExpPrototype.h

@@ -39,6 +39,18 @@ public:
     virtual ~RegExpPrototype() override;
 
 private:
+    static RegexResult do_match(const Regex<ECMA262>&, const StringView&);
+
+    JS_DECLARE_NATIVE_GETTER(dot_all);
+    JS_DECLARE_NATIVE_GETTER(flags);
+    JS_DECLARE_NATIVE_GETTER(global);
+    JS_DECLARE_NATIVE_GETTER(ignore_case);
+    JS_DECLARE_NATIVE_GETTER(multiline);
+    JS_DECLARE_NATIVE_GETTER(source);
+    JS_DECLARE_NATIVE_GETTER(sticky);
+    JS_DECLARE_NATIVE_GETTER(unicode);
+
+    JS_DECLARE_NATIVE_FUNCTION(test);
     JS_DECLARE_NATIVE_FUNCTION(to_string);
 };