[regexp] Change the RegExpParser to operate on raw input arrays

.. instead of a FlatStringReader. This is in preparation for reusing
the regexp parser directly from the JS parser, which uses different
string types (AstRawString instead of heap Strings).

Drive-by: Hide parser internals in the .cc file.

Bug: v8:896
Change-Id: I06bd08f2ef5fd7a5e9812c123d88b89cacf5d864
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3101488
Commit-Queue: Patrick Thier <pthier@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76365}

src/regexp/experimental/experimental.cc

@@ -73,11 +73,10 @@ base::Optional<CompilationResult> CompileImpl(Isolate* isolate,
 
   // Parse and compile the regexp source.
   RegExpCompileData parse_result;
-  FlatStringReader reader(isolate, source);
   DCHECK(!isolate->has_pending_exception());
 
-  bool parse_success =
-      RegExpParser::ParseRegExp(isolate, &zone, &reader, flags, &parse_result);
+  bool parse_success = RegExpParser::ParseRegExpFromHeapString(
+      isolate, &zone, source, flags, &parse_result);
   if (!parse_success) {
     // The pattern was already parsed successfully during initialization, so
     // the only way parsing can fail now is because of stack overflow.

src/regexp/regexp.cc

@@ -180,10 +180,9 @@ MaybeHandle<Object> RegExp::Compile(Isolate* isolate, Handle<JSRegExp> re,
 
   PostponeInterruptsScope postpone(isolate);
   RegExpCompileData parse_result;
-  FlatStringReader reader(isolate, pattern);
   DCHECK(!isolate->has_pending_exception());
-  if (!RegExpParser::ParseRegExp(isolate, &zone, &reader, flags,
-                                 &parse_result)) {
+  if (!RegExpParser::ParseRegExpFromHeapString(isolate, &zone, pattern, flags,
+                                               &parse_result)) {
     // Throw an exception if we fail to parse the pattern.
     return RegExp::ThrowRegExpException(isolate, re, pattern,
                                         parse_result.error);
@@ -507,9 +506,8 @@ bool RegExpImpl::CompileIrregexp(Isolate* isolate, Handle<JSRegExp> re,
   Handle<String> pattern(re->Pattern(), isolate);
   pattern = String::Flatten(isolate, pattern);
   RegExpCompileData compile_data;
-  FlatStringReader reader(isolate, pattern);
-  if (!RegExpParser::ParseRegExp(isolate, &zone, &reader, flags,
-                                 &compile_data)) {
+  if (!RegExpParser::ParseRegExpFromHeapString(isolate, &zone, pattern, flags,
+                                               &compile_data)) {
     // Throw an exception if we fail to parse the pattern.
     // THIS SHOULD NOT HAPPEN. We already pre-parsed it successfully once.
     USE(RegExp::ThrowRegExpException(isolate, re, pattern, compile_data.error));

test/cctest/test-regexp.cc

@@ -63,9 +63,8 @@ static bool CheckParse(const char* input) {
   v8::HandleScope scope(CcTest::isolate());
   Zone zone(isolate->allocator(), ZONE_NAME);
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData result;
-  return v8::internal::RegExpParser::ParseRegExp(isolate, &zone, &reader,
+  return RegExpParser::ParseRegExpFromHeapString(isolate, &zone, str,
                                                  JSRegExp::kNone, &result);
 }
 
@@ -76,11 +75,10 @@ static void CheckParseEq(const char* input, const char* expected,
   v8::HandleScope scope(CcTest::isolate());
   Zone zone(isolate->allocator(), ZONE_NAME);
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData result;
   JSRegExp::Flags flags = JSRegExp::kNone;
   if (unicode) flags |= JSRegExp::kUnicode;
-  CHECK(v8::internal::RegExpParser::ParseRegExp(isolate, &zone, &reader, flags,
+  CHECK(RegExpParser::ParseRegExpFromHeapString(isolate, &zone, str, flags,
                                                 &result));
   CHECK_NOT_NULL(result.tree);
   CHECK(result.error == RegExpError::kNone);
@@ -98,9 +96,8 @@ static bool CheckSimple(const char* input) {
   v8::HandleScope scope(CcTest::isolate());
   Zone zone(isolate->allocator(), ZONE_NAME);
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData result;
-  CHECK(v8::internal::RegExpParser::ParseRegExp(isolate, &zone, &reader,
+  CHECK(RegExpParser::ParseRegExpFromHeapString(isolate, &zone, str,
                                                 JSRegExp::kNone, &result));
   CHECK_NOT_NULL(result.tree);
   CHECK(result.error == RegExpError::kNone);
@@ -118,9 +115,8 @@ static MinMaxPair CheckMinMaxMatch(const char* input) {
   v8::HandleScope scope(CcTest::isolate());
   Zone zone(isolate->allocator(), ZONE_NAME);
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData result;
-  CHECK(v8::internal::RegExpParser::ParseRegExp(isolate, &zone, &reader,
+  CHECK(RegExpParser::ParseRegExpFromHeapString(isolate, &zone, str,
                                                 JSRegExp::kNone, &result));
   CHECK_NOT_NULL(result.tree);
   CHECK(result.error == RegExpError::kNone);
@@ -433,11 +429,10 @@ static void ExpectError(const char* input, const char* expected,
   v8::HandleScope scope(CcTest::isolate());
   Zone zone(isolate->allocator(), ZONE_NAME);
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData result;
   JSRegExp::Flags flags = JSRegExp::kNone;
   if (unicode) flags |= JSRegExp::kUnicode;
-  CHECK(!v8::internal::RegExpParser::ParseRegExp(isolate, &zone, &reader, flags,
+  CHECK(!RegExpParser::ParseRegExpFromHeapString(isolate, &zone, str, flags,
                                                  &result));
   CHECK_NULL(result.tree);
   CHECK(result.error != RegExpError::kNone);
@@ -536,15 +531,15 @@ static RegExpNode* Compile(const char* input, bool multiline, bool unicode,
                            bool is_one_byte, Zone* zone) {
   Isolate* isolate = CcTest::i_isolate();
   Handle<String> str = isolate->factory()->NewStringFromAsciiChecked(input);
-  FlatStringReader reader(isolate, str);
   RegExpCompileData compile_data;
   compile_data.compilation_target = RegExpCompilationTarget::kNative;
   JSRegExp::Flags flags = JSRegExp::kNone;
   if (multiline) flags = JSRegExp::kMultiline;
   if (unicode) flags = JSRegExp::kUnicode;
-  if (!v8::internal::RegExpParser::ParseRegExp(isolate, zone, &reader, flags,
-                                               &compile_data))
+  if (!RegExpParser::ParseRegExpFromHeapString(isolate, zone, str, flags,
+                                               &compile_data)) {
     return nullptr;
+  }
   Handle<String> pattern = isolate->factory()
                                ->NewStringFromUtf8(base::CStrVector(input))
                                .ToHandleChecked();