[regexp] Reduce public API surface

This further reduces the number of things declared in the public
regexp API file, currently still named jsregexp.h.

* Move JSRegExp::Flags convenience functions to regexp-compiler.h.
* Set RegExpImpl methods private if possible (these will later be
  moved to a new hidden impl class).
* Merge RegExpEngine::CompilationResult into RegExpCompileData.
* Move remaining RegExpEngine methods to RegExpImpl and delete
  RegExpEngine.
* Extract RegExpGlobalCache.
* Document a few data structures.

Upcoming CLs will rename RegExpImpl to RegExp and jsregexp.h to
regexp.h. This should then be the only header included from other
directories.

Bug: v8:9359
Change-Id: I78c8f4cca495a2b95735a48b6181583bc3310bdf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1662294Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62218}

BUILD.gn

@@ -2677,7 +2677,6 @@ v8_source_set("v8_base_without_compiler") {
     "src/profiler/tick-sample.h",
     "src/profiler/tracing-cpu-profiler.cc",
     "src/profiler/tracing-cpu-profiler.h",
-    "src/regexp/jsregexp-inl.h",
     "src/regexp/jsregexp.cc",
     "src/regexp/jsregexp.h",
     "src/regexp/property-sequences.cc",

src/regexp/jsregexp-inl.h

-// Copyright 2013 the V8 project authors. All rights reserved.
-// Use of this source code is governed by a BSD-style license that can be
-// found in the LICENSE file.
-
-
-#ifndef V8_REGEXP_JSREGEXP_INL_H_
-#define V8_REGEXP_JSREGEXP_INL_H_
-
-#include "src/objects/js-regexp-inl.h"
-#include "src/objects/objects.h"
-#include "src/regexp/jsregexp.h"
-#include "src/utils/allocation.h"
-
-namespace v8 {
-namespace internal {
-
-
-RegExpImpl::GlobalCache::~GlobalCache() {
-  // Deallocate the register array if we allocated it in the constructor
-  // (as opposed to using the existing jsregexp_static_offsets_vector).
-  if (register_array_size_ > Isolate::kJSRegexpStaticOffsetsVectorSize) {
-    DeleteArray(register_array_);
-  }
-}
-
-
-int32_t* RegExpImpl::GlobalCache::FetchNext() {
-  current_match_index_++;
-  if (current_match_index_ >= num_matches_) {
-    // Current batch of results exhausted.
-    // Fail if last batch was not even fully filled.
-    if (num_matches_ < max_matches_) {
-      num_matches_ = 0;  // Signal failed match.
-      return nullptr;
-    }
-
-    int32_t* last_match =
-        &register_array_[(current_match_index_ - 1) * registers_per_match_];
-    int last_end_index = last_match[1];
-
-    if (regexp_->TypeTag() == JSRegExp::ATOM) {
-      num_matches_ =
-          RegExpImpl::AtomExecRaw(isolate_, regexp_, subject_, last_end_index,
-                                  register_array_, register_array_size_);
-    } else {
-      int last_start_index = last_match[0];
-      if (last_start_index == last_end_index) {
-        // Zero-length match. Advance by one code point.
-        last_end_index = AdvanceZeroLength(last_end_index);
-      }
-      if (last_end_index > subject_->length()) {
-        num_matches_ = 0;  // Signal failed match.
-        return nullptr;
-      }
-      num_matches_ = RegExpImpl::IrregexpExecRaw(
-          isolate_, regexp_, subject_, last_end_index, register_array_,
-          register_array_size_);
-    }
-
-    if (num_matches_ <= 0) return nullptr;
-    current_match_index_ = 0;
-    return register_array_;
-  } else {
-    return &register_array_[current_match_index_ * registers_per_match_];
-  }
-}
-
-
-int32_t* RegExpImpl::GlobalCache::LastSuccessfulMatch() {
-  int index = current_match_index_ * registers_per_match_;
-  if (num_matches_ == 0) {
-    // After a failed match we shift back by one result.
-    index -= registers_per_match_;
-  }
-  return &register_array_[index];
-}
-
-}  // namespace internal
-}  // namespace v8
-
-#endif  // V8_REGEXP_JSREGEXP_INL_H_

src/regexp/regexp-compiler.cc

@@ -7,7 +7,6 @@
 #include "src/diagnostics/code-tracer.h"
 #include "src/execution/isolate.h"
 #include "src/objects/objects-inl.h"
-#include "src/regexp/jsregexp.h"  // TODO(jgruber): Only needed for IgnoreCase.
 #include "src/regexp/regexp-macro-assembler-arch.h"
 #include "src/regexp/regexp-macro-assembler-tracer.h"
 #include "src/strings/unicode-inl.h"

src/regexp/regexp-compiler.h

@@ -45,6 +45,36 @@ constexpr int kPatternTooShortForBoyerMoore = 2;
 
 }  // namespace regexp_compiler_constants
 
+inline bool IgnoreCase(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kIgnoreCase) != 0;
+}
+
+inline bool IsUnicode(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kUnicode) != 0;
+}
+
+inline bool IsSticky(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kSticky) != 0;
+}
+
+inline bool IsGlobal(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kGlobal) != 0;
+}
+
+inline bool DotAll(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kDotAll) != 0;
+}
+
+inline bool Multiline(JSRegExp::Flags flags) {
+  return (flags & JSRegExp::kMultiline) != 0;
+}
+
+inline bool NeedsUnicodeCaseEquivalents(JSRegExp::Flags flags) {
+  // Both unicode and ignore_case flags are set. We need to use ICU to find
+  // the closure over case equivalents.
+  return IsUnicode(flags) && IgnoreCase(flags);
+}
+
 // A set of unsigned integers that behaves especially well on small
 // integers (< 32).  May do zone-allocation.
 class OutSet : public ZoneObject {
@@ -645,6 +675,8 @@ class RegExpCompiler {
       return CompilationResult("RegExp too big");
     }
 
+    bool Succeeded() const { return error_message == nullptr; }
+
     const char* const error_message = nullptr;
     Object code;
     int num_registers = 0;

src/runtime/runtime-regexp.cc

@@ -11,7 +11,8 @@
 #include "src/logging/counters.h"
 #include "src/numbers/conversions-inl.h"
 #include "src/objects/js-array-inl.h"
-#include "src/regexp/jsregexp-inl.h"
+#include "src/objects/js-regexp-inl.h"
+#include "src/regexp/jsregexp.h"
 #include "src/regexp/regexp-utils.h"
 #include "src/runtime/runtime-utils.h"
 #include "src/strings/string-builder-inl.h"
@@ -638,7 +639,7 @@ V8_WARN_UNUSED_RESULT static Object StringReplaceGlobalRegExpWithString(
     }
   }
 
-  RegExpImpl::GlobalCache global_cache(regexp, subject, isolate);
+  RegExpGlobalCache global_cache(regexp, subject, isolate);
   if (global_cache.HasException()) return ReadOnlyRoots(isolate).exception();
 
   int32_t* current_match = global_cache.FetchNext();
@@ -703,7 +704,7 @@ V8_WARN_UNUSED_RESULT static Object StringReplaceGlobalRegExpWithEmptyString(
     }
   }
 
-  RegExpImpl::GlobalCache global_cache(regexp, subject, isolate);
+  RegExpGlobalCache global_cache(regexp, subject, isolate);
   if (global_cache.HasException()) return ReadOnlyRoots(isolate).exception();
 
   int32_t* current_match = global_cache.FetchNext();
@@ -1115,7 +1116,7 @@ static Object SearchRegExpMultiple(Isolate* isolate, Handle<String> subject,
     }
   }
 
-  RegExpImpl::GlobalCache global_cache(regexp, subject, isolate);
+  RegExpGlobalCache global_cache(regexp, subject, isolate);
   if (global_cache.HasException()) return ReadOnlyRoots(isolate).exception();
 
   // Ensured in Runtime_RegExpExecMultiple.

src/runtime/runtime-strings.cc

@@ -10,7 +10,6 @@
 #include "src/objects/objects-inl.h"
 #include "src/objects/slots.h"
 #include "src/objects/smi.h"
-#include "src/regexp/jsregexp-inl.h"
 #include "src/regexp/regexp-utils.h"
 #include "src/runtime/runtime-utils.h"
 #include "src/strings/string-builder-inl.h"

test/cctest/test-regexp.cc

@@ -547,8 +547,8 @@ static RegExpNode* Compile(const char* input, bool multiline, bool unicode,
                                .ToHandleChecked();
   Handle<String> sample_subject =
       isolate->factory()->NewStringFromUtf8(CStrVector("")).ToHandleChecked();
-  RegExpEngine::Compile(isolate, zone, &compile_data, flags, pattern,
-                        sample_subject, is_one_byte);
+  RegExpImpl::CompileForTesting(isolate, zone, &compile_data, flags, pattern,
+                                sample_subject, is_one_byte);
   return compile_data.node;
 }
 
@@ -561,7 +561,7 @@ static void Execute(const char* input, bool multiline, bool unicode,
   USE(node);
 #ifdef DEBUG
   if (dot_output) {
-    RegExpEngine::DotPrint(input, node, false);
+    RegExpImpl::DotPrintForTesting(input, node, false);
   }
 #endif  // DEBUG
 }