From 6beb0cf4cf9b33d7a82275a53422903b8fcad5ca Mon Sep 17 00:00:00 2001
From: Iain Ireland <iireland@mozilla.com>
Date: Mon, 25 Jul 2022 10:40:37 -0700
Subject: [PATCH] [regexp] SpiderMonkey embedding fixes

There are two changes in this patch.

1. We previously added `VerifyRegExpSyntax` in regexp-parser.h to support checking regexp syntax for early errors in SpiderMonkey. Now that V8 is also emitting early errors for regexps (bug v8:896), SpiderMonkey can use the same code as V8.

2. Bug v8:11069 used a std::unordered_map as a cache for range arrays. This is currently the only place in irregexp that can call non-placement new, which SpiderMonkey has a static analysis to detect. Converting this to a ZoneUnorderedMap solves the problem for us, and seems consistent with the rest of irregexp.

Bug: v8:13108
Change-Id: Icedafd7d30fd040760cb0676a7bef8d55853bb93
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3785444
Commit-Queue: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#81988}
---
 src/regexp/regexp-macro-assembler.h | 4 ++--
 src/regexp/regexp-parser.cc         | 8 --------
 src/regexp/regexp-parser.h          | 6 ------
 3 files changed, 2 insertions(+), 16 deletions(-)

diff --git a/src/regexp/regexp-macro-assembler.h b/src/regexp/regexp-macro-assembler.h
index 16a23414e1..c231d112f8 100644
--- a/src/regexp/regexp-macro-assembler.h
+++ b/src/regexp/regexp-macro-assembler.h
@@ -291,7 +291,7 @@ class NativeRegExpMacroAssembler: public RegExpMacroAssembler {
   };
 
   NativeRegExpMacroAssembler(Isolate* isolate, Zone* zone)
-      : RegExpMacroAssembler(isolate, zone) {}
+      : RegExpMacroAssembler(isolate, zone), range_array_cache_(zone) {}
   ~NativeRegExpMacroAssembler() override = default;
 
   // Returns a {Result} sentinel, or the number of successful matches.
@@ -349,7 +349,7 @@ class NativeRegExpMacroAssembler: public RegExpMacroAssembler {
                      const byte* input_end, int* output, int output_size,
                      Isolate* isolate, JSRegExp regexp);
 
-  std::unordered_map<uint32_t, Handle<ByteArray>> range_array_cache_;
+  ZoneUnorderedMap<uint32_t, Handle<ByteArray>> range_array_cache_;
 };
 
 }  // namespace internal
diff --git a/src/regexp/regexp-parser.cc b/src/regexp/regexp-parser.cc
index 5d71527eba..f50a5ffd0a 100644
--- a/src/regexp/regexp-parser.cc
+++ b/src/regexp/regexp-parser.cc
@@ -2367,14 +2367,6 @@ template bool RegExpParser::VerifyRegExpSyntax<base::uc16>(
     Zone*, uintptr_t, const base::uc16*, int, RegExpFlags, RegExpCompileData*,
     const DisallowGarbageCollection&);
 
-// static
-bool RegExpParser::VerifyRegExpSyntax(Isolate* isolate, Zone* zone,
-                                      Handle<String> input, RegExpFlags flags,
-                                      RegExpCompileData* result,
-                                      const DisallowGarbageCollection&) {
-  return ParseRegExpFromHeapString(isolate, zone, input, flags, result);
-}
-
 #undef LAST
 
 }  // namespace internal
diff --git a/src/regexp/regexp-parser.h b/src/regexp/regexp-parser.h
index 4fc6400297..d7ff611dcb 100644
--- a/src/regexp/regexp-parser.h
+++ b/src/regexp/regexp-parser.h
@@ -28,12 +28,6 @@ class V8_EXPORT_PRIVATE RegExpParser : public AllStatic {
                                  const CharT* input, int input_length,
                                  RegExpFlags flags, RegExpCompileData* result,
                                  const DisallowGarbageCollection& no_gc);
-
-  // Used by the SpiderMonkey embedding of irregexp.
-  static bool VerifyRegExpSyntax(Isolate* isolate, Zone* zone,
-                                 Handle<String> input, RegExpFlags flags,
-                                 RegExpCompileData* result,
-                                 const DisallowGarbageCollection& no_gc);
 };
 
 }  // namespace internal