From 0f6ea4612af53b455ee8969761b3606561932234 Mon Sep 17 00:00:00 2001
From: Seonghyun Kim <sh8281.kim@samsung.com>
Date: Thu, 25 Jul 2024 11:18:32 +0900
Subject: [PATCH] Fix memory error on Yarr

If we want to store WTF::String with Vector or HashSet,
we would use another type of allocator

Signed-off-by: Seonghyun Kim <sh8281.kim@samsung.com>
---
 third_party/yarr/HashSet.h    |  33 ++++++
 third_party/yarr/Vector.h     | 185 +++++++++++++++++++++++++++++++++-
 third_party/yarr/YarrParser.h |   6 +-
 3 files changed, 219 insertions(+), 5 deletions(-)

diff --git a/third_party/yarr/HashSet.h b/third_party/yarr/HashSet.h
index 95d304de0..bcc2cd4fe 100644
--- a/third_party/yarr/HashSet.h
+++ b/third_party/yarr/HashSet.h
@@ -57,6 +57,39 @@ class HashSet : public std::unordered_set<Key, std::hash<Key>, std::equal_to<Key
     }
 };
 
+template <typename Key, typename Allocator = GCUtil::gc_malloc_allocator<Key>>
+class GCHashSet : public Escargot::HashSet<Key, std::hash<Key>, std::equal_to<Key>, Allocator> {
+public:
+    struct AddResult {
+        bool isNewEntry;
+    };
+    AddResult add(const Key& k)
+    {
+        AddResult r;
+        r.isNewEntry = Escargot::HashSet<Key, std::hash<Key>, std::equal_to<Key>, Allocator>::insert(k).second;
+        return r;
+    }
+
+    template<typename Other>
+    void formUnion(const Other& other)
+    {
+        for (const auto& value: other) {
+            add(value);
+        }
+    }
+
+    bool contains(const Key& k)
+    {
+        return Escargot::HashSet<Key, std::hash<Key>, std::equal_to<Key>, Allocator>::find(k) != Escargot::HashSet<Key, std::hash<Key>, std::equal_to<Key>, Allocator>::end();
+    }
+
+    bool isEmpty()
+    {
+        return Escargot::HashSet<Key, std::hash<Key>, std::equal_to<Key>, Allocator>::empty();
+    }
+};
+
 } // namespace WTF
 
 using WTF::HashSet;
+using WTF::GCHashSet;
diff --git a/third_party/yarr/Vector.h b/third_party/yarr/Vector.h
index cd0230fde..4e96f82f0 100644
--- a/third_party/yarr/Vector.h
+++ b/third_party/yarr/Vector.h
@@ -30,9 +30,7 @@ class Vector {
 public:
     typedef typename std::vector<T>::iterator iterator;
     typedef typename std::vector<T>::const_iterator const_iterator;
-    std::vector<T> impl;
 
-public:
     Vector() {}
     Vector(const Vector& v)
     {
@@ -208,8 +206,191 @@ class Vector {
         std::fill(begin(), end(), val);
     }
 
+private:
+    std::vector<T> impl;
+};
+
+template <typename T>
+class GCVector {
+public:
+    using iterator = T*;
+    using const_iterator = const T*;
+
+    GCVector() {}
+    GCVector(const GCVector& v)
+    {
+        append(v);
+    }
+
+    GCVector(const T* v, size_t len)
+    {
+        impl.reserve(len);
+        for (size_t i = 0; i < len; i ++) {
+            impl.push_back(v[i]);
+        }
+    }
+
+    GCVector(std::initializer_list<T> list)
+    {
+        impl.reserve(list.size());
+        for (auto& i : list) {
+            impl.push_back(i);
+        }
+    }
+
+    size_t size() const
+    {
+        return impl.size();
+    }
+
+    T& operator[](size_t i)
+    {
+        return impl[i];
+    }
+
+    const T& operator[](size_t i) const
+    {
+        return impl[i];
+    }
+
+    T& at(size_t i)
+    {
+        return impl[i];
+    }
+
+    T* data()
+    {
+        return impl.data();
+    }
+
+    iterator begin()
+    {
+        return impl.begin();
+    }
+
+    iterator end()
+    {
+        return impl.end();
+    }
+
+    const_iterator begin() const
+    {
+        return impl.begin();
+    }
+
+    const_iterator end() const
+    {
+        return impl.end();
+    }
+
+    T& last()
+    {
+        return impl.back();
+    }
+
+    bool isEmpty() const
+    {
+        return impl.empty();
+    }
+
+    template <typename U>
+    void append(const U& u)
+    {
+        impl.push_back(static_cast<T>(u));
+    }
+
+    void append(T&& u)
+    {
+        impl.push_back(std::move(u));
+    }
+
+    template <size_t M>
+    void append(const Vector<T, M>& v)
+    {
+        impl.insert(impl.end(), v.impl.begin(), v.impl.end());
+    }
+
+    void insert(size_t i, const T& t)
+    {
+        impl.insert(impl.begin() + i, t);
+    }
+
+    void remove(size_t i)
+    {
+        impl.erase(impl.begin() + i);
+    }
+
+    void removeLast()
+    {
+        impl.pop_back();
+    }
+
+    void clear()
+    {
+        impl.clear();
+    }
+
+    void grow(size_t s)
+    {
+        impl.resize(s);
+    }
+
+    void shrink(size_t newLength)
+    {
+        ASSERT(newLength <= impl.size());
+        while (impl.size() != newLength) {
+            impl.pop_back();
+        }
+    }
+
+    void shrinkToFit()
+    {
+        impl.shrink_to_fit();
+    }
+
+    size_t capacity() const
+    {
+        return impl.capacity();
+    }
+
+    void reserveInitialCapacity(size_t siz)
+    {
+        impl.reserve(siz);
+    }
+
+    void deleteAllValues()
+    {
+        clear();
+    }
+
+    void reserve(size_t capacity)
+    {
+        impl.reserve(capacity);
+    }
+
+    T takeLast()
+    {
+        T last(*impl.rbegin());
+        impl.pop_back();
+        return last;
+    }
+
+    void fill(const T& val, size_t newSize)
+    {
+        if (size() > newSize)
+            shrink(newSize);
+        else if (newSize > capacity()) {
+            clear();
+            grow(newSize);
+        }
+        std::fill(begin(), end(), val);
+    }
+
+private:
+    Escargot::Vector<T, GCUtil::gc_malloc_allocator<T>> impl;
 };
 
 } // namespace WTF
 
 using WTF::Vector;
+using WTF::GCVector;
diff --git a/third_party/yarr/YarrParser.h b/third_party/yarr/YarrParser.h
index 66753d8bb..ffeb0d682 100644
--- a/third_party/yarr/YarrParser.h
+++ b/third_party/yarr/YarrParser.h
@@ -62,7 +62,7 @@ class Parser {
     };
 
     class NamedCaptureGroups {
-        typedef HashSet<String> GroupNameHashSet;
+        typedef GCHashSet<String> GroupNameHashSet;
 
     public:
         NamedCaptureGroups()
@@ -117,7 +117,7 @@ class Parser {
 
     private:
         GroupNameHashSet m_captureGroupNames;
-        Vector<GroupNameHashSet, 1> m_activeCaptureGroupNames;
+        GCVector<GroupNameHashSet> m_activeCaptureGroupNames;
     };
 
     /*
@@ -2076,7 +2076,7 @@ class Parser {
     bool m_kIdentityEscapeSeen { false };
     Vector<ParenthesesType, 16> m_parenthesesStack;
     NamedCaptureGroups m_namedCaptureGroups;
-    HashSet<String> m_forwardReferenceNames;
+    GCHashSet<String> m_forwardReferenceNames;
 
     // Derived by empirical testing of compile time in PCRE and WREC.
     static constexpr unsigned MAX_PATTERN_SIZE = 1024 * 1024;