Upstream ScriptRunIterator for segmenting text runs by script

Source/platform/fonts/ScriptRunIterator.h

+// Copyright 2015 The Chromium 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 ScriptRunIterator_h
+#define ScriptRunIterator_h
+
+#include "platform/PlatformExport.h"
+#include "wtf/Deque.h"
+#include "wtf/Vector.h"
+#include "wtf/dtoa/utils.h"
+
+#include <unicode/uchar.h>
+#include <unicode/uscript.h>
+
+namespace blink {
+
+class ScriptData;
+
+class PLATFORM_EXPORT ScriptRunIterator {
+public:
+    ScriptRunIterator(const UChar* text, size_t length);
+
+    // This maintains a reference to data. It must exist for the lifetime of
+    // this object. Typically data is a singleton that exists for the life of
+    // the process.
+    ScriptRunIterator(const UChar* text, size_t length, const ScriptData*);
+
+    bool consume(unsigned& limit, UScriptCode&);
+
+private:
+    struct BracketRec {
+        UChar32 ch;
+        UScriptCode script;
+    };
+    void openBracket(UChar32);
+    void closeBracket(UChar32);
+    bool mergeSets();
+    void fixupStack(UScriptCode resolvedScript);
+    bool fetch(size_t* pos, UChar32*);
+
+    UScriptCode resolveCurrentScript() const;
+
+    const UChar* m_text;
+    const size_t m_length;
+
+    Deque<BracketRec> m_brackets;
+    size_t m_bracketsFixupDepth;
+    // Limit max brackets so that the bracket tracking buffer does not grow
+    // excessively large when processing long runs of text.
+    static const int kMaxBrackets = 32;
+
+    Vector<UScriptCode> m_currentSet;
+    Vector<UScriptCode> m_nextSet;
+    Vector<UScriptCode> m_aheadSet;
+
+    UChar32 m_aheadCharacter;
+    size_t m_aheadPos;
+
+    UScriptCode m_commonPreferred;
+
+    const ScriptData* m_scriptData;
+
+    DISALLOW_COPY_AND_ASSIGN(ScriptRunIterator);
+};
+
+class PLATFORM_EXPORT ScriptData {
+protected:
+    ScriptData() = default;
+
+public:
+    virtual ~ScriptData();
+
+    enum PairedBracketType {
+        BracketTypeNone,
+        BracketTypeOpen,
+        BracketTypeClose,
+        BracketTypeCount
+    };
+
+    static const int kMaxScriptCount = 20;
+
+    virtual void getScripts(UChar32, Vector<UScriptCode>& dst) const = 0;
+
+    virtual UChar32 getPairedBracket(UChar32) const = 0;
+
+    virtual PairedBracketType getPairedBracketType(UChar32) const = 0;
+
+private:
+    DISALLOW_COPY_AND_ASSIGN(ScriptData);
+};
+
+class PLATFORM_EXPORT ICUScriptData : public ScriptData {
+public:
+    ~ICUScriptData() override
+    {
+    }
+
+    static const ICUScriptData* instance();
+
+    void getScripts(UChar32, Vector<UScriptCode>& dst) const override;
+
+    UChar32 getPairedBracket(UChar32) const override;
+
+    PairedBracketType getPairedBracketType(UChar32) const override;
+
+private:
+    ICUScriptData()
+    {
+    }
+
+    DISALLOW_COPY_AND_ASSIGN(ICUScriptData);
+};
+}
+
+#endif