Upstream ScriptRunIterator for segmenting text runs by script

Source/platform/fonts/ScriptRunIterator.h

Index: Source/platform/fonts/ScriptRunIterator.h
diff --git a/Source/platform/fonts/ScriptRunIterator.h b/Source/platform/fonts/ScriptRunIterator.h
new file mode 100644
index 0000000000000000000000000000000000000000..59c58feaacfc4193531bb4d2a5aacdc14bb65153
--- /dev/null
+++ b/Source/platform/fonts/ScriptRunIterator.h
@@ -0,0 +1,114 @@
+// 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

[eae, 2015/08/28 21:10:24]

Wrap at 80 col.

+    // object. Typically data is a singleton that exists for the life of the
+    // process.
+    ScriptRunIterator(const UChar* text, size_t length, const ScriptData* data);
+
+    bool consume(unsigned& limit, UScriptCode& script);
+
+private:
+    struct BracketRec {
+        UChar32 ch;
+        UScriptCode script;
+    };
+    void openBracket(UChar32 ch);
+    void closeBracket(UChar32 ch);
+    bool mergeSets();
+    void fixupStack(UScriptCode resolve_code);
+    bool fetch(size_t* pos, UChar32* ch);
+
+    UScriptCode resolveCurrentScript() const;
+
+    const UChar* m_text;
+    const size_t m_length;
+
+    Deque<BracketRec> m_brackets;
+    size_t m_bracketsFixupDepth;
+    static const int kMaxBrackets = 32;

[eae, 2015/08/28 21:10:24]

Could you please add a comment explaining why we limit it to 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 {

[eae, 2015/08/28 21:10:24]

What's the relationship between ScriptRunIterator and ScriptData? Does the
iterator iterate over scripts in a ScriptData object?

+protected:
+    ScriptData() = default;
+
+public:
+    virtual ~ScriptData();
+
+    enum PairedBracketType {
+        NONE,
+        OPEN,
+        CLOSE,
+        COUNT
+    };
+
+    static const int kMaxScriptCount = 20;
+
+    virtual void getScripts(UChar32 ch, Vector<UScriptCode>& dst) const = 0;
+
+    virtual UChar32 getPairedBracket(UChar32 ch) const = 0;
+
+    virtual PairedBracketType getPairedBracketType(UChar32 ch) const = 0;
+
+private:
+    DISALLOW_COPY_AND_ASSIGN(ScriptData);
+};
+
+class PLATFORM_EXPORT ICUScriptData : public ScriptData {
+public:
+    ~ICUScriptData() override
+    {
+    }
+
+    static const ICUScriptData* instance();
+
+    void getScripts(UChar32 ch, Vector<UScriptCode>& dst) const override;
+
+    UChar32 getPairedBracket(UChar32 ch) const override;
+
+    PairedBracketType getPairedBracketType(UChar32 ch) const override;
+
+private:
+    ICUScriptData()
+    {
+    }
+
+    DISALLOW_COPY_AND_ASSIGN(ICUScriptData);
+};
+}
+
+#endif