nuclear@3: /************************************************************************************
nuclear@3: 
nuclear@3: PublicHeader:   OVR.h
nuclear@3: Filename    :   OVR_String.h
nuclear@3: Content     :   String UTF8 string implementation with copy-on-write semantics
nuclear@3:                 (thread-safe for assignment but not modification).
nuclear@3: Created     :   September 19, 2012
nuclear@3: Notes       : 
nuclear@3: 
nuclear@3: Copyright   :   Copyright 2012 Oculus VR, Inc. All Rights reserved.
nuclear@3: 
nuclear@3: Use of this software is subject to the terms of the Oculus license
nuclear@3: agreement provided at the time of installation or download, or which
nuclear@3: otherwise accompanies this software in either electronic or hard copy form.
nuclear@3: 
nuclear@3: ************************************************************************************/
nuclear@3: 
nuclear@3: #ifndef OVR_String_h
nuclear@3: #define OVR_String_h
nuclear@3: 
nuclear@3: #include "OVR_Types.h"
nuclear@3: #include "OVR_Allocator.h"
nuclear@3: #include "OVR_UTF8Util.h"
nuclear@3: #include "OVR_Atomic.h"
nuclear@3: #include "OVR_Std.h"
nuclear@3: #include "OVR_Alg.h"
nuclear@3: 
nuclear@3: namespace OVR {
nuclear@3: 
nuclear@3: // ***** Classes
nuclear@3: 
nuclear@3: class String;
nuclear@3: class StringBuffer;
nuclear@3: 
nuclear@3: 
nuclear@3: //-----------------------------------------------------------------------------------
nuclear@3: // ***** String Class 
nuclear@3: 
nuclear@3: // String is UTF8 based string class with copy-on-write implementation
nuclear@3: // for assignment.
nuclear@3: 
nuclear@3: class String
nuclear@3: {
nuclear@3: protected:
nuclear@3: 
nuclear@3:     enum FlagConstants
nuclear@3:     {
nuclear@3:         //Flag_GetLength      = 0x7FFFFFFF,
nuclear@3:         // This flag is set if GetLength() == GetSize() for a string.
nuclear@3:         // Avoid extra scanning is Substring and indexing logic.
nuclear@3:         Flag_LengthIsSizeShift   = (sizeof(UPInt)*8 - 1)
nuclear@3:     };
nuclear@3: 
nuclear@3: 
nuclear@3:     // Internal structure to hold string data
nuclear@3:     struct DataDesc
nuclear@3:     {
nuclear@3:         // Number of bytes. Will be the same as the number of chars if the characters
nuclear@3:         // are ascii, may not be equal to number of chars in case string data is UTF8.
nuclear@3:         UPInt   Size;       
nuclear@3:         volatile SInt32 RefCount;
nuclear@3:         char    Data[1];
nuclear@3: 
nuclear@3:         void    AddRef()
nuclear@3:         {
nuclear@3:             AtomicOps<SInt32>::ExchangeAdd_NoSync(&RefCount, 1);
nuclear@3:         }
nuclear@3:         // Decrement ref count. This needs to be thread-safe, since
nuclear@3:         // a different thread could have also decremented the ref count.
nuclear@3:         // For example, if u start off with a ref count = 2. Now if u
nuclear@3:         // decrement the ref count and check against 0 in different
nuclear@3:         // statements, a different thread can also decrement the ref count
nuclear@3:         // in between our decrement and checking against 0 and will find
nuclear@3:         // the ref count = 0 and delete the object. This will lead to a crash
nuclear@3:         // when context switches to our thread and we'll be trying to delete
nuclear@3:         // an already deleted object. Hence decrementing the ref count and
nuclear@3:         // checking against 0 needs to made an atomic operation.
nuclear@3:         void    Release()
nuclear@3:         {
nuclear@3:             if ((AtomicOps<SInt32>::ExchangeAdd_NoSync(&RefCount, -1) - 1) == 0)
nuclear@3:                 OVR_FREE(this);
nuclear@3:         }
nuclear@3: 
nuclear@3:         static UPInt GetLengthFlagBit()     { return UPInt(1) << Flag_LengthIsSizeShift; }
nuclear@3:         UPInt       GetSize() const         { return Size & ~GetLengthFlagBit() ; }
nuclear@3:         UPInt       GetLengthFlag()  const  { return Size & GetLengthFlagBit(); }
nuclear@3:         bool        LengthIsSize() const    { return GetLengthFlag() != 0; }
nuclear@3:     };
nuclear@3: 
nuclear@3:     // Heap type of the string is encoded in the lower bits.
nuclear@3:     enum HeapType
nuclear@3:     {
nuclear@3:         HT_Global   = 0,    // Heap is global.
nuclear@3:         HT_Local    = 1,    // SF::String_loc: Heap is determined based on string's address.
nuclear@3:         HT_Dynamic  = 2,    // SF::String_temp: Heap is stored as a part of the class.
nuclear@3:         HT_Mask     = 3
nuclear@3:     };
nuclear@3: 
nuclear@3:     union {
nuclear@3:         DataDesc* pData;
nuclear@3:         UPInt     HeapTypeBits;
nuclear@3:     };
nuclear@3:     typedef union {
nuclear@3:         DataDesc* pData;
nuclear@3:         UPInt     HeapTypeBits;
nuclear@3:     } DataDescUnion;
nuclear@3: 
nuclear@3:     inline HeapType    GetHeapType() const { return (HeapType) (HeapTypeBits & HT_Mask); }
nuclear@3: 
nuclear@3:     inline DataDesc*   GetData() const
nuclear@3:     {
nuclear@3:         DataDescUnion u;
nuclear@3:         u.pData    = pData;
nuclear@3:         u.HeapTypeBits = (u.HeapTypeBits & ~(UPInt)HT_Mask);
nuclear@3:         return u.pData;
nuclear@3:     }
nuclear@3:     
nuclear@3:     inline void        SetData(DataDesc* pdesc)
nuclear@3:     {
nuclear@3:         HeapType ht = GetHeapType();
nuclear@3:         pData = pdesc;
nuclear@3:         OVR_ASSERT((HeapTypeBits & HT_Mask) == 0);
nuclear@3:         HeapTypeBits |= ht;        
nuclear@3:     }
nuclear@3: 
nuclear@3:     
nuclear@3:     DataDesc*   AllocData(UPInt size, UPInt lengthIsSize);
nuclear@3:     DataDesc*   AllocDataCopy1(UPInt size, UPInt lengthIsSize,
nuclear@3:                                const char* pdata, UPInt copySize);
nuclear@3:     DataDesc*   AllocDataCopy2(UPInt size, UPInt lengthIsSize,
nuclear@3:                                const char* pdata1, UPInt copySize1,
nuclear@3:                                const char* pdata2, UPInt copySize2);
nuclear@3: 
nuclear@3:     // Special constructor to avoid data initalization when used in derived class.
nuclear@3:     struct NoConstructor { };
nuclear@3:     String(const NoConstructor&) { }
nuclear@3: 
nuclear@3: public:
nuclear@3: 
nuclear@3:     // For initializing string with dynamic buffer
nuclear@3:     struct InitStruct
nuclear@3:     {
nuclear@3:         virtual ~InitStruct() { }
nuclear@3:         virtual void InitString(char* pbuffer, UPInt size) const = 0;
nuclear@3:     };
nuclear@3: 
nuclear@3: 
nuclear@3:     // Constructors / Destructors.
nuclear@3:     String();
nuclear@3:     String(const char* data);
nuclear@3:     String(const char* data1, const char* pdata2, const char* pdata3 = 0);
nuclear@3:     String(const char* data, UPInt buflen);
nuclear@3:     String(const String& src);
nuclear@3:     String(const StringBuffer& src);
nuclear@3:     String(const InitStruct& src, UPInt size);
nuclear@3:     explicit String(const wchar_t* data);      
nuclear@3: 
nuclear@3:     // Destructor (Captain Obvious guarantees!)
nuclear@3:     ~String()
nuclear@3:     {
nuclear@3:         GetData()->Release();
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Declaration of NullString
nuclear@3:     static DataDesc NullData;
nuclear@3: 
nuclear@3: 
nuclear@3:     // *** General Functions
nuclear@3: 
nuclear@3:     void        Clear();
nuclear@3: 
nuclear@3:     // For casting to a pointer to char.
nuclear@3:     operator const char*() const        { return GetData()->Data; }
nuclear@3:     // Pointer to raw buffer.
nuclear@3:     const char* ToCStr() const          { return GetData()->Data; }
nuclear@3: 
nuclear@3:     // Returns number of bytes
nuclear@3:     UPInt       GetSize() const         { return GetData()->GetSize() ; }
nuclear@3:     // Tells whether or not the string is empty
nuclear@3:     bool        IsEmpty() const         { return GetSize() == 0; }
nuclear@3: 
nuclear@3:     // Returns  number of characters
nuclear@3:     UPInt       GetLength() const;
nuclear@3: 
nuclear@3:     // Returns  character at the specified index
nuclear@3:     UInt32      GetCharAt(UPInt index) const;
nuclear@3:     UInt32      GetFirstCharAt(UPInt index, const char** offset) const;
nuclear@3:     UInt32      GetNextChar(const char** offset) const;
nuclear@3: 
nuclear@3:     // Appends a character
nuclear@3:     void        AppendChar(UInt32 ch);
nuclear@3: 
nuclear@3:     // Append a string
nuclear@3:     void        AppendString(const wchar_t* pstr, SPInt len = -1);
nuclear@3:     void        AppendString(const char* putf8str, SPInt utf8StrSz = -1);
nuclear@3: 
nuclear@3:     // Assigned a string with dynamic data (copied through initializer).
nuclear@3:     void        AssignString(const InitStruct& src, UPInt size);
nuclear@3:     // Assigns string with known size.
nuclear@3:     void        AssignString(const char* putf8str, UPInt size);
nuclear@3: 
nuclear@3:     //  Resize the string to the new size
nuclear@3: //  void        Resize(UPInt _size);
nuclear@3: 
nuclear@3:     // Removes the character at posAt
nuclear@3:     void        Remove(UPInt posAt, SPInt len = 1);
nuclear@3: 
nuclear@3:     // Returns a String that's a substring of this.
nuclear@3:     //  -start is the index of the first UTF8 character you want to include.
nuclear@3:     //  -end is the index one past the last UTF8 character you want to include.
nuclear@3:     String   Substring(UPInt start, UPInt end) const;
nuclear@3: 
nuclear@3:     // Case-conversion
nuclear@3:     String   ToUpper() const;
nuclear@3:     String   ToLower() const;
nuclear@3: 
nuclear@3:     // Inserts substr at posAt
nuclear@3:     String&    Insert (const char* substr, UPInt posAt, SPInt len = -1);
nuclear@3: 
nuclear@3:     // Inserts character at posAt
nuclear@3:     UPInt       InsertCharAt(UInt32 c, UPInt posAt);
nuclear@3: 
nuclear@3:     // Inserts substr at posAt, which is an index of a character (not byte).
nuclear@3:     // Of size is specified, it is in bytes.
nuclear@3: //  String&    Insert(const UInt32* substr, UPInt posAt, SPInt size = -1);
nuclear@3: 
nuclear@3:     // Get Byte index of the character at position = index
nuclear@3:     UPInt       GetByteIndex(UPInt index) const { return (UPInt)UTF8Util::GetByteIndex(index, GetData()->Data); }
nuclear@3: 
nuclear@3:     // Utility: case-insensitive string compare.  stricmp() & strnicmp() are not
nuclear@3:     // ANSI or POSIX, do not seem to appear in Linux.
nuclear@3:     static int OVR_STDCALL   CompareNoCase(const char* a, const char* b);
nuclear@3:     static int OVR_STDCALL   CompareNoCase(const char* a, const char* b, SPInt len);
nuclear@3: 
nuclear@3:     // Hash function, case-insensitive
nuclear@3:     static UPInt OVR_STDCALL BernsteinHashFunctionCIS(const void* pdataIn, UPInt size, UPInt seed = 5381);
nuclear@3: 
nuclear@3:     // Hash function, case-sensitive
nuclear@3:     static UPInt OVR_STDCALL BernsteinHashFunction(const void* pdataIn, UPInt size, UPInt seed = 5381);
nuclear@3: 
nuclear@3: 
nuclear@3:     // ***** File path parsing helper functions.
nuclear@3:     // Implemented in OVR_String_FilePath.cpp.
nuclear@3: 
nuclear@3:     // Absolute paths can star with:
nuclear@3:     //  - protocols:        'file://', 'http://'
nuclear@3:     //  - windows drive:    'c:\'
nuclear@3:     //  - UNC share name:   '\\share'
nuclear@3:     //  - unix root         '/'
nuclear@3:     static bool HasAbsolutePath(const char* path);
nuclear@3:     static bool HasExtension(const char* path);
nuclear@3:     static bool HasProtocol(const char* path);
nuclear@3: 
nuclear@3:     bool    HasAbsolutePath() const { return HasAbsolutePath(ToCStr()); }
nuclear@3:     bool    HasExtension() const    { return HasExtension(ToCStr()); }
nuclear@3:     bool    HasProtocol() const     { return HasProtocol(ToCStr()); }
nuclear@3: 
nuclear@3:     String  GetProtocol() const;    // Returns protocol, if any, with trailing '://'.
nuclear@3:     String  GetPath() const;        // Returns path with trailing '/'.
nuclear@3:     String  GetFilename() const;    // Returns filename, including extension.
nuclear@3:     String  GetExtension() const;   // Returns extension with a dot.
nuclear@3: 
nuclear@3:     void    StripProtocol();        // Strips front protocol, if any, from the string.
nuclear@3:     void    StripExtension();       // Strips off trailing extension.
nuclear@3:     
nuclear@3: 
nuclear@3:     // Operators
nuclear@3:     // Assignment
nuclear@3:     void        operator =  (const char* str);
nuclear@3:     void        operator =  (const wchar_t* str);
nuclear@3:     void        operator =  (const String& src);
nuclear@3:     void        operator =  (const StringBuffer& src);
nuclear@3: 
nuclear@3:     // Addition
nuclear@3:     void        operator += (const String& src);
nuclear@3:     void        operator += (const char* psrc)       { AppendString(psrc); }
nuclear@3:     void        operator += (const wchar_t* psrc)    { AppendString(psrc); }
nuclear@3:     void        operator += (char  ch)               { AppendChar(ch); }
nuclear@3:     String      operator +  (const char* str) const;
nuclear@3:     String      operator +  (const String& src)  const;
nuclear@3: 
nuclear@3:     // Comparison
nuclear@3:     bool        operator == (const String& str) const
nuclear@3:     {
nuclear@3:         return (OVR_strcmp(GetData()->Data, str.GetData()->Data)== 0);
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator != (const String& str) const
nuclear@3:     {
nuclear@3:         return !operator == (str);
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator == (const char* str) const
nuclear@3:     {
nuclear@3:         return OVR_strcmp(GetData()->Data, str) == 0;
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator != (const char* str) const
nuclear@3:     {
nuclear@3:         return !operator == (str);
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator <  (const char* pstr) const
nuclear@3:     {
nuclear@3:         return OVR_strcmp(GetData()->Data, pstr) < 0;
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator <  (const String& str) const
nuclear@3:     {
nuclear@3:         return *this < str.GetData()->Data;
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator >  (const char* pstr) const
nuclear@3:     {
nuclear@3:         return OVR_strcmp(GetData()->Data, pstr) > 0;
nuclear@3:     }
nuclear@3: 
nuclear@3:     bool        operator >  (const String& str) const
nuclear@3:     {
nuclear@3:         return *this > str.GetData()->Data;
nuclear@3:     }
nuclear@3: 
nuclear@3:     int CompareNoCase(const char* pstr) const
nuclear@3:     {
nuclear@3:         return CompareNoCase(GetData()->Data, pstr);
nuclear@3:     }
nuclear@3:     int CompareNoCase(const String& str) const
nuclear@3:     {
nuclear@3:         return CompareNoCase(GetData()->Data, str.ToCStr());
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Accesses raw bytes
nuclear@3:     const char&     operator [] (int index) const
nuclear@3:     {
nuclear@3:         OVR_ASSERT(index >= 0 && (UPInt)index < GetSize());
nuclear@3:         return GetData()->Data[index];
nuclear@3:     }
nuclear@3:     const char&     operator [] (UPInt index) const
nuclear@3:     {
nuclear@3:         OVR_ASSERT(index < GetSize());
nuclear@3:         return GetData()->Data[index];
nuclear@3:     }
nuclear@3: 
nuclear@3: 
nuclear@3:     // Case insensitive keys are used to look up insensitive string in hash tables
nuclear@3:     // for SWF files with version before SWF 7.
nuclear@3:     struct NoCaseKey
nuclear@3:     {   
nuclear@3:         const String* pStr;
nuclear@3:         NoCaseKey(const String &str) : pStr(&str){};
nuclear@3:     };
nuclear@3: 
nuclear@3:     bool    operator == (const NoCaseKey& strKey) const
nuclear@3:     {
nuclear@3:         return (CompareNoCase(ToCStr(), strKey.pStr->ToCStr()) == 0);
nuclear@3:     }
nuclear@3:     bool    operator != (const NoCaseKey& strKey) const
nuclear@3:     {
nuclear@3:         return !(CompareNoCase(ToCStr(), strKey.pStr->ToCStr()) == 0);
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Hash functor used for strings.
nuclear@3:     struct HashFunctor
nuclear@3:     {    
nuclear@3:         UPInt  operator()(const String& data) const
nuclear@3:         {
nuclear@3:             UPInt  size = data.GetSize();
nuclear@3:             return String::BernsteinHashFunction((const char*)data, size);
nuclear@3:         }        
nuclear@3:     };
nuclear@3:     // Case-insensitive hash functor used for strings. Supports additional
nuclear@3:     // lookup based on NoCaseKey.
nuclear@3:     struct NoCaseHashFunctor
nuclear@3:     {    
nuclear@3:         UPInt  operator()(const String& data) const
nuclear@3:         {
nuclear@3:             UPInt  size = data.GetSize();
nuclear@3:             return String::BernsteinHashFunctionCIS((const char*)data, size);
nuclear@3:         }
nuclear@3:         UPInt  operator()(const NoCaseKey& data) const
nuclear@3:         {       
nuclear@3:             UPInt  size = data.pStr->GetSize();
nuclear@3:             return String::BernsteinHashFunctionCIS((const char*)data.pStr->ToCStr(), size);
nuclear@3:         }
nuclear@3:     };
nuclear@3: 
nuclear@3: };
nuclear@3: 
nuclear@3: 
nuclear@3: //-----------------------------------------------------------------------------------
nuclear@3: // ***** String Buffer used for Building Strings
nuclear@3: 
nuclear@3: class StringBuffer
nuclear@3: {
nuclear@3:     char*           pData;
nuclear@3:     UPInt           Size;
nuclear@3:     UPInt           BufferSize;
nuclear@3:     UPInt           GrowSize;    
nuclear@3:     mutable bool    LengthIsSize;    
nuclear@3: 
nuclear@3: public:
nuclear@3: 
nuclear@3:     // Constructors / Destructor.    
nuclear@3:     StringBuffer();
nuclear@3:     explicit StringBuffer(UPInt growSize);
nuclear@3:     StringBuffer(const char* data);
nuclear@3:     StringBuffer(const char* data, UPInt buflen);
nuclear@3:     StringBuffer(const String& src);
nuclear@3:     StringBuffer(const StringBuffer& src);
nuclear@3:     explicit StringBuffer(const wchar_t* data);
nuclear@3:     ~StringBuffer();
nuclear@3:     
nuclear@3: 
nuclear@3:     // Modify grow size used for growing/shrinking the buffer.
nuclear@3:     UPInt       GetGrowSize() const         { return GrowSize; }
nuclear@3:     void        SetGrowSize(UPInt growSize);
nuclear@3:     
nuclear@3: 
nuclear@3:     // *** General Functions
nuclear@3:     // Does not release memory, just sets Size to 0
nuclear@3:     void        Clear();
nuclear@3: 
nuclear@3:     // For casting to a pointer to char.
nuclear@3:     operator const char*() const        { return (pData) ? pData : ""; }
nuclear@3:     // Pointer to raw buffer.
nuclear@3:     const char* ToCStr() const          { return (pData) ? pData : ""; }
nuclear@3: 
nuclear@3:     // Returns number of bytes.
nuclear@3:     UPInt       GetSize() const         { return Size ; }
nuclear@3:     // Tells whether or not the string is empty.
nuclear@3:     bool        IsEmpty() const         { return GetSize() == 0; }
nuclear@3: 
nuclear@3:     // Returns  number of characters
nuclear@3:     UPInt       GetLength() const;
nuclear@3: 
nuclear@3:     // Returns  character at the specified index
nuclear@3:     UInt32      GetCharAt(UPInt index) const;
nuclear@3:     UInt32      GetFirstCharAt(UPInt index, const char** offset) const;
nuclear@3:     UInt32      GetNextChar(const char** offset) const;
nuclear@3: 
nuclear@3: 
nuclear@3:     //  Resize the string to the new size
nuclear@3:     void        Resize(UPInt _size);
nuclear@3:     void        Reserve(UPInt _size);
nuclear@3: 
nuclear@3:     // Appends a character
nuclear@3:     void        AppendChar(UInt32 ch);
nuclear@3: 
nuclear@3:     // Append a string
nuclear@3:     void        AppendString(const wchar_t* pstr, SPInt len = -1);
nuclear@3:     void        AppendString(const char* putf8str, SPInt utf8StrSz = -1);
nuclear@3:     void        AppendFormat(const char* format, ...);
nuclear@3: 
nuclear@3:     // Assigned a string with dynamic data (copied through initializer).
nuclear@3:     //void        AssignString(const InitStruct& src, UPInt size);
nuclear@3: 
nuclear@3:     // Inserts substr at posAt
nuclear@3:     void        Insert (const char* substr, UPInt posAt, SPInt len = -1);
nuclear@3:     // Inserts character at posAt
nuclear@3:     UPInt       InsertCharAt(UInt32 c, UPInt posAt);
nuclear@3: 
nuclear@3:     // Assignment
nuclear@3:     void        operator =  (const char* str);
nuclear@3:     void        operator =  (const wchar_t* str);
nuclear@3:     void        operator =  (const String& src);
nuclear@3: 
nuclear@3:     // Addition
nuclear@3:     void        operator += (const String& src)      { AppendString(src.ToCStr(),src.GetSize()); }
nuclear@3:     void        operator += (const char* psrc)       { AppendString(psrc); }
nuclear@3:     void        operator += (const wchar_t* psrc)    { AppendString(psrc); }
nuclear@3:     void        operator += (char  ch)               { AppendChar(ch); }
nuclear@3:     //String   operator +  (const char* str) const ;
nuclear@3:     //String   operator +  (const String& src)  const ;
nuclear@3: 
nuclear@3:     // Accesses raw bytes
nuclear@3:     char&       operator [] (int index)
nuclear@3:     {
nuclear@3:         OVR_ASSERT(((UPInt)index) < GetSize());
nuclear@3:         return pData[index];
nuclear@3:     }
nuclear@3:     char&       operator [] (UPInt index)
nuclear@3:     {
nuclear@3:         OVR_ASSERT(index < GetSize());
nuclear@3:         return pData[index];
nuclear@3:     }
nuclear@3: 
nuclear@3:     const char&     operator [] (int index) const 
nuclear@3:     {
nuclear@3:         OVR_ASSERT(((UPInt)index) < GetSize());
nuclear@3:         return pData[index];
nuclear@3:     }
nuclear@3:     const char&     operator [] (UPInt index) const
nuclear@3:     {
nuclear@3:         OVR_ASSERT(index < GetSize());
nuclear@3:         return pData[index];
nuclear@3:     }
nuclear@3: };
nuclear@3: 
nuclear@3: 
nuclear@3: //
nuclear@3: // Wrapper for string data. The data must have a guaranteed 
nuclear@3: // lifespan throughout the usage of the wrapper. Not intended for 
nuclear@3: // cached usage. Not thread safe.
nuclear@3: //
nuclear@3: class StringDataPtr
nuclear@3: {
nuclear@3: public:
nuclear@3:     StringDataPtr() : pStr(NULL), Size(0) {}
nuclear@3:     StringDataPtr(const StringDataPtr& p)
nuclear@3:         : pStr(p.pStr), Size(p.Size) {}
nuclear@3:     StringDataPtr(const char* pstr, UPInt sz)
nuclear@3:         : pStr(pstr), Size(sz) {}
nuclear@3:     StringDataPtr(const char* pstr)
nuclear@3:         : pStr(pstr), Size((pstr != NULL) ? OVR_strlen(pstr) : 0) {}
nuclear@3:     explicit StringDataPtr(const String& str)
nuclear@3:         : pStr(str.ToCStr()), Size(str.GetSize()) {}
nuclear@3:     template <typename T, int N> 
nuclear@3:     StringDataPtr(const T (&v)[N])
nuclear@3:         : pStr(v), Size(N) {}
nuclear@3: 
nuclear@3: public:
nuclear@3:     const char* ToCStr() const { return pStr; }
nuclear@3:     UPInt       GetSize() const { return Size; }
nuclear@3:     bool        IsEmpty() const { return GetSize() == 0; }
nuclear@3: 
nuclear@3:     // value is a prefix of this string
nuclear@3:     // Character's values are not compared.
nuclear@3:     bool        IsPrefix(const StringDataPtr& value) const
nuclear@3:     {
nuclear@3:         return ToCStr() == value.ToCStr() && GetSize() >= value.GetSize();
nuclear@3:     }
nuclear@3:     // value is a suffix of this string
nuclear@3:     // Character's values are not compared.
nuclear@3:     bool        IsSuffix(const StringDataPtr& value) const
nuclear@3:     {
nuclear@3:         return ToCStr() <= value.ToCStr() && (End()) == (value.End());
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Find first character.
nuclear@3:     // init_ind - initial index.
nuclear@3:     SPInt       FindChar(char c, UPInt init_ind = 0) const 
nuclear@3:     {
nuclear@3:         for (UPInt i = init_ind; i < GetSize(); ++i)
nuclear@3:             if (pStr[i] == c)
nuclear@3:                 return static_cast<SPInt>(i);
nuclear@3: 
nuclear@3:         return -1; 
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Find last character.
nuclear@3:     // init_ind - initial index.
nuclear@3:     SPInt       FindLastChar(char c, UPInt init_ind = ~0) const 
nuclear@3:     {
nuclear@3:         if (init_ind == (UPInt)~0 || init_ind > GetSize())
nuclear@3:             init_ind = GetSize();
nuclear@3:         else
nuclear@3:             ++init_ind;
nuclear@3: 
nuclear@3:         for (UPInt i = init_ind; i > 0; --i)
nuclear@3:             if (pStr[i - 1] == c)
nuclear@3:                 return static_cast<SPInt>(i - 1);
nuclear@3: 
nuclear@3:         return -1; 
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Create new object and trim size bytes from the left.
nuclear@3:     StringDataPtr  GetTrimLeft(UPInt size) const
nuclear@3:     {
nuclear@3:         // Limit trim size to the size of the string.
nuclear@3:         size = Alg::PMin(GetSize(), size);
nuclear@3: 
nuclear@3:         return StringDataPtr(ToCStr() + size, GetSize() - size);
nuclear@3:     }
nuclear@3:     // Create new object and trim size bytes from the right.
nuclear@3:     StringDataPtr  GetTrimRight(UPInt size) const
nuclear@3:     {
nuclear@3:         // Limit trim to the size of the string.
nuclear@3:         size = Alg::PMin(GetSize(), size);
nuclear@3: 
nuclear@3:         return StringDataPtr(ToCStr(), GetSize() - size);
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Create new object, which contains next token.
nuclear@3:     // Useful for parsing.
nuclear@3:     StringDataPtr GetNextToken(char separator = ':') const
nuclear@3:     {
nuclear@3:         UPInt cur_pos = 0;
nuclear@3:         const char* cur_str = ToCStr();
nuclear@3: 
nuclear@3:         for (; cur_pos < GetSize() && cur_str[cur_pos]; ++cur_pos)
nuclear@3:         {
nuclear@3:             if (cur_str[cur_pos] == separator)
nuclear@3:             {
nuclear@3:                 break;
nuclear@3:             }
nuclear@3:         }
nuclear@3: 
nuclear@3:         return StringDataPtr(ToCStr(), cur_pos);
nuclear@3:     }
nuclear@3: 
nuclear@3:     // Trim size bytes from the left.
nuclear@3:     StringDataPtr& TrimLeft(UPInt size)
nuclear@3:     {
nuclear@3:         // Limit trim size to the size of the string.
nuclear@3:         size = Alg::PMin(GetSize(), size);
nuclear@3:         pStr += size;
nuclear@3:         Size -= size;
nuclear@3: 
nuclear@3:         return *this;
nuclear@3:     }
nuclear@3:     // Trim size bytes from the right.
nuclear@3:     StringDataPtr& TrimRight(UPInt size)
nuclear@3:     {
nuclear@3:         // Limit trim to the size of the string.
nuclear@3:         size = Alg::PMin(GetSize(), size);
nuclear@3:         Size -= size;
nuclear@3: 
nuclear@3:         return *this;
nuclear@3:     }
nuclear@3: 
nuclear@3:     const char* Begin() const { return ToCStr(); }
nuclear@3:     const char* End() const { return ToCStr() + GetSize(); }
nuclear@3: 
nuclear@3:     // Hash functor used string data pointers
nuclear@3:     struct HashFunctor
nuclear@3:     {    
nuclear@3:         UPInt operator()(const StringDataPtr& data) const
nuclear@3:         {
nuclear@3:             return String::BernsteinHashFunction(data.ToCStr(), data.GetSize());
nuclear@3:         }        
nuclear@3:     };
nuclear@3: 
nuclear@3:     bool operator== (const StringDataPtr& data) const 
nuclear@3:     {
nuclear@3:         return (OVR_strncmp(pStr, data.pStr, data.Size) == 0);
nuclear@3:     }
nuclear@3: 
nuclear@3: protected:
nuclear@3:     const char* pStr;
nuclear@3:     UPInt       Size;
nuclear@3: };
nuclear@3: 
nuclear@3: } // OVR
nuclear@3: 
nuclear@3: #endif