include/wx/string.h

   1 ///////////////////////////////////////////////////////////////////////////////
   2 // Name:        string.h
   3 // Purpose:     wxString and wxArrayString classes
   4 // Author:      Vadim Zeitlin
   5 // Modified by:
   6 // Created:     29/01/98
   7 // RCS-ID:      $Id$
   8 // Copyright:   (c) 1998 Vadim Zeitlin <zeitlin@dptmaths.ens-cachan.fr>
   9 // Licence:     wxWindows license
  10 ///////////////////////////////////////////////////////////////////////////////
  11
  12 #ifndef _WX_WXSTRINGH__
  13 #define _WX_WXSTRINGH__
  14
  15 #ifdef __GNUG__
  16 #pragma interface "string.h"
  17 #endif
  18
  19 #ifdef __WXMAC__
  20     #include <ctype.h>
  21 #endif
  22
  23 #include <string.h>
  24 #include <stdio.h>
  25 #include <stdarg.h>
  26 #include <limits.h>
  27 #include <stdlib.h>
  28
  29 #ifdef HAVE_STRINGS_H
  30     #include <strings.h>    // for strcasecmp()
  31 #endif // AIX
  32
  33 #ifndef  WX_PRECOMP
  34   #include "wx/defs.h"
  35
  36   #ifdef    WXSTRING_IS_WXOBJECT
  37     #include "wx/object.h"
  38   #endif
  39 #endif // !PCH
  40
  41 #include "wx/debug.h"
  42 #include "wx/wxchar.h"
  43 #include "wx/buffer.h"
  44
  45 /*
  46     Efficient string class [more or less] compatible with MFC CString,
  47     wxWindows version 1 wxString and std::string and some handy functions
  48     missing from string.h.
  49 */
  50
  51 // ---------------------------------------------------------------------------
  52 // macros
  53 // ---------------------------------------------------------------------------
  54
  55 // compile the std::string compatibility functions if defined
  56 #define   wxSTD_STRING_COMPATIBILITY
  57
  58 // define to derive wxString from wxObject
  59 #ifdef    WXSTRING_IS_WXOBJECT
  60 #undef    WXSTRING_IS_WXOBJECT
  61 #endif
  62
  63 // maximum possible length for a string means "take all string" everywhere
  64 //  (as sizeof(StringData) is unknown here we substract 100)
  65 const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100;
  66
  67 // 'naughty' cast
  68 #define   WXSTRINGCAST (wxChar *)(const wxChar *)
  69 #define   WXCSTRINGCAST (wxChar *)(const wxChar *)
  70 #define   MBSTRINGCAST (char *)(const char *)
  71 #define   WCSTRINGCAST (wchar_t *)(const wchar_t *)
  72
  73 // implementation only
  74 #define   ASSERT_VALID_INDEX(i) wxASSERT( (unsigned)(i) <= Len() )
  75
  76 // ---------------------------------------------------------------------------
  77 // Global functions complementing standard C string library replacements for
  78 // strlen() and portable strcasecmp()
  79 //---------------------------------------------------------------------------
  80 // USE wx* FUNCTIONS IN wx/wxchar.h INSTEAD - THIS IS ONLY FOR BINARY COMPATIBILITY
  81
  82 // checks whether the passed in pointer is NULL and if the string is empty
  83 inline bool WXDLLEXPORT IsEmpty(const char *p) { return (!p || !*p); }
  84
  85 // safe version of strlen() (returns 0 if passed NULL pointer)
  86 inline size_t WXDLLEXPORT Strlen(const char *psz)
  87   { return psz ? strlen(psz) : 0; }
  88
  89 // portable strcasecmp/_stricmp
  90 inline int WXDLLEXPORT Stricmp(const char *psz1, const char *psz2)
  91 {
  92 #if     defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
  93   return _stricmp(psz1, psz2);
  94 #elif     defined(__SC__)
  95   return _stricmp(psz1, psz2);
  96 #elif     defined(__SALFORDC__)
  97   return stricmp(psz1, psz2);
  98 #elif defined(__BORLANDC__)
  99   return stricmp(psz1, psz2);
 100 #elif defined(__WATCOMC__)
 101   return stricmp(psz1, psz2);
 102 #elif   defined(__UNIX__) || defined(__GNUWIN32__)
 103   return strcasecmp(psz1, psz2);
 104 #elif defined(__MWERKS__) && !defined(__INTEL__)
 105   register char c1, c2;
 106   do {
 107     c1 = tolower(*psz1++);
 108     c2 = tolower(*psz2++);
 109   } while ( c1 && (c1 == c2) );
 110
 111   return c1 - c2;
 112 #else
 113   // almost all compilers/libraries provide this function (unfortunately under
 114   // different names), that's why we don't implement our own which will surely
 115   // be more efficient than this code (uncomment to use):
 116   /*
 117     register char c1, c2;
 118     do {
 119       c1 = tolower(*psz1++);
 120       c2 = tolower(*psz2++);
 121     } while ( c1 && (c1 == c2) );
 122
 123     return c1 - c2;
 124   */
 125
 126   #error  "Please define string case-insensitive compare for your OS/compiler"
 127 #endif  // OS/compiler
 128 }
 129
 130 // ----------------------------------------------------------------------------
 131 // global data
 132 // ----------------------------------------------------------------------------
 133
 134 WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString;
 135
 136 // global pointer to empty string
 137 WXDLLEXPORT_DATA(extern const wxChar*) g_szNul;
 138
 139 // return an empty wxString
 140 class WXDLLEXPORT wxString; // not yet defined
 141 inline const wxString& wxGetEmptyString() { return *(wxString *)&g_szNul; }
 142
 143 // ---------------------------------------------------------------------------
 144 // string data prepended with some housekeeping info (used by wxString class),
 145 // is never used directly (but had to be put here to allow inlining)
 146 // ---------------------------------------------------------------------------
 147 struct WXDLLEXPORT wxStringData
 148 {
 149   int     nRefs;        // reference count
 150   size_t  nDataLength,  // actual string length
 151           nAllocLength; // allocated memory size
 152
 153   // mimics declaration 'wxChar data[nAllocLength]'
 154   wxChar* data() const { return (wxChar*)(this + 1); }
 155
 156   // empty string has a special ref count so it's never deleted
 157   bool  IsEmpty()   const { return (nRefs == -1); }
 158   bool  IsShared()  const { return (nRefs > 1);   }
 159
 160   // lock/unlock
 161   void  Lock()   { if ( !IsEmpty() ) nRefs++;                    }
 162   void  Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this);  }
 163
 164   // if we had taken control over string memory (GetWriteBuf), it's
 165   // intentionally put in invalid state
 166   void  Validate(bool b)  { nRefs = (b ? 1 : 0); }
 167   bool  IsValid() const   { return (nRefs != 0); }
 168 };
 169
 170 // ---------------------------------------------------------------------------
 171 // types of multibyte<->Unicode conversions
 172 // ---------------------------------------------------------------------------
 173 #if wxUSE_WCHAR_T
 174 class WXDLLEXPORT wxMBConv
 175 {
 176  public:
 177   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 178   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 179   // No longer inline since BC++ complains.
 180   const wxWCharBuffer cMB2WC(const char *psz) const;
 181   const wxCharBuffer cWC2MB(const wchar_t *psz) const;
 182 #if wxUSE_UNICODE
 183   const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); }
 184   const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); }
 185   const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; }
 186   const wchar_t* cMB2WC(const wchar_t *psz) const { return psz; }
 187 #else
 188   const char* cMB2WX(const char *psz) const { return psz; }
 189   const char* cWX2MB(const char *psz) const { return psz; }
 190   const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); }
 191   const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
 192 #endif
 193 };
 194 WXDLLEXPORT_DATA(extern wxMBConv) wxConvLibc;
 195 #define wxConv_libc wxConvLibc
 196
 197 #define wxANOTHER_MBCONV(type) \
 198 class type : public wxMBConv { \
 199  public: \
 200   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const; \
 201   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const; \
 202 }
 203
 204 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConvFile)) wxConvFile;
 205 #define wxConv_file wxConvFile
 206 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConvUTF7)) wxConvUTF7;
 207 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConvUTF8)) wxConvUTF8;
 208 #define wxConv_UTF8 wxConvUTF8
 209 #if defined(__WXGTK12__)
 210     WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConvGdk)) wxConvGdk;
 211     #define wxConv_gdk wxConvGdk
 212 #endif // GTK > 1.0
 213
 214 class wxCharacterSet;
 215 class WXDLLEXPORT wxCSConv : public wxMBConv
 216 {
 217  private:
 218   wxChar *m_name;
 219   wxCharacterSet *m_cset;
 220   bool m_deferred;
 221   void SetName(const wxChar *charset);
 222  public:
 223   wxCSConv(const wxChar *charset);
 224   virtual ~wxCSConv();
 225   void LoadNow();
 226   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 227   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 228 };
 229
 230 WXDLLEXPORT_DATA(extern wxCSConv) wxConvLocal;
 231 #define wxConv_local wxConvLocal
 232
 233 WXDLLEXPORT_DATA(extern wxMBConv *) wxConvCurrent;
 234 #define wxConv_current wxConvCurrent
 235
 236 // filenames are multibyte on Unix and probably widechar on Windows?
 237 #if defined(__UNIX__) || defined(__BORLANDC__)
 238 #define wxMBFILES 1
 239 #else
 240 #define wxMBFILES 0
 241 #endif
 242
 243 #if wxMBFILES
 244 #define wxFNCONV(name) wxConvFile.cWX2MB(name)
 245 #define FNSTRINGCAST MBSTRINGCAST
 246 #else
 247 #define wxFNCONV(name) name
 248 #define FNSTRINGCAST WXSTRINGCAST
 249 #endif
 250 #else//!wxUSE_WCHAR_T
 251 class WXDLLEXPORT wxMBConv {
 252 public:
 253   const char* cMB2WX(const char *psz) const { return psz; }
 254   const char* cWX2MB(const char *psz) const { return psz; }
 255 };
 256 WXDLLEXPORT_DATA(extern wxMBConv) wxConvLibc, wxConvFile;
 257 #define wxConv_libc wxConvLibc
 258 #define wxConv_file wxConvFile
 259 WXDLLEXPORT_DATA(extern wxMBConv *) wxConvCurrent;
 260 #define wxConv_current wxConvCurrent
 261 #define wxFNCONV(name) name
 262 #define FNSTRINGCAST WXSTRINGCAST
 263 #endif//wxUSE_WCHAR_T
 264
 265 // ---------------------------------------------------------------------------
 266 // This is (yet another one) String class for C++ programmers. It doesn't use
 267 // any of "advanced" C++ features (i.e. templates, exceptions, namespaces...)
 268 // thus you should be able to compile it with practicaly any C++ compiler.
 269 // This class uses copy-on-write technique, i.e. identical strings share the
 270 // same memory as long as neither of them is changed.
 271 //
 272 // This class aims to be as compatible as possible with the new standard
 273 // std::string class, but adds some additional functions and should be at
 274 // least as efficient than the standard implementation.
 275 //
 276 // Performance note: it's more efficient to write functions which take "const
 277 // String&" arguments than "const char *" if you assign the argument to
 278 // another string.
 279 //
 280 // It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5.
 281 //
 282 // To do:
 283 //  - ressource support (string tables in ressources)
 284 //  - more wide character (UNICODE) support
 285 //  - regular expressions support
 286 // ---------------------------------------------------------------------------
 287
 288 #ifdef  WXSTRING_IS_WXOBJECT
 289 class WXDLLEXPORT wxString : public wxObject
 290 {
 291     DECLARE_DYNAMIC_CLASS(wxString)
 292 #else   //WXSTRING_IS_WXOBJECT
 293 class WXDLLEXPORT wxString
 294 {
 295 #endif  //WXSTRING_IS_WXOBJECT
 296
 297 friend class WXDLLEXPORT wxArrayString;
 298
 299   // NB: special care was taken in arranging the member functions in such order
 300   //     that all inline functions can be effectively inlined, verify that all
 301   //     performace critical functions are still inlined if you change order!
 302 private:
 303   // points to data preceded by wxStringData structure with ref count info
 304   wxChar *m_pchData;
 305
 306   // accessor to string data
 307   wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; }
 308
 309   // string (re)initialization functions
 310     // initializes the string to the empty value (must be called only from
 311     // ctors, use Reinit() otherwise)
 312   void Init() { m_pchData = (wxChar *)g_szNul; }
 313     // initializaes the string with (a part of) C-string
 314   void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN);
 315     // as Init, but also frees old data
 316   void Reinit() { GetStringData()->Unlock(); Init(); }
 317
 318   // memory allocation
 319     // allocates memory for string of lenght nLen
 320   void AllocBuffer(size_t nLen);
 321     // copies data to another string
 322   void AllocCopy(wxString&, int, int) const;
 323     // effectively copies data to string
 324   void AssignCopy(size_t, const wxChar *);
 325
 326   // append a (sub)string
 327   void ConcatSelf(int nLen, const wxChar *src);
 328
 329   // functions called before writing to the string: they copy it if there
 330   // are other references to our data (should be the only owner when writing)
 331   void CopyBeforeWrite();
 332   void AllocBeforeWrite(size_t);
 333
 334   // this method is not implemented - there is _no_ conversion from int to
 335   // string, you're doing something wrong if the compiler wants to call it!
 336   //
 337   // try `s << i' or `s.Printf("%d", i)' instead
 338   wxString(int);
 339   wxString(long);
 340
 341 public:
 342   // constructors and destructor
 343     // ctor for an empty string
 344   wxString() { Init(); }
 345     // copy ctor
 346   wxString(const wxString& stringSrc)
 347   {
 348     wxASSERT( stringSrc.GetStringData()->IsValid() );
 349
 350     if ( stringSrc.IsEmpty() ) {
 351       // nothing to do for an empty string
 352       Init();
 353     }
 354     else {
 355       m_pchData = stringSrc.m_pchData;            // share same data
 356       GetStringData()->Lock();                    // => one more copy
 357     }
 358   }
 359     // string containing nRepeat copies of ch
 360   wxString(wxChar ch, size_t nRepeat = 1);
 361     // ctor takes first nLength characters from C string
 362     // (default value of wxSTRING_MAXLEN means take all the string)
 363   wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN)
 364     { InitWith(psz, 0, nLength); }
 365 #if wxUSE_UNICODE
 366     // from multibyte string
 367     // (NB: nLength is right now number of Unicode characters, not
 368     //  characters in psz! So try not to use it yet!)
 369   wxString(const char *psz, wxMBConv& conv = wxConvLibc, size_t nLength = wxSTRING_MAXLEN);
 370     // from wxWCharBuffer (i.e. return from wxGetString)
 371   wxString(const wxWCharBuffer& psz)
 372     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 373 #else
 374     // from C string (for compilers using unsigned char)
 375   wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN)
 376     { InitWith((const char*)psz, 0, nLength); }
 377     // from multibyte string
 378   wxString(const char *psz, wxMBConv& WXUNUSED(conv), size_t nLength = wxSTRING_MAXLEN)
 379     { InitWith(psz, 0, nLength); }
 380 #if wxUSE_WCHAR_T
 381     // from wide (Unicode) string
 382   wxString(const wchar_t *pwz);
 383 #endif
 384     // from wxCharBuffer
 385   wxString(const wxCharBuffer& psz)
 386     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 387 #endif
 388     // dtor is not virtual, this class must not be inherited from!
 389  ~wxString() { GetStringData()->Unlock(); }
 390
 391   // generic attributes & operations
 392     // as standard strlen()
 393   size_t Len() const { return GetStringData()->nDataLength; }
 394     // string contains any characters?
 395   bool IsEmpty() const { return Len() == 0; }
 396     // empty string is "FALSE", so !str will return TRUE
 397   bool operator!() const { return IsEmpty(); }
 398     // empty string contents
 399   void Empty()
 400   {
 401     if ( !IsEmpty() )
 402       Reinit();
 403
 404     // should be empty
 405     wxASSERT( GetStringData()->nDataLength == 0 );
 406   }
 407     // empty the string and free memory
 408   void Clear()
 409   {
 410     if ( !GetStringData()->IsEmpty() )
 411       Reinit();
 412
 413     wxASSERT( GetStringData()->nDataLength == 0 );  // should be empty
 414     wxASSERT( GetStringData()->nAllocLength == 0 ); // and not own any memory
 415   }
 416
 417   // contents test
 418     // Is an ascii value
 419   bool IsAscii() const;
 420     // Is a number
 421   bool IsNumber() const;
 422     // Is a word
 423   bool IsWord() const;
 424
 425   // data access (all indexes are 0 based)
 426     // read access
 427     wxChar  GetChar(size_t n) const
 428       { ASSERT_VALID_INDEX( n );  return m_pchData[n]; }
 429     // read/write access
 430     wxChar& GetWritableChar(size_t n)
 431       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 432     // write access
 433     void  SetChar(size_t n, wxChar ch)
 434       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; }
 435
 436     // get last character
 437     wxChar  Last() const
 438       { wxASSERT( !IsEmpty() ); return m_pchData[Len() - 1]; }
 439     // get writable last character
 440     wxChar& Last()
 441       { wxASSERT( !IsEmpty() ); CopyBeforeWrite(); return m_pchData[Len()-1]; }
 442
 443     // under Unix it is tested with configure, assume it works on other
 444     // platforms (there might be overloading problems if size_t and int are
 445     // the same type)
 446 #if !defined(__UNIX__) || wxUSE_SIZE_T_STRING_OPERATOR
 447     // operator version of GetChar
 448     wxChar  operator[](size_t n) const
 449       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 450 #endif
 451
 452     // operator version of GetChar
 453     wxChar  operator[](int n) const
 454       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 455     // operator version of GetWritableChar
 456     wxChar& operator[](size_t n)
 457       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 458
 459     // implicit conversion to C string
 460     operator const wxChar*() const { return m_pchData; }
 461     // explicit conversion to C string (use this with printf()!)
 462     const wxChar* c_str()   const { return m_pchData; }
 463     // (and this with [wx]Printf()!)
 464     const wxChar* wx_str()  const { return m_pchData; }
 465     //
 466     const wxChar* GetData() const { return m_pchData; }
 467 #if wxUSE_UNICODE
 468     const wxCharBuffer mb_str(wxMBConv& conv = wxConvLibc) const { return conv.cWC2MB(m_pchData); }
 469     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const { return m_pchData; }
 470 #if wxMBFILES
 471     const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
 472 #else
 473     const wxChar* fn_str() const { return m_pchData; }
 474 #endif
 475 #else
 476     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConvLibc ) const { return m_pchData; }
 477 #if wxUSE_WCHAR_T
 478     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 479 #endif
 480     const wxChar* fn_str() const { return m_pchData; }
 481 #endif
 482     // for convenience
 483     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 484
 485   // overloaded assignment
 486     // from another wxString
 487   wxString& operator=(const wxString& stringSrc);
 488     // from a character
 489   wxString& operator=(wxChar ch);
 490     // from a C string
 491   wxString& operator=(const wxChar *psz);
 492 #if wxUSE_UNICODE
 493     // from wxWCharBuffer
 494   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 495 #else
 496     // from another kind of C string
 497   wxString& operator=(const unsigned char* psz);
 498 #if wxUSE_WCHAR_T
 499     // from a wide string
 500   wxString& operator=(const wchar_t *pwz);
 501 #endif
 502     // from wxCharBuffer
 503   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 504 #endif
 505
 506   // string concatenation
 507     // in place concatenation
 508     /*
 509         Concatenate and return the result. Note that the left to right
 510         associativity of << allows to write things like "str << str1 << str2
 511         << ..." (unlike with +=)
 512      */
 513       // string += string
 514   wxString& operator<<(const wxString& s)
 515   {
 516     wxASSERT( s.GetStringData()->IsValid() );
 517
 518     ConcatSelf(s.Len(), s);
 519     return *this;
 520   }
 521       // string += C string
 522   wxString& operator<<(const wxChar *psz)
 523     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 524       // string += char
 525   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 526
 527       // string += string
 528   void operator+=(const wxString& s) { (void)operator<<(s); }
 529       // string += C string
 530   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 531       // string += char
 532   void operator+=(wxChar ch) { (void)operator<<(ch); }
 533
 534       // string += buffer (i.e. from wxGetString)
 535 #if wxUSE_UNICODE
 536   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 537   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 538 #else
 539   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 540   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 541 #endif
 542
 543     // string += C string
 544   wxString& Append(const wxChar* psz)
 545     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 546     // append count copies of given character
 547   wxString& Append(wxChar ch, size_t count = 1u)
 548     { wxString str(ch, count); return *this << str; }
 549
 550     // prepend a string, return the string itself
 551   wxString& Prepend(const wxString& str)
 552     { *this = str + *this; return *this; }
 553
 554     // non-destructive concatenation
 555       //
 556   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 557       //
 558   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 559       //
 560   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 561       //
 562   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 563       //
 564   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 565
 566   // stream-like functions
 567       // insert an int into string
 568   wxString& operator<<(int i);
 569       // insert a float into string
 570   wxString& operator<<(float f);
 571       // insert a double into string
 572   wxString& operator<<(double d);
 573
 574   // string comparison
 575     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 576   int  Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 577     // same as Cmp() but not case-sensitive
 578   int  CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 579     // test for the string equality, either considering case or not
 580     // (if compareWithCase then the case matters)
 581   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 582     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 583
 584   // simple sub-string extraction
 585       // return substring starting at nFirst of length nCount (or till the end
 586       // if nCount = default value)
 587   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 588
 589     // operator version of Mid()
 590   wxString  operator()(size_t start, size_t len) const
 591     { return Mid(start, len); }
 592
 593       // get first nCount characters
 594   wxString Left(size_t nCount) const;
 595       // get last nCount characters
 596   wxString Right(size_t nCount) const;
 597       // get all characters before the first occurence of ch
 598       // (returns the whole string if ch not found)
 599   wxString BeforeFirst(wxChar ch) const;
 600       // get all characters before the last occurence of ch
 601       // (returns empty string if ch not found)
 602   wxString BeforeLast(wxChar ch) const;
 603       // get all characters after the first occurence of ch
 604       // (returns empty string if ch not found)
 605   wxString AfterFirst(wxChar ch) const;
 606       // get all characters after the last occurence of ch
 607       // (returns the whole string if ch not found)
 608   wxString AfterLast(wxChar ch) const;
 609
 610     // for compatibility only, use more explicitly named functions above
 611   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 612   wxString After(wxChar ch) const { return AfterFirst(ch); }
 613
 614   // case conversion
 615       // convert to upper case in place, return the string itself
 616   wxString& MakeUpper();
 617       // convert to upper case, return the copy of the string
 618       // Here's something to remember: BC++ doesn't like returns in inlines.
 619   wxString Upper() const ;
 620       // convert to lower case in place, return the string itself
 621   wxString& MakeLower();
 622       // convert to lower case, return the copy of the string
 623   wxString Lower() const ;
 624
 625   // trimming/padding whitespace (either side) and truncating
 626       // remove spaces from left or from right (default) side
 627   wxString& Trim(bool bFromRight = TRUE);
 628       // add nCount copies chPad in the beginning or at the end (default)
 629   wxString& Pad(size_t nCount, wxChar chPad = _T(' '), bool bFromRight = TRUE);
 630       // truncate string to given length
 631   wxString& Truncate(size_t uiLen);
 632
 633   // searching and replacing
 634       // searching (return starting index, or -1 if not found)
 635   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 636       // searching (return starting index, or -1 if not found)
 637   int Find(const wxChar *pszSub) const;               // like strstr
 638       // replace first (or all of bReplaceAll) occurences of substring with
 639       // another string, returns the number of replacements made
 640   size_t Replace(const wxChar *szOld,
 641                  const wxChar *szNew,
 642                  bool bReplaceAll = TRUE);
 643
 644     // check if the string contents matches a mask containing '*' and '?'
 645   bool Matches(const wxChar *szMask) const;
 646
 647   // formated input/output
 648     // as sprintf(), returns the number of characters written or < 0 on error
 649   int Printf(const wxChar *pszFormat, ...);
 650     // as vprintf(), returns the number of characters written or < 0 on error
 651   int PrintfV(const wxChar* pszFormat, va_list argptr);
 652
 653   // raw access to string memory
 654     // ensure that string has space for at least nLen characters
 655     // only works if the data of this string is not shared
 656   void Alloc(size_t nLen);
 657     // minimize the string's memory
 658     // only works if the data of this string is not shared
 659   void Shrink();
 660     // get writable buffer of at least nLen bytes. Unget() *must* be called
 661     // a.s.a.p. to put string back in a reasonable state!
 662   wxChar *GetWriteBuf(size_t nLen);
 663     // call this immediately after GetWriteBuf() has been used
 664   void UngetWriteBuf();
 665
 666   // wxWindows version 1 compatibility functions
 667
 668   // use Mid()
 669   wxString SubString(size_t from, size_t to) const
 670       { return Mid(from, (to - from + 1)); }
 671     // values for second parameter of CompareTo function
 672   enum caseCompare {exact, ignoreCase};
 673     // values for first parameter of Strip function
 674   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 675
 676     // use Printf()
 677   int sprintf(const wxChar *pszFormat, ...);
 678
 679     // use Cmp()
 680   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 681     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 682
 683     // use Len
 684   size_t Length() const { return Len(); }
 685     // Count the number of characters
 686   int Freq(wxChar ch) const;
 687     // use MakeLower
 688   void LowerCase() { MakeLower(); }
 689     // use MakeUpper
 690   void UpperCase() { MakeUpper(); }
 691     // use Trim except that it doesn't change this string
 692   wxString Strip(stripType w = trailing) const;
 693
 694     // use Find (more general variants not yet supported)
 695   size_t Index(const wxChar* psz) const { return Find(psz); }
 696   size_t Index(wxChar ch)         const { return Find(ch);  }
 697     // use Truncate
 698   wxString& Remove(size_t pos) { return Truncate(pos); }
 699   wxString& RemoveLast() { return Truncate(Len() - 1); }
 700
 701   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 702
 703     // use Find()
 704   int First( const wxChar ch ) const { return Find(ch); }
 705   int First( const wxChar* psz ) const { return Find(psz); }
 706   int First( const wxString &str ) const { return Find(str); }
 707   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 708   bool Contains(const wxString& str) const { return Find(str) != -1; }
 709
 710     // use IsEmpty()
 711   bool IsNull() const { return IsEmpty(); }
 712
 713 #ifdef  wxSTD_STRING_COMPATIBILITY
 714   // std::string compatibility functions
 715
 716   // an 'invalid' value for string index
 717   static const size_t npos;
 718
 719   // constructors
 720     // take nLen chars starting at nPos
 721   wxString(const wxString& str, size_t nPos, size_t nLen)
 722   {
 723     wxASSERT( str.GetStringData()->IsValid() );
 724     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 725   }
 726     // take all characters from pStart to pEnd
 727   wxString(const void *pStart, const void *pEnd);
 728
 729   // lib.string.capacity
 730     // return the length of the string
 731   size_t size() const { return Len(); }
 732     // return the length of the string
 733   size_t length() const { return Len(); }
 734     // return the maximum size of the string
 735   size_t max_size() const { return wxSTRING_MAXLEN; }
 736     // resize the string, filling the space with c if c != 0
 737   void resize(size_t nSize, wxChar ch = _T('\0'));
 738     // delete the contents of the string
 739   void clear() { Empty(); }
 740     // returns true if the string is empty
 741   bool empty() const { return IsEmpty(); }
 742
 743   // lib.string.access
 744     // return the character at position n
 745   wxChar at(size_t n) const { return GetChar(n); }
 746     // returns the writable character at position n
 747   wxChar& at(size_t n) { return GetWritableChar(n); }
 748
 749   // lib.string.modifiers
 750     // append a string
 751   wxString& append(const wxString& str)
 752     { *this += str; return *this; }
 753     // append elements str[pos], ..., str[pos+n]
 754   wxString& append(const wxString& str, size_t pos, size_t n)
 755     { ConcatSelf(n, str.c_str() + pos); return *this; }
 756     // append first n (or all if n == npos) characters of sz
 757   wxString& append(const wxChar *sz, size_t n = npos)
 758     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 759
 760     // append n copies of ch
 761   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 762
 763     // same as `this_string = str'
 764   wxString& assign(const wxString& str) { return (*this) = str; }
 765     // same as ` = str[pos..pos + n]
 766   wxString& assign(const wxString& str, size_t pos, size_t n)
 767     { return *this = wxString((const wxChar *)str + pos, n); }
 768     // same as `= first n (or all if n == npos) characters of sz'
 769   wxString& assign(const wxChar *sz, size_t n = npos)
 770     { return *this = wxString(sz, n); }
 771     // same as `= n copies of ch'
 772   wxString& assign(size_t n, wxChar ch)
 773     { return *this = wxString(ch, n); }
 774
 775     // insert another string
 776   wxString& insert(size_t nPos, const wxString& str);
 777     // insert n chars of str starting at nStart (in str)
 778   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 779     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 780
 781     // insert first n (or all if n == npos) characters of sz
 782   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 783     { return insert(nPos, wxString(sz, n)); }
 784     // insert n copies of ch
 785   wxString& insert(size_t nPos, size_t n, wxChar ch)
 786     { return insert(nPos, wxString(ch, n)); }
 787
 788     // delete characters from nStart to nStart + nLen
 789   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 790
 791     // replaces the substring of length nLen starting at nStart
 792   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 793     // replaces the substring with nCount copies of ch
 794   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 795     // replaces a substring with another substring
 796   wxString& replace(size_t nStart, size_t nLen,
 797                     const wxString& str, size_t nStart2, size_t nLen2);
 798     // replaces the substring with first nCount chars of sz
 799     wxString& replace(size_t nStart, size_t nLen,
 800                       const wxChar* sz, size_t nCount);
 801
 802     // swap two strings
 803   void swap(wxString& str);
 804
 805     // All find() functions take the nStart argument which specifies the
 806     // position to start the search on, the default value is 0. All functions
 807     // return npos if there were no match.
 808
 809     // find a substring
 810   size_t find(const wxString& str, size_t nStart = 0) const;
 811
 812   // VC++ 1.5 can't cope with this syntax.
 813 #if !defined(__VISUALC__) || defined(__WIN32__)
 814     // find first n characters of sz
 815   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 816 #endif
 817
 818   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 819 #if !defined(__BORLANDC__)
 820     // find the first occurence of character ch after nStart
 821   size_t find(wxChar ch, size_t nStart = 0) const;
 822 #endif
 823     // rfind() family is exactly like find() but works right to left
 824
 825     // as find, but from the end
 826   size_t rfind(const wxString& str, size_t nStart = npos) const;
 827
 828   // VC++ 1.5 can't cope with this syntax.
 829 #if !defined(__VISUALC__) || defined(__WIN32__)
 830     // as find, but from the end
 831   size_t rfind(const wxChar* sz, size_t nStart = npos,
 832           size_t n = npos) const;
 833     // as find, but from the end
 834   size_t rfind(wxChar ch, size_t nStart = npos) const;
 835 #endif
 836
 837     // find first/last occurence of any character in the set
 838
 839     //
 840   size_t find_first_of(const wxString& str, size_t nStart = 0) const;
 841     //
 842   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 843     // same as find(char, size_t)
 844   size_t find_first_of(wxChar c, size_t nStart = 0) const;
 845     //
 846   size_t find_last_of (const wxString& str, size_t nStart = npos) const;
 847     //
 848   size_t find_last_of (const wxChar* s, size_t nStart = npos) const;
 849     // same as rfind(char, size_t)
 850   size_t find_last_of (wxChar c, size_t nStart = npos) const;
 851
 852     // find first/last occurence of any character not in the set
 853
 854     //
 855   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const;
 856     //
 857   size_t find_first_not_of(const wxChar* s, size_t nStart = 0) const;
 858     //
 859   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 860     //
 861   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 862     //
 863   size_t find_last_not_of(const wxChar* s, size_t nStart = npos) const;
 864     //
 865   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 866
 867     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 868     // equal or greater than the compare() argument.
 869
 870     // just like strcmp()
 871   int compare(const wxString& str) const { return Cmp(str); }
 872     // comparison with a substring
 873   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 874     // comparison of 2 substrings
 875   int compare(size_t nStart, size_t nLen,
 876               const wxString& str, size_t nStart2, size_t nLen2) const;
 877     // just like strcmp()
 878   int compare(const wxChar* sz) const { return Cmp(sz); }
 879     // substring comparison with first nCount characters of sz
 880   int compare(size_t nStart, size_t nLen,
 881               const wxChar* sz, size_t nCount = npos) const;
 882
 883   // substring extraction
 884   wxString substr(size_t nStart = 0, size_t nLen = npos) const;
 885 #endif // wxSTD_STRING_COMPATIBILITY
 886 };
 887
 888 // ----------------------------------------------------------------------------
 889 // The string array uses it's knowledge of internal structure of the wxString
 890 // class to optimize string storage. Normally, we would store pointers to
 891 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 892 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 893 // really all we need to turn such pointer into a string!
 894 //
 895 // Of course, it can be called a dirty hack, but we use twice less memory and
 896 // this approach is also more speed efficient, so it's probably worth it.
 897 //
 898 // Usage notes: when a string is added/inserted, a new copy of it is created,
 899 // so the original string may be safely deleted. When a string is retrieved
 900 // from the array (operator[] or Item() method), a reference is returned.
 901 // ----------------------------------------------------------------------------
 902 class WXDLLEXPORT wxArrayString
 903 {
 904 public:
 905   // type of function used by wxArrayString::Sort()
 906   typedef int (*CompareFunction)(const wxString& first,
 907                                  const wxString& second);
 908
 909   // constructors and destructor
 910     // default ctor
 911   wxArrayString();
 912     // copy ctor
 913   wxArrayString(const wxArrayString& array);
 914     // assignment operator
 915   wxArrayString& operator=(const wxArrayString& src);
 916     // not virtual, this class should not be derived from
 917  ~wxArrayString();
 918
 919   // memory management
 920     // empties the list, but doesn't release memory
 921   void Empty();
 922     // empties the list and releases memory
 923   void Clear();
 924     // preallocates memory for given number of items
 925   void Alloc(size_t nCount);
 926     // minimzes the memory usage (by freeing all extra memory)
 927   void Shrink();
 928
 929   // simple accessors
 930     // number of elements in the array
 931   size_t GetCount() const { return m_nCount; }
 932     // is it empty?
 933   bool IsEmpty() const { return m_nCount == 0; }
 934     // number of elements in the array (GetCount is preferred API)
 935   size_t Count() const { return m_nCount; }
 936
 937   // items access (range checking is done in debug version)
 938     // get item at position uiIndex
 939   wxString& Item(size_t nIndex) const
 940     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 941     // same as Item()
 942   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 943     // get last item
 944   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 945
 946   // item management
 947     // Search the element in the array, starting from the beginning if
 948     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 949     // sensitive (default). Returns index of the first item matched or
 950     // wxNOT_FOUND
 951   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 952     // add new element at the end
 953   void Add(const wxString& str);
 954     // add new element at given position
 955   void Insert(const wxString& str, size_t uiIndex);
 956     // remove first item matching this value
 957   void Remove(const wxChar *sz);
 958     // remove item by index
 959   void Remove(size_t nIndex);
 960
 961   // sorting
 962     // sort array elements in alphabetical order (or reversed alphabetical
 963     // order if reverseOrder parameter is TRUE)
 964   void Sort(bool reverseOrder = FALSE);
 965     // sort array elements using specified comparaison function
 966   void Sort(CompareFunction compareFunction);
 967
 968 private:
 969   void    Grow();     // makes array bigger if needed
 970   void    Free();     // free the string stored
 971
 972   void    DoSort();   // common part of all Sort() variants
 973
 974   size_t  m_nSize,    // current size of the array
 975           m_nCount;   // current number of elements
 976
 977   wxChar  **m_pItems;   // pointer to data
 978 };
 979
 980 // ---------------------------------------------------------------------------
 981 // wxString comparison functions: operator versions are always case sensitive
 982 // ---------------------------------------------------------------------------
 983 //
 984 inline bool operator==(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) == 0); }
 985 //
 986 inline bool operator==(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) == 0); }
 987 //
 988 inline bool operator==(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) == 0); }
 989 //
 990 inline bool operator!=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) != 0); }
 991 //
 992 inline bool operator!=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) != 0); }
 993 //
 994 inline bool operator!=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) != 0); }
 995 //
 996 inline bool operator< (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) < 0); }
 997 //
 998 inline bool operator< (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <  0); }
 999 //
1000 inline bool operator< (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >  0); }
1001 //
1002 inline bool operator> (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >  0); }
1003 //
1004 inline bool operator> (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >  0); }
1005 //
1006 inline bool operator> (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <  0); }
1007 //
1008 inline bool operator<=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) <= 0); }
1009 //
1010 inline bool operator<=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <= 0); }
1011 //
1012 inline bool operator<=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >= 0); }
1013 //
1014 inline bool operator>=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >= 0); }
1015 //
1016 inline bool operator>=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >= 0); }
1017 //
1018 inline bool operator>=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <= 0); }
1019
1020 wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
1021 wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
1022 wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
1023 wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
1024 wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
1025 #if wxUSE_UNICODE
1026 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxWCharBuffer& buf)
1027 { return string + (const wchar_t *)buf; }
1028 inline wxString WXDLLEXPORT operator+(const wxWCharBuffer& buf, const wxString& string)
1029 { return (const wchar_t *)buf + string; }
1030 #else
1031 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxCharBuffer& buf)
1032 { return string + (const char *)buf; }
1033 inline wxString WXDLLEXPORT operator+(const wxCharBuffer& buf, const wxString& string)
1034 { return (const char *)buf + string; }
1035 #endif
1036
1037 // ---------------------------------------------------------------------------
1038 // Implementation only from here until the end of file
1039 // ---------------------------------------------------------------------------
1040
1041 #ifdef wxSTD_STRING_COMPATIBILITY
1042
1043 #include "wx/ioswrap.h"
1044
1045 WXDLLEXPORT istream& operator>>(istream& is, wxString& str);
1046
1047 #endif  // wxSTD_STRING_COMPATIBILITY
1048
1049 #endif  // _WX_WXSTRINGH__