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   MBSTRINGCAST (char *)(const char *)
  70 #define   WCSTRINGCAST (wchar_t *)(const wchar_t *)
  71
  72 // implementation only
  73 #define   ASSERT_VALID_INDEX(i) wxASSERT( (unsigned)(i) <= Len() )
  74
  75 // ---------------------------------------------------------------------------
  76 // Global functions complementing standard C string library replacements for
  77 // strlen() and portable strcasecmp()
  78 //---------------------------------------------------------------------------
  79 // USE wx* FUNCTIONS IN wx/wxchar.h INSTEAD - THIS IS ONLY FOR BINARY COMPATIBILITY
  80
  81 // checks whether the passed in pointer is NULL and if the string is empty
  82 inline bool WXDLLEXPORT IsEmpty(const char *p) { return (!p || !*p); }
  83
  84 // safe version of strlen() (returns 0 if passed NULL pointer)
  85 inline size_t WXDLLEXPORT Strlen(const char *psz)
  86   { return psz ? strlen(psz) : 0; }
  87
  88 // portable strcasecmp/_stricmp
  89 inline int WXDLLEXPORT Stricmp(const char *psz1, const char *psz2)
  90 {
  91 #if     defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
  92   return _stricmp(psz1, psz2);
  93 #elif     defined(__SC__)
  94   return _stricmp(psz1, psz2);
  95 #elif     defined(__SALFORDC__)
  96   return stricmp(psz1, psz2);
  97 #elif defined(__BORLANDC__)
  98   return stricmp(psz1, psz2);
  99 #elif defined(__WATCOMC__)
 100   return stricmp(psz1, psz2);
 101 #elif   defined(__UNIX__) || defined(__GNUWIN32__)
 102   return strcasecmp(psz1, psz2);
 103 #elif defined(__MWERKS__) && !defined(__INTEL__)
 104   register char c1, c2;
 105   do {
 106     c1 = tolower(*psz1++);
 107     c2 = tolower(*psz2++);
 108   } while ( c1 && (c1 == c2) );
 109
 110   return c1 - c2;
 111 #else
 112   // almost all compilers/libraries provide this function (unfortunately under
 113   // different names), that's why we don't implement our own which will surely
 114   // be more efficient than this code (uncomment to use):
 115   /*
 116     register char c1, c2;
 117     do {
 118       c1 = tolower(*psz1++);
 119       c2 = tolower(*psz2++);
 120     } while ( c1 && (c1 == c2) );
 121
 122     return c1 - c2;
 123   */
 124
 125   #error  "Please define string case-insensitive compare for your OS/compiler"
 126 #endif  // OS/compiler
 127 }
 128
 129 // ----------------------------------------------------------------------------
 130 // global data
 131 // ----------------------------------------------------------------------------
 132
 133 WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString;
 134
 135 // global pointer to empty string
 136 WXDLLEXPORT_DATA(extern const wxChar*) g_szNul;
 137
 138 // return an empty wxString
 139 class WXDLLEXPORT wxString; // not yet defined
 140 inline const wxString& wxGetEmptyString() { return *(wxString *)&g_szNul; }
 141
 142 // ---------------------------------------------------------------------------
 143 // string data prepended with some housekeeping info (used by wxString class),
 144 // is never used directly (but had to be put here to allow inlining)
 145 // ---------------------------------------------------------------------------
 146 struct WXDLLEXPORT wxStringData
 147 {
 148   int     nRefs;        // reference count
 149   size_t  nDataLength,  // actual string length
 150           nAllocLength; // allocated memory size
 151
 152   // mimics declaration 'wxChar data[nAllocLength]'
 153   wxChar* data() const { return (wxChar*)(this + 1); }
 154
 155   // empty string has a special ref count so it's never deleted
 156   bool  IsEmpty()   const { return (nRefs == -1); }
 157   bool  IsShared()  const { return (nRefs > 1);   }
 158
 159   // lock/unlock
 160   void  Lock()   { if ( !IsEmpty() ) nRefs++;                    }
 161   void  Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this);  }
 162
 163   // if we had taken control over string memory (GetWriteBuf), it's
 164   // intentionally put in invalid state
 165   void  Validate(bool b)  { nRefs = (b ? 1 : 0); }
 166   bool  IsValid() const   { return (nRefs != 0); }
 167 };
 168
 169 // ---------------------------------------------------------------------------
 170 // types of multibyte<->Unicode conversions
 171 // ---------------------------------------------------------------------------
 172 class WXDLLEXPORT wxMBConv
 173 {
 174  public:
 175   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 176   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 177   const wxWCharBuffer cMB2WC(const char *psz) const
 178     {
 179       size_t nLen = MB2WC((wchar_t *) NULL, psz, 0);
 180       wxWCharBuffer buf(nLen);
 181       MB2WC(WCSTRINGCAST buf, psz, nLen);
 182       return buf;
 183     }
 184   const wxCharBuffer cWC2MB(const wchar_t *psz) const
 185     {
 186       size_t nLen = WC2MB((char *) NULL, psz, 0);
 187       wxCharBuffer buf(nLen);
 188       WC2MB(MBSTRINGCAST buf, psz, nLen);
 189       return buf;
 190     }
 191 #if wxUSE_UNICODE
 192   const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); }
 193   const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); }
 194   const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; }
 195   const wchar_t* cMB2WC(const wchar_t *psz) const { return psz; }
 196 #else
 197   const char* cMB2WX(const char *psz) const { return psz; }
 198   const char* cWX2MB(const char *psz) const { return psz; }
 199   const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); }
 200   const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
 201 #endif
 202 };
 203 WXDLLEXPORT_DATA(extern wxMBConv) wxConv_libc;
 204
 205 #define wxANOTHER_MBCONV(type) \
 206 class type : public wxMBConv { \
 207  public: \
 208   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const; \
 209   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const; \
 210 }
 211
 212 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_file)) wxConv_file;
 213 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_UTF7)) wxConv_UTF7;
 214 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_UTF8)) wxConv_UTF8;
 215 #if defined(__WXGTK__) && (GTK_MINOR_VERSION > 0)
 216     WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_gdk)) wxConv_gdk;
 217 #endif // GTK > 1.0
 218
 219 class wxCharacterSet;
 220 class WXDLLEXPORT wxCSConv : public wxMBConv
 221 {
 222  private:
 223   wxChar *m_name;
 224   wxCharacterSet *m_cset;
 225   bool m_deferred;
 226   void SetName(const wxChar *charset);
 227  public:
 228   wxCSConv(const wxChar *charset);
 229   virtual ~wxCSConv();
 230   void LoadNow();
 231   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 232   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 233 };
 234
 235 WXDLLEXPORT_DATA(extern wxCSConv) wxConv_local;
 236
 237 WXDLLEXPORT_DATA(extern wxMBConv *) wxConv_current;
 238
 239 // filenames are multibyte on Unix and probably widechar on Windows?
 240 #ifdef __UNIX__
 241 #define wxMBFILES 1
 242 #else
 243 #define wxMBFILES 0
 244 #endif
 245
 246 #if wxMBFILES
 247 #define wxFNCONV(name) wxConv_file.cWX2MB(name)
 248 #define FNSTRINGCAST MBSTRINGCAST
 249 #else
 250 #define wxFNCONV(name) name
 251 #define FNSTRINGCAST WXSTRINGCAST
 252 #endif
 253
 254 // ---------------------------------------------------------------------------
 255 // This is (yet another one) String class for C++ programmers. It doesn't use
 256 // any of "advanced" C++ features (i.e. templates, exceptions, namespaces...)
 257 // thus you should be able to compile it with practicaly any C++ compiler.
 258 // This class uses copy-on-write technique, i.e. identical strings share the
 259 // same memory as long as neither of them is changed.
 260 //
 261 // This class aims to be as compatible as possible with the new standard
 262 // std::string class, but adds some additional functions and should be at
 263 // least as efficient than the standard implementation.
 264 //
 265 // Performance note: it's more efficient to write functions which take "const
 266 // String&" arguments than "const char *" if you assign the argument to
 267 // another string.
 268 //
 269 // It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5.
 270 //
 271 // To do:
 272 //  - ressource support (string tables in ressources)
 273 //  - more wide character (UNICODE) support
 274 //  - regular expressions support
 275 // ---------------------------------------------------------------------------
 276
 277 #ifdef  WXSTRING_IS_WXOBJECT
 278 class WXDLLEXPORT wxString : public wxObject
 279 {
 280     DECLARE_DYNAMIC_CLASS(wxString)
 281 #else   //WXSTRING_IS_WXOBJECT
 282 class WXDLLEXPORT wxString
 283 {
 284 #endif  //WXSTRING_IS_WXOBJECT
 285
 286 friend class WXDLLEXPORT wxArrayString;
 287
 288   // NB: special care was taken in arranging the member functions in such order
 289   //     that all inline functions can be effectively inlined, verify that all
 290   //     performace critical functions are still inlined if you change order!
 291 private:
 292   // points to data preceded by wxStringData structure with ref count info
 293   wxChar *m_pchData;
 294
 295   // accessor to string data
 296   wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; }
 297
 298   // string (re)initialization functions
 299     // initializes the string to the empty value (must be called only from
 300     // ctors, use Reinit() otherwise)
 301   void Init() { m_pchData = (wxChar *)g_szNul; }
 302     // initializaes the string with (a part of) C-string
 303   void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN);
 304     // as Init, but also frees old data
 305   void Reinit() { GetStringData()->Unlock(); Init(); }
 306
 307   // memory allocation
 308     // allocates memory for string of lenght nLen
 309   void AllocBuffer(size_t nLen);
 310     // copies data to another string
 311   void AllocCopy(wxString&, int, int) const;
 312     // effectively copies data to string
 313   void AssignCopy(size_t, const wxChar *);
 314
 315   // append a (sub)string
 316   void ConcatSelf(int nLen, const wxChar *src);
 317
 318   // functions called before writing to the string: they copy it if there
 319   // are other references to our data (should be the only owner when writing)
 320   void CopyBeforeWrite();
 321   void AllocBeforeWrite(size_t);
 322
 323   // this method is not implemented - there is _no_ conversion from int to
 324   // string, you're doing something wrong if the compiler wants to call it!
 325   //
 326   // try `s << i' or `s.Printf("%d", i)' instead
 327   wxString(int);
 328   wxString(long);
 329
 330 public:
 331   // constructors and destructor
 332     // ctor for an empty string
 333   wxString() { Init(); }
 334     // copy ctor
 335   wxString(const wxString& stringSrc)
 336   {
 337     wxASSERT( stringSrc.GetStringData()->IsValid() );
 338
 339     if ( stringSrc.IsEmpty() ) {
 340       // nothing to do for an empty string
 341       Init();
 342     }
 343     else {
 344       m_pchData = stringSrc.m_pchData;            // share same data
 345       GetStringData()->Lock();                    // => one more copy
 346     }
 347   }
 348     // string containing nRepeat copies of ch
 349   wxString(wxChar ch, size_t nRepeat = 1);
 350     // ctor takes first nLength characters from C string
 351     // (default value of wxSTRING_MAXLEN means take all the string)
 352   wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN)
 353     { InitWith(psz, 0, nLength); }
 354 #if wxUSE_UNICODE
 355     // from multibyte string
 356     // (NB: nLength is right now number of Unicode characters, not
 357     //  characters in psz! So try not to use it yet!)
 358   wxString(const char *psz, wxMBConv& conv = wxConv_libc, size_t nLength = wxSTRING_MAXLEN);
 359     // from wxWCharBuffer (i.e. return from wxGetString)
 360   wxString(const wxWCharBuffer& psz)
 361     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 362 #else
 363     // from C string (for compilers using unsigned char)
 364   wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN)
 365     { InitWith((const char*)psz, 0, nLength); }
 366     // from multibyte string
 367   wxString(const char *psz, wxMBConv& WXUNUSED(conv), size_t nLength = wxSTRING_MAXLEN)
 368     { InitWith(psz, 0, nLength); }
 369     // from wide (Unicode) string
 370   wxString(const wchar_t *pwz);
 371     // from wxCharBuffer
 372   wxString(const wxCharBuffer& psz)
 373     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 374 #endif
 375     // dtor is not virtual, this class must not be inherited from!
 376  ~wxString() { GetStringData()->Unlock(); }
 377
 378   // generic attributes & operations
 379     // as standard strlen()
 380   size_t Len() const { return GetStringData()->nDataLength; }
 381     // string contains any characters?
 382   bool IsEmpty() const { return Len() == 0; }
 383     // empty string is "FALSE", so !str will return TRUE
 384   bool operator!() const { return IsEmpty(); }
 385     // empty string contents
 386   void Empty()
 387   {
 388     if ( !IsEmpty() )
 389       Reinit();
 390
 391     // should be empty
 392     wxASSERT( GetStringData()->nDataLength == 0 );
 393   }
 394     // empty the string and free memory
 395   void Clear()
 396   {
 397     if ( !GetStringData()->IsEmpty() )
 398       Reinit();
 399
 400     wxASSERT( GetStringData()->nDataLength == 0 );  // should be empty
 401     wxASSERT( GetStringData()->nAllocLength == 0 ); // and not own any memory
 402   }
 403
 404   // contents test
 405     // Is an ascii value
 406   bool IsAscii() const;
 407     // Is a number
 408   bool IsNumber() const;
 409     // Is a word
 410   bool IsWord() const;
 411
 412   // data access (all indexes are 0 based)
 413     // read access
 414     wxChar  GetChar(size_t n) const
 415       { ASSERT_VALID_INDEX( n );  return m_pchData[n]; }
 416     // read/write access
 417     wxChar& GetWritableChar(size_t n)
 418       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 419     // write access
 420     void  SetChar(size_t n, wxChar ch)
 421       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; }
 422
 423     // get last character
 424     wxChar  Last() const
 425       { wxASSERT( !IsEmpty() ); return m_pchData[Len() - 1]; }
 426     // get writable last character
 427     wxChar& Last()
 428       { wxASSERT( !IsEmpty() ); CopyBeforeWrite(); return m_pchData[Len()-1]; }
 429
 430     // under Unix it is tested with configure, assume it works on other
 431     // platforms (there might be overloading problems if size_t and int are
 432     // the same type)
 433 #if !defined(__UNIX__) || wxUSE_SIZE_T_STRING_OPERATOR
 434     // operator version of GetChar
 435     wxChar  operator[](size_t n) const
 436       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 437 #endif
 438
 439     // operator version of GetChar
 440     wxChar  operator[](int n) const
 441       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 442     // operator version of GetWritableChar
 443     wxChar& operator[](size_t n)
 444       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 445
 446     // implicit conversion to C string
 447     operator const wxChar*() const { return m_pchData; }
 448     // explicit conversion to C string (use this with printf()!)
 449     const wxChar* c_str()   const { return m_pchData; }
 450     //
 451     const wxChar* GetData() const { return m_pchData; }
 452 #if wxUSE_UNICODE
 453     const wxCharBuffer mb_str(wxMBConv& conv = wxConv_libc) const { return conv.cWC2MB(m_pchData); }
 454     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConv_libc) const { return m_pchData; }
 455 #if wxMBFILES
 456     const wxCharBuffer fn_str() const { return mb_str(wxConv_file); }
 457 #else
 458     const wxChar* fn_str() const { return m_pchData; }
 459 #endif
 460 #else
 461     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConv_libc ) const { return m_pchData; }
 462     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 463     const wxChar* fn_str() const { return m_pchData; }
 464 #endif
 465     // for convenience
 466     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConv_current); }
 467
 468   // overloaded assignment
 469     // from another wxString
 470   wxString& operator=(const wxString& stringSrc);
 471     // from a character
 472   wxString& operator=(wxChar ch);
 473     // from a C string
 474   wxString& operator=(const wxChar *psz);
 475 #if wxUSE_UNICODE
 476     // from wxWCharBuffer
 477   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 478 #else
 479     // from another kind of C string
 480   wxString& operator=(const unsigned char* psz);
 481     // from a wide string
 482   wxString& operator=(const wchar_t *pwz);
 483     // from wxCharBuffer
 484   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 485 #endif
 486
 487   // string concatenation
 488     // in place concatenation
 489     /*
 490         Concatenate and return the result. Note that the left to right
 491         associativity of << allows to write things like "str << str1 << str2
 492         << ..." (unlike with +=)
 493      */
 494       // string += string
 495   wxString& operator<<(const wxString& s)
 496   {
 497     wxASSERT( s.GetStringData()->IsValid() );
 498
 499     ConcatSelf(s.Len(), s);
 500     return *this;
 501   }
 502       // string += C string
 503   wxString& operator<<(const wxChar *psz)
 504     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 505       // string += char
 506   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 507
 508       // string += string
 509   void operator+=(const wxString& s) { (void)operator<<(s); }
 510       // string += C string
 511   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 512       // string += char
 513   void operator+=(wxChar ch) { (void)operator<<(ch); }
 514
 515       // string += buffer (i.e. from wxGetString)
 516 #if wxUSE_UNICODE
 517   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 518   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 519 #else
 520   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 521   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 522 #endif
 523
 524     // string += C string
 525   wxString& Append(const wxChar* psz)
 526     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 527     // append count copies of given character
 528   wxString& Append(wxChar ch, size_t count = 1u)
 529     { wxString str(ch, count); return *this << str; }
 530
 531     // prepend a string, return the string itself
 532   wxString& Prepend(const wxString& str)
 533     { *this = str + *this; return *this; }
 534
 535     // non-destructive concatenation
 536       //
 537   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 538       //
 539   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 540       //
 541   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 542       //
 543   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 544       //
 545   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 546
 547   // stream-like functions
 548       // insert an int into string
 549   wxString& operator<<(int i);
 550       // insert a float into string
 551   wxString& operator<<(float f);
 552       // insert a double into string
 553   wxString& operator<<(double d);
 554
 555   // string comparison
 556     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 557   int  Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 558     // same as Cmp() but not case-sensitive
 559   int  CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 560     // test for the string equality, either considering case or not
 561     // (if compareWithCase then the case matters)
 562   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 563     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 564
 565   // simple sub-string extraction
 566       // return substring starting at nFirst of length nCount (or till the end
 567       // if nCount = default value)
 568   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 569
 570     // operator version of Mid()
 571   wxString  operator()(size_t start, size_t len) const
 572     { return Mid(start, len); }
 573
 574       // get first nCount characters
 575   wxString Left(size_t nCount) const;
 576       // get last nCount characters
 577   wxString Right(size_t nCount) const;
 578       // get all characters before the first occurence of ch
 579       // (returns the whole string if ch not found)
 580   wxString BeforeFirst(wxChar ch) const;
 581       // get all characters before the last occurence of ch
 582       // (returns empty string if ch not found)
 583   wxString BeforeLast(wxChar ch) const;
 584       // get all characters after the first occurence of ch
 585       // (returns empty string if ch not found)
 586   wxString AfterFirst(wxChar ch) const;
 587       // get all characters after the last occurence of ch
 588       // (returns the whole string if ch not found)
 589   wxString AfterLast(wxChar ch) const;
 590
 591     // for compatibility only, use more explicitly named functions above
 592   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 593   wxString After(wxChar ch) const { return AfterFirst(ch); }
 594
 595   // case conversion
 596       // convert to upper case in place, return the string itself
 597   wxString& MakeUpper();
 598       // convert to upper case, return the copy of the string
 599       // Here's something to remember: BC++ doesn't like returns in inlines.
 600   wxString Upper() const ;
 601       // convert to lower case in place, return the string itself
 602   wxString& MakeLower();
 603       // convert to lower case, return the copy of the string
 604   wxString Lower() const ;
 605
 606   // trimming/padding whitespace (either side) and truncating
 607       // remove spaces from left or from right (default) side
 608   wxString& Trim(bool bFromRight = TRUE);
 609       // add nCount copies chPad in the beginning or at the end (default)
 610   wxString& Pad(size_t nCount, wxChar chPad = _T(' '), bool bFromRight = TRUE);
 611       // truncate string to given length
 612   wxString& Truncate(size_t uiLen);
 613
 614   // searching and replacing
 615       // searching (return starting index, or -1 if not found)
 616   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 617       // searching (return starting index, or -1 if not found)
 618   int Find(const wxChar *pszSub) const;               // like strstr
 619       // replace first (or all of bReplaceAll) occurences of substring with
 620       // another string, returns the number of replacements made
 621   size_t Replace(const wxChar *szOld,
 622                  const wxChar *szNew,
 623                  bool bReplaceAll = TRUE);
 624
 625     // check if the string contents matches a mask containing '*' and '?'
 626   bool Matches(const wxChar *szMask) const;
 627
 628   // formated input/output
 629     // as sprintf(), returns the number of characters written or < 0 on error
 630   int Printf(const wxChar *pszFormat, ...);
 631     // as vprintf(), returns the number of characters written or < 0 on error
 632   int PrintfV(const wxChar* pszFormat, va_list argptr);
 633
 634   // raw access to string memory
 635     // ensure that string has space for at least nLen characters
 636     // only works if the data of this string is not shared
 637   void Alloc(size_t nLen);
 638     // minimize the string's memory
 639     // only works if the data of this string is not shared
 640   void Shrink();
 641     // get writable buffer of at least nLen bytes. Unget() *must* be called
 642     // a.s.a.p. to put string back in a reasonable state!
 643   wxChar *GetWriteBuf(size_t nLen);
 644     // call this immediately after GetWriteBuf() has been used
 645   void UngetWriteBuf();
 646
 647   // wxWindows version 1 compatibility functions
 648
 649   // use Mid()
 650   wxString SubString(size_t from, size_t to) const
 651       { return Mid(from, (to - from + 1)); }
 652     // values for second parameter of CompareTo function
 653   enum caseCompare {exact, ignoreCase};
 654     // values for first parameter of Strip function
 655   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 656
 657     // use Printf()
 658   int sprintf(const wxChar *pszFormat, ...);
 659
 660     // use Cmp()
 661   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 662     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 663
 664     // use Len
 665   size_t Length() const { return Len(); }
 666     // Count the number of characters
 667   int Freq(wxChar ch) const;
 668     // use MakeLower
 669   void LowerCase() { MakeLower(); }
 670     // use MakeUpper
 671   void UpperCase() { MakeUpper(); }
 672     // use Trim except that it doesn't change this string
 673   wxString Strip(stripType w = trailing) const;
 674
 675     // use Find (more general variants not yet supported)
 676   size_t Index(const wxChar* psz) const { return Find(psz); }
 677   size_t Index(wxChar ch)         const { return Find(ch);  }
 678     // use Truncate
 679   wxString& Remove(size_t pos) { return Truncate(pos); }
 680   wxString& RemoveLast() { return Truncate(Len() - 1); }
 681
 682   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 683
 684     // use Find()
 685   int First( const wxChar ch ) const { return Find(ch); }
 686   int First( const wxChar* psz ) const { return Find(psz); }
 687   int First( const wxString &str ) const { return Find(str); }
 688   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 689   bool Contains(const wxString& str) const { return Find(str) != -1; }
 690
 691     // use IsEmpty()
 692   bool IsNull() const { return IsEmpty(); }
 693
 694 #ifdef  wxSTD_STRING_COMPATIBILITY
 695   // std::string compatibility functions
 696
 697   // an 'invalid' value for string index
 698   static const size_t npos;
 699
 700   // constructors
 701     // take nLen chars starting at nPos
 702   wxString(const wxString& str, size_t nPos, size_t nLen)
 703   {
 704     wxASSERT( str.GetStringData()->IsValid() );
 705     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 706   }
 707     // take all characters from pStart to pEnd
 708   wxString(const void *pStart, const void *pEnd);
 709
 710   // lib.string.capacity
 711     // return the length of the string
 712   size_t size() const { return Len(); }
 713     // return the length of the string
 714   size_t length() const { return Len(); }
 715     // return the maximum size of the string
 716   size_t max_size() const { return wxSTRING_MAXLEN; }
 717     // resize the string, filling the space with c if c != 0
 718   void resize(size_t nSize, wxChar ch = _T('\0'));
 719     // delete the contents of the string
 720   void clear() { Empty(); }
 721     // returns true if the string is empty
 722   bool empty() const { return IsEmpty(); }
 723
 724   // lib.string.access
 725     // return the character at position n
 726   wxChar at(size_t n) const { return GetChar(n); }
 727     // returns the writable character at position n
 728   wxChar& at(size_t n) { return GetWritableChar(n); }
 729
 730   // lib.string.modifiers
 731     // append a string
 732   wxString& append(const wxString& str)
 733     { *this += str; return *this; }
 734     // append elements str[pos], ..., str[pos+n]
 735   wxString& append(const wxString& str, size_t pos, size_t n)
 736     { ConcatSelf(n, str.c_str() + pos); return *this; }
 737     // append first n (or all if n == npos) characters of sz
 738   wxString& append(const wxChar *sz, size_t n = npos)
 739     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 740
 741     // append n copies of ch
 742   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 743
 744     // same as `this_string = str'
 745   wxString& assign(const wxString& str) { return (*this) = str; }
 746     // same as ` = str[pos..pos + n]
 747   wxString& assign(const wxString& str, size_t pos, size_t n)
 748     { return *this = wxString((const wxChar *)str + pos, n); }
 749     // same as `= first n (or all if n == npos) characters of sz'
 750   wxString& assign(const wxChar *sz, size_t n = npos)
 751     { return *this = wxString(sz, n); }
 752     // same as `= n copies of ch'
 753   wxString& assign(size_t n, wxChar ch)
 754     { return *this = wxString(ch, n); }
 755
 756     // insert another string
 757   wxString& insert(size_t nPos, const wxString& str);
 758     // insert n chars of str starting at nStart (in str)
 759   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 760     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 761
 762     // insert first n (or all if n == npos) characters of sz
 763   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 764     { return insert(nPos, wxString(sz, n)); }
 765     // insert n copies of ch
 766   wxString& insert(size_t nPos, size_t n, wxChar ch)
 767     { return insert(nPos, wxString(ch, n)); }
 768
 769     // delete characters from nStart to nStart + nLen
 770   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 771
 772     // replaces the substring of length nLen starting at nStart
 773   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 774     // replaces the substring with nCount copies of ch
 775   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 776     // replaces a substring with another substring
 777   wxString& replace(size_t nStart, size_t nLen,
 778                     const wxString& str, size_t nStart2, size_t nLen2);
 779     // replaces the substring with first nCount chars of sz
 780     wxString& replace(size_t nStart, size_t nLen,
 781                       const wxChar* sz, size_t nCount);
 782
 783     // swap two strings
 784   void swap(wxString& str);
 785
 786     // All find() functions take the nStart argument which specifies the
 787     // position to start the search on, the default value is 0. All functions
 788     // return npos if there were no match.
 789
 790     // find a substring
 791   size_t find(const wxString& str, size_t nStart = 0) const;
 792
 793   // VC++ 1.5 can't cope with this syntax.
 794 #if !defined(__VISUALC__) || defined(__WIN32__)
 795     // find first n characters of sz
 796   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 797 #endif
 798
 799   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 800 #if !defined(__BORLANDC__)
 801     // find the first occurence of character ch after nStart
 802   size_t find(wxChar ch, size_t nStart = 0) const;
 803 #endif
 804     // rfind() family is exactly like find() but works right to left
 805
 806     // as find, but from the end
 807   size_t rfind(const wxString& str, size_t nStart = npos) const;
 808
 809   // VC++ 1.5 can't cope with this syntax.
 810 #if !defined(__VISUALC__) || defined(__WIN32__)
 811     // as find, but from the end
 812   size_t rfind(const wxChar* sz, size_t nStart = npos,
 813           size_t n = npos) const;
 814     // as find, but from the end
 815   size_t rfind(wxChar ch, size_t nStart = npos) const;
 816 #endif
 817
 818     // find first/last occurence of any character in the set
 819
 820     //
 821   size_t find_first_of(const wxString& str, size_t nStart = 0) const;
 822     //
 823   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 824     // same as find(char, size_t)
 825   size_t find_first_of(wxChar c, size_t nStart = 0) const;
 826     //
 827   size_t find_last_of (const wxString& str, size_t nStart = npos) const;
 828     //
 829   size_t find_last_of (const wxChar* s, size_t nStart = npos) const;
 830     // same as rfind(char, size_t)
 831   size_t find_last_of (wxChar c, size_t nStart = npos) const;
 832
 833     // find first/last occurence of any character not in the set
 834
 835     //
 836   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const;
 837     //
 838   size_t find_first_not_of(const wxChar* s, size_t nStart = 0) const;
 839     //
 840   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 841     //
 842   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 843     //
 844   size_t find_last_not_of(const wxChar* s, size_t nStart = npos) const;
 845     //
 846   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 847
 848     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 849     // equal or greater than the compare() argument.
 850
 851     // just like strcmp()
 852   int compare(const wxString& str) const { return Cmp(str); }
 853     // comparison with a substring
 854   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 855     // comparison of 2 substrings
 856   int compare(size_t nStart, size_t nLen,
 857               const wxString& str, size_t nStart2, size_t nLen2) const;
 858     // just like strcmp()
 859   int compare(const wxChar* sz) const { return Cmp(sz); }
 860     // substring comparison with first nCount characters of sz
 861   int compare(size_t nStart, size_t nLen,
 862               const wxChar* sz, size_t nCount = npos) const;
 863
 864   // substring extraction
 865   wxString substr(size_t nStart = 0, size_t nLen = npos) const;
 866 #endif // wxSTD_STRING_COMPATIBILITY
 867 };
 868
 869 // ----------------------------------------------------------------------------
 870 // The string array uses it's knowledge of internal structure of the wxString
 871 // class to optimize string storage. Normally, we would store pointers to
 872 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 873 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 874 // really all we need to turn such pointer into a string!
 875 //
 876 // Of course, it can be called a dirty hack, but we use twice less memory and
 877 // this approach is also more speed efficient, so it's probably worth it.
 878 //
 879 // Usage notes: when a string is added/inserted, a new copy of it is created,
 880 // so the original string may be safely deleted. When a string is retrieved
 881 // from the array (operator[] or Item() method), a reference is returned.
 882 // ----------------------------------------------------------------------------
 883 class WXDLLEXPORT wxArrayString
 884 {
 885 public:
 886   // type of function used by wxArrayString::Sort()
 887   typedef int (*CompareFunction)(const wxString& first,
 888                                  const wxString& second);
 889
 890   // constructors and destructor
 891     // default ctor
 892   wxArrayString();
 893     // copy ctor
 894   wxArrayString(const wxArrayString& array);
 895     // assignment operator
 896   wxArrayString& operator=(const wxArrayString& src);
 897     // not virtual, this class should not be derived from
 898  ~wxArrayString();
 899
 900   // memory management
 901     // empties the list, but doesn't release memory
 902   void Empty();
 903     // empties the list and releases memory
 904   void Clear();
 905     // preallocates memory for given number of items
 906   void Alloc(size_t nCount);
 907     // minimzes the memory usage (by freeing all extra memory)
 908   void Shrink();
 909
 910   // simple accessors
 911     // number of elements in the array
 912   size_t GetCount() const { return m_nCount; }
 913     // is it empty?
 914   bool IsEmpty() const { return m_nCount == 0; }
 915     // number of elements in the array (GetCount is preferred API)
 916   size_t Count() const { return m_nCount; }
 917
 918   // items access (range checking is done in debug version)
 919     // get item at position uiIndex
 920   wxString& Item(size_t nIndex) const
 921     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 922     // same as Item()
 923   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 924     // get last item
 925   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 926
 927   // item management
 928     // Search the element in the array, starting from the beginning if
 929     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 930     // sensitive (default). Returns index of the first item matched or
 931     // wxNOT_FOUND
 932   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 933     // add new element at the end
 934   void Add(const wxString& str);
 935     // add new element at given position
 936   void Insert(const wxString& str, size_t uiIndex);
 937     // remove first item matching this value
 938   void Remove(const wxChar *sz);
 939     // remove item by index
 940   void Remove(size_t nIndex);
 941
 942   // sorting
 943     // sort array elements in alphabetical order (or reversed alphabetical
 944     // order if reverseOrder parameter is TRUE)
 945   void Sort(bool reverseOrder = FALSE);
 946     // sort array elements using specified comparaison function
 947   void Sort(CompareFunction compareFunction);
 948
 949 private:
 950   void    Grow();     // makes array bigger if needed
 951   void    Free();     // free the string stored
 952
 953   void    DoSort();   // common part of all Sort() variants
 954
 955   size_t  m_nSize,    // current size of the array
 956           m_nCount;   // current number of elements
 957
 958   wxChar  **m_pItems;   // pointer to data
 959 };
 960
 961 // ---------------------------------------------------------------------------
 962 // wxString comparison functions: operator versions are always case sensitive
 963 // ---------------------------------------------------------------------------
 964 //
 965 inline bool operator==(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) == 0); }
 966 //
 967 inline bool operator==(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) == 0); }
 968 //
 969 inline bool operator==(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) == 0); }
 970 //
 971 inline bool operator!=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) != 0); }
 972 //
 973 inline bool operator!=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) != 0); }
 974 //
 975 inline bool operator!=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) != 0); }
 976 //
 977 inline bool operator< (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) < 0); }
 978 //
 979 inline bool operator< (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <  0); }
 980 //
 981 inline bool operator< (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >  0); }
 982 //
 983 inline bool operator> (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >  0); }
 984 //
 985 inline bool operator> (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >  0); }
 986 //
 987 inline bool operator> (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <  0); }
 988 //
 989 inline bool operator<=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) <= 0); }
 990 //
 991 inline bool operator<=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <= 0); }
 992 //
 993 inline bool operator<=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >= 0); }
 994 //
 995 inline bool operator>=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >= 0); }
 996 //
 997 inline bool operator>=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >= 0); }
 998 //
 999 inline bool operator>=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <= 0); }
1000
1001 wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
1002 wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
1003 wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
1004 wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
1005 wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
1006 #if wxUSE_UNICODE
1007 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxWCharBuffer& buf)
1008 { return string + (const wchar_t *)buf; }
1009 inline wxString WXDLLEXPORT operator+(const wxWCharBuffer& buf, const wxString& string)
1010 { return (const wchar_t *)buf + string; }
1011 #else
1012 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxCharBuffer& buf)
1013 { return string + (const char *)buf; }
1014 inline wxString WXDLLEXPORT operator+(const wxCharBuffer& buf, const wxString& string)
1015 { return (const char *)buf + string; }
1016 #endif
1017
1018 // ---------------------------------------------------------------------------
1019 // Implementation only from here until the end of file
1020 // ---------------------------------------------------------------------------
1021
1022 #ifdef wxSTD_STRING_COMPATIBILITY
1023
1024 #include "wx/ioswrap.h"
1025
1026 WXDLLEXPORT istream& operator>>(istream& is, wxString& str);
1027
1028 #endif  // wxSTD_STRING_COMPATIBILITY
1029
1030 #endif  // _WX_WXSTRINGH__