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 /*
  13     Efficient string class [more or less] compatible with MFC CString,
  14     wxWindows version 1 wxString and std::string and some handy functions
  15     missing from string.h.
  16 */
  17
  18 #ifndef _WX_WXSTRINGH__
  19 #define _WX_WXSTRINGH__
  20
  21 #ifdef __GNUG__
  22     #pragma interface "string.h"
  23 #endif
  24
  25 // ----------------------------------------------------------------------------
  26 // conditinal compilation
  27 // ----------------------------------------------------------------------------
  28
  29 // compile the std::string compatibility functions if defined
  30 #define   wxSTD_STRING_COMPATIBILITY
  31
  32 // define to derive wxString from wxObject (deprecated!)
  33 #ifdef WXSTRING_IS_WXOBJECT
  34     #undef WXSTRING_IS_WXOBJECT
  35 #endif
  36
  37 // ----------------------------------------------------------------------------
  38 // headers
  39 // ----------------------------------------------------------------------------
  40
  41 #ifdef __WXMAC__
  42     #include <ctype.h>
  43 #endif
  44
  45 #ifdef __EMX__
  46     #include <std.h>
  47 #endif
  48
  49 #include <string.h>
  50 #include <stdio.h>
  51 #include <stdarg.h>
  52 #include <limits.h>
  53 #include <stdlib.h>
  54
  55 #ifdef HAVE_STRINGS_H
  56     #include <strings.h>    // for strcasecmp()
  57 #endif // AIX
  58
  59 #include "wx/defs.h"        // everybody should include this
  60 #include "wx/debug.h"       // for wxASSERT()
  61 #include "wx/wxchar.h"      // for wxChar
  62 #include "wx/buffer.h"      // for wxCharBuffer
  63 #include "wx/strconv.h"     // for wxConvertXXX() macros and wxMBConv classes
  64
  65 #ifndef  WX_PRECOMP
  66   #ifdef    WXSTRING_IS_WXOBJECT
  67     #include "wx/object.h" // base class
  68   #endif
  69 #endif // !PCH
  70
  71 // ---------------------------------------------------------------------------
  72 // macros
  73 // ---------------------------------------------------------------------------
  74
  75 // 'naughty' cast
  76 #define   WXSTRINGCAST (wxChar *)(const wxChar *)
  77 #define   wxCSTRINGCAST (wxChar *)(const wxChar *)
  78 #define   wxMBSTRINGCAST (char *)(const char *)
  79 #define   wxWCSTRINGCAST (wchar_t *)(const wchar_t *)
  80
  81 // implementation only
  82 #define   ASSERT_VALID_INDEX(i) wxASSERT( (unsigned)(i) <= Len() )
  83
  84 // ----------------------------------------------------------------------------
  85 // constants
  86 // ----------------------------------------------------------------------------
  87
  88 // maximum possible length for a string means "take all string" everywhere
  89 //  (as sizeof(StringData) is unknown here, we substract 100)
  90 const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100;
  91
  92 // ----------------------------------------------------------------------------
  93 // global data
  94 // ----------------------------------------------------------------------------
  95
  96 // global pointer to empty string
  97 WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString;
  98
  99 // ---------------------------------------------------------------------------
 100 // global functions complementing standard C string library replacements for
 101 // strlen() and portable strcasecmp()
 102 //---------------------------------------------------------------------------
 103
 104 // Use wxXXX() functions from wxchar.h instead! These functions are for
 105 // backwards compatibility only.
 106
 107 // checks whether the passed in pointer is NULL and if the string is empty
 108 inline bool WXDLLEXPORT IsEmpty(const char *p) { return (!p || !*p); }
 109
 110 // safe version of strlen() (returns 0 if passed NULL pointer)
 111 inline size_t WXDLLEXPORT Strlen(const char *psz)
 112   { return psz ? strlen(psz) : 0; }
 113
 114 // portable strcasecmp/_stricmp
 115 inline int WXDLLEXPORT Stricmp(const char *psz1, const char *psz2)
 116 {
 117 #if defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
 118   return _stricmp(psz1, psz2);
 119 #elif defined(__SC__)
 120   return _stricmp(psz1, psz2);
 121 #elif defined(__SALFORDC__)
 122   return stricmp(psz1, psz2);
 123 #elif defined(__BORLANDC__)
 124   return stricmp(psz1, psz2);
 125 #elif defined(__WATCOMC__)
 126   return stricmp(psz1, psz2);
 127 #elif defined(__EMX__)
 128   return stricmp(psz1, psz2);
 129 #elif defined(__WXPM__)
 130   return stricmp(psz1, psz2);
 131 #elif defined(__UNIX__) || defined(__GNUWIN32__)
 132   return strcasecmp(psz1, psz2);
 133 #elif defined(__MWERKS__) && !defined(__INTEL__)
 134   register char c1, c2;
 135   do {
 136     c1 = tolower(*psz1++);
 137     c2 = tolower(*psz2++);
 138   } while ( c1 && (c1 == c2) );
 139
 140   return c1 - c2;
 141 #else
 142   // almost all compilers/libraries provide this function (unfortunately under
 143   // different names), that's why we don't implement our own which will surely
 144   // be more efficient than this code (uncomment to use):
 145   /*
 146     register char c1, c2;
 147     do {
 148       c1 = tolower(*psz1++);
 149       c2 = tolower(*psz2++);
 150     } while ( c1 && (c1 == c2) );
 151
 152     return c1 - c2;
 153   */
 154
 155   #error  "Please define string case-insensitive compare for your OS/compiler"
 156 #endif  // OS/compiler
 157 }
 158
 159 // return an empty wxString
 160 class WXDLLEXPORT wxString; // not yet defined
 161 inline const wxString& wxGetEmptyString() { return *(wxString *)&wxEmptyString; }
 162
 163 // ---------------------------------------------------------------------------
 164 // string data prepended with some housekeeping info (used by wxString class),
 165 // is never used directly (but had to be put here to allow inlining)
 166 // ---------------------------------------------------------------------------
 167
 168 struct WXDLLEXPORT wxStringData
 169 {
 170   int     nRefs;        // reference count
 171   size_t  nDataLength,  // actual string length
 172           nAllocLength; // allocated memory size
 173
 174   // mimics declaration 'wxChar data[nAllocLength]'
 175   wxChar* data() const { return (wxChar*)(this + 1); }
 176
 177   // empty string has a special ref count so it's never deleted
 178   bool  IsEmpty()   const { return (nRefs == -1); }
 179   bool  IsShared()  const { return (nRefs > 1);   }
 180
 181   // lock/unlock
 182   void  Lock()   { if ( !IsEmpty() ) nRefs++;                    }
 183   void  Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this);  }
 184
 185   // if we had taken control over string memory (GetWriteBuf), it's
 186   // intentionally put in invalid state
 187   void  Validate(bool b)  { nRefs = (b ? 1 : 0); }
 188   bool  IsValid() const   { return (nRefs != 0); }
 189 };
 190
 191 // ---------------------------------------------------------------------------
 192 // This is (yet another one) String class for C++ programmers. It doesn't use
 193 // any of "advanced" C++ features (i.e. templates, exceptions, namespaces...)
 194 // thus you should be able to compile it with practicaly any C++ compiler.
 195 // This class uses copy-on-write technique, i.e. identical strings share the
 196 // same memory as long as neither of them is changed.
 197 //
 198 // This class aims to be as compatible as possible with the new standard
 199 // std::string class, but adds some additional functions and should be at
 200 // least as efficient than the standard implementation.
 201 //
 202 // Performance note: it's more efficient to write functions which take "const
 203 // String&" arguments than "const char *" if you assign the argument to
 204 // another string.
 205 //
 206 // It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5.
 207 //
 208 // To do:
 209 //  - ressource support (string tables in ressources)
 210 //  - more wide character (UNICODE) support
 211 //  - regular expressions support
 212 // ---------------------------------------------------------------------------
 213
 214 #ifdef  WXSTRING_IS_WXOBJECT
 215 class WXDLLEXPORT wxString : public wxObject
 216 {
 217     DECLARE_DYNAMIC_CLASS(wxString)
 218 #else   //WXSTRING_IS_WXOBJECT
 219 class WXDLLEXPORT wxString
 220 {
 221 #endif  //WXSTRING_IS_WXOBJECT
 222
 223 friend class WXDLLEXPORT wxArrayString;
 224
 225   // NB: special care was taken in arranging the member functions in such order
 226   //     that all inline functions can be effectively inlined, verify that all
 227   //     performace critical functions are still inlined if you change order!
 228 private:
 229   // points to data preceded by wxStringData structure with ref count info
 230   wxChar *m_pchData;
 231
 232   // accessor to string data
 233   wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; }
 234
 235   // string (re)initialization functions
 236     // initializes the string to the empty value (must be called only from
 237     // ctors, use Reinit() otherwise)
 238   void Init() { m_pchData = (wxChar *)wxEmptyString; }
 239     // initializaes the string with (a part of) C-string
 240   void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN);
 241     // as Init, but also frees old data
 242   void Reinit() { GetStringData()->Unlock(); Init(); }
 243
 244   // memory allocation
 245     // allocates memory for string of lenght nLen
 246   void AllocBuffer(size_t nLen);
 247     // copies data to another string
 248   void AllocCopy(wxString&, int, int) const;
 249     // effectively copies data to string
 250   void AssignCopy(size_t, const wxChar *);
 251
 252   // append a (sub)string
 253   void ConcatSelf(int nLen, const wxChar *src);
 254
 255   // functions called before writing to the string: they copy it if there
 256   // are other references to our data (should be the only owner when writing)
 257   void CopyBeforeWrite();
 258   void AllocBeforeWrite(size_t);
 259
 260   // this method is not implemented - there is _no_ conversion from int to
 261   // string, you're doing something wrong if the compiler wants to call it!
 262   //
 263   // try `s << i' or `s.Printf("%d", i)' instead
 264   wxString(int);
 265   wxString(long);
 266
 267 public:
 268   // constructors and destructor
 269     // ctor for an empty string
 270   wxString() { Init(); }
 271     // copy ctor
 272   wxString(const wxString& stringSrc)
 273   {
 274     wxASSERT( stringSrc.GetStringData()->IsValid() );
 275
 276     if ( stringSrc.IsEmpty() ) {
 277       // nothing to do for an empty string
 278       Init();
 279     }
 280     else {
 281       m_pchData = stringSrc.m_pchData;            // share same data
 282       GetStringData()->Lock();                    // => one more copy
 283     }
 284   }
 285     // string containing nRepeat copies of ch
 286   wxString(wxChar ch, size_t nRepeat = 1);
 287     // ctor takes first nLength characters from C string
 288     // (default value of wxSTRING_MAXLEN means take all the string)
 289   wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN)
 290     { InitWith(psz, 0, nLength); }
 291
 292 #if wxUSE_UNICODE
 293     // from multibyte string
 294     // (NB: nLength is right now number of Unicode characters, not
 295     //  characters in psz! So try not to use it yet!)
 296   wxString(const char *psz, wxMBConv& conv = wxConvLibc, size_t nLength = wxSTRING_MAXLEN);
 297     // from wxWCharBuffer (i.e. return from wxGetString)
 298   wxString(const wxWCharBuffer& psz)
 299     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 300 #else // ANSI
 301     // from C string (for compilers using unsigned char)
 302   wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN)
 303     { InitWith((const char*)psz, 0, nLength); }
 304     // from multibyte string
 305   wxString(const char *psz, wxMBConv& WXUNUSED(conv), size_t nLength = wxSTRING_MAXLEN)
 306     { InitWith(psz, 0, nLength); }
 307
 308 #if wxUSE_WCHAR_T
 309     // from wide (Unicode) string
 310   wxString(const wchar_t *pwz);
 311 #endif // !wxUSE_WCHAR_T
 312
 313     // from wxCharBuffer
 314   wxString(const wxCharBuffer& psz)
 315     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 316 #endif // Unicode/ANSI
 317
 318     // dtor is not virtual, this class must not be inherited from!
 319  ~wxString() { GetStringData()->Unlock(); }
 320
 321   // generic attributes & operations
 322     // as standard strlen()
 323   size_t Len() const { return GetStringData()->nDataLength; }
 324     // string contains any characters?
 325   bool IsEmpty() const { return Len() == 0; }
 326     // empty string is "FALSE", so !str will return TRUE
 327   bool operator!() const { return IsEmpty(); }
 328     // empty string contents
 329   void Empty()
 330   {
 331     if ( !IsEmpty() )
 332       Reinit();
 333
 334     // should be empty
 335     wxASSERT( GetStringData()->nDataLength == 0 );
 336   }
 337     // empty the string and free memory
 338   void Clear()
 339   {
 340     if ( !GetStringData()->IsEmpty() )
 341       Reinit();
 342
 343     wxASSERT( GetStringData()->nDataLength == 0 );  // should be empty
 344     wxASSERT( GetStringData()->nAllocLength == 0 ); // and not own any memory
 345   }
 346
 347   // contents test
 348     // Is an ascii value
 349   bool IsAscii() const;
 350     // Is a number
 351   bool IsNumber() const;
 352     // Is a word
 353   bool IsWord() const;
 354
 355   // data access (all indexes are 0 based)
 356     // read access
 357     wxChar  GetChar(size_t n) const
 358       { ASSERT_VALID_INDEX( n );  return m_pchData[n]; }
 359     // read/write access
 360     wxChar& GetWritableChar(size_t n)
 361       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 362     // write access
 363     void  SetChar(size_t n, wxChar ch)
 364       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; }
 365
 366     // get last character
 367     wxChar  Last() const
 368       { wxASSERT( !IsEmpty() ); return m_pchData[Len() - 1]; }
 369     // get writable last character
 370     wxChar& Last()
 371       { wxASSERT( !IsEmpty() ); CopyBeforeWrite(); return m_pchData[Len()-1]; }
 372
 373     // operator version of GetChar
 374     wxChar  operator[](size_t n) const
 375       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 376
 377     // operator version of GetChar
 378     wxChar  operator[](int n) const
 379       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 380     // This is a rather ugly hack, but needed to resolve overloading
 381     // conflicts on the AXP architecture:
 382 #ifdef __alpha__
 383     // operator version of GetChar
 384     wxChar  operator[](unsigned int n) const
 385       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 386 #endif
 387     // operator version of GetWritableChar
 388     wxChar& operator[](size_t n)
 389       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 390
 391     // implicit conversion to C string
 392     operator const wxChar*() const { return m_pchData; }
 393     // explicit conversion to C string (use this with printf()!)
 394     const wxChar* c_str()   const { return m_pchData; }
 395     // (and this with [wx]Printf()!)
 396     const wxChar* wx_str()  const { return m_pchData; }
 397     // identical to c_str()
 398     const wxChar* GetData() const { return m_pchData; }
 399
 400     // conversions with (possible) format convertions: have to return a
 401     // buffer with temporary data
 402 #if wxUSE_UNICODE
 403     const wxCharBuffer mb_str(wxMBConv& conv = wxConvLibc) const { return conv.cWC2MB(m_pchData); }
 404     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 405
 406     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const { return m_pchData; }
 407
 408 #if wxMBFILES
 409     const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
 410 #else // !wxMBFILES
 411     const wxChar* fn_str() const { return m_pchData; }
 412 #endif // wxMBFILES/!wxMBFILES
 413 #else // ANSI
 414 #if wxUSE_MULTIBYTE
 415     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const
 416         { return m_pchData; }
 417     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 418 #else // !mmultibyte
 419     const wxChar* mb_str() const { return m_pchData; }
 420     const wxWX2MBbuf mbc_str() const { return mb_str(); }
 421 #endif // multibyte/!multibyte
 422 #if wxUSE_WCHAR_T
 423     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 424 #endif // wxUSE_WCHAR_T
 425     const wxChar* fn_str() const { return m_pchData; }
 426 #endif // Unicode/ANSI
 427
 428   // overloaded assignment
 429     // from another wxString
 430   wxString& operator=(const wxString& stringSrc);
 431     // from a character
 432   wxString& operator=(wxChar ch);
 433     // from a C string
 434   wxString& operator=(const wxChar *psz);
 435 #if wxUSE_UNICODE
 436     // from wxWCharBuffer
 437   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 438 #else // ANSI
 439     // from another kind of C string
 440   wxString& operator=(const unsigned char* psz);
 441 #if wxUSE_WCHAR_T
 442     // from a wide string
 443   wxString& operator=(const wchar_t *pwz);
 444 #endif
 445     // from wxCharBuffer
 446   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 447 #endif // Unicode/ANSI
 448
 449   // string concatenation
 450     // in place concatenation
 451     /*
 452         Concatenate and return the result. Note that the left to right
 453         associativity of << allows to write things like "str << str1 << str2
 454         << ..." (unlike with +=)
 455      */
 456       // string += string
 457   wxString& operator<<(const wxString& s)
 458   {
 459     wxASSERT( s.GetStringData()->IsValid() );
 460
 461     ConcatSelf(s.Len(), s);
 462     return *this;
 463   }
 464       // string += C string
 465   wxString& operator<<(const wxChar *psz)
 466     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 467       // string += char
 468   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 469
 470       // string += string
 471   void operator+=(const wxString& s) { (void)operator<<(s); }
 472       // string += C string
 473   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 474       // string += char
 475   void operator+=(wxChar ch) { (void)operator<<(ch); }
 476
 477       // string += buffer (i.e. from wxGetString)
 478 #if wxUSE_UNICODE
 479   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 480   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 481 #else
 482   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 483   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 484 #endif
 485
 486     // string += C string
 487   wxString& Append(const wxChar* psz)
 488     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 489     // append count copies of given character
 490   wxString& Append(wxChar ch, size_t count = 1u)
 491     { wxString str(ch, count); return *this << str; }
 492
 493     // prepend a string, return the string itself
 494   wxString& Prepend(const wxString& str)
 495     { *this = str + *this; return *this; }
 496
 497     // non-destructive concatenation
 498       //
 499   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 500       //
 501   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 502       //
 503   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 504       //
 505   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 506       //
 507   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 508
 509   // stream-like functions
 510       // insert an int into string
 511   wxString& operator<<(int i);
 512       // insert a float into string
 513   wxString& operator<<(float f);
 514       // insert a double into string
 515   wxString& operator<<(double d);
 516
 517   // string comparison
 518     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 519   int Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 520     // same as Cmp() but not case-sensitive
 521   int CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 522     // test for the string equality, either considering case or not
 523     // (if compareWithCase then the case matters)
 524   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 525     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 526     // comparison with a signle character: returns TRUE if equal
 527   bool IsSameAs(wxChar c, bool compareWithCase = TRUE) const
 528     {
 529       return (Len() == 1) && (compareWithCase ? GetChar(0u) == c
 530                               : wxToupper(GetChar(0u)) == wxToupper(c));
 531     }
 532
 533   // simple sub-string extraction
 534       // return substring starting at nFirst of length nCount (or till the end
 535       // if nCount = default value)
 536   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 537
 538     // operator version of Mid()
 539   wxString  operator()(size_t start, size_t len) const
 540     { return Mid(start, len); }
 541
 542       // get first nCount characters
 543   wxString Left(size_t nCount) const;
 544       // get last nCount characters
 545   wxString Right(size_t nCount) const;
 546       // get all characters before the first occurence of ch
 547       // (returns the whole string if ch not found)
 548   wxString BeforeFirst(wxChar ch) const;
 549       // get all characters before the last occurence of ch
 550       // (returns empty string if ch not found)
 551   wxString BeforeLast(wxChar ch) const;
 552       // get all characters after the first occurence of ch
 553       // (returns empty string if ch not found)
 554   wxString AfterFirst(wxChar ch) const;
 555       // get all characters after the last occurence of ch
 556       // (returns the whole string if ch not found)
 557   wxString AfterLast(wxChar ch) const;
 558
 559     // for compatibility only, use more explicitly named functions above
 560   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 561   wxString After(wxChar ch) const { return AfterFirst(ch); }
 562
 563   // case conversion
 564       // convert to upper case in place, return the string itself
 565   wxString& MakeUpper();
 566       // convert to upper case, return the copy of the string
 567       // Here's something to remember: BC++ doesn't like returns in inlines.
 568   wxString Upper() const ;
 569       // convert to lower case in place, return the string itself
 570   wxString& MakeLower();
 571       // convert to lower case, return the copy of the string
 572   wxString Lower() const ;
 573
 574   // trimming/padding whitespace (either side) and truncating
 575       // remove spaces from left or from right (default) side
 576   wxString& Trim(bool bFromRight = TRUE);
 577       // add nCount copies chPad in the beginning or at the end (default)
 578   wxString& Pad(size_t nCount, wxChar chPad = wxT(' '), bool bFromRight = TRUE);
 579       // truncate string to given length
 580   wxString& Truncate(size_t uiLen);
 581
 582   // searching and replacing
 583       // searching (return starting index, or -1 if not found)
 584   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 585       // searching (return starting index, or -1 if not found)
 586   int Find(const wxChar *pszSub) const;               // like strstr
 587       // replace first (or all of bReplaceAll) occurences of substring with
 588       // another string, returns the number of replacements made
 589   size_t Replace(const wxChar *szOld,
 590                  const wxChar *szNew,
 591                  bool bReplaceAll = TRUE);
 592
 593     // check if the string contents matches a mask containing '*' and '?'
 594   bool Matches(const wxChar *szMask) const;
 595
 596   // formated input/output
 597     // as sprintf(), returns the number of characters written or < 0 on error
 598   int Printf(const wxChar *pszFormat, ...);
 599     // as vprintf(), returns the number of characters written or < 0 on error
 600   int PrintfV(const wxChar* pszFormat, va_list argptr);
 601
 602   // raw access to string memory
 603     // ensure that string has space for at least nLen characters
 604     // only works if the data of this string is not shared
 605   void Alloc(size_t nLen);
 606     // minimize the string's memory
 607     // only works if the data of this string is not shared
 608   void Shrink();
 609     // get writable buffer of at least nLen bytes. Unget() *must* be called
 610     // a.s.a.p. to put string back in a reasonable state!
 611   wxChar *GetWriteBuf(size_t nLen);
 612     // call this immediately after GetWriteBuf() has been used
 613   void UngetWriteBuf();
 614
 615   // wxWindows version 1 compatibility functions
 616
 617   // use Mid()
 618   wxString SubString(size_t from, size_t to) const
 619       { return Mid(from, (to - from + 1)); }
 620     // values for second parameter of CompareTo function
 621   enum caseCompare {exact, ignoreCase};
 622     // values for first parameter of Strip function
 623   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 624
 625     // use Printf()
 626   int sprintf(const wxChar *pszFormat, ...);
 627
 628     // use Cmp()
 629   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 630     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 631
 632     // use Len
 633   size_t Length() const { return Len(); }
 634     // Count the number of characters
 635   int Freq(wxChar ch) const;
 636     // use MakeLower
 637   void LowerCase() { MakeLower(); }
 638     // use MakeUpper
 639   void UpperCase() { MakeUpper(); }
 640     // use Trim except that it doesn't change this string
 641   wxString Strip(stripType w = trailing) const;
 642
 643     // use Find (more general variants not yet supported)
 644   size_t Index(const wxChar* psz) const { return Find(psz); }
 645   size_t Index(wxChar ch)         const { return Find(ch);  }
 646     // use Truncate
 647   wxString& Remove(size_t pos) { return Truncate(pos); }
 648   wxString& RemoveLast() { return Truncate(Len() - 1); }
 649
 650   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 651
 652     // use Find()
 653   int First( const wxChar ch ) const { return Find(ch); }
 654   int First( const wxChar* psz ) const { return Find(psz); }
 655   int First( const wxString &str ) const { return Find(str); }
 656   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 657   bool Contains(const wxString& str) const { return Find(str) != -1; }
 658
 659     // use IsEmpty()
 660   bool IsNull() const { return IsEmpty(); }
 661
 662 #ifdef  wxSTD_STRING_COMPATIBILITY
 663   // std::string compatibility functions
 664
 665   // standard types
 666   typedef wxChar value_type;
 667   typedef const value_type *const_iterator;
 668
 669   // an 'invalid' value for string index
 670   static const size_t npos;
 671
 672   // constructors
 673     // take nLen chars starting at nPos
 674   wxString(const wxString& str, size_t nPos, size_t nLen)
 675   {
 676     wxASSERT( str.GetStringData()->IsValid() );
 677     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 678   }
 679     // take all characters from pStart to pEnd
 680   wxString(const void *pStart, const void *pEnd);
 681
 682   // lib.string.capacity
 683     // return the length of the string
 684   size_t size() const { return Len(); }
 685     // return the length of the string
 686   size_t length() const { return Len(); }
 687     // return the maximum size of the string
 688   size_t max_size() const { return wxSTRING_MAXLEN; }
 689     // resize the string, filling the space with c if c != 0
 690   void resize(size_t nSize, wxChar ch = wxT('\0'));
 691     // delete the contents of the string
 692   void clear() { Empty(); }
 693     // returns true if the string is empty
 694   bool empty() const { return IsEmpty(); }
 695
 696   // lib.string.access
 697     // return the character at position n
 698   wxChar at(size_t n) const { return GetChar(n); }
 699     // returns the writable character at position n
 700   wxChar& at(size_t n) { return GetWritableChar(n); }
 701
 702     // first valid index position
 703   const_iterator begin() const { return wx_str(); }
 704     // position one after the last valid one
 705   const_iterator end() const { return wx_str() + length(); }
 706
 707   // lib.string.modifiers
 708     // append a string
 709   wxString& append(const wxString& str)
 710     { *this += str; return *this; }
 711     // append elements str[pos], ..., str[pos+n]
 712   wxString& append(const wxString& str, size_t pos, size_t n)
 713     { ConcatSelf(n, str.c_str() + pos); return *this; }
 714     // append first n (or all if n == npos) characters of sz
 715   wxString& append(const wxChar *sz, size_t n = npos)
 716     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 717
 718     // append n copies of ch
 719   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 720
 721     // same as `this_string = str'
 722   wxString& assign(const wxString& str) { return (*this) = str; }
 723     // same as ` = str[pos..pos + n]
 724   wxString& assign(const wxString& str, size_t pos, size_t n)
 725     { return *this = wxString((const wxChar *)str + pos, n); }
 726     // same as `= first n (or all if n == npos) characters of sz'
 727   wxString& assign(const wxChar *sz, size_t n = npos)
 728     { return *this = wxString(sz, n); }
 729     // same as `= n copies of ch'
 730   wxString& assign(size_t n, wxChar ch)
 731     { return *this = wxString(ch, n); }
 732
 733     // insert another string
 734   wxString& insert(size_t nPos, const wxString& str);
 735     // insert n chars of str starting at nStart (in str)
 736   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 737     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 738
 739     // insert first n (or all if n == npos) characters of sz
 740   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 741     { return insert(nPos, wxString(sz, n)); }
 742     // insert n copies of ch
 743   wxString& insert(size_t nPos, size_t n, wxChar ch)
 744     { return insert(nPos, wxString(ch, n)); }
 745
 746     // delete characters from nStart to nStart + nLen
 747   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 748
 749     // replaces the substring of length nLen starting at nStart
 750   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 751     // replaces the substring with nCount copies of ch
 752   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 753     // replaces a substring with another substring
 754   wxString& replace(size_t nStart, size_t nLen,
 755                     const wxString& str, size_t nStart2, size_t nLen2);
 756     // replaces the substring with first nCount chars of sz
 757     wxString& replace(size_t nStart, size_t nLen,
 758                       const wxChar* sz, size_t nCount);
 759
 760     // swap two strings
 761   void swap(wxString& str);
 762
 763     // All find() functions take the nStart argument which specifies the
 764     // position to start the search on, the default value is 0. All functions
 765     // return npos if there were no match.
 766
 767     // find a substring
 768   size_t find(const wxString& str, size_t nStart = 0) const;
 769
 770   // VC++ 1.5 can't cope with this syntax.
 771 #if !defined(__VISUALC__) || defined(__WIN32__)
 772     // find first n characters of sz
 773   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 774 #endif
 775
 776   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 777 #if !defined(__BORLANDC__)
 778     // find the first occurence of character ch after nStart
 779   size_t find(wxChar ch, size_t nStart = 0) const;
 780 #endif
 781     // rfind() family is exactly like find() but works right to left
 782
 783     // as find, but from the end
 784   size_t rfind(const wxString& str, size_t nStart = npos) const;
 785
 786   // VC++ 1.5 can't cope with this syntax.
 787 #if !defined(__VISUALC__) || defined(__WIN32__)
 788     // as find, but from the end
 789   size_t rfind(const wxChar* sz, size_t nStart = npos,
 790           size_t n = npos) const;
 791     // as find, but from the end
 792   size_t rfind(wxChar ch, size_t nStart = npos) const;
 793 #endif
 794
 795     // find first/last occurence of any character in the set
 796
 797     // as strpbrk() but starts at nStart, returns npos if not found
 798   size_t find_first_of(const wxString& str, size_t nStart = 0) const
 799     { return find_first_of(str.c_str(), nStart); }
 800     // same as above
 801   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 802     // same as find(char, size_t)
 803   size_t find_first_of(wxChar c, size_t nStart = 0) const
 804     { return find(c, nStart); }
 805     // find the last (starting from nStart) char from str in this string
 806   size_t find_last_of (const wxString& str, size_t nStart = npos) const
 807     { return find_last_of(str.c_str(), nStart); }
 808     // same as above
 809   size_t find_last_of (const wxChar* sz, size_t nStart = npos) const;
 810     // same as above
 811   size_t find_last_of(wxChar c, size_t nStart = npos) const
 812     { return rfind(c, nStart); }
 813
 814     // find first/last occurence of any character not in the set
 815
 816     // as strspn() (starting from nStart), returns npos on failure
 817   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
 818     { return find_first_not_of(str.c_str(), nStart); }
 819     // same as above
 820   size_t find_first_not_of(const wxChar* sz, size_t nStart = 0) const;
 821     // same as above
 822   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 823     //  as strcspn()
 824   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 825     // same as above
 826   size_t find_last_not_of(const wxChar* sz, size_t nStart = npos) const;
 827     // same as above
 828   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 829
 830     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 831     // equal or greater than the compare() argument.
 832
 833     // just like strcmp()
 834   int compare(const wxString& str) const { return Cmp(str); }
 835     // comparison with a substring
 836   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 837     // comparison of 2 substrings
 838   int compare(size_t nStart, size_t nLen,
 839               const wxString& str, size_t nStart2, size_t nLen2) const;
 840     // just like strcmp()
 841   int compare(const wxChar* sz) const { return Cmp(sz); }
 842     // substring comparison with first nCount characters of sz
 843   int compare(size_t nStart, size_t nLen,
 844               const wxChar* sz, size_t nCount = npos) const;
 845
 846   // substring extraction
 847   wxString substr(size_t nStart = 0, size_t nLen = npos) const;
 848 #endif // wxSTD_STRING_COMPATIBILITY
 849 };
 850
 851 // ----------------------------------------------------------------------------
 852 // The string array uses it's knowledge of internal structure of the wxString
 853 // class to optimize string storage. Normally, we would store pointers to
 854 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 855 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 856 // really all we need to turn such pointer into a string!
 857 //
 858 // Of course, it can be called a dirty hack, but we use twice less memory and
 859 // this approach is also more speed efficient, so it's probably worth it.
 860 //
 861 // Usage notes: when a string is added/inserted, a new copy of it is created,
 862 // so the original string may be safely deleted. When a string is retrieved
 863 // from the array (operator[] or Item() method), a reference is returned.
 864 // ----------------------------------------------------------------------------
 865
 866 class WXDLLEXPORT wxArrayString
 867 {
 868 public:
 869   // type of function used by wxArrayString::Sort()
 870   typedef int (*CompareFunction)(const wxString& first,
 871                                  const wxString& second);
 872
 873   // constructors and destructor
 874     // default ctor: if autoSort is TRUE, the array is always sorted (in
 875     // alphabetical order)
 876   wxArrayString(bool autoSort = FALSE);
 877     // copy ctor
 878   wxArrayString(const wxArrayString& array);
 879     // assignment operator
 880   wxArrayString& operator=(const wxArrayString& src);
 881     // not virtual, this class should not be derived from
 882  ~wxArrayString();
 883
 884   // memory management
 885     // empties the list, but doesn't release memory
 886   void Empty();
 887     // empties the list and releases memory
 888   void Clear();
 889     // preallocates memory for given number of items
 890   void Alloc(size_t nCount);
 891     // minimzes the memory usage (by freeing all extra memory)
 892   void Shrink();
 893
 894   // simple accessors
 895     // number of elements in the array
 896   size_t GetCount() const { return m_nCount; }
 897     // is it empty?
 898   bool IsEmpty() const { return m_nCount == 0; }
 899     // number of elements in the array (GetCount is preferred API)
 900   size_t Count() const { return m_nCount; }
 901
 902   // items access (range checking is done in debug version)
 903     // get item at position uiIndex
 904   wxString& Item(size_t nIndex) const
 905     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 906     // same as Item()
 907   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 908     // get last item
 909   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 910
 911   // item management
 912     // Search the element in the array, starting from the beginning if
 913     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 914     // sensitive (default). Returns index of the first item matched or
 915     // wxNOT_FOUND
 916   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 917     // add new element at the end (if the array is not sorted), return its
 918     // index
 919   size_t Add(const wxString& str);
 920     // add new element at given position
 921   void Insert(const wxString& str, size_t uiIndex);
 922     // remove first item matching this value
 923   void Remove(const wxChar *sz);
 924     // remove item by index
 925   void Remove(size_t nIndex);
 926
 927   // sorting
 928     // sort array elements in alphabetical order (or reversed alphabetical
 929     // order if reverseOrder parameter is TRUE)
 930   void Sort(bool reverseOrder = FALSE);
 931     // sort array elements using specified comparaison function
 932   void Sort(CompareFunction compareFunction);
 933
 934 protected:
 935   void Copy(const wxArrayString& src);  // copies the contents of another array
 936
 937 private:
 938   void Grow();                          // makes array bigger if needed
 939   void Free();                          // free all the strings stored
 940
 941   void DoSort();                        // common part of all Sort() variants
 942
 943   size_t  m_nSize,    // current size of the array
 944           m_nCount;   // current number of elements
 945
 946   wxChar  **m_pItems; // pointer to data
 947
 948   bool    m_autoSort; // if TRUE, keep the array always sorted
 949 };
 950
 951 class WXDLLEXPORT wxSortedArrayString : public wxArrayString
 952 {
 953 public:
 954   wxSortedArrayString() : wxArrayString(TRUE)
 955     { }
 956   wxSortedArrayString(const wxArrayString& array) : wxArrayString(TRUE)
 957     { Copy(array); }
 958 };
 959
 960 // ---------------------------------------------------------------------------
 961 // wxString comparison functions: operator versions are always case sensitive
 962 // ---------------------------------------------------------------------------
 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 // comparison with char
1002 inline bool operator==(wxChar c, const wxString& s) { return s.IsSameAs(c); }
1003 inline bool operator==(const wxString& s, wxChar c) { return s.IsSameAs(c); }
1004 inline bool operator!=(wxChar c, const wxString& s) { return !s.IsSameAs(c); }
1005 inline bool operator!=(const wxString& s, wxChar c) { return !s.IsSameAs(c); }
1006
1007 #if wxUSE_UNICODE
1008 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
1009     { return (s1.Cmp((const wchar_t *)s2) == 0); }
1010 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
1011     { return (s2.Cmp((const wchar_t *)s1) == 0); }
1012 #else
1013 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
1014     { return (s1.Cmp((const char *)s2) == 0); }
1015 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
1016     { return (s2.Cmp((const char *)s1) == 0); }
1017 #endif
1018
1019 wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
1020 wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
1021 wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
1022 wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
1023 wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
1024 #if wxUSE_UNICODE
1025 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxWCharBuffer& buf)
1026     { return string + (const wchar_t *)buf; }
1027 inline wxString WXDLLEXPORT operator+(const wxWCharBuffer& buf, const wxString& string)
1028     { return (const wchar_t *)buf + string; }
1029 #else
1030 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxCharBuffer& buf)
1031     { return string + (const char *)buf; }
1032 inline wxString WXDLLEXPORT operator+(const wxCharBuffer& buf, const wxString& string)
1033     { return (const char *)buf + string; }
1034 #endif
1035
1036 // ---------------------------------------------------------------------------
1037 // Implementation only from here until the end of file
1038 // ---------------------------------------------------------------------------
1039
1040 // don't pollute the library user's name space
1041 #undef ASSERT_VALID_INDEX
1042
1043 #if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM
1044
1045 #include "wx/ioswrap.h"
1046
1047 WXDLLEXPORT istream& operator>>(istream&, wxString&);
1048 WXDLLEXPORT ostream& operator<<(ostream&, const wxString&);
1049
1050 #endif  // wxSTD_STRING_COMPATIBILITY
1051
1052 #endif  // _WX_WXSTRINGH__