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     // under Unix it is tested with configure, assume it works on other
 374     // platforms (there might be overloading problems if size_t and int are
 375     // the same type)
 376 #if !defined(__UNIX__) || wxUSE_SIZE_T_STRING_OPERATOR
 377     // operator version of GetChar
 378     wxChar  operator[](size_t n) const
 379       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 380 #endif
 381
 382     // operator version of GetChar
 383     wxChar  operator[](int n) const
 384       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 385     // operator version of GetWritableChar
 386     wxChar& operator[](size_t n)
 387       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 388
 389     // implicit conversion to C string
 390     operator const wxChar*() const { return m_pchData; }
 391     // explicit conversion to C string (use this with printf()!)
 392     const wxChar* c_str()   const { return m_pchData; }
 393     // (and this with [wx]Printf()!)
 394     const wxChar* wx_str()  const { return m_pchData; }
 395     // identical to c_str()
 396     const wxChar* GetData() const { return m_pchData; }
 397
 398     // conversions with (possible) format convertions: have to return a
 399     // buffer with temporary data
 400 #if wxUSE_UNICODE
 401     const wxCharBuffer mb_str(wxMBConv& conv = wxConvLibc) const { return conv.cWC2MB(m_pchData); }
 402     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 403
 404     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const { return m_pchData; }
 405
 406 #if wxMBFILES
 407     const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
 408 #else // !wxMBFILES
 409     const wxChar* fn_str() const { return m_pchData; }
 410 #endif // wxMBFILES/!wxMBFILES
 411 #else // ANSI
 412 #if wxUSE_MULTIBYTE
 413     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const
 414         { return m_pchData; }
 415     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 416 #else // !mmultibyte
 417     const wxChar* mb_str() const { return m_pchData; }
 418     const wxWX2MBbuf mbc_str() const { return mb_str(); }
 419 #endif // multibyte/!multibyte
 420 #if wxUSE_WCHAR_T
 421     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 422 #endif // wxUSE_WCHAR_T
 423     const wxChar* fn_str() const { return m_pchData; }
 424 #endif // Unicode/ANSI
 425
 426   // overloaded assignment
 427     // from another wxString
 428   wxString& operator=(const wxString& stringSrc);
 429     // from a character
 430   wxString& operator=(wxChar ch);
 431     // from a C string
 432   wxString& operator=(const wxChar *psz);
 433 #if wxUSE_UNICODE
 434     // from wxWCharBuffer
 435   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 436 #else // ANSI
 437     // from another kind of C string
 438   wxString& operator=(const unsigned char* psz);
 439 #if wxUSE_WCHAR_T
 440     // from a wide string
 441   wxString& operator=(const wchar_t *pwz);
 442 #endif
 443     // from wxCharBuffer
 444   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 445 #endif // Unicode/ANSI
 446
 447   // string concatenation
 448     // in place concatenation
 449     /*
 450         Concatenate and return the result. Note that the left to right
 451         associativity of << allows to write things like "str << str1 << str2
 452         << ..." (unlike with +=)
 453      */
 454       // string += string
 455   wxString& operator<<(const wxString& s)
 456   {
 457     wxASSERT( s.GetStringData()->IsValid() );
 458
 459     ConcatSelf(s.Len(), s);
 460     return *this;
 461   }
 462       // string += C string
 463   wxString& operator<<(const wxChar *psz)
 464     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 465       // string += char
 466   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 467
 468       // string += string
 469   void operator+=(const wxString& s) { (void)operator<<(s); }
 470       // string += C string
 471   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 472       // string += char
 473   void operator+=(wxChar ch) { (void)operator<<(ch); }
 474
 475       // string += buffer (i.e. from wxGetString)
 476 #if wxUSE_UNICODE
 477   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 478   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 479 #else
 480   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 481   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 482 #endif
 483
 484     // string += C string
 485   wxString& Append(const wxChar* psz)
 486     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 487     // append count copies of given character
 488   wxString& Append(wxChar ch, size_t count = 1u)
 489     { wxString str(ch, count); return *this << str; }
 490
 491     // prepend a string, return the string itself
 492   wxString& Prepend(const wxString& str)
 493     { *this = str + *this; return *this; }
 494
 495     // non-destructive concatenation
 496       //
 497   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 498       //
 499   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 500       //
 501   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 502       //
 503   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 504       //
 505   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 506
 507   // stream-like functions
 508       // insert an int into string
 509   wxString& operator<<(int i);
 510       // insert a float into string
 511   wxString& operator<<(float f);
 512       // insert a double into string
 513   wxString& operator<<(double d);
 514
 515   // string comparison
 516     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 517   int Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 518     // same as Cmp() but not case-sensitive
 519   int CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 520     // test for the string equality, either considering case or not
 521     // (if compareWithCase then the case matters)
 522   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 523     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 524     // comparison with a signle character: returns TRUE if equal
 525   bool IsSameAs(wxChar c, bool compareWithCase = TRUE) const
 526     {
 527       return (Len() == 1) && (compareWithCase ? GetChar(0u) == c
 528                               : wxToupper(GetChar(0u)) == wxToupper(c));
 529     }
 530
 531   // simple sub-string extraction
 532       // return substring starting at nFirst of length nCount (or till the end
 533       // if nCount = default value)
 534   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 535
 536     // operator version of Mid()
 537   wxString  operator()(size_t start, size_t len) const
 538     { return Mid(start, len); }
 539
 540       // get first nCount characters
 541   wxString Left(size_t nCount) const;
 542       // get last nCount characters
 543   wxString Right(size_t nCount) const;
 544       // get all characters before the first occurence of ch
 545       // (returns the whole string if ch not found)
 546   wxString BeforeFirst(wxChar ch) const;
 547       // get all characters before the last occurence of ch
 548       // (returns empty string if ch not found)
 549   wxString BeforeLast(wxChar ch) const;
 550       // get all characters after the first occurence of ch
 551       // (returns empty string if ch not found)
 552   wxString AfterFirst(wxChar ch) const;
 553       // get all characters after the last occurence of ch
 554       // (returns the whole string if ch not found)
 555   wxString AfterLast(wxChar ch) const;
 556
 557     // for compatibility only, use more explicitly named functions above
 558   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 559   wxString After(wxChar ch) const { return AfterFirst(ch); }
 560
 561   // case conversion
 562       // convert to upper case in place, return the string itself
 563   wxString& MakeUpper();
 564       // convert to upper case, return the copy of the string
 565       // Here's something to remember: BC++ doesn't like returns in inlines.
 566   wxString Upper() const ;
 567       // convert to lower case in place, return the string itself
 568   wxString& MakeLower();
 569       // convert to lower case, return the copy of the string
 570   wxString Lower() const ;
 571
 572   // trimming/padding whitespace (either side) and truncating
 573       // remove spaces from left or from right (default) side
 574   wxString& Trim(bool bFromRight = TRUE);
 575       // add nCount copies chPad in the beginning or at the end (default)
 576   wxString& Pad(size_t nCount, wxChar chPad = T(' '), bool bFromRight = TRUE);
 577       // truncate string to given length
 578   wxString& Truncate(size_t uiLen);
 579
 580   // searching and replacing
 581       // searching (return starting index, or -1 if not found)
 582   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 583       // searching (return starting index, or -1 if not found)
 584   int Find(const wxChar *pszSub) const;               // like strstr
 585       // replace first (or all of bReplaceAll) occurences of substring with
 586       // another string, returns the number of replacements made
 587   size_t Replace(const wxChar *szOld,
 588                  const wxChar *szNew,
 589                  bool bReplaceAll = TRUE);
 590
 591     // check if the string contents matches a mask containing '*' and '?'
 592   bool Matches(const wxChar *szMask) const;
 593
 594   // formated input/output
 595     // as sprintf(), returns the number of characters written or < 0 on error
 596   int Printf(const wxChar *pszFormat, ...);
 597     // as vprintf(), returns the number of characters written or < 0 on error
 598   int PrintfV(const wxChar* pszFormat, va_list argptr);
 599
 600   // raw access to string memory
 601     // ensure that string has space for at least nLen characters
 602     // only works if the data of this string is not shared
 603   void Alloc(size_t nLen);
 604     // minimize the string's memory
 605     // only works if the data of this string is not shared
 606   void Shrink();
 607     // get writable buffer of at least nLen bytes. Unget() *must* be called
 608     // a.s.a.p. to put string back in a reasonable state!
 609   wxChar *GetWriteBuf(size_t nLen);
 610     // call this immediately after GetWriteBuf() has been used
 611   void UngetWriteBuf();
 612
 613   // wxWindows version 1 compatibility functions
 614
 615   // use Mid()
 616   wxString SubString(size_t from, size_t to) const
 617       { return Mid(from, (to - from + 1)); }
 618     // values for second parameter of CompareTo function
 619   enum caseCompare {exact, ignoreCase};
 620     // values for first parameter of Strip function
 621   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 622
 623     // use Printf()
 624   int sprintf(const wxChar *pszFormat, ...);
 625
 626     // use Cmp()
 627   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 628     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 629
 630     // use Len
 631   size_t Length() const { return Len(); }
 632     // Count the number of characters
 633   int Freq(wxChar ch) const;
 634     // use MakeLower
 635   void LowerCase() { MakeLower(); }
 636     // use MakeUpper
 637   void UpperCase() { MakeUpper(); }
 638     // use Trim except that it doesn't change this string
 639   wxString Strip(stripType w = trailing) const;
 640
 641     // use Find (more general variants not yet supported)
 642   size_t Index(const wxChar* psz) const { return Find(psz); }
 643   size_t Index(wxChar ch)         const { return Find(ch);  }
 644     // use Truncate
 645   wxString& Remove(size_t pos) { return Truncate(pos); }
 646   wxString& RemoveLast() { return Truncate(Len() - 1); }
 647
 648   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 649
 650     // use Find()
 651   int First( const wxChar ch ) const { return Find(ch); }
 652   int First( const wxChar* psz ) const { return Find(psz); }
 653   int First( const wxString &str ) const { return Find(str); }
 654   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 655   bool Contains(const wxString& str) const { return Find(str) != -1; }
 656
 657     // use IsEmpty()
 658   bool IsNull() const { return IsEmpty(); }
 659
 660 #ifdef  wxSTD_STRING_COMPATIBILITY
 661   // std::string compatibility functions
 662
 663   // standard types
 664   typedef wxChar value_type;
 665   typedef const value_type *const_iterator;
 666
 667   // an 'invalid' value for string index
 668   static const size_t npos;
 669
 670   // constructors
 671     // take nLen chars starting at nPos
 672   wxString(const wxString& str, size_t nPos, size_t nLen)
 673   {
 674     wxASSERT( str.GetStringData()->IsValid() );
 675     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 676   }
 677     // take all characters from pStart to pEnd
 678   wxString(const void *pStart, const void *pEnd);
 679
 680   // lib.string.capacity
 681     // return the length of the string
 682   size_t size() const { return Len(); }
 683     // return the length of the string
 684   size_t length() const { return Len(); }
 685     // return the maximum size of the string
 686   size_t max_size() const { return wxSTRING_MAXLEN; }
 687     // resize the string, filling the space with c if c != 0
 688   void resize(size_t nSize, wxChar ch = T('\0'));
 689     // delete the contents of the string
 690   void clear() { Empty(); }
 691     // returns true if the string is empty
 692   bool empty() const { return IsEmpty(); }
 693
 694   // lib.string.access
 695     // return the character at position n
 696   wxChar at(size_t n) const { return GetChar(n); }
 697     // returns the writable character at position n
 698   wxChar& at(size_t n) { return GetWritableChar(n); }
 699
 700     // first valid index position
 701   const_iterator begin() const { return wx_str(); }
 702     // position one after the last valid one
 703   const_iterator end() const { return wx_str() + length(); }
 704
 705   // lib.string.modifiers
 706     // append a string
 707   wxString& append(const wxString& str)
 708     { *this += str; return *this; }
 709     // append elements str[pos], ..., str[pos+n]
 710   wxString& append(const wxString& str, size_t pos, size_t n)
 711     { ConcatSelf(n, str.c_str() + pos); return *this; }
 712     // append first n (or all if n == npos) characters of sz
 713   wxString& append(const wxChar *sz, size_t n = npos)
 714     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 715
 716     // append n copies of ch
 717   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 718
 719     // same as `this_string = str'
 720   wxString& assign(const wxString& str) { return (*this) = str; }
 721     // same as ` = str[pos..pos + n]
 722   wxString& assign(const wxString& str, size_t pos, size_t n)
 723     { return *this = wxString((const wxChar *)str + pos, n); }
 724     // same as `= first n (or all if n == npos) characters of sz'
 725   wxString& assign(const wxChar *sz, size_t n = npos)
 726     { return *this = wxString(sz, n); }
 727     // same as `= n copies of ch'
 728   wxString& assign(size_t n, wxChar ch)
 729     { return *this = wxString(ch, n); }
 730
 731     // insert another string
 732   wxString& insert(size_t nPos, const wxString& str);
 733     // insert n chars of str starting at nStart (in str)
 734   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 735     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 736
 737     // insert first n (or all if n == npos) characters of sz
 738   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 739     { return insert(nPos, wxString(sz, n)); }
 740     // insert n copies of ch
 741   wxString& insert(size_t nPos, size_t n, wxChar ch)
 742     { return insert(nPos, wxString(ch, n)); }
 743
 744     // delete characters from nStart to nStart + nLen
 745   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 746
 747     // replaces the substring of length nLen starting at nStart
 748   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 749     // replaces the substring with nCount copies of ch
 750   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 751     // replaces a substring with another substring
 752   wxString& replace(size_t nStart, size_t nLen,
 753                     const wxString& str, size_t nStart2, size_t nLen2);
 754     // replaces the substring with first nCount chars of sz
 755     wxString& replace(size_t nStart, size_t nLen,
 756                       const wxChar* sz, size_t nCount);
 757
 758     // swap two strings
 759   void swap(wxString& str);
 760
 761     // All find() functions take the nStart argument which specifies the
 762     // position to start the search on, the default value is 0. All functions
 763     // return npos if there were no match.
 764
 765     // find a substring
 766   size_t find(const wxString& str, size_t nStart = 0) const;
 767
 768   // VC++ 1.5 can't cope with this syntax.
 769 #if !defined(__VISUALC__) || defined(__WIN32__)
 770     // find first n characters of sz
 771   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 772 #endif
 773
 774   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 775 #if !defined(__BORLANDC__)
 776     // find the first occurence of character ch after nStart
 777   size_t find(wxChar ch, size_t nStart = 0) const;
 778 #endif
 779     // rfind() family is exactly like find() but works right to left
 780
 781     // as find, but from the end
 782   size_t rfind(const wxString& str, size_t nStart = npos) const;
 783
 784   // VC++ 1.5 can't cope with this syntax.
 785 #if !defined(__VISUALC__) || defined(__WIN32__)
 786     // as find, but from the end
 787   size_t rfind(const wxChar* sz, size_t nStart = npos,
 788           size_t n = npos) const;
 789     // as find, but from the end
 790   size_t rfind(wxChar ch, size_t nStart = npos) const;
 791 #endif
 792
 793     // find first/last occurence of any character in the set
 794
 795     // as strpbrk() but starts at nStart, returns npos if not found
 796   size_t find_first_of(const wxString& str, size_t nStart = 0) const
 797     { return find_first_of(str.c_str(), nStart); }
 798     // same as above
 799   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 800     // same as find(char, size_t)
 801   size_t find_first_of(wxChar c, size_t nStart = 0) const
 802     { return find(c, nStart); }
 803     // find the last (starting from nStart) char from str in this string
 804   size_t find_last_of (const wxString& str, size_t nStart = npos) const
 805     { return find_last_of(str.c_str(), nStart); }
 806     // same as above
 807   size_t find_last_of (const wxChar* sz, size_t nStart = npos) const;
 808     // same as above
 809   size_t find_last_of(wxChar c, size_t nStart = npos) const
 810     { return rfind(c, nStart); }
 811
 812     // find first/last occurence of any character not in the set
 813
 814     // as strspn() (starting from nStart), returns npos on failure
 815   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
 816     { return find_first_not_of(str.c_str(), nStart); }
 817     // same as above
 818   size_t find_first_not_of(const wxChar* sz, size_t nStart = 0) const;
 819     // same as above
 820   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 821     //  as strcspn()
 822   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 823     // same as above
 824   size_t find_last_not_of(const wxChar* sz, size_t nStart = npos) const;
 825     // same as above
 826   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 827
 828     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 829     // equal or greater than the compare() argument.
 830
 831     // just like strcmp()
 832   int compare(const wxString& str) const { return Cmp(str); }
 833     // comparison with a substring
 834   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 835     // comparison of 2 substrings
 836   int compare(size_t nStart, size_t nLen,
 837               const wxString& str, size_t nStart2, size_t nLen2) const;
 838     // just like strcmp()
 839   int compare(const wxChar* sz) const { return Cmp(sz); }
 840     // substring comparison with first nCount characters of sz
 841   int compare(size_t nStart, size_t nLen,
 842               const wxChar* sz, size_t nCount = npos) const;
 843
 844   // substring extraction
 845   wxString substr(size_t nStart = 0, size_t nLen = npos) const;
 846 #endif // wxSTD_STRING_COMPATIBILITY
 847 };
 848
 849 // ----------------------------------------------------------------------------
 850 // The string array uses it's knowledge of internal structure of the wxString
 851 // class to optimize string storage. Normally, we would store pointers to
 852 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 853 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 854 // really all we need to turn such pointer into a string!
 855 //
 856 // Of course, it can be called a dirty hack, but we use twice less memory and
 857 // this approach is also more speed efficient, so it's probably worth it.
 858 //
 859 // Usage notes: when a string is added/inserted, a new copy of it is created,
 860 // so the original string may be safely deleted. When a string is retrieved
 861 // from the array (operator[] or Item() method), a reference is returned.
 862 // ----------------------------------------------------------------------------
 863
 864 class WXDLLEXPORT wxArrayString
 865 {
 866 public:
 867   // type of function used by wxArrayString::Sort()
 868   typedef int (*CompareFunction)(const wxString& first,
 869                                  const wxString& second);
 870
 871   // constructors and destructor
 872     // default ctor
 873   wxArrayString();
 874     // copy ctor
 875   wxArrayString(const wxArrayString& array);
 876     // assignment operator
 877   wxArrayString& operator=(const wxArrayString& src);
 878     // not virtual, this class should not be derived from
 879  ~wxArrayString();
 880
 881   // memory management
 882     // empties the list, but doesn't release memory
 883   void Empty();
 884     // empties the list and releases memory
 885   void Clear();
 886     // preallocates memory for given number of items
 887   void Alloc(size_t nCount);
 888     // minimzes the memory usage (by freeing all extra memory)
 889   void Shrink();
 890
 891   // simple accessors
 892     // number of elements in the array
 893   size_t GetCount() const { return m_nCount; }
 894     // is it empty?
 895   bool IsEmpty() const { return m_nCount == 0; }
 896     // number of elements in the array (GetCount is preferred API)
 897   size_t Count() const { return m_nCount; }
 898
 899   // items access (range checking is done in debug version)
 900     // get item at position uiIndex
 901   wxString& Item(size_t nIndex) const
 902     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 903     // same as Item()
 904   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 905     // get last item
 906   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 907
 908   // item management
 909     // Search the element in the array, starting from the beginning if
 910     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 911     // sensitive (default). Returns index of the first item matched or
 912     // wxNOT_FOUND
 913   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 914     // add new element at the end
 915   void Add(const wxString& str);
 916     // add new element at given position
 917   void Insert(const wxString& str, size_t uiIndex);
 918     // remove first item matching this value
 919   void Remove(const wxChar *sz);
 920     // remove item by index
 921   void Remove(size_t nIndex);
 922
 923   // sorting
 924     // sort array elements in alphabetical order (or reversed alphabetical
 925     // order if reverseOrder parameter is TRUE)
 926   void Sort(bool reverseOrder = FALSE);
 927     // sort array elements using specified comparaison function
 928   void Sort(CompareFunction compareFunction);
 929
 930 private:
 931   void    Grow();     // makes array bigger if needed
 932   void    Free();     // free the string stored
 933
 934   void    DoSort();   // common part of all Sort() variants
 935
 936   size_t  m_nSize,    // current size of the array
 937           m_nCount;   // current number of elements
 938
 939   wxChar  **m_pItems;   // pointer to data
 940 };
 941
 942 // ---------------------------------------------------------------------------
 943 // wxString comparison functions: operator versions are always case sensitive
 944 // ---------------------------------------------------------------------------
 945
 946 //
 947 inline bool operator==(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) == 0); }
 948 //
 949 inline bool operator==(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) == 0); }
 950 //
 951 inline bool operator==(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) == 0); }
 952 //
 953 inline bool operator!=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) != 0); }
 954 //
 955 inline bool operator!=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) != 0); }
 956 //
 957 inline bool operator!=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) != 0); }
 958 //
 959 inline bool operator< (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) < 0); }
 960 //
 961 inline bool operator< (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <  0); }
 962 //
 963 inline bool operator< (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >  0); }
 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 // comparison with char
 984 inline bool operator==(wxChar c, const wxString& s) { return s.IsSameAs(c); }
 985 inline bool operator==(const wxString& s, wxChar c) { return s.IsSameAs(c); }
 986 inline bool operator!=(wxChar c, const wxString& s) { return !s.IsSameAs(c); }
 987 inline bool operator!=(const wxString& s, wxChar c) { return !s.IsSameAs(c); }
 988
 989 #if wxUSE_UNICODE
 990 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
 991     { return (s1.Cmp((const wchar_t *)s2) == 0); }
 992 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
 993     { return (s2.Cmp((const wchar_t *)s1) == 0); }
 994 #else
 995 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
 996     { return (s1.Cmp((const char *)s2) == 0); }
 997 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
 998     { return (s2.Cmp((const char *)s1) == 0); }
 999 #endif
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 // don't pollute the library user's name space
1023 #undef ASSERT_VALID_INDEX
1024
1025 #if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM
1026
1027 #include "wx/ioswrap.h"
1028
1029 WXDLLEXPORT istream& operator>>(istream& is, wxString& str);
1030
1031 #endif  // wxSTD_STRING_COMPATIBILITY
1032
1033 #endif  // _WX_WXSTRINGH__