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 #if defined(__WXMAC__) || defined(__VISAGECPP__)
  42     #include <ctype.h>
  43 #endif
  44
  45 #ifdef __EMX__
  46     #include <std.h>
  47 #endif
  48
  49 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
  50    // problem in VACPP V4 with including stdlib.h multiple times
  51    // strconv includes it anyway
  52 #  include <stdio.h>
  53 #  include <string.h>
  54 #  include <stdarg.h>
  55 #  include <limits.h>
  56 #else
  57 #  include <string.h>
  58 #  include <stdio.h>
  59 #  include <stdarg.h>
  60 #  include <limits.h>
  61 #  include <stdlib.h>
  62 #endif
  63
  64 #ifdef HAVE_STRINGS_H
  65     #include <strings.h>    // for strcasecmp()
  66 #endif // AIX
  67
  68 #include "wx/defs.h"        // everybody should include this
  69 #include "wx/debug.h"       // for wxASSERT()
  70 #include "wx/wxchar.h"      // for wxChar
  71 #include "wx/buffer.h"      // for wxCharBuffer
  72 #include "wx/strconv.h"     // for wxConvertXXX() macros and wxMBConv classes
  73
  74 #ifndef  WX_PRECOMP
  75   #ifdef    WXSTRING_IS_WXOBJECT
  76     #include "wx/object.h" // base class
  77   #endif
  78 #endif // !PCH
  79
  80 // ---------------------------------------------------------------------------
  81 // macros
  82 // ---------------------------------------------------------------------------
  83
  84 // 'naughty' cast
  85 #define   WXSTRINGCAST (wxChar *)(const wxChar *)
  86 #define   wxCSTRINGCAST (wxChar *)(const wxChar *)
  87 #define   wxMBSTRINGCAST (char *)(const char *)
  88 #define   wxWCSTRINGCAST (wchar_t *)(const wchar_t *)
  89
  90 // implementation only
  91 #define   ASSERT_VALID_INDEX(i) wxASSERT( (unsigned)(i) <= Len() )
  92
  93 // ----------------------------------------------------------------------------
  94 // constants
  95 // ----------------------------------------------------------------------------
  96
  97 // maximum possible length for a string means "take all string" everywhere
  98 //  (as sizeof(StringData) is unknown here, we substract 100)
  99 const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100;
 100
 101 // ----------------------------------------------------------------------------
 102 // global data
 103 // ----------------------------------------------------------------------------
 104
 105 // global pointer to empty string
 106 WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString;
 107
 108 // ---------------------------------------------------------------------------
 109 // global functions complementing standard C string library replacements for
 110 // strlen() and portable strcasecmp()
 111 //---------------------------------------------------------------------------
 112
 113 // Use wxXXX() functions from wxchar.h instead! These functions are for
 114 // backwards compatibility only.
 115
 116 // checks whether the passed in pointer is NULL and if the string is empty
 117 inline bool WXDLLEXPORT IsEmpty(const char *p) { return (!p || !*p); }
 118
 119 // safe version of strlen() (returns 0 if passed NULL pointer)
 120 inline size_t WXDLLEXPORT Strlen(const char *psz)
 121   { return psz ? strlen(psz) : 0; }
 122
 123 // portable strcasecmp/_stricmp
 124 inline int WXDLLEXPORT Stricmp(const char *psz1, const char *psz2)
 125 {
 126 #if defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
 127   return _stricmp(psz1, psz2);
 128 #elif defined(__SC__)
 129   return _stricmp(psz1, psz2);
 130 #elif defined(__SALFORDC__)
 131   return stricmp(psz1, psz2);
 132 #elif defined(__BORLANDC__)
 133   return stricmp(psz1, psz2);
 134 #elif defined(__WATCOMC__)
 135   return stricmp(psz1, psz2);
 136 #elif defined(__EMX__)
 137   return stricmp(psz1, psz2);
 138 #elif defined(__WXPM__)
 139   return stricmp(psz1, psz2);
 140 #elif defined(__UNIX__) || defined(__GNUWIN32__)
 141   return strcasecmp(psz1, psz2);
 142 #elif defined(__MWERKS__) && !defined(__INTEL__)
 143   register char c1, c2;
 144   do {
 145     c1 = tolower(*psz1++);
 146     c2 = tolower(*psz2++);
 147   } while ( c1 && (c1 == c2) );
 148
 149   return c1 - c2;
 150 #else
 151   // almost all compilers/libraries provide this function (unfortunately under
 152   // different names), that's why we don't implement our own which will surely
 153   // be more efficient than this code (uncomment to use):
 154   /*
 155     register char c1, c2;
 156     do {
 157       c1 = tolower(*psz1++);
 158       c2 = tolower(*psz2++);
 159     } while ( c1 && (c1 == c2) );
 160
 161     return c1 - c2;
 162   */
 163
 164   #error  "Please define string case-insensitive compare for your OS/compiler"
 165 #endif  // OS/compiler
 166 }
 167
 168 // wxSnprintf() is like snprintf() if it's available and sprintf() (always
 169 // available, but dangerous!) if not
 170 extern int WXDLLEXPORT wxSnprintf(wxChar *buf, size_t len,
 171                                   const wxChar *format, ...);
 172
 173 // and wxVsnprintf() is like vsnprintf() or vsprintf()
 174 extern int WXDLLEXPORT wxVsnprintf(wxChar *buf, size_t len,
 175                                    const wxChar *format, va_list argptr);
 176
 177 // return an empty wxString
 178 class WXDLLEXPORT wxString; // not yet defined
 179 inline const wxString& wxGetEmptyString() { return *(wxString *)&wxEmptyString; }
 180
 181 // ---------------------------------------------------------------------------
 182 // string data prepended with some housekeeping info (used by wxString class),
 183 // is never used directly (but had to be put here to allow inlining)
 184 // ---------------------------------------------------------------------------
 185
 186 struct WXDLLEXPORT wxStringData
 187 {
 188   int     nRefs;        // reference count
 189   size_t  nDataLength,  // actual string length
 190           nAllocLength; // allocated memory size
 191
 192   // mimics declaration 'wxChar data[nAllocLength]'
 193   wxChar* data() const { return (wxChar*)(this + 1); }
 194
 195   // empty string has a special ref count so it's never deleted
 196   bool  IsEmpty()   const { return (nRefs == -1); }
 197   bool  IsShared()  const { return (nRefs > 1);   }
 198
 199   // lock/unlock
 200   void  Lock()   { if ( !IsEmpty() ) nRefs++;                    }
 201   void  Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this);  }
 202
 203   // if we had taken control over string memory (GetWriteBuf), it's
 204   // intentionally put in invalid state
 205   void  Validate(bool b)  { nRefs = (b ? 1 : 0); }
 206   bool  IsValid() const   { return (nRefs != 0); }
 207 };
 208
 209 // ---------------------------------------------------------------------------
 210 // This is (yet another one) String class for C++ programmers. It doesn't use
 211 // any of "advanced" C++ features (i.e. templates, exceptions, namespaces...)
 212 // thus you should be able to compile it with practicaly any C++ compiler.
 213 // This class uses copy-on-write technique, i.e. identical strings share the
 214 // same memory as long as neither of them is changed.
 215 //
 216 // This class aims to be as compatible as possible with the new standard
 217 // std::string class, but adds some additional functions and should be at
 218 // least as efficient than the standard implementation.
 219 //
 220 // Performance note: it's more efficient to write functions which take "const
 221 // String&" arguments than "const char *" if you assign the argument to
 222 // another string.
 223 //
 224 // It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5.
 225 //
 226 // To do:
 227 //  - ressource support (string tables in ressources)
 228 //  - more wide character (UNICODE) support
 229 //  - regular expressions support
 230 // ---------------------------------------------------------------------------
 231
 232 #ifdef  WXSTRING_IS_WXOBJECT
 233 class WXDLLEXPORT wxString : public wxObject
 234 {
 235     DECLARE_DYNAMIC_CLASS(wxString)
 236 #else   //WXSTRING_IS_WXOBJECT
 237 class WXDLLEXPORT wxString
 238 {
 239 #endif  //WXSTRING_IS_WXOBJECT
 240
 241 friend class WXDLLEXPORT wxArrayString;
 242
 243   // NB: special care was taken in arranging the member functions in such order
 244   //     that all inline functions can be effectively inlined, verify that all
 245   //     performace critical functions are still inlined if you change order!
 246 private:
 247   // points to data preceded by wxStringData structure with ref count info
 248   wxChar *m_pchData;
 249
 250   // accessor to string data
 251   wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; }
 252
 253   // string (re)initialization functions
 254     // initializes the string to the empty value (must be called only from
 255     // ctors, use Reinit() otherwise)
 256   void Init() { m_pchData = (wxChar *)wxEmptyString; }
 257     // initializaes the string with (a part of) C-string
 258   void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN);
 259     // as Init, but also frees old data
 260   void Reinit() { GetStringData()->Unlock(); Init(); }
 261
 262   // memory allocation
 263     // allocates memory for string of lenght nLen
 264   void AllocBuffer(size_t nLen);
 265     // copies data to another string
 266   void AllocCopy(wxString&, int, int) const;
 267     // effectively copies data to string
 268   void AssignCopy(size_t, const wxChar *);
 269
 270   // append a (sub)string
 271   void ConcatSelf(int nLen, const wxChar *src);
 272
 273   // functions called before writing to the string: they copy it if there
 274   // are other references to our data (should be the only owner when writing)
 275   void CopyBeforeWrite();
 276   void AllocBeforeWrite(size_t);
 277
 278   // this method is not implemented - there is _no_ conversion from int to
 279   // string, you're doing something wrong if the compiler wants to call it!
 280   //
 281   // try `s << i' or `s.Printf("%d", i)' instead
 282   wxString(int);
 283   wxString(long);
 284
 285 public:
 286   // constructors and destructor
 287     // ctor for an empty string
 288   wxString() { Init(); }
 289     // copy ctor
 290   wxString(const wxString& stringSrc)
 291   {
 292     wxASSERT( stringSrc.GetStringData()->IsValid() );
 293
 294     if ( stringSrc.IsEmpty() ) {
 295       // nothing to do for an empty string
 296       Init();
 297     }
 298     else {
 299       m_pchData = stringSrc.m_pchData;            // share same data
 300       GetStringData()->Lock();                    // => one more copy
 301     }
 302   }
 303     // string containing nRepeat copies of ch
 304   wxString(wxChar ch, size_t nRepeat = 1);
 305     // ctor takes first nLength characters from C string
 306     // (default value of wxSTRING_MAXLEN means take all the string)
 307   wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN)
 308     { InitWith(psz, 0, nLength); }
 309
 310 #if wxUSE_UNICODE
 311     // from multibyte string
 312     // (NB: nLength is right now number of Unicode characters, not
 313     //  characters in psz! So try not to use it yet!)
 314   wxString(const char *psz, wxMBConv& conv = wxConvLibc, size_t nLength = wxSTRING_MAXLEN);
 315     // from wxWCharBuffer (i.e. return from wxGetString)
 316   wxString(const wxWCharBuffer& psz)
 317     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 318 #else // ANSI
 319     // from C string (for compilers using unsigned char)
 320   wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN)
 321     { InitWith((const char*)psz, 0, nLength); }
 322     // from multibyte string
 323   wxString(const char *psz, wxMBConv& WXUNUSED(conv) , size_t nLength = wxSTRING_MAXLEN)
 324     { InitWith(psz, 0, nLength); }
 325
 326 #if wxUSE_WCHAR_T
 327     // from wide (Unicode) string
 328   wxString(const wchar_t *pwz);
 329 #endif // !wxUSE_WCHAR_T
 330
 331     // from wxCharBuffer
 332   wxString(const wxCharBuffer& psz)
 333     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 334 #endif // Unicode/ANSI
 335
 336     // dtor is not virtual, this class must not be inherited from!
 337  ~wxString() { GetStringData()->Unlock(); }
 338
 339   // generic attributes & operations
 340     // as standard strlen()
 341   size_t Len() const { return GetStringData()->nDataLength; }
 342     // string contains any characters?
 343   bool IsEmpty() const { return Len() == 0; }
 344     // empty string is "FALSE", so !str will return TRUE
 345   bool operator!() const { return IsEmpty(); }
 346     // empty string contents
 347   void Empty()
 348   {
 349     if ( !IsEmpty() )
 350       Reinit();
 351
 352     // should be empty
 353     wxASSERT( GetStringData()->nDataLength == 0 );
 354   }
 355     // empty the string and free memory
 356   void Clear()
 357   {
 358     if ( !GetStringData()->IsEmpty() )
 359       Reinit();
 360
 361     wxASSERT( GetStringData()->nDataLength == 0 );  // should be empty
 362     wxASSERT( GetStringData()->nAllocLength == 0 ); // and not own any memory
 363   }
 364
 365   // contents test
 366     // Is an ascii value
 367   bool IsAscii() const;
 368     // Is a number
 369   bool IsNumber() const;
 370     // Is a word
 371   bool IsWord() const;
 372
 373   // data access (all indexes are 0 based)
 374     // read access
 375     wxChar  GetChar(size_t n) const
 376       { ASSERT_VALID_INDEX( n );  return m_pchData[n]; }
 377     // read/write access
 378     wxChar& GetWritableChar(size_t n)
 379       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 380     // write access
 381     void  SetChar(size_t n, wxChar ch)
 382       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; }
 383
 384     // get last character
 385     wxChar  Last() const
 386       { wxASSERT( !IsEmpty() ); return m_pchData[Len() - 1]; }
 387     // get writable last character
 388     wxChar& Last()
 389       { wxASSERT( !IsEmpty() ); CopyBeforeWrite(); return m_pchData[Len()-1]; }
 390
 391     // operator version of GetChar
 392     wxChar  operator[](size_t n) const
 393       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 394
 395     // operator version of GetChar
 396     wxChar  operator[](int n) const
 397       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 398 #ifdef __alpha__
 399     // operator version of GetChar
 400     wxChar  operator[](unsigned int n) const
 401       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 402 #endif
 403
 404     // operator version of GetWriteableChar
 405     wxChar& operator[](size_t n)
 406       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 407 #ifdef __alpha__
 408     // operator version of GetWriteableChar
 409     wxChar&  operator[](unsigned int n)
 410       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 411 #endif
 412
 413     // implicit conversion to C string
 414     operator const wxChar*() const { return m_pchData; }
 415     // explicit conversion to C string (use this with printf()!)
 416     const wxChar* c_str()   const { return m_pchData; }
 417     // (and this with [wx]Printf()!)
 418     const wxChar* wx_str()  const { return m_pchData; }
 419     // identical to c_str()
 420     const wxChar* GetData() const { return m_pchData; }
 421
 422     // conversions with (possible) format convertions: have to return a
 423     // buffer with temporary data
 424 #if wxUSE_UNICODE
 425     const wxCharBuffer mb_str(wxMBConv& conv = wxConvLibc) const { return conv.cWC2MB(m_pchData); }
 426     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 427
 428     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const { return m_pchData; }
 429
 430 #if wxMBFILES
 431     const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
 432 #else // !wxMBFILES
 433     const wxChar* fn_str() const { return m_pchData; }
 434 #endif // wxMBFILES/!wxMBFILES
 435 #else // ANSI
 436 #if wxUSE_MULTIBYTE
 437     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConvLibc) const
 438         { return m_pchData; }
 439     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 440 #else // !mmultibyte
 441     const wxChar* mb_str() const { return m_pchData; }
 442     const wxWX2MBbuf mbc_str() const { return mb_str(); }
 443 #endif // multibyte/!multibyte
 444 #if wxUSE_WCHAR_T
 445     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 446 #endif // wxUSE_WCHAR_T
 447     const wxChar* fn_str() const { return m_pchData; }
 448 #endif // Unicode/ANSI
 449
 450   // overloaded assignment
 451     // from another wxString
 452   wxString& operator=(const wxString& stringSrc);
 453     // from a character
 454   wxString& operator=(wxChar ch);
 455     // from a C string
 456   wxString& operator=(const wxChar *psz);
 457 #if wxUSE_UNICODE
 458     // from wxWCharBuffer
 459   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 460 #else // ANSI
 461     // from another kind of C string
 462   wxString& operator=(const unsigned char* psz);
 463 #if wxUSE_WCHAR_T
 464     // from a wide string
 465   wxString& operator=(const wchar_t *pwz);
 466 #endif
 467     // from wxCharBuffer
 468   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 469 #endif // Unicode/ANSI
 470
 471   // string concatenation
 472     // in place concatenation
 473     /*
 474         Concatenate and return the result. Note that the left to right
 475         associativity of << allows to write things like "str << str1 << str2
 476         << ..." (unlike with +=)
 477      */
 478       // string += string
 479   wxString& operator<<(const wxString& s)
 480   {
 481     wxASSERT( s.GetStringData()->IsValid() );
 482
 483     ConcatSelf(s.Len(), s);
 484     return *this;
 485   }
 486       // string += C string
 487   wxString& operator<<(const wxChar *psz)
 488     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 489       // string += char
 490   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 491
 492       // string += string
 493   void operator+=(const wxString& s) { (void)operator<<(s); }
 494       // string += C string
 495   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 496       // string += char
 497   void operator+=(wxChar ch) { (void)operator<<(ch); }
 498
 499       // string += buffer (i.e. from wxGetString)
 500 #if wxUSE_UNICODE
 501   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 502   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 503 #else
 504   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 505   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 506 #endif
 507
 508     // string += C string
 509   wxString& Append(const wxChar* psz)
 510     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 511     // append count copies of given character
 512   wxString& Append(wxChar ch, size_t count = 1u)
 513     { wxString str(ch, count); return *this << str; }
 514
 515     // prepend a string, return the string itself
 516   wxString& Prepend(const wxString& str)
 517     { *this = str + *this; return *this; }
 518
 519     // non-destructive concatenation
 520       //
 521   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 522       //
 523   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 524       //
 525   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 526       //
 527   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 528       //
 529   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 530
 531   // stream-like functions
 532       // insert an int into string
 533   wxString& operator<<(int i);
 534       // insert a float into string
 535   wxString& operator<<(float f);
 536       // insert a double into string
 537   wxString& operator<<(double d);
 538
 539   // string comparison
 540     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 541   int Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 542     // same as Cmp() but not case-sensitive
 543   int CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 544     // test for the string equality, either considering case or not
 545     // (if compareWithCase then the case matters)
 546   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 547     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 548     // comparison with a signle character: returns TRUE if equal
 549   bool IsSameAs(wxChar c, bool compareWithCase = TRUE) const
 550     {
 551       return (Len() == 1) && (compareWithCase ? GetChar(0u) == c
 552                               : wxToupper(GetChar(0u)) == wxToupper(c));
 553     }
 554
 555   // simple sub-string extraction
 556       // return substring starting at nFirst of length nCount (or till the end
 557       // if nCount = default value)
 558   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 559
 560     // operator version of Mid()
 561   wxString  operator()(size_t start, size_t len) const
 562     { return Mid(start, len); }
 563
 564       // get first nCount characters
 565   wxString Left(size_t nCount) const;
 566       // get last nCount characters
 567   wxString Right(size_t nCount) const;
 568       // get all characters before the first occurence of ch
 569       // (returns the whole string if ch not found)
 570   wxString BeforeFirst(wxChar ch) const;
 571       // get all characters before the last occurence of ch
 572       // (returns empty string if ch not found)
 573   wxString BeforeLast(wxChar ch) const;
 574       // get all characters after the first occurence of ch
 575       // (returns empty string if ch not found)
 576   wxString AfterFirst(wxChar ch) const;
 577       // get all characters after the last occurence of ch
 578       // (returns the whole string if ch not found)
 579   wxString AfterLast(wxChar ch) const;
 580
 581     // for compatibility only, use more explicitly named functions above
 582   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 583   wxString After(wxChar ch) const { return AfterFirst(ch); }
 584
 585   // case conversion
 586       // convert to upper case in place, return the string itself
 587   wxString& MakeUpper();
 588       // convert to upper case, return the copy of the string
 589       // Here's something to remember: BC++ doesn't like returns in inlines.
 590   wxString Upper() const ;
 591       // convert to lower case in place, return the string itself
 592   wxString& MakeLower();
 593       // convert to lower case, return the copy of the string
 594   wxString Lower() const ;
 595
 596   // trimming/padding whitespace (either side) and truncating
 597       // remove spaces from left or from right (default) side
 598   wxString& Trim(bool bFromRight = TRUE);
 599       // add nCount copies chPad in the beginning or at the end (default)
 600   wxString& Pad(size_t nCount, wxChar chPad = wxT(' '), bool bFromRight = TRUE);
 601       // truncate string to given length
 602   wxString& Truncate(size_t uiLen);
 603
 604   // searching and replacing
 605       // searching (return starting index, or -1 if not found)
 606   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 607       // searching (return starting index, or -1 if not found)
 608   int Find(const wxChar *pszSub) const;               // like strstr
 609       // replace first (or all of bReplaceAll) occurences of substring with
 610       // another string, returns the number of replacements made
 611   size_t Replace(const wxChar *szOld,
 612                  const wxChar *szNew,
 613                  bool bReplaceAll = TRUE);
 614
 615     // check if the string contents matches a mask containing '*' and '?'
 616   bool Matches(const wxChar *szMask) const;
 617
 618   // formated input/output
 619     // as sprintf(), returns the number of characters written or < 0 on error
 620   int Printf(const wxChar *pszFormat, ...);
 621     // as vprintf(), returns the number of characters written or < 0 on error
 622   int PrintfV(const wxChar* pszFormat, va_list argptr);
 623
 624   // raw access to string memory
 625     // ensure that string has space for at least nLen characters
 626     // only works if the data of this string is not shared
 627   void Alloc(size_t nLen);
 628     // minimize the string's memory
 629     // only works if the data of this string is not shared
 630   void Shrink();
 631     // get writable buffer of at least nLen bytes. Unget() *must* be called
 632     // a.s.a.p. to put string back in a reasonable state!
 633   wxChar *GetWriteBuf(size_t nLen);
 634     // call this immediately after GetWriteBuf() has been used
 635   void UngetWriteBuf();
 636
 637   // wxWindows version 1 compatibility functions
 638
 639   // use Mid()
 640   wxString SubString(size_t from, size_t to) const
 641       { return Mid(from, (to - from + 1)); }
 642     // values for second parameter of CompareTo function
 643   enum caseCompare {exact, ignoreCase};
 644     // values for first parameter of Strip function
 645   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 646
 647     // use Printf()
 648   int sprintf(const wxChar *pszFormat, ...);
 649
 650     // use Cmp()
 651   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 652     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 653
 654     // use Len
 655   size_t Length() const { return Len(); }
 656     // Count the number of characters
 657   int Freq(wxChar ch) const;
 658     // use MakeLower
 659   void LowerCase() { MakeLower(); }
 660     // use MakeUpper
 661   void UpperCase() { MakeUpper(); }
 662     // use Trim except that it doesn't change this string
 663   wxString Strip(stripType w = trailing) const;
 664
 665     // use Find (more general variants not yet supported)
 666   size_t Index(const wxChar* psz) const { return Find(psz); }
 667   size_t Index(wxChar ch)         const { return Find(ch);  }
 668     // use Truncate
 669   wxString& Remove(size_t pos) { return Truncate(pos); }
 670   wxString& RemoveLast() { return Truncate(Len() - 1); }
 671
 672   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 673
 674     // use Find()
 675   int First( const wxChar ch ) const { return Find(ch); }
 676   int First( const wxChar* psz ) const { return Find(psz); }
 677   int First( const wxString &str ) const { return Find(str); }
 678   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 679   bool Contains(const wxString& str) const { return Find(str) != -1; }
 680
 681     // use IsEmpty()
 682   bool IsNull() const { return IsEmpty(); }
 683
 684 #ifdef  wxSTD_STRING_COMPATIBILITY
 685   // std::string compatibility functions
 686
 687   // standard types
 688   typedef wxChar value_type;
 689   typedef const value_type *const_iterator;
 690
 691   // an 'invalid' value for string index
 692   static const size_t npos;
 693
 694   // constructors
 695     // take nLen chars starting at nPos
 696   wxString(const wxString& str, size_t nPos, size_t nLen)
 697   {
 698     wxASSERT( str.GetStringData()->IsValid() );
 699     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 700   }
 701     // take all characters from pStart to pEnd
 702   wxString(const void *pStart, const void *pEnd);
 703
 704   // lib.string.capacity
 705     // return the length of the string
 706   size_t size() const { return Len(); }
 707     // return the length of the string
 708   size_t length() const { return Len(); }
 709     // return the maximum size of the string
 710   size_t max_size() const { return wxSTRING_MAXLEN; }
 711     // resize the string, filling the space with c if c != 0
 712   void resize(size_t nSize, wxChar ch = wxT('\0'));
 713     // delete the contents of the string
 714   void clear() { Empty(); }
 715     // returns true if the string is empty
 716   bool empty() const { return IsEmpty(); }
 717     // inform string about planned change in size
 718   void reserve(size_t size) { Alloc(size); }
 719
 720   // lib.string.access
 721     // return the character at position n
 722   wxChar at(size_t n) const { return GetChar(n); }
 723     // returns the writable character at position n
 724   wxChar& at(size_t n) { return GetWritableChar(n); }
 725
 726     // first valid index position
 727   const_iterator begin() const { return wx_str(); }
 728     // position one after the last valid one
 729   const_iterator end() const { return wx_str() + length(); }
 730
 731   // lib.string.modifiers
 732     // append a string
 733   wxString& append(const wxString& str)
 734     { *this += str; return *this; }
 735     // append elements str[pos], ..., str[pos+n]
 736   wxString& append(const wxString& str, size_t pos, size_t n)
 737     { ConcatSelf(n, str.c_str() + pos); return *this; }
 738     // append first n (or all if n == npos) characters of sz
 739   wxString& append(const wxChar *sz, size_t n = npos)
 740     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 741
 742     // append n copies of ch
 743   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 744
 745     // same as `this_string = str'
 746   wxString& assign(const wxString& str) { return (*this) = str; }
 747     // same as ` = str[pos..pos + n]
 748   wxString& assign(const wxString& str, size_t pos, size_t n)
 749     { return *this = wxString((const wxChar *)str + pos, n); }
 750     // same as `= first n (or all if n == npos) characters of sz'
 751   wxString& assign(const wxChar *sz, size_t n = npos)
 752     { return *this = wxString(sz, n); }
 753     // same as `= n copies of ch'
 754   wxString& assign(size_t n, wxChar ch)
 755     { return *this = wxString(ch, n); }
 756
 757     // insert another string
 758   wxString& insert(size_t nPos, const wxString& str);
 759     // insert n chars of str starting at nStart (in str)
 760   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 761     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 762
 763     // insert first n (or all if n == npos) characters of sz
 764   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 765     { return insert(nPos, wxString(sz, n)); }
 766     // insert n copies of ch
 767   wxString& insert(size_t nPos, size_t n, wxChar ch)
 768     { return insert(nPos, wxString(ch, n)); }
 769
 770     // delete characters from nStart to nStart + nLen
 771   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 772
 773     // replaces the substring of length nLen starting at nStart
 774   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 775     // replaces the substring with nCount copies of ch
 776   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 777     // replaces a substring with another substring
 778   wxString& replace(size_t nStart, size_t nLen,
 779                     const wxString& str, size_t nStart2, size_t nLen2);
 780     // replaces the substring with first nCount chars of sz
 781     wxString& replace(size_t nStart, size_t nLen,
 782                       const wxChar* sz, size_t nCount);
 783
 784     // swap two strings
 785   void swap(wxString& str);
 786
 787     // All find() functions take the nStart argument which specifies the
 788     // position to start the search on, the default value is 0. All functions
 789     // return npos if there were no match.
 790
 791     // find a substring
 792   size_t find(const wxString& str, size_t nStart = 0) const;
 793
 794   // VC++ 1.5 can't cope with this syntax.
 795 #if !defined(__VISUALC__) || defined(__WIN32__)
 796     // find first n characters of sz
 797   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 798 #endif
 799
 800   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 801 #if !defined(__BORLANDC__)
 802     // find the first occurence of character ch after nStart
 803   size_t find(wxChar ch, size_t nStart = 0) const;
 804 #endif
 805     // rfind() family is exactly like find() but works right to left
 806
 807     // as find, but from the end
 808   size_t rfind(const wxString& str, size_t nStart = npos) const;
 809
 810   // VC++ 1.5 can't cope with this syntax.
 811 #if !defined(__VISUALC__) || defined(__WIN32__)
 812     // as find, but from the end
 813   size_t rfind(const wxChar* sz, size_t nStart = npos,
 814           size_t n = npos) const;
 815     // as find, but from the end
 816   size_t rfind(wxChar ch, size_t nStart = npos) const;
 817 #endif
 818
 819     // find first/last occurence of any character in the set
 820
 821     // as strpbrk() but starts at nStart, returns npos if not found
 822   size_t find_first_of(const wxString& str, size_t nStart = 0) const
 823     { return find_first_of(str.c_str(), nStart); }
 824     // same as above
 825   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 826     // same as find(char, size_t)
 827   size_t find_first_of(wxChar c, size_t nStart = 0) const
 828     { return find(c, nStart); }
 829     // find the last (starting from nStart) char from str in this string
 830   size_t find_last_of (const wxString& str, size_t nStart = npos) const
 831     { return find_last_of(str.c_str(), nStart); }
 832     // same as above
 833   size_t find_last_of (const wxChar* sz, size_t nStart = npos) const;
 834     // same as above
 835   size_t find_last_of(wxChar c, size_t nStart = npos) const
 836     { return rfind(c, nStart); }
 837
 838     // find first/last occurence of any character not in the set
 839
 840     // as strspn() (starting from nStart), returns npos on failure
 841   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
 842     { return find_first_not_of(str.c_str(), nStart); }
 843     // same as above
 844   size_t find_first_not_of(const wxChar* sz, size_t nStart = 0) const;
 845     // same as above
 846   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 847     //  as strcspn()
 848   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 849     // same as above
 850   size_t find_last_not_of(const wxChar* sz, size_t nStart = npos) const;
 851     // same as above
 852   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 853
 854     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 855     // equal or greater than the compare() argument.
 856
 857     // just like strcmp()
 858   int compare(const wxString& str) const { return Cmp(str); }
 859     // comparison with a substring
 860   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 861     // comparison of 2 substrings
 862   int compare(size_t nStart, size_t nLen,
 863               const wxString& str, size_t nStart2, size_t nLen2) const;
 864     // just like strcmp()
 865   int compare(const wxChar* sz) const { return Cmp(sz); }
 866     // substring comparison with first nCount characters of sz
 867   int compare(size_t nStart, size_t nLen,
 868               const wxChar* sz, size_t nCount = npos) const;
 869
 870   // substring extraction
 871   wxString substr(size_t nStart = 0, size_t nLen = npos) const
 872     { return Mid(nStart, nLen); }
 873 #endif // wxSTD_STRING_COMPATIBILITY
 874 };
 875
 876 // ----------------------------------------------------------------------------
 877 // The string array uses it's knowledge of internal structure of the wxString
 878 // class to optimize string storage. Normally, we would store pointers to
 879 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 880 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 881 // really all we need to turn such pointer into a string!
 882 //
 883 // Of course, it can be called a dirty hack, but we use twice less memory and
 884 // this approach is also more speed efficient, so it's probably worth it.
 885 //
 886 // Usage notes: when a string is added/inserted, a new copy of it is created,
 887 // so the original string may be safely deleted. When a string is retrieved
 888 // from the array (operator[] or Item() method), a reference is returned.
 889 // ----------------------------------------------------------------------------
 890
 891 class WXDLLEXPORT wxArrayString
 892 {
 893 public:
 894   // type of function used by wxArrayString::Sort()
 895   typedef int (*CompareFunction)(const wxString& first,
 896                                  const wxString& second);
 897
 898   // constructors and destructor
 899     // default ctor: if autoSort is TRUE, the array is always sorted (in
 900     // alphabetical order)
 901   wxArrayString(bool autoSort = FALSE);
 902     // copy ctor
 903   wxArrayString(const wxArrayString& array);
 904     // assignment operator
 905   wxArrayString& operator=(const wxArrayString& src);
 906     // not virtual, this class should not be derived from
 907  ~wxArrayString();
 908
 909   // memory management
 910     // empties the list, but doesn't release memory
 911   void Empty();
 912     // empties the list and releases memory
 913   void Clear();
 914     // preallocates memory for given number of items
 915   void Alloc(size_t nCount);
 916     // minimzes the memory usage (by freeing all extra memory)
 917   void Shrink();
 918
 919   // simple accessors
 920     // number of elements in the array
 921   size_t GetCount() const { return m_nCount; }
 922     // is it empty?
 923   bool IsEmpty() const { return m_nCount == 0; }
 924     // number of elements in the array (GetCount is preferred API)
 925   size_t Count() const { return m_nCount; }
 926
 927   // items access (range checking is done in debug version)
 928     // get item at position uiIndex
 929   wxString& Item(size_t nIndex) const
 930     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 931     // same as Item()
 932   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 933     // get last item
 934   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 935
 936   // item management
 937     // Search the element in the array, starting from the beginning if
 938     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 939     // sensitive (default). Returns index of the first item matched or
 940     // wxNOT_FOUND
 941   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 942     // add new element at the end (if the array is not sorted), return its
 943     // index
 944   size_t Add(const wxString& str);
 945     // add new element at given position
 946   void Insert(const wxString& str, size_t uiIndex);
 947     // remove first item matching this value
 948   void Remove(const wxChar *sz);
 949     // remove item by index
 950   void Remove(size_t nIndex);
 951
 952   // sorting
 953     // sort array elements in alphabetical order (or reversed alphabetical
 954     // order if reverseOrder parameter is TRUE)
 955   void Sort(bool reverseOrder = FALSE);
 956     // sort array elements using specified comparaison function
 957   void Sort(CompareFunction compareFunction);
 958
 959 protected:
 960   void Copy(const wxArrayString& src);  // copies the contents of another array
 961
 962 private:
 963   void Grow();                          // makes array bigger if needed
 964   void Free();                          // free all the strings stored
 965
 966   void DoSort();                        // common part of all Sort() variants
 967
 968   size_t  m_nSize,    // current size of the array
 969           m_nCount;   // current number of elements
 970
 971   wxChar  **m_pItems; // pointer to data
 972
 973   bool    m_autoSort; // if TRUE, keep the array always sorted
 974 };
 975
 976 class WXDLLEXPORT wxSortedArrayString : public wxArrayString
 977 {
 978 public:
 979   wxSortedArrayString() : wxArrayString(TRUE)
 980     { }
 981   wxSortedArrayString(const wxArrayString& array) : wxArrayString(TRUE)
 982     { Copy(array); }
 983 };
 984
 985 // ---------------------------------------------------------------------------
 986 // wxString comparison functions: operator versions are always case sensitive
 987 // ---------------------------------------------------------------------------
 988
 989 //
 990 inline bool operator==(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) == 0); }
 991 //
 992 inline bool operator==(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) == 0); }
 993 //
 994 inline bool operator==(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) == 0); }
 995 //
 996 inline bool operator!=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) != 0); }
 997 //
 998 inline bool operator!=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) != 0); }
 999 //
1000 inline bool operator!=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) != 0); }
1001 //
1002 inline bool operator< (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) < 0); }
1003 //
1004 inline bool operator< (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <  0); }
1005 //
1006 inline bool operator< (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >  0); }
1007 //
1008 inline bool operator> (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >  0); }
1009 //
1010 inline bool operator> (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >  0); }
1011 //
1012 inline bool operator> (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <  0); }
1013 //
1014 inline bool operator<=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) <= 0); }
1015 //
1016 inline bool operator<=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <= 0); }
1017 //
1018 inline bool operator<=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >= 0); }
1019 //
1020 inline bool operator>=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >= 0); }
1021 //
1022 inline bool operator>=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >= 0); }
1023 //
1024 inline bool operator>=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <= 0); }
1025
1026 // comparison with char
1027 inline bool operator==(wxChar c, const wxString& s) { return s.IsSameAs(c); }
1028 inline bool operator==(const wxString& s, wxChar c) { return s.IsSameAs(c); }
1029 inline bool operator!=(wxChar c, const wxString& s) { return !s.IsSameAs(c); }
1030 inline bool operator!=(const wxString& s, wxChar c) { return !s.IsSameAs(c); }
1031
1032 #if wxUSE_UNICODE
1033 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
1034     { return (s1.Cmp((const wchar_t *)s2) == 0); }
1035 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
1036     { return (s2.Cmp((const wchar_t *)s1) == 0); }
1037 #else
1038 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
1039     { return (s1.Cmp((const char *)s2) == 0); }
1040 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
1041     { return (s2.Cmp((const char *)s1) == 0); }
1042 #endif
1043
1044 wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
1045 wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
1046 wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
1047 wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
1048 wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
1049 #if wxUSE_UNICODE
1050 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxWCharBuffer& buf)
1051     { return string + (const wchar_t *)buf; }
1052 inline wxString WXDLLEXPORT operator+(const wxWCharBuffer& buf, const wxString& string)
1053     { return (const wchar_t *)buf + string; }
1054 #else
1055 inline wxString WXDLLEXPORT operator+(const wxString& string, const wxCharBuffer& buf)
1056     { return string + (const char *)buf; }
1057 inline wxString WXDLLEXPORT operator+(const wxCharBuffer& buf, const wxString& string)
1058     { return (const char *)buf + string; }
1059 #endif
1060
1061 // ---------------------------------------------------------------------------
1062 // Implementation only from here until the end of file
1063 // ---------------------------------------------------------------------------
1064
1065 // don't pollute the library user's name space
1066 #undef ASSERT_VALID_INDEX
1067
1068 #if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM
1069
1070 #include "wx/ioswrap.h"
1071
1072 WXDLLEXPORT istream& operator>>(istream&, wxString&);
1073 WXDLLEXPORT ostream& operator<<(ostream&, const wxString&);
1074
1075 #endif  // wxSTD_STRING_COMPATIBILITY
1076
1077 #endif  // _WX_WXSTRINGH__