include/wx/string.h

   1 ///////////////////////////////////////////////////////////////////////////////
   2 // Name:        string.h
   3 // Purpose:     wxString and wxArrayString classes
   4 // Author:      Vadim Zeitlin
   5 // Modified by:
   6 // Created:     29/01/98
   7 // RCS-ID:      $Id$
   8 // Copyright:   (c) 1998 Vadim Zeitlin <zeitlin@dptmaths.ens-cachan.fr>
   9 // Licence:     wxWindows license
  10 ///////////////////////////////////////////////////////////////////////////////
  11
  12 #ifndef _WX_WXSTRINGH__
  13 #define _WX_WXSTRINGH__
  14
  15 #ifdef __GNUG__
  16 #pragma interface "string.h"
  17 #endif
  18
  19 #ifdef __WXMAC__
  20     #include <ctype.h>
  21 #endif
  22
  23 #include <string.h>
  24 #include <stdio.h>
  25 #include <stdarg.h>
  26 #include <limits.h>
  27 #include <stdlib.h>
  28
  29 #ifdef HAVE_STRINGS_H
  30     #include <strings.h>    // for strcasecmp()
  31 #endif // AIX
  32
  33 #ifndef  WX_PRECOMP
  34   #include "wx/defs.h"
  35
  36   #ifdef    WXSTRING_IS_WXOBJECT
  37     #include "wx/object.h"
  38   #endif
  39 #endif // !PCH
  40
  41 #include "wx/debug.h"
  42 #include "wx/wxchar.h"
  43 #include "wx/buffer.h"
  44
  45 /*
  46     Efficient string class [more or less] compatible with MFC CString,
  47     wxWindows version 1 wxString and std::string and some handy functions
  48     missing from string.h.
  49 */
  50
  51 // ---------------------------------------------------------------------------
  52 // macros
  53 // ---------------------------------------------------------------------------
  54
  55 // compile the std::string compatibility functions if defined
  56 #define   wxSTD_STRING_COMPATIBILITY
  57
  58 // define to derive wxString from wxObject
  59 #ifdef    WXSTRING_IS_WXOBJECT
  60 #undef    WXSTRING_IS_WXOBJECT
  61 #endif
  62
  63 // maximum possible length for a string means "take all string" everywhere
  64 //  (as sizeof(StringData) is unknown here we substract 100)
  65 const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100;
  66
  67 // 'naughty' cast
  68 #define   WXSTRINGCAST (wxChar *)(const wxChar *)
  69 #define   MBSTRINGCAST (char *)(const char *)
  70 #define   WCSTRINGCAST (wchar_t *)(const wchar_t *)
  71
  72 // implementation only
  73 #define   ASSERT_VALID_INDEX(i) wxASSERT( (unsigned)(i) <= Len() )
  74
  75 // ---------------------------------------------------------------------------
  76 // Global functions complementing standard C string library replacements for
  77 // strlen() and portable strcasecmp()
  78 //---------------------------------------------------------------------------
  79 // USE wx* FUNCTIONS IN wx/wxchar.h INSTEAD - THIS IS ONLY FOR BINARY COMPATIBILITY
  80
  81 // checks whether the passed in pointer is NULL and if the string is empty
  82 inline bool WXDLLEXPORT IsEmpty(const char *p) { return (!p || !*p); }
  83
  84 // safe version of strlen() (returns 0 if passed NULL pointer)
  85 inline size_t WXDLLEXPORT Strlen(const char *psz)
  86   { return psz ? strlen(psz) : 0; }
  87
  88 // portable strcasecmp/_stricmp
  89 inline int WXDLLEXPORT Stricmp(const char *psz1, const char *psz2)
  90 {
  91 #if     defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
  92   return _stricmp(psz1, psz2);
  93 #elif     defined(__SC__)
  94   return _stricmp(psz1, psz2);
  95 #elif     defined(__SALFORDC__)
  96   return stricmp(psz1, psz2);
  97 #elif defined(__BORLANDC__)
  98   return stricmp(psz1, psz2);
  99 #elif defined(__WATCOMC__)
 100   return stricmp(psz1, psz2);
 101 #elif   defined(__UNIX__) || defined(__GNUWIN32__)
 102   return strcasecmp(psz1, psz2);
 103 #elif defined(__MWERKS__) && !defined(__INTEL__)
 104   register char c1, c2;
 105   do {
 106     c1 = tolower(*psz1++);
 107     c2 = tolower(*psz2++);
 108   } while ( c1 && (c1 == c2) );
 109
 110   return c1 - c2;
 111 #else
 112   // almost all compilers/libraries provide this function (unfortunately under
 113   // different names), that's why we don't implement our own which will surely
 114   // be more efficient than this code (uncomment to use):
 115   /*
 116     register char c1, c2;
 117     do {
 118       c1 = tolower(*psz1++);
 119       c2 = tolower(*psz2++);
 120     } while ( c1 && (c1 == c2) );
 121
 122     return c1 - c2;
 123   */
 124
 125   #error  "Please define string case-insensitive compare for your OS/compiler"
 126 #endif  // OS/compiler
 127 }
 128
 129 // ----------------------------------------------------------------------------
 130 // global data
 131 // ----------------------------------------------------------------------------
 132
 133 WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString;
 134
 135 // global pointer to empty string
 136 WXDLLEXPORT_DATA(extern const wxChar*) g_szNul;
 137
 138 // return an empty wxString
 139 class WXDLLEXPORT wxString; // not yet defined
 140 inline const wxString& wxGetEmptyString() { return *(wxString *)&g_szNul; }
 141
 142 // ---------------------------------------------------------------------------
 143 // string data prepended with some housekeeping info (used by wxString class),
 144 // is never used directly (but had to be put here to allow inlining)
 145 // ---------------------------------------------------------------------------
 146 struct WXDLLEXPORT wxStringData
 147 {
 148   int     nRefs;        // reference count
 149   size_t  nDataLength,  // actual string length
 150           nAllocLength; // allocated memory size
 151
 152   // mimics declaration 'wxChar data[nAllocLength]'
 153   wxChar* data() const { return (wxChar*)(this + 1); }
 154
 155   // empty string has a special ref count so it's never deleted
 156   bool  IsEmpty()   const { return (nRefs == -1); }
 157   bool  IsShared()  const { return (nRefs > 1);   }
 158
 159   // lock/unlock
 160   void  Lock()   { if ( !IsEmpty() ) nRefs++;                    }
 161   void  Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this);  }
 162
 163   // if we had taken control over string memory (GetWriteBuf), it's
 164   // intentionally put in invalid state
 165   void  Validate(bool b)  { nRefs = (b ? 1 : 0); }
 166   bool  IsValid() const   { return (nRefs != 0); }
 167 };
 168
 169 // ---------------------------------------------------------------------------
 170 // types of multibyte<->Unicode conversions
 171 // ---------------------------------------------------------------------------
 172 class WXDLLEXPORT wxMBConv
 173 {
 174  public:
 175   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 176   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 177   const wxWCharBuffer cMB2WC(const char *psz) const
 178     {
 179       size_t nLen = MB2WC((wchar_t *) NULL, psz, 0);
 180       wxWCharBuffer buf(nLen);
 181       MB2WC(WCSTRINGCAST buf, psz, nLen);
 182       return buf;
 183     }
 184   const wxCharBuffer cWC2MB(const wchar_t *psz) const
 185     {
 186       size_t nLen = WC2MB((char *) NULL, psz, 0);
 187       wxCharBuffer buf(nLen);
 188       WC2MB(MBSTRINGCAST buf, psz, nLen);
 189       return buf;
 190     }
 191 #if wxUSE_UNICODE
 192   const wxWCharBuffer cMB2WX(const char *psz) const { return cMB2WC(psz); }
 193   const wxCharBuffer cWX2MB(const wchar_t *psz) const { return cWC2MB(psz); }
 194   const wchar_t* cWC2WX(const wchar_t *psz) const { return psz; }
 195   const wchar_t* cMB2WC(const wchar_t *psz) const { return psz; }
 196 #else
 197   const char* cMB2WX(const char *psz) const { return psz; }
 198   const char* cWX2MB(const char *psz) const { return psz; }
 199   const wxCharBuffer cWC2WX(const wchar_t *psz) const { return cWC2MB(psz); }
 200   const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
 201 #endif
 202 };
 203 WXDLLEXPORT_DATA(extern wxMBConv) wxConv_libc;
 204
 205 #define wxANOTHER_MBCONV(type) \
 206 class type : public wxMBConv { \
 207  public: \
 208   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const; \
 209   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const; \
 210 }
 211
 212 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_file)) wxConv_file;
 213 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_UTF7)) wxConv_UTF7;
 214 WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_UTF8)) wxConv_UTF8;
 215 #if defined(__WXGTK__) && (GTK_MINOR_VERSION > 0)
 216     WXDLLEXPORT_DATA(extern wxANOTHER_MBCONV(wxMBConv_gdk)) wxConv_gdk;
 217 #endif // GTK > 1.0
 218
 219 class wxCharacterSet;
 220 class WXDLLEXPORT wxCSConv : public wxMBConv
 221 {
 222  private:
 223   wxCharacterSet *cset;
 224  public:
 225   wxCSConv(const wxChar *charset);
 226   virtual ~wxCSConv(void);
 227   virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
 228   virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
 229 };
 230
 231 WXDLLEXPORT_DATA(extern wxCSConv) wxConv_local;
 232
 233 WXDLLEXPORT_DATA(extern wxMBConv *) wxConv_current;
 234
 235 // filenames are multibyte on Unix and probably widechar on Windows?
 236 #ifdef __UNIX__
 237 #define wxMBFILES 1
 238 #else
 239 #define wxMBFILES 0
 240 #endif
 241
 242 #if wxMBFILES
 243 #define wxFNCONV(name) wxConv_file.cWX2MB(name)
 244 #define FNSTRINGCAST MBSTRINGCAST
 245 #else
 246 #define wxFNCONV(name) name
 247 #define FNSTRINGCAST WXSTRINGCAST
 248 #endif
 249
 250 // ---------------------------------------------------------------------------
 251 // This is (yet another one) String class for C++ programmers. It doesn't use
 252 // any of "advanced" C++ features (i.e. templates, exceptions, namespaces...)
 253 // thus you should be able to compile it with practicaly any C++ compiler.
 254 // This class uses copy-on-write technique, i.e. identical strings share the
 255 // same memory as long as neither of them is changed.
 256 //
 257 // This class aims to be as compatible as possible with the new standard
 258 // std::string class, but adds some additional functions and should be at
 259 // least as efficient than the standard implementation.
 260 //
 261 // Performance note: it's more efficient to write functions which take "const
 262 // String&" arguments than "const char *" if you assign the argument to
 263 // another string.
 264 //
 265 // It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5.
 266 //
 267 // To do:
 268 //  - ressource support (string tables in ressources)
 269 //  - more wide character (UNICODE) support
 270 //  - regular expressions support
 271 // ---------------------------------------------------------------------------
 272
 273 #ifdef  WXSTRING_IS_WXOBJECT
 274 class WXDLLEXPORT wxString : public wxObject
 275 {
 276     DECLARE_DYNAMIC_CLASS(wxString)
 277 #else   //WXSTRING_IS_WXOBJECT
 278 class WXDLLEXPORT wxString
 279 {
 280 #endif  //WXSTRING_IS_WXOBJECT
 281
 282 friend class WXDLLEXPORT wxArrayString;
 283
 284   // NB: special care was taken in arranging the member functions in such order
 285   //     that all inline functions can be effectively inlined, verify that all
 286   //     performace critical functions are still inlined if you change order!
 287 private:
 288   // points to data preceded by wxStringData structure with ref count info
 289   wxChar *m_pchData;
 290
 291   // accessor to string data
 292   wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; }
 293
 294   // string (re)initialization functions
 295     // initializes the string to the empty value (must be called only from
 296     // ctors, use Reinit() otherwise)
 297   void Init() { m_pchData = (wxChar *)g_szNul; }
 298     // initializaes the string with (a part of) C-string
 299   void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN);
 300     // as Init, but also frees old data
 301   void Reinit() { GetStringData()->Unlock(); Init(); }
 302
 303   // memory allocation
 304     // allocates memory for string of lenght nLen
 305   void AllocBuffer(size_t nLen);
 306     // copies data to another string
 307   void AllocCopy(wxString&, int, int) const;
 308     // effectively copies data to string
 309   void AssignCopy(size_t, const wxChar *);
 310
 311   // append a (sub)string
 312   void ConcatSelf(int nLen, const wxChar *src);
 313
 314   // functions called before writing to the string: they copy it if there
 315   // are other references to our data (should be the only owner when writing)
 316   void CopyBeforeWrite();
 317   void AllocBeforeWrite(size_t);
 318
 319   // this method is not implemented - there is _no_ conversion from int to
 320   // string, you're doing something wrong if the compiler wants to call it!
 321   //
 322   // try `s << i' or `s.Printf("%d", i)' instead
 323   wxString(int);
 324   wxString(long);
 325
 326 public:
 327   // constructors and destructor
 328     // ctor for an empty string
 329   wxString() { Init(); }
 330     // copy ctor
 331   wxString(const wxString& stringSrc)
 332   {
 333     wxASSERT( stringSrc.GetStringData()->IsValid() );
 334
 335     if ( stringSrc.IsEmpty() ) {
 336       // nothing to do for an empty string
 337       Init();
 338     }
 339     else {
 340       m_pchData = stringSrc.m_pchData;            // share same data
 341       GetStringData()->Lock();                    // => one more copy
 342     }
 343   }
 344     // string containing nRepeat copies of ch
 345   wxString(wxChar ch, size_t nRepeat = 1);
 346     // ctor takes first nLength characters from C string
 347     // (default value of wxSTRING_MAXLEN means take all the string)
 348   wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN)
 349     { InitWith(psz, 0, nLength); }
 350 #if wxUSE_UNICODE
 351     // from multibyte string
 352     // (NB: nLength is right now number of Unicode characters, not
 353     //  characters in psz! So try not to use it yet!)
 354   wxString(const char *psz, wxMBConv& conv = wxConv_libc, size_t nLength = wxSTRING_MAXLEN);
 355     // from wxWCharBuffer (i.e. return from wxGetString)
 356   wxString(const wxWCharBuffer& psz)
 357     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 358 #else
 359     // from C string (for compilers using unsigned char)
 360   wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN)
 361     { InitWith((const char*)psz, 0, nLength); }
 362     // from wide (Unicode) string
 363   wxString(const wchar_t *pwz);
 364     // from wxCharBuffer
 365   wxString(const wxCharBuffer& psz)
 366     { InitWith(psz, 0, wxSTRING_MAXLEN); }
 367 #endif
 368     // dtor is not virtual, this class must not be inherited from!
 369  ~wxString() { GetStringData()->Unlock(); }
 370
 371   // generic attributes & operations
 372     // as standard strlen()
 373   size_t Len() const { return GetStringData()->nDataLength; }
 374     // string contains any characters?
 375   bool IsEmpty() const { return Len() == 0; }
 376     // empty string is "FALSE", so !str will return TRUE
 377   bool operator!() const { return IsEmpty(); }
 378     // empty string contents
 379   void Empty()
 380   {
 381     if ( !IsEmpty() )
 382       Reinit();
 383
 384     // should be empty
 385     wxASSERT( GetStringData()->nDataLength == 0 );
 386   }
 387     // empty the string and free memory
 388   void Clear()
 389   {
 390     if ( !GetStringData()->IsEmpty() )
 391       Reinit();
 392
 393     wxASSERT( GetStringData()->nDataLength == 0 );  // should be empty
 394     wxASSERT( GetStringData()->nAllocLength == 0 ); // and not own any memory
 395   }
 396
 397   // contents test
 398     // Is an ascii value
 399   bool IsAscii() const;
 400     // Is a number
 401   bool IsNumber() const;
 402     // Is a word
 403   bool IsWord() const;
 404
 405   // data access (all indexes are 0 based)
 406     // read access
 407     wxChar  GetChar(size_t n) const
 408       { ASSERT_VALID_INDEX( n );  return m_pchData[n]; }
 409     // read/write access
 410     wxChar& GetWritableChar(size_t n)
 411       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 412     // write access
 413     void  SetChar(size_t n, wxChar ch)
 414       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; }
 415
 416     // get last character
 417     wxChar  Last() const
 418       { wxASSERT( !IsEmpty() ); return m_pchData[Len() - 1]; }
 419     // get writable last character
 420     wxChar& Last()
 421       { wxASSERT( !IsEmpty() ); CopyBeforeWrite(); return m_pchData[Len()-1]; }
 422
 423     // under Unix it is tested with configure, assume it works on other
 424     // platforms (there might be overloading problems if size_t and int are
 425     // the same type)
 426 #if !defined(__UNIX__) || wxUSE_SIZE_T_STRING_OPERATOR
 427     // operator version of GetChar
 428     wxChar  operator[](size_t n) const
 429       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 430 #endif
 431
 432     // operator version of GetChar
 433     wxChar  operator[](int n) const
 434       { ASSERT_VALID_INDEX( n ); return m_pchData[n]; }
 435     // operator version of GetWritableChar
 436     wxChar& operator[](size_t n)
 437       { ASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; }
 438
 439     // implicit conversion to C string
 440     operator const wxChar*() const { return m_pchData; }
 441     // explicit conversion to C string (use this with printf()!)
 442     const wxChar* c_str()   const { return m_pchData; }
 443     //
 444     const wxChar* GetData() const { return m_pchData; }
 445 #if wxUSE_UNICODE
 446     const wxCharBuffer mb_str(wxMBConv& conv = wxConv_libc) const { return conv.cWC2MB(m_pchData); }
 447     const wxChar* wc_str(wxMBConv& WXUNUSED(conv) = wxConv_libc) const { return m_pchData; }
 448 #if wxMBFILES
 449     const wxCharBuffer fn_str() const { return mb_str(wxConv_file); }
 450 #else
 451     const wxChar* fn_str() const { return m_pchData; }
 452 #endif
 453 #else
 454     const wxChar* mb_str(wxMBConv& WXUNUSED(conv) = wxConv_libc ) const { return m_pchData; }
 455     const wxWCharBuffer wc_str(wxMBConv& conv) const { return conv.cMB2WC(m_pchData); }
 456     const wxChar* fn_str() const { return m_pchData; }
 457 #endif
 458     // for convenience
 459     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConv_current); }
 460
 461   // overloaded assignment
 462     // from another wxString
 463   wxString& operator=(const wxString& stringSrc);
 464     // from a character
 465   wxString& operator=(wxChar ch);
 466     // from a C string
 467   wxString& operator=(const wxChar *psz);
 468 #if wxUSE_UNICODE
 469     // from wxWCharBuffer
 470   wxString& operator=(const wxWCharBuffer& psz) { return operator=((const wchar_t *)psz); }
 471 #else
 472     // from another kind of C string
 473   wxString& operator=(const unsigned char* psz);
 474     // from a wide string
 475   wxString& operator=(const wchar_t *pwz);
 476     // from wxCharBuffer
 477   wxString& operator=(const wxCharBuffer& psz) { return operator=((const char *)psz); }
 478 #endif
 479
 480   // string concatenation
 481     // in place concatenation
 482     /*
 483         Concatenate and return the result. Note that the left to right
 484         associativity of << allows to write things like "str << str1 << str2
 485         << ..." (unlike with +=)
 486      */
 487       // string += string
 488   wxString& operator<<(const wxString& s)
 489   {
 490     wxASSERT( s.GetStringData()->IsValid() );
 491
 492     ConcatSelf(s.Len(), s);
 493     return *this;
 494   }
 495       // string += C string
 496   wxString& operator<<(const wxChar *psz)
 497     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 498       // string += char
 499   wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; }
 500
 501       // string += string
 502   void operator+=(const wxString& s) { (void)operator<<(s); }
 503       // string += C string
 504   void operator+=(const wxChar *psz) { (void)operator<<(psz); }
 505       // string += char
 506   void operator+=(wxChar ch) { (void)operator<<(ch); }
 507
 508       // string += buffer (i.e. from wxGetString)
 509 #if wxUSE_UNICODE
 510   wxString& operator<<(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); return *this; }
 511   void operator+=(const wxWCharBuffer& s) { (void)operator<<((const wchar_t *)s); }
 512 #else
 513   wxString& operator<<(const wxCharBuffer& s) { (void)operator<<((const char *)s); return *this; }
 514   void operator+=(const wxCharBuffer& s) { (void)operator<<((const char *)s); }
 515 #endif
 516
 517     // string += C string
 518   wxString& Append(const wxChar* psz)
 519     { ConcatSelf(wxStrlen(psz), psz); return *this; }
 520     // append count copies of given character
 521   wxString& Append(wxChar ch, size_t count = 1u)
 522     { wxString str(ch, count); return *this << str; }
 523
 524     // prepend a string, return the string itself
 525   wxString& Prepend(const wxString& str)
 526     { *this = str + *this; return *this; }
 527
 528     // non-destructive concatenation
 529       //
 530   friend wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 531       //
 532   friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 533       //
 534   friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 535       //
 536   friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 537       //
 538   friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 539
 540   // stream-like functions
 541       // insert an int into string
 542   wxString& operator<<(int i);
 543       // insert a float into string
 544   wxString& operator<<(float f);
 545       // insert a double into string
 546   wxString& operator<<(double d);
 547
 548   // string comparison
 549     // case-sensitive comparison (returns a value < 0, = 0 or > 0)
 550   int  Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); }
 551     // same as Cmp() but not case-sensitive
 552   int  CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); }
 553     // test for the string equality, either considering case or not
 554     // (if compareWithCase then the case matters)
 555   bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const
 556     { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; }
 557
 558   // simple sub-string extraction
 559       // return substring starting at nFirst of length nCount (or till the end
 560       // if nCount = default value)
 561   wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const;
 562
 563     // operator version of Mid()
 564   wxString  operator()(size_t start, size_t len) const
 565     { return Mid(start, len); }
 566
 567       // get first nCount characters
 568   wxString Left(size_t nCount) const;
 569       // get last nCount characters
 570   wxString Right(size_t nCount) const;
 571       // get all characters before the first occurence of ch
 572       // (returns the whole string if ch not found)
 573   wxString BeforeFirst(wxChar ch) const;
 574       // get all characters before the last occurence of ch
 575       // (returns empty string if ch not found)
 576   wxString BeforeLast(wxChar ch) const;
 577       // get all characters after the first occurence of ch
 578       // (returns empty string if ch not found)
 579   wxString AfterFirst(wxChar ch) const;
 580       // get all characters after the last occurence of ch
 581       // (returns the whole string if ch not found)
 582   wxString AfterLast(wxChar ch) const;
 583
 584     // for compatibility only, use more explicitly named functions above
 585   wxString Before(wxChar ch) const { return BeforeLast(ch); }
 586   wxString After(wxChar ch) const { return AfterFirst(ch); }
 587
 588   // case conversion
 589       // convert to upper case in place, return the string itself
 590   wxString& MakeUpper();
 591       // convert to upper case, return the copy of the string
 592       // Here's something to remember: BC++ doesn't like returns in inlines.
 593   wxString Upper() const ;
 594       // convert to lower case in place, return the string itself
 595   wxString& MakeLower();
 596       // convert to lower case, return the copy of the string
 597   wxString Lower() const ;
 598
 599   // trimming/padding whitespace (either side) and truncating
 600       // remove spaces from left or from right (default) side
 601   wxString& Trim(bool bFromRight = TRUE);
 602       // add nCount copies chPad in the beginning or at the end (default)
 603   wxString& Pad(size_t nCount, wxChar chPad = _T(' '), bool bFromRight = TRUE);
 604       // truncate string to given length
 605   wxString& Truncate(size_t uiLen);
 606
 607   // searching and replacing
 608       // searching (return starting index, or -1 if not found)
 609   int Find(wxChar ch, bool bFromEnd = FALSE) const;   // like strchr/strrchr
 610       // searching (return starting index, or -1 if not found)
 611   int Find(const wxChar *pszSub) const;               // like strstr
 612       // replace first (or all of bReplaceAll) occurences of substring with
 613       // another string, returns the number of replacements made
 614   size_t Replace(const wxChar *szOld,
 615                  const wxChar *szNew,
 616                  bool bReplaceAll = TRUE);
 617
 618     // check if the string contents matches a mask containing '*' and '?'
 619   bool Matches(const wxChar *szMask) const;
 620
 621   // formated input/output
 622     // as sprintf(), returns the number of characters written or < 0 on error
 623   int Printf(const wxChar *pszFormat, ...);
 624     // as vprintf(), returns the number of characters written or < 0 on error
 625   int PrintfV(const wxChar* pszFormat, va_list argptr);
 626
 627   // raw access to string memory
 628     // ensure that string has space for at least nLen characters
 629     // only works if the data of this string is not shared
 630   void Alloc(size_t nLen);
 631     // minimize the string's memory
 632     // only works if the data of this string is not shared
 633   void Shrink();
 634     // get writable buffer of at least nLen bytes. Unget() *must* be called
 635     // a.s.a.p. to put string back in a reasonable state!
 636   wxChar *GetWriteBuf(size_t nLen);
 637     // call this immediately after GetWriteBuf() has been used
 638   void UngetWriteBuf();
 639
 640   // wxWindows version 1 compatibility functions
 641
 642   // use Mid()
 643   wxString SubString(size_t from, size_t to) const
 644       { return Mid(from, (to - from + 1)); }
 645     // values for second parameter of CompareTo function
 646   enum caseCompare {exact, ignoreCase};
 647     // values for first parameter of Strip function
 648   enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
 649
 650     // use Printf()
 651   int sprintf(const wxChar *pszFormat, ...);
 652
 653     // use Cmp()
 654   inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
 655     { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
 656
 657     // use Len
 658   size_t Length() const { return Len(); }
 659     // Count the number of characters
 660   int Freq(wxChar ch) const;
 661     // use MakeLower
 662   void LowerCase() { MakeLower(); }
 663     // use MakeUpper
 664   void UpperCase() { MakeUpper(); }
 665     // use Trim except that it doesn't change this string
 666   wxString Strip(stripType w = trailing) const;
 667
 668     // use Find (more general variants not yet supported)
 669   size_t Index(const wxChar* psz) const { return Find(psz); }
 670   size_t Index(wxChar ch)         const { return Find(ch);  }
 671     // use Truncate
 672   wxString& Remove(size_t pos) { return Truncate(pos); }
 673   wxString& RemoveLast() { return Truncate(Len() - 1); }
 674
 675   wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); }
 676
 677     // use Find()
 678   int First( const wxChar ch ) const { return Find(ch); }
 679   int First( const wxChar* psz ) const { return Find(psz); }
 680   int First( const wxString &str ) const { return Find(str); }
 681   int Last( const wxChar ch ) const { return Find(ch, TRUE); }
 682   bool Contains(const wxString& str) const { return Find(str) != -1; }
 683
 684     // use IsEmpty()
 685   bool IsNull() const { return IsEmpty(); }
 686
 687 #ifdef  wxSTD_STRING_COMPATIBILITY
 688   // std::string compatibility functions
 689
 690   // an 'invalid' value for string index
 691   static const size_t npos;
 692
 693   // constructors
 694     // take nLen chars starting at nPos
 695   wxString(const wxString& str, size_t nPos, size_t nLen)
 696   {
 697     wxASSERT( str.GetStringData()->IsValid() );
 698     InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen);
 699   }
 700     // take all characters from pStart to pEnd
 701   wxString(const void *pStart, const void *pEnd);
 702
 703   // lib.string.capacity
 704     // return the length of the string
 705   size_t size() const { return Len(); }
 706     // return the length of the string
 707   size_t length() const { return Len(); }
 708     // return the maximum size of the string
 709   size_t max_size() const { return wxSTRING_MAXLEN; }
 710     // resize the string, filling the space with c if c != 0
 711   void resize(size_t nSize, wxChar ch = _T('\0'));
 712     // delete the contents of the string
 713   void clear() { Empty(); }
 714     // returns true if the string is empty
 715   bool empty() const { return IsEmpty(); }
 716
 717   // lib.string.access
 718     // return the character at position n
 719   wxChar at(size_t n) const { return GetChar(n); }
 720     // returns the writable character at position n
 721   wxChar& at(size_t n) { return GetWritableChar(n); }
 722
 723   // lib.string.modifiers
 724     // append a string
 725   wxString& append(const wxString& str)
 726     { *this += str; return *this; }
 727     // append elements str[pos], ..., str[pos+n]
 728   wxString& append(const wxString& str, size_t pos, size_t n)
 729     { ConcatSelf(n, str.c_str() + pos); return *this; }
 730     // append first n (or all if n == npos) characters of sz
 731   wxString& append(const wxChar *sz, size_t n = npos)
 732     { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; }
 733
 734     // append n copies of ch
 735   wxString& append(size_t n, wxChar ch) { return Pad(n, ch); }
 736
 737     // same as `this_string = str'
 738   wxString& assign(const wxString& str) { return (*this) = str; }
 739     // same as ` = str[pos..pos + n]
 740   wxString& assign(const wxString& str, size_t pos, size_t n)
 741     { return *this = wxString((const wxChar *)str + pos, n); }
 742     // same as `= first n (or all if n == npos) characters of sz'
 743   wxString& assign(const wxChar *sz, size_t n = npos)
 744     { return *this = wxString(sz, n); }
 745     // same as `= n copies of ch'
 746   wxString& assign(size_t n, wxChar ch)
 747     { return *this = wxString(ch, n); }
 748
 749     // insert another string
 750   wxString& insert(size_t nPos, const wxString& str);
 751     // insert n chars of str starting at nStart (in str)
 752   wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
 753     { return insert(nPos, wxString((const wxChar *)str + nStart, n)); }
 754
 755     // insert first n (or all if n == npos) characters of sz
 756   wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos)
 757     { return insert(nPos, wxString(sz, n)); }
 758     // insert n copies of ch
 759   wxString& insert(size_t nPos, size_t n, wxChar ch)
 760     { return insert(nPos, wxString(ch, n)); }
 761
 762     // delete characters from nStart to nStart + nLen
 763   wxString& erase(size_t nStart = 0, size_t nLen = npos);
 764
 765     // replaces the substring of length nLen starting at nStart
 766   wxString& replace(size_t nStart, size_t nLen, const wxChar* sz);
 767     // replaces the substring with nCount copies of ch
 768   wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch);
 769     // replaces a substring with another substring
 770   wxString& replace(size_t nStart, size_t nLen,
 771                     const wxString& str, size_t nStart2, size_t nLen2);
 772     // replaces the substring with first nCount chars of sz
 773     wxString& replace(size_t nStart, size_t nLen,
 774                       const wxChar* sz, size_t nCount);
 775
 776     // swap two strings
 777   void swap(wxString& str);
 778
 779     // All find() functions take the nStart argument which specifies the
 780     // position to start the search on, the default value is 0. All functions
 781     // return npos if there were no match.
 782
 783     // find a substring
 784   size_t find(const wxString& str, size_t nStart = 0) const;
 785
 786   // VC++ 1.5 can't cope with this syntax.
 787 #if !defined(__VISUALC__) || defined(__WIN32__)
 788     // find first n characters of sz
 789   size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const;
 790 #endif
 791
 792   // Gives a duplicate symbol (presumably a case-insensitivity problem)
 793 #if !defined(__BORLANDC__)
 794     // find the first occurence of character ch after nStart
 795   size_t find(wxChar ch, size_t nStart = 0) const;
 796 #endif
 797     // rfind() family is exactly like find() but works right to left
 798
 799     // as find, but from the end
 800   size_t rfind(const wxString& str, size_t nStart = npos) const;
 801
 802   // VC++ 1.5 can't cope with this syntax.
 803 #if !defined(__VISUALC__) || defined(__WIN32__)
 804     // as find, but from the end
 805   size_t rfind(const wxChar* sz, size_t nStart = npos,
 806           size_t n = npos) const;
 807     // as find, but from the end
 808   size_t rfind(wxChar ch, size_t nStart = npos) const;
 809 #endif
 810
 811     // find first/last occurence of any character in the set
 812
 813     //
 814   size_t find_first_of(const wxString& str, size_t nStart = 0) const;
 815     //
 816   size_t find_first_of(const wxChar* sz, size_t nStart = 0) const;
 817     // same as find(char, size_t)
 818   size_t find_first_of(wxChar c, size_t nStart = 0) const;
 819     //
 820   size_t find_last_of (const wxString& str, size_t nStart = npos) const;
 821     //
 822   size_t find_last_of (const wxChar* s, size_t nStart = npos) const;
 823     // same as rfind(char, size_t)
 824   size_t find_last_of (wxChar c, size_t nStart = npos) const;
 825
 826     // find first/last occurence of any character not in the set
 827
 828     //
 829   size_t find_first_not_of(const wxString& str, size_t nStart = 0) const;
 830     //
 831   size_t find_first_not_of(const wxChar* s, size_t nStart = 0) const;
 832     //
 833   size_t find_first_not_of(wxChar ch, size_t nStart = 0) const;
 834     //
 835   size_t find_last_not_of(const wxString& str, size_t nStart=npos) const;
 836     //
 837   size_t find_last_not_of(const wxChar* s, size_t nStart = npos) const;
 838     //
 839   size_t find_last_not_of(wxChar ch, size_t nStart = npos) const;
 840
 841     // All compare functions return -1, 0 or 1 if the [sub]string is less,
 842     // equal or greater than the compare() argument.
 843
 844     // just like strcmp()
 845   int compare(const wxString& str) const { return Cmp(str); }
 846     // comparison with a substring
 847   int compare(size_t nStart, size_t nLen, const wxString& str) const;
 848     // comparison of 2 substrings
 849   int compare(size_t nStart, size_t nLen,
 850               const wxString& str, size_t nStart2, size_t nLen2) const;
 851     // just like strcmp()
 852   int compare(const wxChar* sz) const { return Cmp(sz); }
 853     // substring comparison with first nCount characters of sz
 854   int compare(size_t nStart, size_t nLen,
 855               const wxChar* sz, size_t nCount = npos) const;
 856
 857   // substring extraction
 858   wxString substr(size_t nStart = 0, size_t nLen = npos) const;
 859 #endif // wxSTD_STRING_COMPATIBILITY
 860 };
 861
 862 // ----------------------------------------------------------------------------
 863 // The string array uses it's knowledge of internal structure of the wxString
 864 // class to optimize string storage. Normally, we would store pointers to
 865 // string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is
 866 // sizeof(char *)) we store these pointers instead. The cast to "wxString *" is
 867 // really all we need to turn such pointer into a string!
 868 //
 869 // Of course, it can be called a dirty hack, but we use twice less memory and
 870 // this approach is also more speed efficient, so it's probably worth it.
 871 //
 872 // Usage notes: when a string is added/inserted, a new copy of it is created,
 873 // so the original string may be safely deleted. When a string is retrieved
 874 // from the array (operator[] or Item() method), a reference is returned.
 875 // ----------------------------------------------------------------------------
 876 class WXDLLEXPORT wxArrayString
 877 {
 878 public:
 879   // type of function used by wxArrayString::Sort()
 880   typedef int (*CompareFunction)(const wxString& first,
 881                                  const wxString& second);
 882
 883   // constructors and destructor
 884     // default ctor
 885   wxArrayString();
 886     // copy ctor
 887   wxArrayString(const wxArrayString& array);
 888     // assignment operator
 889   wxArrayString& operator=(const wxArrayString& src);
 890     // not virtual, this class should not be derived from
 891  ~wxArrayString();
 892
 893   // memory management
 894     // empties the list, but doesn't release memory
 895   void Empty();
 896     // empties the list and releases memory
 897   void Clear();
 898     // preallocates memory for given number of items
 899   void Alloc(size_t nCount);
 900     // minimzes the memory usage (by freeing all extra memory)
 901   void Shrink();
 902
 903   // simple accessors
 904     // number of elements in the array
 905   size_t GetCount() const { return m_nCount; }
 906     // is it empty?
 907   bool IsEmpty() const { return m_nCount == 0; }
 908     // number of elements in the array (GetCount is preferred API)
 909   size_t Count() const { return m_nCount; }
 910
 911   // items access (range checking is done in debug version)
 912     // get item at position uiIndex
 913   wxString& Item(size_t nIndex) const
 914     { wxASSERT( nIndex < m_nCount ); return *(wxString *)&(m_pItems[nIndex]); }
 915     // same as Item()
 916   wxString& operator[](size_t nIndex) const { return Item(nIndex); }
 917     // get last item
 918   wxString& Last() const { wxASSERT( !IsEmpty() ); return Item(Count() - 1); }
 919
 920   // item management
 921     // Search the element in the array, starting from the beginning if
 922     // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case
 923     // sensitive (default). Returns index of the first item matched or
 924     // wxNOT_FOUND
 925   int  Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const;
 926     // add new element at the end
 927   void Add(const wxString& str);
 928     // add new element at given position
 929   void Insert(const wxString& str, size_t uiIndex);
 930     // remove first item matching this value
 931   void Remove(const wxChar *sz);
 932     // remove item by index
 933   void Remove(size_t nIndex);
 934
 935   // sorting
 936     // sort array elements in alphabetical order (or reversed alphabetical
 937     // order if reverseOrder parameter is TRUE)
 938   void Sort(bool reverseOrder = FALSE);
 939     // sort array elements using specified comparaison function
 940   void Sort(CompareFunction compareFunction);
 941
 942 private:
 943   void    Grow();     // makes array bigger if needed
 944   void    Free();     // free the string stored
 945
 946   void    DoSort();   // common part of all Sort() variants
 947
 948   size_t  m_nSize,    // current size of the array
 949           m_nCount;   // current number of elements
 950
 951   wxChar  **m_pItems;   // pointer to data
 952 };
 953
 954 // ---------------------------------------------------------------------------
 955 // wxString comparison functions: operator versions are always case sensitive
 956 // ---------------------------------------------------------------------------
 957 //
 958 inline bool operator==(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) == 0); }
 959 //
 960 inline bool operator==(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) == 0); }
 961 //
 962 inline bool operator==(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) == 0); }
 963 //
 964 inline bool operator!=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) != 0); }
 965 //
 966 inline bool operator!=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) != 0); }
 967 //
 968 inline bool operator!=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) != 0); }
 969 //
 970 inline bool operator< (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) < 0); }
 971 //
 972 inline bool operator< (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <  0); }
 973 //
 974 inline bool operator< (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >  0); }
 975 //
 976 inline bool operator> (const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >  0); }
 977 //
 978 inline bool operator> (const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >  0); }
 979 //
 980 inline bool operator> (const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <  0); }
 981 //
 982 inline bool operator<=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) <= 0); }
 983 //
 984 inline bool operator<=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) <= 0); }
 985 //
 986 inline bool operator<=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) >= 0); }
 987 //
 988 inline bool operator>=(const wxString& s1, const wxString& s2) { return (s1.Cmp(s2) >= 0); }
 989 //
 990 inline bool operator>=(const wxString& s1, const wxChar  * s2) { return (s1.Cmp(s2) >= 0); }
 991 //
 992 inline bool operator>=(const wxChar  * s1, const wxString& s2) { return (s2.Cmp(s1) <= 0); }
 993
 994 wxString WXDLLEXPORT operator+(const wxString& string1,  const wxString& string2);
 995 wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch);
 996 wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string);
 997 wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz);
 998 wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string);
 999
1000 // ---------------------------------------------------------------------------
1001 // Implementation only from here until the end of file
1002 // ---------------------------------------------------------------------------
1003
1004 #ifdef wxSTD_STRING_COMPATIBILITY
1005
1006 #include "wx/ioswrap.h"
1007
1008 WXDLLEXPORT istream& operator>>(istream& is, wxString& str);
1009
1010 #endif  // wxSTD_STRING_COMPATIBILITY
1011
1012 #endif  // _WX_WXSTRINGH__