src/common/string.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        string.cpp
   3 // Purpose:     wxString class
   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 #ifdef __GNUG__
  13 #pragma implementation "string.h"
  14 #endif
  15
  16 /*
  17  * About ref counting:
  18  *  1) all empty strings use g_strEmpty, nRefs = -1 (set in Init())
  19  *  2) AllocBuffer() sets nRefs to 1, Lock() increments it by one
  20  *  3) Unlock() decrements nRefs and frees memory if it goes to 0
  21  */
  22
  23 // ===========================================================================
  24 // headers, declarations, constants
  25 // ===========================================================================
  26
  27 // For compilers that support precompilation, includes "wx.h".
  28 #include "wx/wxprec.h"
  29
  30 #ifdef __BORLANDC__
  31 #pragma hdrstop
  32 #endif
  33
  34 #ifndef WX_PRECOMP
  35   #include "wx/defs.h"
  36   #include "wx/string.h"
  37   #include "wx/intl.h"
  38 #endif
  39
  40 #include <ctype.h>
  41 #include <string.h>
  42 #include <stdlib.h>
  43
  44 // This probably isn't right, what should it be Vadim?
  45 // Otherwise we end up with no wxVsprintf defined.
  46 #ifdef __WXMOTIF__
  47 #define HAVE_VPRINTF
  48 #endif
  49
  50 #ifdef wxUSE_WCSRTOMBS
  51   #include <wchar.h>    // for wcsrtombs(), see comments where it's used
  52 #endif // GNU
  53
  54 #ifdef  WXSTRING_IS_WXOBJECT
  55   IMPLEMENT_DYNAMIC_CLASS(wxString, wxObject)
  56 #endif  //WXSTRING_IS_WXOBJECT
  57
  58 // allocating extra space for each string consumes more memory but speeds up
  59 // the concatenation operations (nLen is the current string's length)
  60 // NB: EXTRA_ALLOC must be >= 0!
  61 #define EXTRA_ALLOC       (19 - nLen % 16)
  62
  63 // ---------------------------------------------------------------------------
  64 // static class variables definition
  65 // ---------------------------------------------------------------------------
  66
  67 #ifdef  STD_STRING_COMPATIBILITY
  68   const size_t wxString::npos = STRING_MAXLEN;
  69 #endif
  70
  71 // ----------------------------------------------------------------------------
  72 // static data
  73 // ----------------------------------------------------------------------------
  74
  75 // for an empty string, GetStringData() will return this address: this
  76 // structure has the same layout as wxStringData and it's data() method will
  77 // return the empty string (dummy pointer)
  78 static const struct
  79 {
  80   wxStringData data;
  81   char dummy;
  82 } g_strEmpty = { {-1, 0, 0}, '\0' };
  83
  84 // empty C style string: points to 'string data' byte of g_strEmpty
  85 extern const char *g_szNul = &g_strEmpty.dummy;
  86
  87 // ----------------------------------------------------------------------------
  88 // conditional compilation
  89 // ----------------------------------------------------------------------------
  90
  91 // we want to find out if the current platform supports vsnprintf()-like
  92 // function: for Unix this is done with configure, for Windows we test the
  93 // compiler explicitly.
  94 #ifdef __WXMSW__
  95     #ifdef _MSC_VER
  96         #define wxVsprintf     _vsnprintf
  97     #endif
  98 #else   // !Windows
  99     #ifdef HAVE_VSNPRINTF
 100         #define wxVsprintf       vsnprintf
 101     #endif
 102 #endif  // Windows/!Windows
 103
 104 #ifndef wxVsprintf
 105     // in this case we'll use vsprintf() (which is ANSI and thus should be
 106     // always available), but it's unsafe because it doesn't check for buffer
 107     // size - so give a warning
 108     #define wxVsprintf(buffer,len,format,argptr) vsprintf(buffer,format, argptr)
 109 #ifndef __SC__
 110     #pragma message("Using sprintf() because no snprintf()-like function defined")
 111 #endif
 112 #endif
 113
 114 // ----------------------------------------------------------------------------
 115 // global functions
 116 // ----------------------------------------------------------------------------
 117
 118 #ifdef  STD_STRING_COMPATIBILITY
 119
 120 // MS Visual C++ version 5.0 provides the new STL headers as well as the old
 121 // iostream ones.
 122 //
 123 // ATTN: you can _not_ use both of these in the same program!
 124 #if wxUSE_IOSTREAMH
 125 #include <iostream.h>
 126 #define   NAMESPACE
 127 #else
 128 #include <iostream>
 129 #  ifdef _MSC_VER
 130       using namespace std;
 131 #  endif
 132 // for msvc (bcc50+ also) you don't need these NAMESPACE defines,
 133 // using namespace std; takes care of that.
 134 #define   NAMESPACE   std::
 135 #endif
 136
 137 #ifdef __WXMSW__
 138     #ifdef _MSC_VER
 139         #define wxVsprintf     _vsnprintf
 140     #endif
 141 #else
 142     #if defined ( HAVE_VSNPRINTF )
 143         #define wxVsprintf       vsnprintf
 144     #endif
 145 #endif
 146
 147 #ifndef wxVsprintf
 148     // vsprintf() is ANSI so we can always use it, but it's unsafe!
 149     #define wxVsprintf(buffer,len,format,argptr) vsprintf(buffer,format, argptr)
 150     #pragma message("Using sprintf() because no snprintf()-like function defined")
 151 #endif
 152
 153 NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str))
 154 {
 155 #if 0
 156   int w = is.width(0);
 157   if ( is.ipfx(0) ) {
 158     NAMESPACE streambuf *sb = is.rdbuf();
 159     str.erase();
 160     while ( true ) {
 161       int ch = sb->sbumpc ();
 162       if ( ch == EOF ) {
 163         is.setstate(NAMESPACE ios::eofbit);
 164         break;
 165       }
 166       else if ( isspace(ch) ) {
 167         sb->sungetc();
 168         break;
 169       }
 170
 171       str += ch;
 172       if ( --w == 1 )
 173         break;
 174     }
 175   }
 176
 177   is.isfx();
 178   if ( str.length() == 0 )
 179     is.setstate(NAMESPACE ios::failbit);
 180 #endif
 181   return is;
 182 }
 183
 184 #endif  //std::string compatibility
 185
 186 // ----------------------------------------------------------------------------
 187 // private classes
 188 // ----------------------------------------------------------------------------
 189
 190 // this small class is used to gather statistics for performance tuning
 191 //#define WXSTRING_STATISTICS
 192 #ifdef  WXSTRING_STATISTICS
 193   class Averager
 194   {
 195   public:
 196     Averager(const char *sz) { m_sz = sz; m_nTotal = m_nCount = 0; }
 197    ~Averager()
 198    { printf("wxString: average %s = %f\n", m_sz, ((float)m_nTotal)/m_nCount); }
 199
 200     void Add(size_t n) { m_nTotal += n; m_nCount++; }
 201
 202   private:
 203     size_t m_nCount, m_nTotal;
 204     const char *m_sz;
 205   } g_averageLength("allocation size"),
 206     g_averageSummandLength("summand length"),
 207     g_averageConcatHit("hit probability in concat"),
 208     g_averageInitialLength("initial string length");
 209
 210   #define STATISTICS_ADD(av, val) g_average##av.Add(val)
 211 #else
 212   #define STATISTICS_ADD(av, val)
 213 #endif // WXSTRING_STATISTICS
 214
 215 // ===========================================================================
 216 // wxString class core
 217 // ===========================================================================
 218
 219 // ---------------------------------------------------------------------------
 220 // construction
 221 // ---------------------------------------------------------------------------
 222
 223 // constructs string of <nLength> copies of character <ch>
 224 wxString::wxString(char ch, size_t nLength)
 225 {
 226   Init();
 227
 228   if ( nLength > 0 ) {
 229     AllocBuffer(nLength);
 230
 231     wxASSERT( sizeof(char) == 1 );  // can't use memset if not
 232
 233     memset(m_pchData, ch, nLength);
 234   }
 235 }
 236
 237 // takes nLength elements of psz starting at nPos
 238 void wxString::InitWith(const char *psz, size_t nPos, size_t nLength)
 239 {
 240   Init();
 241
 242   wxASSERT( nPos <= Strlen(psz) );
 243
 244   if ( nLength == STRING_MAXLEN )
 245     nLength = Strlen(psz + nPos);
 246
 247   STATISTICS_ADD(InitialLength, nLength);
 248
 249   if ( nLength > 0 ) {
 250     // trailing '\0' is written in AllocBuffer()
 251     AllocBuffer(nLength);
 252     memcpy(m_pchData, psz + nPos, nLength*sizeof(char));
 253   }
 254 }
 255
 256 // the same as previous constructor, but for compilers using unsigned char
 257 wxString::wxString(const unsigned char* psz, size_t nLength)
 258 {
 259   InitWith((const char *)psz, 0, nLength);
 260 }
 261
 262 #ifdef  STD_STRING_COMPATIBILITY
 263
 264 // poor man's iterators are "void *" pointers
 265 wxString::wxString(const void *pStart, const void *pEnd)
 266 {
 267   InitWith((const char *)pStart, 0,
 268            (const char *)pEnd - (const char *)pStart);
 269 }
 270
 271 #endif  //std::string compatibility
 272
 273 // from wide string
 274 wxString::wxString(const wchar_t *pwz)
 275 {
 276   // first get necessary size
 277
 278   // NB: GNU libc5 wcstombs() is completely broken, don't use it (it doesn't
 279   //     honor the 3rd parameter, thus it will happily crash here).
 280 #ifdef wxUSE_WCSRTOMBS
 281   // don't know if it's really needed (or if we can pass NULL), but better safe
 282   // than quick
 283   mbstate_t mbstate;
 284   size_t nLen = wcsrtombs((char *) NULL, &pwz, 0, &mbstate);
 285 #else  // !GNU libc
 286   size_t nLen = wcstombs((char *) NULL, pwz, 0);
 287 #endif // GNU
 288
 289   // empty?
 290   if ( nLen != 0 ) {
 291     AllocBuffer(nLen);
 292     wcstombs(m_pchData, pwz, nLen);
 293   }
 294   else {
 295     Init();
 296   }
 297 }
 298
 299 // ---------------------------------------------------------------------------
 300 // memory allocation
 301 // ---------------------------------------------------------------------------
 302
 303 // allocates memory needed to store a C string of length nLen
 304 void wxString::AllocBuffer(size_t nLen)
 305 {
 306   wxASSERT( nLen >  0         );    //
 307   wxASSERT( nLen <= INT_MAX-1 );    // max size (enough room for 1 extra)
 308
 309   STATISTICS_ADD(Length, nLen);
 310
 311   // allocate memory:
 312   // 1) one extra character for '\0' termination
 313   // 2) sizeof(wxStringData) for housekeeping info
 314   wxStringData* pData = (wxStringData*)
 315     malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(char));
 316   pData->nRefs        = 1;
 317   pData->nDataLength  = nLen;
 318   pData->nAllocLength = nLen + EXTRA_ALLOC;
 319   m_pchData           = pData->data();  // data starts after wxStringData
 320   m_pchData[nLen]     = '\0';
 321 }
 322
 323 // must be called before changing this string
 324 void wxString::CopyBeforeWrite()
 325 {
 326   wxStringData* pData = GetStringData();
 327
 328   if ( pData->IsShared() ) {
 329     pData->Unlock();                // memory not freed because shared
 330     size_t nLen = pData->nDataLength;
 331     AllocBuffer(nLen);
 332     memcpy(m_pchData, pData->data(), nLen*sizeof(char));
 333   }
 334
 335   wxASSERT( !GetStringData()->IsShared() );  // we must be the only owner
 336 }
 337
 338 // must be called before replacing contents of this string
 339 void wxString::AllocBeforeWrite(size_t nLen)
 340 {
 341   wxASSERT( nLen != 0 );  // doesn't make any sense
 342
 343   // must not share string and must have enough space
 344   wxStringData* pData = GetStringData();
 345   if ( pData->IsShared() || (nLen > pData->nAllocLength) ) {
 346     // can't work with old buffer, get new one
 347     pData->Unlock();
 348     AllocBuffer(nLen);
 349   }
 350   else {
 351     // update the string length
 352     pData->nDataLength = nLen;
 353   }
 354
 355   wxASSERT( !GetStringData()->IsShared() );  // we must be the only owner
 356 }
 357
 358 // allocate enough memory for nLen characters
 359 void wxString::Alloc(size_t nLen)
 360 {
 361   wxStringData *pData = GetStringData();
 362   if ( pData->nAllocLength <= nLen ) {
 363     if ( pData->IsEmpty() ) {
 364       nLen += EXTRA_ALLOC;
 365
 366       wxStringData* pData = (wxStringData*)
 367         malloc(sizeof(wxStringData) + (nLen + 1)*sizeof(char));
 368       pData->nRefs = 1;
 369       pData->nDataLength = 0;
 370       pData->nAllocLength = nLen;
 371       m_pchData = pData->data();  // data starts after wxStringData
 372       m_pchData[0u] = '\0';
 373     }
 374     else if ( pData->IsShared() ) {
 375       pData->Unlock();                // memory not freed because shared
 376       size_t nOldLen = pData->nDataLength;
 377       AllocBuffer(nLen);
 378       memcpy(m_pchData, pData->data(), nOldLen*sizeof(char));
 379     }
 380     else {
 381       nLen += EXTRA_ALLOC;
 382
 383       wxStringData *p = (wxStringData *)
 384         realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(char));
 385
 386       if ( p == NULL ) {
 387         // @@@ what to do on memory error?
 388         return;
 389       }
 390
 391       // it's not important if the pointer changed or not (the check for this
 392       // is not faster than assigning to m_pchData in all cases)
 393       p->nAllocLength = nLen;
 394       m_pchData = p->data();
 395     }
 396   }
 397   //else: we've already got enough
 398 }
 399
 400 // shrink to minimal size (releasing extra memory)
 401 void wxString::Shrink()
 402 {
 403   wxStringData *pData = GetStringData();
 404
 405   // this variable is unused in release build, so avoid the compiler warning by
 406   // just not declaring it
 407 #ifdef __WXDEBUG__
 408   void *p =
 409 #endif
 410   realloc(pData, sizeof(wxStringData) + (pData->nDataLength + 1)*sizeof(char));
 411
 412   wxASSERT( p != NULL );  // can't free memory?
 413   wxASSERT( p == pData ); // we're decrementing the size - block shouldn't move!
 414 }
 415
 416 // get the pointer to writable buffer of (at least) nLen bytes
 417 char *wxString::GetWriteBuf(size_t nLen)
 418 {
 419   AllocBeforeWrite(nLen);
 420
 421   wxASSERT( GetStringData()->nRefs == 1 );
 422   GetStringData()->Validate(FALSE);
 423
 424   return m_pchData;
 425 }
 426
 427 // put string back in a reasonable state after GetWriteBuf
 428 void wxString::UngetWriteBuf()
 429 {
 430   GetStringData()->nDataLength = strlen(m_pchData);
 431   GetStringData()->Validate(TRUE);
 432 }
 433
 434 // ---------------------------------------------------------------------------
 435 // data access
 436 // ---------------------------------------------------------------------------
 437
 438 // all functions are inline in string.h
 439
 440 // ---------------------------------------------------------------------------
 441 // assignment operators
 442 // ---------------------------------------------------------------------------
 443
 444 // helper function: does real copy
 445 void wxString::AssignCopy(size_t nSrcLen, const char *pszSrcData)
 446 {
 447   if ( nSrcLen == 0 ) {
 448     Reinit();
 449   }
 450   else {
 451     AllocBeforeWrite(nSrcLen);
 452     memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(char));
 453     GetStringData()->nDataLength = nSrcLen;
 454     m_pchData[nSrcLen] = '\0';
 455   }
 456 }
 457
 458 // assigns one string to another
 459 wxString& wxString::operator=(const wxString& stringSrc)
 460 {
 461   wxASSERT( stringSrc.GetStringData()->IsValid() );
 462
 463   // don't copy string over itself
 464   if ( m_pchData != stringSrc.m_pchData ) {
 465     if ( stringSrc.GetStringData()->IsEmpty() ) {
 466       Reinit();
 467     }
 468     else {
 469       // adjust references
 470       GetStringData()->Unlock();
 471       m_pchData = stringSrc.m_pchData;
 472       GetStringData()->Lock();
 473     }
 474   }
 475
 476   return *this;
 477 }
 478
 479 // assigns a single character
 480 wxString& wxString::operator=(char ch)
 481 {
 482   AssignCopy(1, &ch);
 483   return *this;
 484 }
 485
 486 // assigns C string
 487 wxString& wxString::operator=(const char *psz)
 488 {
 489   AssignCopy(Strlen(psz), psz);
 490   return *this;
 491 }
 492
 493 // same as 'signed char' variant
 494 wxString& wxString::operator=(const unsigned char* psz)
 495 {
 496   *this = (const char *)psz;
 497   return *this;
 498 }
 499
 500 wxString& wxString::operator=(const wchar_t *pwz)
 501 {
 502   wxString str(pwz);
 503   *this = str;
 504   return *this;
 505 }
 506
 507 // ---------------------------------------------------------------------------
 508 // string concatenation
 509 // ---------------------------------------------------------------------------
 510
 511 // add something to this string
 512 void wxString::ConcatSelf(int nSrcLen, const char *pszSrcData)
 513 {
 514   STATISTICS_ADD(SummandLength, nSrcLen);
 515
 516   // concatenating an empty string is a NOP
 517   if ( nSrcLen > 0 ) {
 518     wxStringData *pData = GetStringData();
 519     size_t nLen = pData->nDataLength;
 520     size_t nNewLen = nLen + nSrcLen;
 521
 522     // alloc new buffer if current is too small
 523     if ( pData->IsShared() ) {
 524       STATISTICS_ADD(ConcatHit, 0);
 525
 526       // we have to allocate another buffer
 527       wxStringData* pOldData = GetStringData();
 528       AllocBuffer(nNewLen);
 529       memcpy(m_pchData, pOldData->data(), nLen*sizeof(char));
 530       pOldData->Unlock();
 531     }
 532     else if ( nNewLen > pData->nAllocLength ) {
 533       STATISTICS_ADD(ConcatHit, 0);
 534
 535       // we have to grow the buffer
 536       Alloc(nNewLen);
 537     }
 538     else {
 539       STATISTICS_ADD(ConcatHit, 1);
 540
 541       // the buffer is already big enough
 542     }
 543
 544     // should be enough space
 545     wxASSERT( nNewLen <= GetStringData()->nAllocLength );
 546
 547     // fast concatenation - all is done in our buffer
 548     memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(char));
 549
 550     m_pchData[nNewLen] = '\0';              // put terminating '\0'
 551     GetStringData()->nDataLength = nNewLen; // and fix the length
 552   }
 553   //else: the string to append was empty
 554 }
 555
 556 /*
 557  * concatenation functions come in 5 flavours:
 558  *  string + string
 559  *  char   + string      and      string + char
 560  *  C str  + string      and      string + C str
 561  */
 562
 563 wxString operator+(const wxString& string1, const wxString& string2)
 564 {
 565   wxASSERT( string1.GetStringData()->IsValid() );
 566   wxASSERT( string2.GetStringData()->IsValid() );
 567
 568   wxString s = string1;
 569   s += string2;
 570
 571   return s;
 572 }
 573
 574 wxString operator+(const wxString& string, char ch)
 575 {
 576   wxASSERT( string.GetStringData()->IsValid() );
 577
 578   wxString s = string;
 579   s += ch;
 580
 581   return s;
 582 }
 583
 584 wxString operator+(char ch, const wxString& string)
 585 {
 586   wxASSERT( string.GetStringData()->IsValid() );
 587
 588   wxString s = ch;
 589   s += string;
 590
 591   return s;
 592 }
 593
 594 wxString operator+(const wxString& string, const char *psz)
 595 {
 596   wxASSERT( string.GetStringData()->IsValid() );
 597
 598   wxString s;
 599   s.Alloc(Strlen(psz) + string.Len());
 600   s = string;
 601   s += psz;
 602
 603   return s;
 604 }
 605
 606 wxString operator+(const char *psz, const wxString& string)
 607 {
 608   wxASSERT( string.GetStringData()->IsValid() );
 609
 610   wxString s;
 611   s.Alloc(Strlen(psz) + string.Len());
 612   s = psz;
 613   s += string;
 614
 615   return s;
 616 }
 617
 618 // ===========================================================================
 619 // other common string functions
 620 // ===========================================================================
 621
 622 // ---------------------------------------------------------------------------
 623 // simple sub-string extraction
 624 // ---------------------------------------------------------------------------
 625
 626 // helper function: clone the data attached to this string
 627 void wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const
 628 {
 629   if ( nCopyLen == 0 ) {
 630     dest.Init();
 631   }
 632   else {
 633     dest.AllocBuffer(nCopyLen);
 634     memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(char));
 635   }
 636 }
 637
 638 // extract string of length nCount starting at nFirst
 639 wxString wxString::Mid(size_t nFirst, size_t nCount) const
 640 {
 641   wxStringData *pData = GetStringData();
 642   size_t nLen = pData->nDataLength;
 643
 644   // default value of nCount is STRING_MAXLEN and means "till the end"
 645   if ( nCount == STRING_MAXLEN )
 646   {
 647     nCount = nLen - nFirst;
 648   }
 649
 650   // out-of-bounds requests return sensible things
 651   if ( nFirst + nCount > nLen )
 652   {
 653     nCount = nLen - nFirst;
 654   }
 655
 656   if ( nFirst > nLen )
 657   {
 658     // AllocCopy() will return empty string
 659     nCount = 0;
 660   }
 661
 662   wxString dest;
 663   AllocCopy(dest, nCount, nFirst);
 664
 665   return dest;
 666 }
 667
 668 // extract nCount last (rightmost) characters
 669 wxString wxString::Right(size_t nCount) const
 670 {
 671   if ( nCount > (size_t)GetStringData()->nDataLength )
 672     nCount = GetStringData()->nDataLength;
 673
 674   wxString dest;
 675   AllocCopy(dest, nCount, GetStringData()->nDataLength - nCount);
 676   return dest;
 677 }
 678
 679 // get all characters after the last occurence of ch
 680 // (returns the whole string if ch not found)
 681 wxString wxString::Right(char ch) const
 682 {
 683   wxString str;
 684   int iPos = Find(ch, TRUE);
 685   if ( iPos == NOT_FOUND )
 686     str = *this;
 687   else
 688     str = c_str() + iPos + 1;
 689
 690   return str;
 691 }
 692
 693 // extract nCount first (leftmost) characters
 694 wxString wxString::Left(size_t nCount) const
 695 {
 696   if ( nCount > (size_t)GetStringData()->nDataLength )
 697     nCount = GetStringData()->nDataLength;
 698
 699   wxString dest;
 700   AllocCopy(dest, nCount, 0);
 701   return dest;
 702 }
 703
 704 // get all characters before the first occurence of ch
 705 // (returns the whole string if ch not found)
 706 wxString wxString::Left(char ch) const
 707 {
 708   wxString str;
 709   for ( const char *pc = m_pchData; *pc != '\0' && *pc != ch; pc++ )
 710     str += *pc;
 711
 712   return str;
 713 }
 714
 715 /// get all characters before the last occurence of ch
 716 /// (returns empty string if ch not found)
 717 wxString wxString::Before(char ch) const
 718 {
 719   wxString str;
 720   int iPos = Find(ch, TRUE);
 721   if ( iPos != NOT_FOUND && iPos != 0 )
 722     str = wxString(c_str(), iPos);
 723
 724   return str;
 725 }
 726
 727 /// get all characters after the first occurence of ch
 728 /// (returns empty string if ch not found)
 729 wxString wxString::After(char ch) const
 730 {
 731   wxString str;
 732   int iPos = Find(ch);
 733   if ( iPos != NOT_FOUND )
 734     str = c_str() + iPos + 1;
 735
 736   return str;
 737 }
 738
 739 // replace first (or all) occurences of some substring with another one
 740 size_t wxString::Replace(const char *szOld, const char *szNew, bool bReplaceAll)
 741 {
 742   size_t uiCount = 0;   // count of replacements made
 743
 744   size_t uiOldLen = Strlen(szOld);
 745
 746   wxString strTemp;
 747   const char *pCurrent = m_pchData;
 748   const char *pSubstr;
 749   while ( *pCurrent != '\0' ) {
 750     pSubstr = strstr(pCurrent, szOld);
 751     if ( pSubstr == NULL ) {
 752       // strTemp is unused if no replacements were made, so avoid the copy
 753       if ( uiCount == 0 )
 754         return 0;
 755
 756       strTemp += pCurrent;    // copy the rest
 757       break;                  // exit the loop
 758     }
 759     else {
 760       // take chars before match
 761       strTemp.ConcatSelf(pSubstr - pCurrent, pCurrent);
 762       strTemp += szNew;
 763       pCurrent = pSubstr + uiOldLen;  // restart after match
 764
 765       uiCount++;
 766
 767       // stop now?
 768       if ( !bReplaceAll ) {
 769         strTemp += pCurrent;    // copy the rest
 770         break;                  // exit the loop
 771       }
 772     }
 773   }
 774
 775   // only done if there were replacements, otherwise would have returned above
 776   *this = strTemp;
 777
 778   return uiCount;
 779 }
 780
 781 bool wxString::IsAscii() const
 782 {
 783   const char *s = (const char*) *this;
 784   while(*s){
 785     if(!isascii(*s)) return(FALSE);
 786     s++;
 787   }
 788   return(TRUE);
 789 }
 790
 791 bool wxString::IsWord() const
 792 {
 793   const char *s = (const char*) *this;
 794   while(*s){
 795     if(!isalpha(*s)) return(FALSE);
 796     s++;
 797   }
 798   return(TRUE);
 799 }
 800
 801 bool wxString::IsNumber() const
 802 {
 803   const char *s = (const char*) *this;
 804   while(*s){
 805     if(!isdigit(*s)) return(FALSE);
 806     s++;
 807   }
 808   return(TRUE);
 809 }
 810
 811 wxString wxString::Strip(stripType w) const
 812 {
 813     wxString s = *this;
 814     if ( w & leading ) s.Trim(FALSE);
 815     if ( w & trailing ) s.Trim(TRUE);
 816     return s;
 817 }
 818
 819 // ---------------------------------------------------------------------------
 820 // case conversion
 821 // ---------------------------------------------------------------------------
 822
 823 wxString& wxString::MakeUpper()
 824 {
 825   CopyBeforeWrite();
 826
 827   for ( char *p = m_pchData; *p; p++ )
 828     *p = (char)toupper(*p);
 829
 830   return *this;
 831 }
 832
 833 wxString& wxString::MakeLower()
 834 {
 835   CopyBeforeWrite();
 836
 837   for ( char *p = m_pchData; *p; p++ )
 838     *p = (char)tolower(*p);
 839
 840   return *this;
 841 }
 842
 843 // ---------------------------------------------------------------------------
 844 // trimming and padding
 845 // ---------------------------------------------------------------------------
 846
 847 // trims spaces (in the sense of isspace) from left or right side
 848 wxString& wxString::Trim(bool bFromRight)
 849 {
 850   // first check if we're going to modify the string at all
 851   if ( !IsEmpty() &&
 852        (
 853         (bFromRight && isspace(GetChar(Len() - 1))) ||
 854         (!bFromRight && isspace(GetChar(0u)))
 855        )
 856      )
 857   {
 858     // ok, there is at least one space to trim
 859     CopyBeforeWrite();
 860
 861     if ( bFromRight )
 862     {
 863       // find last non-space character
 864       char *psz = m_pchData + GetStringData()->nDataLength - 1;
 865       while ( isspace(*psz) && (psz >= m_pchData) )
 866         psz--;
 867
 868       // truncate at trailing space start
 869       *++psz = '\0';
 870       GetStringData()->nDataLength = psz - m_pchData;
 871     }
 872     else
 873     {
 874       // find first non-space character
 875       const char *psz = m_pchData;
 876       while ( isspace(*psz) )
 877         psz++;
 878
 879       // fix up data and length
 880       int nDataLength = GetStringData()->nDataLength - (psz - m_pchData);
 881       memmove(m_pchData, psz, (nDataLength + 1)*sizeof(char));
 882       GetStringData()->nDataLength = nDataLength;
 883     }
 884   }
 885
 886   return *this;
 887 }
 888
 889 // adds nCount characters chPad to the string from either side
 890 wxString& wxString::Pad(size_t nCount, char chPad, bool bFromRight)
 891 {
 892   wxString s(chPad, nCount);
 893
 894   if ( bFromRight )
 895     *this += s;
 896   else
 897   {
 898     s += *this;
 899     *this = s;
 900   }
 901
 902   return *this;
 903 }
 904
 905 // truncate the string
 906 wxString& wxString::Truncate(size_t uiLen)
 907 {
 908   if ( uiLen < Len() ) {
 909     CopyBeforeWrite();
 910
 911     *(m_pchData + uiLen) = '\0';
 912     GetStringData()->nDataLength = uiLen;
 913   }
 914   //else: nothing to do, string is already short enough
 915
 916   return *this;
 917 }
 918
 919 // ---------------------------------------------------------------------------
 920 // finding (return NOT_FOUND if not found and index otherwise)
 921 // ---------------------------------------------------------------------------
 922
 923 // find a character
 924 int wxString::Find(char ch, bool bFromEnd) const
 925 {
 926   const char *psz = bFromEnd ? strrchr(m_pchData, ch) : strchr(m_pchData, ch);
 927
 928   return (psz == NULL) ? NOT_FOUND : psz - m_pchData;
 929 }
 930
 931 // find a sub-string (like strstr)
 932 int wxString::Find(const char *pszSub) const
 933 {
 934   const char *psz = strstr(m_pchData, pszSub);
 935
 936   return (psz == NULL) ? NOT_FOUND : psz - m_pchData;
 937 }
 938
 939 // ---------------------------------------------------------------------------
 940 // stream-like operators
 941 // ---------------------------------------------------------------------------
 942 wxString& wxString::operator<<(int i)
 943 {
 944     wxString res;
 945     res.Printf("%d", i);
 946
 947     return (*this) << res;
 948 }
 949
 950 wxString& wxString::operator<<(float f)
 951 {
 952     wxString res;
 953     res.Printf("%f", f);
 954
 955     return (*this) << res;
 956 }
 957
 958 wxString& wxString::operator<<(double d)
 959 {
 960     wxString res;
 961     res.Printf("%g", d);
 962
 963     return (*this) << res;
 964 }
 965
 966 // ---------------------------------------------------------------------------
 967 // formatted output
 968 // ---------------------------------------------------------------------------
 969 int wxString::Printf(const char *pszFormat, ...)
 970 {
 971   va_list argptr;
 972   va_start(argptr, pszFormat);
 973
 974   int iLen = PrintfV(pszFormat, argptr);
 975
 976   va_end(argptr);
 977
 978   return iLen;
 979 }
 980
 981 int wxString::PrintfV(const char* pszFormat, va_list argptr)
 982 {
 983   // static buffer to avoid dynamic memory allocation each time
 984   static char s_szScratch[1024];
 985
 986   // NB: wxVsprintf() may return either less than the buffer size or -1 if there
 987   //     is not enough place depending on implementation
 988   int iLen = wxVsprintf(s_szScratch, WXSIZEOF(s_szScratch), pszFormat, argptr);
 989   char *buffer;
 990   if ( iLen < (int)WXSIZEOF(s_szScratch) ) {
 991     buffer = s_szScratch;
 992   }
 993   else {
 994       int size = WXSIZEOF(s_szScratch) * 2;
 995       buffer = (char *)malloc(size);
 996       while ( buffer != NULL ) {
 997           iLen = wxVsprintf(buffer, WXSIZEOF(s_szScratch), pszFormat, argptr);
 998           if ( iLen < size ) {
 999               // ok, there was enough space
1000               break;
1001           }
1002
1003           // still not enough, double it again
1004           buffer = (char *)realloc(buffer, size *= 2);
1005       }
1006
1007       if ( !buffer ) {
1008           // out of memory
1009           return -1;
1010       }
1011   }
1012
1013   AllocBeforeWrite(iLen);
1014   strcpy(m_pchData, buffer);
1015
1016   if ( buffer != s_szScratch )
1017       free(buffer);
1018
1019   return iLen;
1020 }
1021
1022 // ----------------------------------------------------------------------------
1023 // misc other operations
1024 // ----------------------------------------------------------------------------
1025 bool wxString::Matches(const char *pszMask) const
1026 {
1027   // check char by char
1028   const char *pszTxt;
1029   for ( pszTxt = c_str(); *pszMask != '\0'; pszMask++, pszTxt++ ) {
1030     switch ( *pszMask ) {
1031       case '?':
1032         if ( *pszTxt == '\0' )
1033           return FALSE;
1034
1035         pszTxt++;
1036         pszMask++;
1037         break;
1038
1039       case '*':
1040         {
1041           // ignore special chars immediately following this one
1042           while ( *pszMask == '*' || *pszMask == '?' )
1043             pszMask++;
1044
1045           // if there is nothing more, match
1046           if ( *pszMask == '\0' )
1047             return TRUE;
1048
1049           // are there any other metacharacters in the mask?
1050           size_t uiLenMask;
1051           const char *pEndMask = strpbrk(pszMask, "*?");
1052
1053           if ( pEndMask != NULL ) {
1054             // we have to match the string between two metachars
1055             uiLenMask = pEndMask - pszMask;
1056           }
1057           else {
1058             // we have to match the remainder of the string
1059             uiLenMask = strlen(pszMask);
1060           }
1061
1062           wxString strToMatch(pszMask, uiLenMask);
1063           const char* pMatch = strstr(pszTxt, strToMatch);
1064           if ( pMatch == NULL )
1065             return FALSE;
1066
1067           // -1 to compensate "++" in the loop
1068           pszTxt = pMatch + uiLenMask - 1;
1069           pszMask += uiLenMask - 1;
1070         }
1071         break;
1072
1073       default:
1074         if ( *pszMask != *pszTxt )
1075           return FALSE;
1076         break;
1077     }
1078   }
1079
1080   // match only if nothing left
1081   return *pszTxt == '\0';
1082 }
1083
1084 // Count the number of chars
1085 int wxString::Freq(char ch) const
1086 {
1087     int count = 0;
1088     int len = Len();
1089     for (int i = 0; i < len; i++)
1090     {
1091         if (GetChar(i) == ch)
1092             count ++;
1093     }
1094     return count;
1095 }
1096
1097 // ---------------------------------------------------------------------------
1098 // standard C++ library string functions
1099 // ---------------------------------------------------------------------------
1100 #ifdef  STD_STRING_COMPATIBILITY
1101
1102 wxString& wxString::insert(size_t nPos, const wxString& str)
1103 {
1104   wxASSERT( str.GetStringData()->IsValid() );
1105   wxASSERT( nPos <= Len() );
1106
1107   if ( !str.IsEmpty() ) {
1108     wxString strTmp;
1109     char *pc = strTmp.GetWriteBuf(Len() + str.Len());
1110     strncpy(pc, c_str(), nPos);
1111     strcpy(pc + nPos, str);
1112     strcpy(pc + nPos + str.Len(), c_str() + nPos);
1113     strTmp.UngetWriteBuf();
1114     *this = strTmp;
1115   }
1116
1117   return *this;
1118 }
1119
1120 size_t wxString::find(const wxString& str, size_t nStart) const
1121 {
1122   wxASSERT( str.GetStringData()->IsValid() );
1123   wxASSERT( nStart <= Len() );
1124
1125   const char *p = strstr(c_str() + nStart, str);
1126
1127   return p == NULL ? npos : p - c_str();
1128 }
1129
1130 // VC++ 1.5 can't cope with the default argument in the header.
1131 #if ! (defined(_MSC_VER) && !defined(__WIN32__))
1132 size_t wxString::find(const char* sz, size_t nStart, size_t n) const
1133 {
1134   return find(wxString(sz, n == npos ? 0 : n), nStart);
1135 }
1136 #endif
1137
1138 // Gives a duplicate symbol (presumably a case-insensitivity problem)
1139 #if !defined(__BORLANDC__)
1140 size_t wxString::find(char ch, size_t nStart) const
1141 {
1142   wxASSERT( nStart <= Len() );
1143
1144   const char *p = strchr(c_str() + nStart, ch);
1145
1146   return p == NULL ? npos : p - c_str();
1147 }
1148 #endif
1149
1150 size_t wxString::rfind(const wxString& str, size_t nStart) const
1151 {
1152   wxASSERT( str.GetStringData()->IsValid() );
1153   wxASSERT( nStart <= Len() );
1154
1155   // # could be quicker than that
1156   const char *p = c_str() + (nStart == npos ? Len() : nStart);
1157   while ( p >= c_str() + str.Len() ) {
1158     if ( strncmp(p - str.Len(), str, str.Len()) == 0 )
1159       return p - str.Len() - c_str();
1160     p--;
1161   }
1162
1163   return npos;
1164 }
1165
1166 // VC++ 1.5 can't cope with the default argument in the header.
1167 #if ! (defined(_MSC_VER) && !defined(__WIN32__))
1168 size_t wxString::rfind(const char* sz, size_t nStart, size_t n) const
1169 {
1170   return rfind(wxString(sz, n == npos ? 0 : n), nStart);
1171 }
1172
1173 size_t wxString::rfind(char ch, size_t nStart) const
1174 {
1175   wxASSERT( nStart <= Len() );
1176
1177   const char *p = strrchr(c_str() + nStart, ch);
1178
1179   return p == NULL ? npos : p - c_str();
1180 }
1181 #endif
1182
1183 wxString wxString::substr(size_t nStart, size_t nLen) const
1184 {
1185   // npos means 'take all'
1186   if ( nLen == npos )
1187     nLen = 0;
1188
1189   wxASSERT( nStart + nLen <= Len() );
1190
1191   return wxString(c_str() + nStart, nLen == npos ? 0 : nLen);
1192 }
1193
1194 wxString& wxString::erase(size_t nStart, size_t nLen)
1195 {
1196   wxString strTmp(c_str(), nStart);
1197   if ( nLen != npos ) {
1198     wxASSERT( nStart + nLen <= Len() );
1199
1200     strTmp.append(c_str() + nStart + nLen);
1201   }
1202
1203   *this = strTmp;
1204   return *this;
1205 }
1206
1207 wxString& wxString::replace(size_t nStart, size_t nLen, const char *sz)
1208 {
1209   wxASSERT( nStart + nLen <= Strlen(sz) );
1210
1211   wxString strTmp;
1212   if ( nStart != 0 )
1213     strTmp.append(c_str(), nStart);
1214   strTmp += sz;
1215   strTmp.append(c_str() + nStart + nLen);
1216
1217   *this = strTmp;
1218   return *this;
1219 }
1220
1221 wxString& wxString::replace(size_t nStart, size_t nLen, size_t nCount, char ch)
1222 {
1223   return replace(nStart, nLen, wxString(ch, nCount));
1224 }
1225
1226 wxString& wxString::replace(size_t nStart, size_t nLen,
1227                             const wxString& str, size_t nStart2, size_t nLen2)
1228 {
1229   return replace(nStart, nLen, str.substr(nStart2, nLen2));
1230 }
1231
1232 wxString& wxString::replace(size_t nStart, size_t nLen,
1233                         const char* sz, size_t nCount)
1234 {
1235   return replace(nStart, nLen, wxString(sz, nCount));
1236 }
1237
1238 #endif  //std::string compatibility
1239
1240 // ============================================================================
1241 // ArrayString
1242 // ============================================================================
1243
1244 // size increment = max(50% of current size, ARRAY_MAXSIZE_INCREMENT)
1245 #define   ARRAY_MAXSIZE_INCREMENT       4096
1246 #ifndef   ARRAY_DEFAULT_INITIAL_SIZE    // also defined in dynarray.h
1247   #define   ARRAY_DEFAULT_INITIAL_SIZE    (16)
1248 #endif
1249
1250 #define   STRING(p)   ((wxString *)(&(p)))
1251
1252 // ctor
1253 wxArrayString::wxArrayString()
1254 {
1255   m_nSize  =
1256   m_nCount = 0;
1257   m_pItems = (char **) NULL;
1258 }
1259
1260 // copy ctor
1261 wxArrayString::wxArrayString(const wxArrayString& src)
1262 {
1263   m_nSize  =
1264   m_nCount = 0;
1265   m_pItems = (char **) NULL;
1266
1267   *this = src;
1268 }
1269
1270 // assignment operator
1271 wxArrayString& wxArrayString::operator=(const wxArrayString& src)
1272 {
1273   if ( m_nSize > 0 )
1274     Clear();
1275
1276   if ( src.m_nCount > ARRAY_DEFAULT_INITIAL_SIZE )
1277     Alloc(src.m_nCount);
1278
1279   // we can't just copy the pointers here because otherwise we would share
1280   // the strings with another array
1281   for ( size_t n = 0; n < src.m_nCount; n++ )
1282     Add(src[n]);
1283
1284   if ( m_nCount != 0 )
1285     memcpy(m_pItems, src.m_pItems, m_nCount*sizeof(char *));
1286
1287   return *this;
1288 }
1289
1290 // grow the array
1291 void wxArrayString::Grow()
1292 {
1293   // only do it if no more place
1294   if( m_nCount == m_nSize ) {
1295     if( m_nSize == 0 ) {
1296       // was empty, alloc some memory
1297       m_nSize = ARRAY_DEFAULT_INITIAL_SIZE;
1298       m_pItems = new char *[m_nSize];
1299     }
1300     else {
1301       // otherwise when it's called for the first time, nIncrement would be 0
1302       // and the array would never be expanded
1303       wxASSERT( ARRAY_DEFAULT_INITIAL_SIZE != 0 );
1304
1305       // add 50% but not too much
1306       size_t nIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE
1307                           ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1;
1308       if ( nIncrement > ARRAY_MAXSIZE_INCREMENT )
1309         nIncrement = ARRAY_MAXSIZE_INCREMENT;
1310       m_nSize += nIncrement;
1311       char **pNew = new char *[m_nSize];
1312
1313       // copy data to new location
1314       memcpy(pNew, m_pItems, m_nCount*sizeof(char *));
1315
1316       // delete old memory (but do not release the strings!)
1317       wxDELETEA(m_pItems);
1318
1319       m_pItems = pNew;
1320     }
1321   }
1322 }
1323
1324 void wxArrayString::Free()
1325 {
1326   for ( size_t n = 0; n < m_nCount; n++ ) {
1327     STRING(m_pItems[n])->GetStringData()->Unlock();
1328   }
1329 }
1330
1331 // deletes all the strings from the list
1332 void wxArrayString::Empty()
1333 {
1334   Free();
1335
1336   m_nCount = 0;
1337 }
1338
1339 // as Empty, but also frees memory
1340 void wxArrayString::Clear()
1341 {
1342   Free();
1343
1344   m_nSize  =
1345   m_nCount = 0;
1346
1347   wxDELETEA(m_pItems);
1348 }
1349
1350 // dtor
1351 wxArrayString::~wxArrayString()
1352 {
1353   Free();
1354
1355   wxDELETEA(m_pItems);
1356 }
1357
1358 // pre-allocates memory (frees the previous data!)
1359 void wxArrayString::Alloc(size_t nSize)
1360 {
1361   wxASSERT( nSize > 0 );
1362
1363   // only if old buffer was not big enough
1364   if ( nSize > m_nSize ) {
1365     Free();
1366     wxDELETEA(m_pItems);
1367     m_pItems = new char *[nSize];
1368     m_nSize  = nSize;
1369   }
1370
1371   m_nCount = 0;
1372 }
1373
1374 // searches the array for an item (forward or backwards)
1375 int wxArrayString::Index(const char *sz, bool bCase, bool bFromEnd) const
1376 {
1377   if ( bFromEnd ) {
1378     if ( m_nCount > 0 ) {
1379       size_t ui = m_nCount;
1380       do {
1381         if ( STRING(m_pItems[--ui])->IsSameAs(sz, bCase) )
1382           return ui;
1383       }
1384       while ( ui != 0 );
1385     }
1386   }
1387   else {
1388     for( size_t ui = 0; ui < m_nCount; ui++ ) {
1389       if( STRING(m_pItems[ui])->IsSameAs(sz, bCase) )
1390         return ui;
1391     }
1392   }
1393
1394   return NOT_FOUND;
1395 }
1396
1397 // add item at the end
1398 void wxArrayString::Add(const wxString& str)
1399 {
1400   wxASSERT( str.GetStringData()->IsValid() );
1401
1402   Grow();
1403
1404   // the string data must not be deleted!
1405   str.GetStringData()->Lock();
1406   m_pItems[m_nCount++] = (char *)str.c_str();
1407 }
1408
1409 // add item at the given position
1410 void wxArrayString::Insert(const wxString& str, size_t nIndex)
1411 {
1412   wxASSERT( str.GetStringData()->IsValid() );
1413
1414   wxCHECK_RET( nIndex <= m_nCount, ("bad index in wxArrayString::Insert") );
1415
1416   Grow();
1417
1418   memmove(&m_pItems[nIndex + 1], &m_pItems[nIndex],
1419           (m_nCount - nIndex)*sizeof(char *));
1420
1421   str.GetStringData()->Lock();
1422   m_pItems[nIndex] = (char *)str.c_str();
1423
1424   m_nCount++;
1425 }
1426
1427 // removes item from array (by index)
1428 void wxArrayString::Remove(size_t nIndex)
1429 {
1430   wxCHECK_RET( nIndex <= m_nCount, _("bad index in wxArrayString::Remove") );
1431
1432   // release our lock
1433   Item(nIndex).GetStringData()->Unlock();
1434
1435   memmove(&m_pItems[nIndex], &m_pItems[nIndex + 1],
1436           (m_nCount - nIndex - 1)*sizeof(char *));
1437   m_nCount--;
1438 }
1439
1440 // removes item from array (by value)
1441 void wxArrayString::Remove(const char *sz)
1442 {
1443   int iIndex = Index(sz);
1444
1445   wxCHECK_RET( iIndex != NOT_FOUND,
1446                _("removing inexistent element in wxArrayString::Remove") );
1447
1448   Remove(iIndex);
1449 }
1450
1451 // sort array elements using passed comparaison function
1452
1453 void wxArrayString::Sort(bool WXUNUSED(bCase), bool WXUNUSED(bReverse) )
1454 {
1455   //@@@@ TO DO
1456   //qsort(m_pItems, m_nCount, sizeof(char *), fCmp);
1457 }