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