]> git.saurik.com Git - wxWidgets.git/blob - src/common/string.cpp
unicode support
[wxWidgets.git] / 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 licence
10 /////////////////////////////////////////////////////////////////////////////
11
12 #if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
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 #include "wx/thread.h"
39 #endif
40
41 #include <ctype.h>
42 #include <string.h>
43 #include <stdlib.h>
44
45 #ifdef __SALFORDC__
46 #include <clib.h>
47 #endif
48
49 // allocating extra space for each string consumes more memory but speeds up
50 // the concatenation operations (nLen is the current string's length)
51 // NB: EXTRA_ALLOC must be >= 0!
52 #define EXTRA_ALLOC (19 - nLen % 16)
53
54 // ---------------------------------------------------------------------------
55 // static class variables definition
56 // ---------------------------------------------------------------------------
57
58 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
59 // must define this static for VA or else you get multiply defined symbols
60 // everywhere
61 const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100;
62 #endif // Visual Age
63
64 #if !wxUSE_STL
65 const size_t wxStringBase::npos = wxSTRING_MAXLEN;
66 #endif
67
68 // ----------------------------------------------------------------------------
69 // static data
70 // ----------------------------------------------------------------------------
71
72 #if wxUSE_STL
73
74 extern const wxChar WXDLLIMPEXP_BASE *wxEmptyString = _T("");
75
76 #else
77
78 // for an empty string, GetStringData() will return this address: this
79 // structure has the same layout as wxStringData and it's data() method will
80 // return the empty string (dummy pointer)
81 static const struct
82 {
83 wxStringData data;
84 wxChar dummy;
85 } g_strEmpty = { {-1, 0, 0}, wxT('\0') };
86
87 // empty C style string: points to 'string data' byte of g_strEmpty
88 extern const wxChar WXDLLIMPEXP_BASE *wxEmptyString = &g_strEmpty.dummy;
89
90 #endif
91
92 // ----------------------------------------------------------------------------
93 // global functions
94 // ----------------------------------------------------------------------------
95
96 #if wxUSE_STD_IOSTREAM
97
98 // MS Visual C++ version 5.0 provides the new STL headers as well as the old
99 // iostream ones.
100 //
101 // ATTN: you can _not_ use both of these in the same program!
102
103 #include <iostream>
104
105 wxSTD istream& operator>>(wxSTD istream& is, wxString& WXUNUSED(str))
106 {
107 #if 0
108 int w = is.width(0);
109 if ( is.ipfx(0) ) {
110 streambuf *sb = is.rdbuf();
111 str.erase();
112 while ( true ) {
113 int ch = sb->sbumpc ();
114 if ( ch == EOF ) {
115 is.setstate(ios::eofbit);
116 break;
117 }
118 else if ( isspace(ch) ) {
119 sb->sungetc();
120 break;
121 }
122
123 str += ch;
124 if ( --w == 1 )
125 break;
126 }
127 }
128
129 is.isfx();
130 if ( str.length() == 0 )
131 is.setstate(ios::failbit);
132 #endif
133 return is;
134 }
135
136 wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str)
137 {
138 os << str.c_str();
139 return os;
140 }
141
142 #endif // wxUSE_STD_IOSTREAM
143
144 // ----------------------------------------------------------------------------
145 // private classes
146 // ----------------------------------------------------------------------------
147
148 // this small class is used to gather statistics for performance tuning
149 //#define WXSTRING_STATISTICS
150 #ifdef WXSTRING_STATISTICS
151 class Averager
152 {
153 public:
154 Averager(const wxChar *sz) { m_sz = sz; m_nTotal = m_nCount = 0; }
155 ~Averager()
156 { wxPrintf("wxString: average %s = %f\n", m_sz, ((float)m_nTotal)/m_nCount); }
157
158 void Add(size_t n) { m_nTotal += n; m_nCount++; }
159
160 private:
161 size_t m_nCount, m_nTotal;
162 const wxChar *m_sz;
163 } g_averageLength("allocation size"),
164 g_averageSummandLength("summand length"),
165 g_averageConcatHit("hit probability in concat"),
166 g_averageInitialLength("initial string length");
167
168 #define STATISTICS_ADD(av, val) g_average##av.Add(val)
169 #else
170 #define STATISTICS_ADD(av, val)
171 #endif // WXSTRING_STATISTICS
172
173 #if !wxUSE_STL
174
175 // ===========================================================================
176 // wxStringData class deallocation
177 // ===========================================================================
178
179 #if defined(__VISUALC__) && defined(_MT) && !defined(_DLL)
180 # pragma message (__FILE__ ": building with Multithreaded non DLL runtime has a performance impact on wxString!")
181 void wxStringData::Free()
182 {
183 free(this);
184 }
185 #endif
186
187 // ===========================================================================
188 // wxStringBase
189 // ===========================================================================
190
191 // takes nLength elements of psz starting at nPos
192 void wxStringBase::InitWith(const wxChar *psz, size_t nPos, size_t nLength)
193 {
194 Init();
195
196 // if the length is not given, assume the string to be NUL terminated
197 if ( nLength == npos ) {
198 wxASSERT_MSG( nPos <= wxStrlen(psz), _T("index out of bounds") );
199
200 nLength = wxStrlen(psz + nPos);
201 }
202
203 STATISTICS_ADD(InitialLength, nLength);
204
205 if ( nLength > 0 ) {
206 // trailing '\0' is written in AllocBuffer()
207 if ( !AllocBuffer(nLength) ) {
208 wxFAIL_MSG( _T("out of memory in wxStringBase::InitWith") );
209 return;
210 }
211 memcpy(m_pchData, psz + nPos, nLength*sizeof(wxChar));
212 }
213 }
214
215 // poor man's iterators are "void *" pointers
216 wxStringBase::wxStringBase(const void *pStart, const void *pEnd)
217 {
218 InitWith((const wxChar *)pStart, 0,
219 (const wxChar *)pEnd - (const wxChar *)pStart);
220 }
221
222 wxStringBase::wxStringBase(size_type n, wxChar ch)
223 {
224 Init();
225 append(n, ch);
226 }
227
228 // ---------------------------------------------------------------------------
229 // memory allocation
230 // ---------------------------------------------------------------------------
231
232 // allocates memory needed to store a C string of length nLen
233 bool wxStringBase::AllocBuffer(size_t nLen)
234 {
235 // allocating 0 sized buffer doesn't make sense, all empty strings should
236 // reuse g_strEmpty
237 wxASSERT( nLen > 0 );
238
239 // make sure that we don't overflow
240 wxASSERT( nLen < (INT_MAX / sizeof(wxChar)) -
241 (sizeof(wxStringData) + EXTRA_ALLOC + 1) );
242
243 STATISTICS_ADD(Length, nLen);
244
245 // allocate memory:
246 // 1) one extra character for '\0' termination
247 // 2) sizeof(wxStringData) for housekeeping info
248 wxStringData* pData = (wxStringData*)
249 malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(wxChar));
250
251 if ( pData == NULL ) {
252 // allocation failures are handled by the caller
253 return FALSE;
254 }
255
256 pData->nRefs = 1;
257 pData->nDataLength = nLen;
258 pData->nAllocLength = nLen + EXTRA_ALLOC;
259 m_pchData = pData->data(); // data starts after wxStringData
260 m_pchData[nLen] = wxT('\0');
261 return TRUE;
262 }
263
264 // must be called before changing this string
265 bool wxStringBase::CopyBeforeWrite()
266 {
267 wxStringData* pData = GetStringData();
268
269 if ( pData->IsShared() ) {
270 pData->Unlock(); // memory not freed because shared
271 size_t nLen = pData->nDataLength;
272 if ( !AllocBuffer(nLen) ) {
273 // allocation failures are handled by the caller
274 return FALSE;
275 }
276 memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar));
277 }
278
279 wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner
280
281 return TRUE;
282 }
283
284 // must be called before replacing contents of this string
285 bool wxStringBase::AllocBeforeWrite(size_t nLen)
286 {
287 wxASSERT( nLen != 0 ); // doesn't make any sense
288
289 // must not share string and must have enough space
290 wxStringData* pData = GetStringData();
291 if ( pData->IsShared() || pData->IsEmpty() ) {
292 // can't work with old buffer, get new one
293 pData->Unlock();
294 if ( !AllocBuffer(nLen) ) {
295 // allocation failures are handled by the caller
296 return FALSE;
297 }
298 }
299 else {
300 if ( nLen > pData->nAllocLength ) {
301 // realloc the buffer instead of calling malloc() again, this is more
302 // efficient
303 STATISTICS_ADD(Length, nLen);
304
305 nLen += EXTRA_ALLOC;
306
307 pData = (wxStringData*)
308 realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar));
309
310 if ( pData == NULL ) {
311 // allocation failures are handled by the caller
312 // keep previous data since reallocation failed
313 return FALSE;
314 }
315
316 pData->nAllocLength = nLen;
317 m_pchData = pData->data();
318 }
319
320 // now we have enough space, just update the string length
321 pData->nDataLength = nLen;
322 }
323
324 wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner
325
326 return TRUE;
327 }
328
329 wxStringBase& wxStringBase::append(size_t n, wxChar ch)
330 {
331 size_type len = length();
332
333 if ( !CopyBeforeWrite() || !Alloc(len + n) ) {
334 wxFAIL_MSG( _T("out of memory in wxStringBase::append") );
335 }
336 GetStringData()->nDataLength = len + n;
337 m_pchData[len + n] = '\0';
338 for ( size_t i = 0; i < n; ++i )
339 m_pchData[len + i] = ch;
340 return *this;
341 }
342
343 void wxStringBase::resize(size_t nSize, wxChar ch)
344 {
345 size_t len = length();
346
347 if ( nSize < len )
348 {
349 erase(begin() + nSize, end());
350 }
351 else if ( nSize > len )
352 {
353 append(nSize - len, ch);
354 }
355 //else: we have exactly the specified length, nothing to do
356 }
357
358 // allocate enough memory for nLen characters
359 bool wxStringBase::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(wxChar));
368
369 if ( pData == NULL ) {
370 // allocation failure handled by caller
371 return FALSE;
372 }
373
374 pData->nRefs = 1;
375 pData->nDataLength = 0;
376 pData->nAllocLength = nLen;
377 m_pchData = pData->data(); // data starts after wxStringData
378 m_pchData[0u] = wxT('\0');
379 }
380 else if ( pData->IsShared() ) {
381 pData->Unlock(); // memory not freed because shared
382 size_t nOldLen = pData->nDataLength;
383 if ( !AllocBuffer(nLen) ) {
384 // allocation failure handled by caller
385 return FALSE;
386 }
387 memcpy(m_pchData, pData->data(), nOldLen*sizeof(wxChar));
388 }
389 else {
390 nLen += EXTRA_ALLOC;
391
392 pData = (wxStringData *)
393 realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar));
394
395 if ( pData == NULL ) {
396 // allocation failure handled by caller
397 // keep previous data since reallocation failed
398 return FALSE;
399 }
400
401 // it's not important if the pointer changed or not (the check for this
402 // is not faster than assigning to m_pchData in all cases)
403 pData->nAllocLength = nLen;
404 m_pchData = pData->data();
405 }
406 }
407 //else: we've already got enough
408 return TRUE;
409 }
410
411 wxStringBase::iterator wxStringBase::erase(iterator it)
412 {
413 size_type idx = it - begin();
414 erase(idx, 1);
415 return begin() + idx;
416 }
417
418 wxStringBase& wxStringBase::erase(size_t nStart, size_t nLen)
419 {
420 wxASSERT(nStart <= length());
421 size_t strLen = length() - nStart;
422 // delete nLen or up to the end of the string characters
423 nLen = strLen < nLen ? strLen : nLen;
424 wxString strTmp(c_str(), nStart);
425 strTmp.append(c_str() + nStart + nLen, length() - nStart - nLen);
426
427 swap(strTmp);
428 return *this;
429 }
430
431 wxStringBase& wxStringBase::insert(size_t nPos, const wxChar *sz, size_t n)
432 {
433 wxASSERT( nPos <= length() );
434
435 if ( n == npos ) n = wxStrlen(sz);
436 if ( n == 0 ) return *this;
437
438 if ( !CopyBeforeWrite() || !Alloc(length() + n) ) {
439 wxFAIL_MSG( _T("out of memory in wxStringBase::insert") );
440 }
441
442 memmove(m_pchData + nPos + n, m_pchData + nPos,
443 (length() - nPos) * sizeof(wxChar));
444 memcpy(m_pchData + nPos, sz, n * sizeof(wxChar));
445 GetStringData()->nDataLength = length() + n;
446 m_pchData[length()] = '\0';
447
448 return *this;
449 }
450
451 void wxStringBase::swap(wxStringBase& str)
452 {
453 wxChar* tmp = str.m_pchData;
454 str.m_pchData = m_pchData;
455 m_pchData = tmp;
456 }
457
458 size_t wxStringBase::find(const wxStringBase& str, size_t nStart) const
459 {
460 wxASSERT( str.GetStringData()->IsValid() );
461 wxASSERT( nStart <= length() );
462
463 const wxChar *p = wxStrstr(c_str() + nStart, str.c_str());
464
465 return p == NULL ? npos : p - c_str();
466 }
467
468 size_t wxStringBase::find(const wxChar* sz, size_t nStart, size_t n) const
469 {
470 return find(wxStringBase(sz, n), nStart);
471 }
472
473 size_t wxStringBase::find(wxChar ch, size_t nStart) const
474 {
475 wxASSERT( nStart <= length() );
476
477 const wxChar *p = wxStrchr(c_str() + nStart, ch);
478
479 return p == NULL ? npos : p - c_str();
480 }
481
482 size_t wxStringBase::rfind(const wxStringBase& str, size_t nStart) const
483 {
484 wxASSERT( str.GetStringData()->IsValid() );
485 wxASSERT( nStart == npos || nStart <= length() );
486
487 if ( length() >= str.length() )
488 {
489 // avoids a corner case later
490 if ( length() == 0 && str.length() == 0 )
491 return 0;
492
493 // "top" is the point where search starts from
494 size_t top = length() - str.length();
495
496 if ( nStart == npos )
497 nStart = length() - 1;
498 if ( nStart < top )
499 top = nStart;
500
501 const wxChar *cursor = c_str() + top;
502 do
503 {
504 if ( memcmp(cursor, str.c_str(),
505 str.length() * sizeof(wxChar)) == 0 )
506 {
507 return cursor - c_str();
508 }
509 } while ( cursor-- > c_str() );
510 }
511
512 return npos;
513 }
514
515 size_t wxStringBase::rfind(const wxChar* sz, size_t nStart, size_t n) const
516 {
517 return rfind(wxStringBase(sz, n), nStart);
518 }
519
520 size_t wxStringBase::rfind(wxChar ch, size_t nStart) const
521 {
522 if ( nStart == npos )
523 {
524 nStart = length();
525 }
526 else
527 {
528 wxASSERT( nStart <= length() );
529 }
530
531 const wxChar *actual;
532 for ( actual = c_str() + ( nStart == npos ? length() : nStart + 1 );
533 actual > c_str(); --actual )
534 {
535 if ( *(actual - 1) == ch )
536 return (actual - 1) - c_str();
537 }
538
539 return npos;
540 }
541
542 size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart) const
543 {
544 const wxChar *start = c_str() + nStart;
545 const wxChar *firstOf = wxStrpbrk(start, sz);
546 if ( firstOf )
547 return firstOf - c_str();
548 else
549 return npos;
550 }
551
552 size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart,
553 size_t n) const
554 {
555 return find_first_of(wxStringBase(sz, n), nStart);
556 }
557
558 size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart) const
559 {
560 if ( nStart == npos )
561 {
562 nStart = length() - 1;
563 }
564 else
565 {
566 wxASSERT_MSG( nStart <= length(),
567 _T("invalid index in find_last_of()") );
568 }
569
570 for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p )
571 {
572 if ( wxStrchr(sz, *p) )
573 return p - c_str();
574 }
575
576 return npos;
577 }
578
579 size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart,
580 size_t n) const
581 {
582 return find_last_of(wxStringBase(sz, n), nStart);
583 }
584
585 size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart) const
586 {
587 if ( nStart == npos )
588 {
589 nStart = length();
590 }
591 else
592 {
593 wxASSERT( nStart <= length() );
594 }
595
596 size_t nAccept = wxStrspn(c_str() + nStart, sz);
597 if ( nAccept >= length() - nStart )
598 return npos;
599 else
600 return nStart + nAccept;
601 }
602
603 size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart,
604 size_t n) const
605 {
606 return find_first_not_of(wxStringBase(sz, n), nStart);
607 }
608
609 size_t wxStringBase::find_first_not_of(wxChar ch, size_t nStart) const
610 {
611 wxASSERT( nStart <= length() );
612
613 for ( const wxChar *p = c_str() + nStart; *p; p++ )
614 {
615 if ( *p != ch )
616 return p - c_str();
617 }
618
619 return npos;
620 }
621
622 size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart) const
623 {
624 if ( nStart == npos )
625 {
626 nStart = length() - 1;
627 }
628 else
629 {
630 wxASSERT( nStart <= length() );
631 }
632
633 for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p )
634 {
635 if ( !wxStrchr(sz, *p) )
636 return p - c_str();
637 }
638
639 return npos;
640 }
641
642 size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart,
643 size_t n) const
644 {
645 return find_last_not_of(wxStringBase(sz, n), nStart);
646 }
647
648 size_t wxStringBase::find_last_not_of(wxChar ch, size_t nStart) const
649 {
650 if ( nStart == npos )
651 {
652 nStart = length() - 1;
653 }
654 else
655 {
656 wxASSERT( nStart <= length() );
657 }
658
659 for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p )
660 {
661 if ( *p != ch )
662 return p - c_str();
663 }
664
665 return npos;
666 }
667
668 wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen,
669 const wxChar *sz)
670 {
671 wxASSERT_MSG( nStart <= length(),
672 _T("index out of bounds in wxStringBase::replace") );
673 size_t strLen = length() - nStart;
674 nLen = strLen < nLen ? strLen : nLen;
675
676 wxStringBase strTmp;
677 strTmp.reserve(length()); // micro optimisation to avoid multiple mem allocs
678
679 if ( nStart != 0 )
680 strTmp.append(c_str(), nStart);
681 strTmp.append(sz);
682 strTmp.append(c_str() + nStart + nLen);
683
684 swap(strTmp);
685 return *this;
686 }
687
688 wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen,
689 size_t nCount, wxChar ch)
690 {
691 return replace(nStart, nLen, wxStringBase(ch, nCount).c_str());
692 }
693
694 wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen,
695 const wxStringBase& str,
696 size_t nStart2, size_t nLen2)
697 {
698 return replace(nStart, nLen, str.substr(nStart2, nLen2));
699 }
700
701 wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen,
702 const wxChar* sz, size_t nCount)
703 {
704 return replace(nStart, nLen, wxStringBase(sz, nCount).c_str());
705 }
706
707 wxStringBase wxStringBase::substr(size_t nStart, size_t nLen) const
708 {
709 if ( nLen == npos )
710 nLen = length() - nStart;
711 return wxStringBase(*this, nStart, nLen);
712 }
713
714 // assigns one string to another
715 wxStringBase& wxStringBase::operator=(const wxStringBase& stringSrc)
716 {
717 wxASSERT( stringSrc.GetStringData()->IsValid() );
718
719 // don't copy string over itself
720 if ( m_pchData != stringSrc.m_pchData ) {
721 if ( stringSrc.GetStringData()->IsEmpty() ) {
722 Reinit();
723 }
724 else {
725 // adjust references
726 GetStringData()->Unlock();
727 m_pchData = stringSrc.m_pchData;
728 GetStringData()->Lock();
729 }
730 }
731
732 return *this;
733 }
734
735 // assigns a single character
736 wxStringBase& wxStringBase::operator=(wxChar ch)
737 {
738 if ( !AssignCopy(1, &ch) ) {
739 wxFAIL_MSG( _T("out of memory in wxStringBase::operator=(wxChar)") );
740 }
741 return *this;
742 }
743
744 // assigns C string
745 wxStringBase& wxStringBase::operator=(const wxChar *psz)
746 {
747 if ( !AssignCopy(wxStrlen(psz), psz) ) {
748 wxFAIL_MSG( _T("out of memory in wxStringBase::operator=(const wxChar *)") );
749 }
750 return *this;
751 }
752
753 // helper function: does real copy
754 bool wxStringBase::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData)
755 {
756 if ( nSrcLen == 0 ) {
757 Reinit();
758 }
759 else {
760 if ( !AllocBeforeWrite(nSrcLen) ) {
761 // allocation failure handled by caller
762 return FALSE;
763 }
764 memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar));
765 GetStringData()->nDataLength = nSrcLen;
766 m_pchData[nSrcLen] = wxT('\0');
767 }
768 return TRUE;
769 }
770
771 // ---------------------------------------------------------------------------
772 // string concatenation
773 // ---------------------------------------------------------------------------
774
775 // add something to this string
776 bool wxStringBase::ConcatSelf(size_t nSrcLen, const wxChar *pszSrcData,
777 size_t nMaxLen)
778 {
779 STATISTICS_ADD(SummandLength, nSrcLen);
780
781 nSrcLen = nSrcLen < nMaxLen ? nSrcLen : nMaxLen;
782
783 // concatenating an empty string is a NOP
784 if ( nSrcLen > 0 ) {
785 wxStringData *pData = GetStringData();
786 size_t nLen = pData->nDataLength;
787 size_t nNewLen = nLen + nSrcLen;
788
789 // alloc new buffer if current is too small
790 if ( pData->IsShared() ) {
791 STATISTICS_ADD(ConcatHit, 0);
792
793 // we have to allocate another buffer
794 wxStringData* pOldData = GetStringData();
795 if ( !AllocBuffer(nNewLen) ) {
796 // allocation failure handled by caller
797 return FALSE;
798 }
799 memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar));
800 pOldData->Unlock();
801 }
802 else if ( nNewLen > pData->nAllocLength ) {
803 STATISTICS_ADD(ConcatHit, 0);
804
805 reserve(nNewLen);
806 // we have to grow the buffer
807 if ( capacity() < nNewLen ) {
808 // allocation failure handled by caller
809 return FALSE;
810 }
811 }
812 else {
813 STATISTICS_ADD(ConcatHit, 1);
814
815 // the buffer is already big enough
816 }
817
818 // should be enough space
819 wxASSERT( nNewLen <= GetStringData()->nAllocLength );
820
821 // fast concatenation - all is done in our buffer
822 memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar));
823
824 m_pchData[nNewLen] = wxT('\0'); // put terminating '\0'
825 GetStringData()->nDataLength = nNewLen; // and fix the length
826 }
827 //else: the string to append was empty
828 return TRUE;
829 }
830
831 // ---------------------------------------------------------------------------
832 // simple sub-string extraction
833 // ---------------------------------------------------------------------------
834
835 // helper function: clone the data attached to this string
836 bool wxStringBase::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const
837 {
838 if ( nCopyLen == 0 ) {
839 dest.Init();
840 }
841 else {
842 if ( !dest.AllocBuffer(nCopyLen) ) {
843 // allocation failure handled by caller
844 return FALSE;
845 }
846 memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(wxChar));
847 }
848 return TRUE;
849 }
850
851 #endif // !wxUSE_STL
852
853 #if !wxUSE_STL || !defined(HAVE_STD_STRING_COMPARE)
854
855 #if !wxUSE_STL
856 #define STRINGCLASS wxStringBase
857 #else
858 #define STRINGCLASS wxString
859 #endif
860
861 static inline int wxDoCmp(const wxChar* s1, size_t l1,
862 const wxChar* s2, size_t l2)
863 {
864 if( l1 == l2 )
865 return wxStrncmp(s1, s2, l1);
866 else if( l1 < l2 )
867 {
868 int ret = wxStrncmp(s1, s2, l1);
869 return ret == 0 ? -1 : ret;
870 }
871 else if( l1 > l2 )
872 {
873 int ret = wxStrncmp(s1, s2, l2);
874 return ret == 0 ? +1 : ret;
875 }
876
877 wxFAIL; // must never get there
878 return 0; // quiet compilers
879 }
880
881 #if wxUSE_STL
882
883 int STRINGCLASS::compare(const wxStringBase& str) const
884 {
885 return ::wxDoCmp(data(), length(), str.data(), str.length());
886 }
887
888 #endif
889
890 int STRINGCLASS::compare(size_t nStart, size_t nLen,
891 const wxStringBase& str) const
892 {
893 wxASSERT(nStart <= length());
894 size_type strLen = length() - nStart;
895 nLen = strLen < nLen ? strLen : nLen;
896 return ::wxDoCmp(data() + nStart, nLen, str.data(), str.length());
897 }
898
899 int STRINGCLASS::compare(size_t nStart, size_t nLen,
900 const wxStringBase& str,
901 size_t nStart2, size_t nLen2) const
902 {
903 wxASSERT(nStart <= length());
904 wxASSERT(nStart2 <= str.length());
905 size_type strLen = length() - nStart,
906 strLen2 = str.length() - nStart2;
907 nLen = strLen < nLen ? strLen : nLen;
908 nLen2 = strLen2 < nLen2 ? strLen2 : nLen2;
909 return ::wxDoCmp(data() + nStart, nLen, str.data() + nStart2, nLen2);
910 }
911
912 #if wxUSE_STL
913
914 int STRINGCLASS::compare(const wxChar* sz) const
915 {
916 size_t nLen = wxStrlen(sz);
917 return ::wxDoCmp(data(), length(), sz, nLen);
918 }
919
920 #endif
921
922 int STRINGCLASS::compare(size_t nStart, size_t nLen,
923 const wxChar* sz, size_t nCount) const
924 {
925 wxASSERT(nStart <= length());
926 size_type strLen = length() - nStart;
927 nLen = strLen < nLen ? strLen : nLen;
928 if( nCount == npos )
929 nCount = wxStrlen(sz);
930
931 return ::wxDoCmp(data() + nStart, nLen, sz, nCount);
932 }
933
934 #undef STRINGCLASS
935
936 #endif // !wxUSE_STL || !defined(HAVE_STD_STRING_COMPARE)
937
938 // ===========================================================================
939 // wxString class core
940 // ===========================================================================
941
942 // ---------------------------------------------------------------------------
943 // construction
944 // ---------------------------------------------------------------------------
945
946 #if wxUSE_UNICODE
947
948 // from multibyte string
949 wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength)
950 {
951 // first get the size of the buffer we need
952 size_t nLen;
953 if ( psz )
954 {
955 // calculate the needed size ourselves or use the provided one
956 nLen = nLength == npos ? conv.MB2WC(NULL, psz, 0) : nLength;
957 }
958 else
959 {
960 // nothing to convert
961 nLen = 0;
962 }
963
964 // anything to do?
965 if ( (nLen != 0) && (nLen != (size_t)-1) )
966 {
967 if ( !Alloc(nLen) )
968 {
969 wxFAIL_MSG( _T("out of memory in wxString::wxString") );
970 }
971 else
972 {
973 wxWCharBuffer buf(nLen + 1);
974 // MB2WC wants the buffer size, not the string length hence +1
975 nLen = conv.MB2WC(buf.data(), psz, nLen + 1);
976
977 if ( nLen != (size_t)-1 )
978 {
979 // initialized ok, set the real length as nLength specified by
980 // the caller could be greater than the real string length
981 assign(buf.data(), nLen);
982 return;
983 }
984 //else: the conversion failed -- leave the string empty (what else?)
985 }
986 }
987 }
988
989 #else // ANSI
990
991 #if wxUSE_WCHAR_T
992 // from wide string
993 wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength)
994 {
995 // first get the size of the buffer we need
996 size_t nLen;
997 if ( pwz )
998 {
999 // calculate the needed size ourselves or use the provided one
1000 nLen = nLength == npos ? conv.WC2MB(NULL, pwz, 0) : nLength;
1001 }
1002 else
1003 {
1004 // nothing to convert
1005 nLen = 0;
1006 }
1007
1008 // anything to do?
1009 if ( (nLen != 0) && (nLen != (size_t)-1) )
1010 {
1011 if ( !Alloc(nLen) )
1012 {
1013 wxFAIL_MSG( _T("out of memory in wxString::wxString") );
1014 }
1015 else
1016 {
1017 wxCharBuffer buf(nLen);
1018 // WC2MB wants the buffer size, not the string length
1019 if ( conv.WC2MB(buf.data(), pwz, nLen + 1) != (size_t)-1 )
1020 {
1021 // initialized ok
1022 assign(buf.data(), nLen);
1023 return;
1024 }
1025 //else: the conversion failed -- leave the string empty (what else?)
1026 }
1027 }
1028
1029 // leave empty
1030 }
1031 #endif // wxUSE_WCHAR_T
1032
1033 #endif // Unicode/ANSI
1034
1035 // shrink to minimal size (releasing extra memory)
1036 bool wxString::Shrink()
1037 {
1038 wxString tmp(begin(), end());
1039 swap(tmp);
1040 return tmp.length() == length();
1041 }
1042
1043 #if !wxUSE_STL
1044 // get the pointer to writable buffer of (at least) nLen bytes
1045 wxChar *wxString::GetWriteBuf(size_t nLen)
1046 {
1047 if ( !AllocBeforeWrite(nLen) ) {
1048 // allocation failure handled by caller
1049 return NULL;
1050 }
1051
1052 wxASSERT( GetStringData()->nRefs == 1 );
1053 GetStringData()->Validate(FALSE);
1054
1055 return m_pchData;
1056 }
1057
1058 // put string back in a reasonable state after GetWriteBuf
1059 void wxString::UngetWriteBuf()
1060 {
1061 GetStringData()->nDataLength = wxStrlen(m_pchData);
1062 GetStringData()->Validate(TRUE);
1063 }
1064
1065 void wxString::UngetWriteBuf(size_t nLen)
1066 {
1067 GetStringData()->nDataLength = nLen;
1068 GetStringData()->Validate(TRUE);
1069 }
1070 #endif
1071
1072 // ---------------------------------------------------------------------------
1073 // data access
1074 // ---------------------------------------------------------------------------
1075
1076 // all functions are inline in string.h
1077
1078 // ---------------------------------------------------------------------------
1079 // assignment operators
1080 // ---------------------------------------------------------------------------
1081
1082 #if !wxUSE_UNICODE
1083
1084 // same as 'signed char' variant
1085 wxString& wxString::operator=(const unsigned char* psz)
1086 {
1087 *this = (const char *)psz;
1088 return *this;
1089 }
1090
1091 #if wxUSE_WCHAR_T
1092 wxString& wxString::operator=(const wchar_t *pwz)
1093 {
1094 wxString str(pwz);
1095 swap(str);
1096 return *this;
1097 }
1098 #endif
1099
1100 #endif
1101
1102 /*
1103 * concatenation functions come in 5 flavours:
1104 * string + string
1105 * char + string and string + char
1106 * C str + string and string + C str
1107 */
1108
1109 wxString operator+(const wxString& str1, const wxString& str2)
1110 {
1111 #if !wxUSE_STL
1112 wxASSERT( str1.GetStringData()->IsValid() );
1113 wxASSERT( str2.GetStringData()->IsValid() );
1114 #endif
1115
1116 wxString s = str1;
1117 s += str2;
1118
1119 return s;
1120 }
1121
1122 wxString operator+(const wxString& str, wxChar ch)
1123 {
1124 #if !wxUSE_STL
1125 wxASSERT( str.GetStringData()->IsValid() );
1126 #endif
1127
1128 wxString s = str;
1129 s += ch;
1130
1131 return s;
1132 }
1133
1134 wxString operator+(wxChar ch, const wxString& str)
1135 {
1136 #if !wxUSE_STL
1137 wxASSERT( str.GetStringData()->IsValid() );
1138 #endif
1139
1140 wxString s = ch;
1141 s += str;
1142
1143 return s;
1144 }
1145
1146 wxString operator+(const wxString& str, const wxChar *psz)
1147 {
1148 #if !wxUSE_STL
1149 wxASSERT( str.GetStringData()->IsValid() );
1150 #endif
1151
1152 wxString s;
1153 if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) {
1154 wxFAIL_MSG( _T("out of memory in wxString::operator+") );
1155 }
1156 s = str;
1157 s += psz;
1158
1159 return s;
1160 }
1161
1162 wxString operator+(const wxChar *psz, const wxString& str)
1163 {
1164 #if !wxUSE_STL
1165 wxASSERT( str.GetStringData()->IsValid() );
1166 #endif
1167
1168 wxString s;
1169 if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) {
1170 wxFAIL_MSG( _T("out of memory in wxString::operator+") );
1171 }
1172 s = psz;
1173 s += str;
1174
1175 return s;
1176 }
1177
1178 // ===========================================================================
1179 // other common string functions
1180 // ===========================================================================
1181
1182 #if wxUSE_UNICODE
1183
1184 wxString wxString::FromAscii(const char *ascii)
1185 {
1186 if (!ascii)
1187 return wxEmptyString;
1188
1189 size_t len = strlen( ascii );
1190 wxString res;
1191
1192 if ( len )
1193 {
1194 wxStringBuffer buf(res, len);
1195
1196 wchar_t *dest = buf;
1197
1198 for ( ;; )
1199 {
1200 if ( (*dest++ = (wchar_t)(unsigned char)*ascii++) == L'\0' )
1201 break;
1202 }
1203 }
1204
1205 return res;
1206 }
1207
1208 wxString wxString::FromAscii(const char ascii)
1209 {
1210 // What do we do with '\0' ?
1211
1212 wxString res;
1213 res += (wchar_t)(unsigned char) ascii;
1214
1215 return res;
1216 }
1217
1218 const wxCharBuffer wxString::ToAscii() const
1219 {
1220 // this will allocate enough space for the terminating NUL too
1221 wxCharBuffer buffer(length());
1222
1223 signed char *dest = (signed char *)buffer.data();
1224
1225 const wchar_t *pwc = c_str();
1226 for ( ;; )
1227 {
1228 *dest++ = *pwc > SCHAR_MAX ? '_' : *pwc;
1229
1230 // the output string can't have embedded NULs anyhow, so we can safely
1231 // stop at first of them even if we do have any
1232 if ( !*pwc++ )
1233 break;
1234 }
1235
1236 return buffer;
1237 }
1238
1239 #endif // Unicode
1240
1241 // extract string of length nCount starting at nFirst
1242 wxString wxString::Mid(size_t nFirst, size_t nCount) const
1243 {
1244 size_t nLen = length();
1245
1246 // default value of nCount is npos and means "till the end"
1247 if ( nCount == npos )
1248 {
1249 nCount = nLen - nFirst;
1250 }
1251
1252 // out-of-bounds requests return sensible things
1253 if ( nFirst + nCount > nLen )
1254 {
1255 nCount = nLen - nFirst;
1256 }
1257
1258 if ( nFirst > nLen )
1259 {
1260 // AllocCopy() will return empty string
1261 nCount = 0;
1262 }
1263
1264 wxString dest(*this, nFirst, nCount);
1265 if ( dest.length() != nCount ) {
1266 wxFAIL_MSG( _T("out of memory in wxString::Mid") );
1267 }
1268
1269 return dest;
1270 }
1271
1272 // check that the string starts with prefix and return the rest of the string
1273 // in the provided pointer if it is not NULL, otherwise return FALSE
1274 bool wxString::StartsWith(const wxChar *prefix, wxString *rest) const
1275 {
1276 wxASSERT_MSG( prefix, _T("invalid parameter in wxString::StartsWith") );
1277
1278 // first check if the beginning of the string matches the prefix: note
1279 // that we don't have to check that we don't run out of this string as
1280 // when we reach the terminating NUL, either prefix string ends too (and
1281 // then it's ok) or we break out of the loop because there is no match
1282 const wxChar *p = c_str();
1283 while ( *prefix )
1284 {
1285 if ( *prefix++ != *p++ )
1286 {
1287 // no match
1288 return FALSE;
1289 }
1290 }
1291
1292 if ( rest )
1293 {
1294 // put the rest of the string into provided pointer
1295 *rest = p;
1296 }
1297
1298 return TRUE;
1299 }
1300
1301 // extract nCount last (rightmost) characters
1302 wxString wxString::Right(size_t nCount) const
1303 {
1304 if ( nCount > length() )
1305 nCount = length();
1306
1307 wxString dest(*this, length() - nCount, nCount);
1308 if ( dest.length() != nCount ) {
1309 wxFAIL_MSG( _T("out of memory in wxString::Right") );
1310 }
1311 return dest;
1312 }
1313
1314 // get all characters after the last occurence of ch
1315 // (returns the whole string if ch not found)
1316 wxString wxString::AfterLast(wxChar ch) const
1317 {
1318 wxString str;
1319 int iPos = Find(ch, TRUE);
1320 if ( iPos == wxNOT_FOUND )
1321 str = *this;
1322 else
1323 str = c_str() + iPos + 1;
1324
1325 return str;
1326 }
1327
1328 // extract nCount first (leftmost) characters
1329 wxString wxString::Left(size_t nCount) const
1330 {
1331 if ( nCount > length() )
1332 nCount = length();
1333
1334 wxString dest(*this, 0, nCount);
1335 if ( dest.length() != nCount ) {
1336 wxFAIL_MSG( _T("out of memory in wxString::Left") );
1337 }
1338 return dest;
1339 }
1340
1341 // get all characters before the first occurence of ch
1342 // (returns the whole string if ch not found)
1343 wxString wxString::BeforeFirst(wxChar ch) const
1344 {
1345 int iPos = Find(ch);
1346 if ( iPos == wxNOT_FOUND ) iPos = length();
1347 return wxString(*this, 0, iPos);
1348 }
1349
1350 /// get all characters before the last occurence of ch
1351 /// (returns empty string if ch not found)
1352 wxString wxString::BeforeLast(wxChar ch) const
1353 {
1354 wxString str;
1355 int iPos = Find(ch, TRUE);
1356 if ( iPos != wxNOT_FOUND && iPos != 0 )
1357 str = wxString(c_str(), iPos);
1358
1359 return str;
1360 }
1361
1362 /// get all characters after the first occurence of ch
1363 /// (returns empty string if ch not found)
1364 wxString wxString::AfterFirst(wxChar ch) const
1365 {
1366 wxString str;
1367 int iPos = Find(ch);
1368 if ( iPos != wxNOT_FOUND )
1369 str = c_str() + iPos + 1;
1370
1371 return str;
1372 }
1373
1374 // replace first (or all) occurences of some substring with another one
1375 size_t
1376 wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplaceAll)
1377 {
1378 // if we tried to replace an empty string we'd enter an infinite loop below
1379 wxCHECK_MSG( szOld && *szOld && szNew, 0,
1380 _T("wxString::Replace(): invalid parameter") );
1381
1382 size_t uiCount = 0; // count of replacements made
1383
1384 size_t uiOldLen = wxStrlen(szOld);
1385
1386 wxString strTemp;
1387 const wxChar *pCurrent = c_str();
1388 const wxChar *pSubstr;
1389 while ( *pCurrent != wxT('\0') ) {
1390 pSubstr = wxStrstr(pCurrent, szOld);
1391 if ( pSubstr == NULL ) {
1392 // strTemp is unused if no replacements were made, so avoid the copy
1393 if ( uiCount == 0 )
1394 return 0;
1395
1396 strTemp += pCurrent; // copy the rest
1397 break; // exit the loop
1398 }
1399 else {
1400 // take chars before match
1401 size_type len = strTemp.length();
1402 strTemp.append(pCurrent, pSubstr - pCurrent);
1403 if ( strTemp.length() != (size_t)(len + pSubstr - pCurrent) ) {
1404 wxFAIL_MSG( _T("out of memory in wxString::Replace") );
1405 return 0;
1406 }
1407 strTemp += szNew;
1408 pCurrent = pSubstr + uiOldLen; // restart after match
1409
1410 uiCount++;
1411
1412 // stop now?
1413 if ( !bReplaceAll ) {
1414 strTemp += pCurrent; // copy the rest
1415 break; // exit the loop
1416 }
1417 }
1418 }
1419
1420 // only done if there were replacements, otherwise would have returned above
1421 swap(strTemp);
1422
1423 return uiCount;
1424 }
1425
1426 bool wxString::IsAscii() const
1427 {
1428 const wxChar *s = (const wxChar*) *this;
1429 while(*s){
1430 if(!isascii(*s)) return(FALSE);
1431 s++;
1432 }
1433 return(TRUE);
1434 }
1435
1436 bool wxString::IsWord() const
1437 {
1438 const wxChar *s = (const wxChar*) *this;
1439 while(*s){
1440 if(!wxIsalpha(*s)) return(FALSE);
1441 s++;
1442 }
1443 return(TRUE);
1444 }
1445
1446 bool wxString::IsNumber() const
1447 {
1448 const wxChar *s = (const wxChar*) *this;
1449 if (wxStrlen(s))
1450 if ((s[0] == '-') || (s[0] == '+')) s++;
1451 while(*s){
1452 if(!wxIsdigit(*s)) return(FALSE);
1453 s++;
1454 }
1455 return(TRUE);
1456 }
1457
1458 wxString wxString::Strip(stripType w) const
1459 {
1460 wxString s = *this;
1461 if ( w & leading ) s.Trim(FALSE);
1462 if ( w & trailing ) s.Trim(TRUE);
1463 return s;
1464 }
1465
1466 // ---------------------------------------------------------------------------
1467 // case conversion
1468 // ---------------------------------------------------------------------------
1469
1470 wxString& wxString::MakeUpper()
1471 {
1472 for ( iterator it = begin(), en = end(); it != en; ++it )
1473 *it = (wxChar)wxToupper(*it);
1474
1475 return *this;
1476 }
1477
1478 wxString& wxString::MakeLower()
1479 {
1480 for ( iterator it = begin(), en = end(); it != en; ++it )
1481 *it = (wxChar)wxTolower(*it);
1482
1483 return *this;
1484 }
1485
1486 // ---------------------------------------------------------------------------
1487 // trimming and padding
1488 // ---------------------------------------------------------------------------
1489
1490 // some compilers (VC++ 6.0 not to name them) return TRUE for a call to
1491 // isspace('ê') in the C locale which seems to be broken to me, but we have to
1492 // live with this by checking that the character is a 7 bit one - even if this
1493 // may fail to detect some spaces (I don't know if Unicode doesn't have
1494 // space-like symbols somewhere except in the first 128 chars), it is arguably
1495 // still better than trimming away accented letters
1496 inline int wxSafeIsspace(wxChar ch) { return (ch < 127) && wxIsspace(ch); }
1497
1498 // trims spaces (in the sense of isspace) from left or right side
1499 wxString& wxString::Trim(bool bFromRight)
1500 {
1501 // first check if we're going to modify the string at all
1502 if ( !IsEmpty() &&
1503 (
1504 (bFromRight && wxSafeIsspace(GetChar(Len() - 1))) ||
1505 (!bFromRight && wxSafeIsspace(GetChar(0u)))
1506 )
1507 )
1508 {
1509 if ( bFromRight )
1510 {
1511 // find last non-space character
1512 iterator psz = begin() + length() - 1;
1513 while ( wxSafeIsspace(*psz) && (psz >= begin()) )
1514 psz--;
1515
1516 // truncate at trailing space start
1517 *++psz = wxT('\0');
1518 erase(psz, end());
1519 }
1520 else
1521 {
1522 // find first non-space character
1523 iterator psz = begin();
1524 while ( wxSafeIsspace(*psz) )
1525 psz++;
1526
1527 // fix up data and length
1528 erase(begin(), psz);
1529 }
1530 }
1531
1532 return *this;
1533 }
1534
1535 // adds nCount characters chPad to the string from either side
1536 wxString& wxString::Pad(size_t nCount, wxChar chPad, bool bFromRight)
1537 {
1538 wxString s(chPad, nCount);
1539
1540 if ( bFromRight )
1541 *this += s;
1542 else
1543 {
1544 s += *this;
1545 swap(s);
1546 }
1547
1548 return *this;
1549 }
1550
1551 // truncate the string
1552 wxString& wxString::Truncate(size_t uiLen)
1553 {
1554 if ( uiLen < Len() ) {
1555 erase(begin() + uiLen, end());
1556 }
1557 //else: nothing to do, string is already short enough
1558
1559 return *this;
1560 }
1561
1562 // ---------------------------------------------------------------------------
1563 // finding (return wxNOT_FOUND if not found and index otherwise)
1564 // ---------------------------------------------------------------------------
1565
1566 // find a character
1567 int wxString::Find(wxChar ch, bool bFromEnd) const
1568 {
1569 size_type idx = bFromEnd ? find_last_of(ch) : find_first_of(ch);
1570
1571 return (idx == npos) ? wxNOT_FOUND : (int)idx;
1572 }
1573
1574 // find a sub-string (like strstr)
1575 int wxString::Find(const wxChar *pszSub) const
1576 {
1577 size_type idx = find(pszSub);
1578
1579 return (idx == npos) ? wxNOT_FOUND : (int)idx;
1580 }
1581
1582 // ----------------------------------------------------------------------------
1583 // conversion to numbers
1584 // ----------------------------------------------------------------------------
1585
1586 bool wxString::ToLong(long *val, int base) const
1587 {
1588 wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToLong") );
1589 wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );
1590
1591 const wxChar *start = c_str();
1592 wxChar *end;
1593 *val = wxStrtol(start, &end, base);
1594
1595 // return TRUE only if scan was stopped by the terminating NUL and if the
1596 // string was not empty to start with
1597 return !*end && (end != start);
1598 }
1599
1600 bool wxString::ToULong(unsigned long *val, int base) const
1601 {
1602 wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToULong") );
1603 wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );
1604
1605 const wxChar *start = c_str();
1606 wxChar *end;
1607 *val = wxStrtoul(start, &end, base);
1608
1609 // return TRUE only if scan was stopped by the terminating NUL and if the
1610 // string was not empty to start with
1611 return !*end && (end != start);
1612 }
1613
1614 bool wxString::ToDouble(double *val) const
1615 {
1616 wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToDouble") );
1617
1618 const wxChar *start = c_str();
1619 wxChar *end;
1620 *val = wxStrtod(start, &end);
1621
1622 // return TRUE only if scan was stopped by the terminating NUL and if the
1623 // string was not empty to start with
1624 return !*end && (end != start);
1625 }
1626
1627 // ---------------------------------------------------------------------------
1628 // formatted output
1629 // ---------------------------------------------------------------------------
1630
1631 /* static */
1632 wxString wxString::Format(const wxChar *pszFormat, ...)
1633 {
1634 va_list argptr;
1635 va_start(argptr, pszFormat);
1636
1637 wxString s;
1638 s.PrintfV(pszFormat, argptr);
1639
1640 va_end(argptr);
1641
1642 return s;
1643 }
1644
1645 /* static */
1646 wxString wxString::FormatV(const wxChar *pszFormat, va_list argptr)
1647 {
1648 wxString s;
1649 s.PrintfV(pszFormat, argptr);
1650 return s;
1651 }
1652
1653 int wxString::Printf(const wxChar *pszFormat, ...)
1654 {
1655 va_list argptr;
1656 va_start(argptr, pszFormat);
1657
1658 int iLen = PrintfV(pszFormat, argptr);
1659
1660 va_end(argptr);
1661
1662 return iLen;
1663 }
1664
1665 int wxString::PrintfV(const wxChar* pszFormat, va_list argptr)
1666 {
1667 int size = 1024;
1668 int len;
1669
1670 for ( ;; )
1671 {
1672 {
1673 wxStringBuffer tmp(*this, size + 1);
1674 wxChar* buf = tmp;
1675
1676 if ( !buf )
1677 {
1678 // out of memory
1679 return -1;
1680 }
1681
1682 len = wxVsnprintf(buf, size, pszFormat, argptr);
1683
1684 // some implementations of vsnprintf() don't NUL terminate
1685 // the string if there is not enough space for it so
1686 // always do it manually
1687 buf[size] = _T('\0');
1688 }
1689
1690 // vsnprintf() may return either -1 (traditional Unix behaviour) or the
1691 // total number of characters which would have been written if the
1692 // buffer were large enough
1693 if ( len >= 0 && len <= size )
1694 {
1695 // ok, there was enough space
1696 break;
1697 }
1698
1699 // still not enough, double it again
1700 size *= 2;
1701 }
1702
1703 // we could have overshot
1704 Shrink();
1705
1706 return Len();
1707 }
1708
1709 // ----------------------------------------------------------------------------
1710 // misc other operations
1711 // ----------------------------------------------------------------------------
1712
1713 // returns TRUE if the string matches the pattern which may contain '*' and
1714 // '?' metacharacters (as usual, '?' matches any character and '*' any number
1715 // of them)
1716 bool wxString::Matches(const wxChar *pszMask) const
1717 {
1718 // I disable this code as it doesn't seem to be faster (in fact, it seems
1719 // to be much slower) than the old, hand-written code below and using it
1720 // here requires always linking with libregex even if the user code doesn't
1721 // use it
1722 #if 0 // wxUSE_REGEX
1723 // first translate the shell-like mask into a regex
1724 wxString pattern;
1725 pattern.reserve(wxStrlen(pszMask));
1726
1727 pattern += _T('^');
1728 while ( *pszMask )
1729 {
1730 switch ( *pszMask )
1731 {
1732 case _T('?'):
1733 pattern += _T('.');
1734 break;
1735
1736 case _T('*'):
1737 pattern += _T(".*");
1738 break;
1739
1740 case _T('^'):
1741 case _T('.'):
1742 case _T('$'):
1743 case _T('('):
1744 case _T(')'):
1745 case _T('|'):
1746 case _T('+'):
1747 case _T('\\'):
1748 // these characters are special in a RE, quote them
1749 // (however note that we don't quote '[' and ']' to allow
1750 // using them for Unix shell like matching)
1751 pattern += _T('\\');
1752 // fall through
1753
1754 default:
1755 pattern += *pszMask;
1756 }
1757
1758 pszMask++;
1759 }
1760 pattern += _T('$');
1761
1762 // and now use it
1763 return wxRegEx(pattern, wxRE_NOSUB | wxRE_EXTENDED).Matches(c_str());
1764 #else // !wxUSE_REGEX
1765 // TODO: this is, of course, awfully inefficient...
1766
1767 // the char currently being checked
1768 const wxChar *pszTxt = c_str();
1769
1770 // the last location where '*' matched
1771 const wxChar *pszLastStarInText = NULL;
1772 const wxChar *pszLastStarInMask = NULL;
1773
1774 match:
1775 for ( ; *pszMask != wxT('\0'); pszMask++, pszTxt++ ) {
1776 switch ( *pszMask ) {
1777 case wxT('?'):
1778 if ( *pszTxt == wxT('\0') )
1779 return FALSE;
1780
1781 // pszTxt and pszMask will be incremented in the loop statement
1782
1783 break;
1784
1785 case wxT('*'):
1786 {
1787 // remember where we started to be able to backtrack later
1788 pszLastStarInText = pszTxt;
1789 pszLastStarInMask = pszMask;
1790
1791 // ignore special chars immediately following this one
1792 // (should this be an error?)
1793 while ( *pszMask == wxT('*') || *pszMask == wxT('?') )
1794 pszMask++;
1795
1796 // if there is nothing more, match
1797 if ( *pszMask == wxT('\0') )
1798 return TRUE;
1799
1800 // are there any other metacharacters in the mask?
1801 size_t uiLenMask;
1802 const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?"));
1803
1804 if ( pEndMask != NULL ) {
1805 // we have to match the string between two metachars
1806 uiLenMask = pEndMask - pszMask;
1807 }
1808 else {
1809 // we have to match the remainder of the string
1810 uiLenMask = wxStrlen(pszMask);
1811 }
1812
1813 wxString strToMatch(pszMask, uiLenMask);
1814 const wxChar* pMatch = wxStrstr(pszTxt, strToMatch);
1815 if ( pMatch == NULL )
1816 return FALSE;
1817
1818 // -1 to compensate "++" in the loop
1819 pszTxt = pMatch + uiLenMask - 1;
1820 pszMask += uiLenMask - 1;
1821 }
1822 break;
1823
1824 default:
1825 if ( *pszMask != *pszTxt )
1826 return FALSE;
1827 break;
1828 }
1829 }
1830
1831 // match only if nothing left
1832 if ( *pszTxt == wxT('\0') )
1833 return TRUE;
1834
1835 // if we failed to match, backtrack if we can
1836 if ( pszLastStarInText ) {
1837 pszTxt = pszLastStarInText + 1;
1838 pszMask = pszLastStarInMask;
1839
1840 pszLastStarInText = NULL;
1841
1842 // don't bother resetting pszLastStarInMask, it's unnecessary
1843
1844 goto match;
1845 }
1846
1847 return FALSE;
1848 #endif // wxUSE_REGEX/!wxUSE_REGEX
1849 }
1850
1851 // Count the number of chars
1852 int wxString::Freq(wxChar ch) const
1853 {
1854 int count = 0;
1855 int len = Len();
1856 for (int i = 0; i < len; i++)
1857 {
1858 if (GetChar(i) == ch)
1859 count ++;
1860 }
1861 return count;
1862 }
1863
1864 // convert to upper case, return the copy of the string
1865 wxString wxString::Upper() const
1866 { wxString s(*this); return s.MakeUpper(); }
1867
1868 // convert to lower case, return the copy of the string
1869 wxString wxString::Lower() const { wxString s(*this); return s.MakeLower(); }
1870
1871 int wxString::sprintf(const wxChar *pszFormat, ...)
1872 {
1873 va_list argptr;
1874 va_start(argptr, pszFormat);
1875 int iLen = PrintfV(pszFormat, argptr);
1876 va_end(argptr);
1877 return iLen;
1878 }
1879
1880 // ============================================================================
1881 // ArrayString
1882 // ============================================================================
1883
1884 #include "wx/arrstr.h"
1885
1886 #if !wxUSE_STL
1887
1888 // size increment = min(50% of current size, ARRAY_MAXSIZE_INCREMENT)
1889 #define ARRAY_MAXSIZE_INCREMENT 4096
1890
1891 #ifndef ARRAY_DEFAULT_INITIAL_SIZE // also defined in dynarray.h
1892 #define ARRAY_DEFAULT_INITIAL_SIZE (16)
1893 #endif
1894
1895 #define STRING(p) ((wxString *)(&(p)))
1896
1897 // ctor
1898 void wxArrayString::Init(bool autoSort)
1899 {
1900 m_nSize =
1901 m_nCount = 0;
1902 m_pItems = (wxChar **) NULL;
1903 m_autoSort = autoSort;
1904 }
1905
1906 // copy ctor
1907 wxArrayString::wxArrayString(const wxArrayString& src)
1908 {
1909 Init(src.m_autoSort);
1910
1911 *this = src;
1912 }
1913
1914 // assignment operator
1915 wxArrayString& wxArrayString::operator=(const wxArrayString& src)
1916 {
1917 if ( m_nSize > 0 )
1918 Clear();
1919
1920 Copy(src);
1921
1922 m_autoSort = src.m_autoSort;
1923
1924 return *this;
1925 }
1926
1927 void wxArrayString::Copy(const wxArrayString& src)
1928 {
1929 if ( src.m_nCount > ARRAY_DEFAULT_INITIAL_SIZE )
1930 Alloc(src.m_nCount);
1931
1932 for ( size_t n = 0; n < src.m_nCount; n++ )
1933 Add(src[n]);
1934 }
1935
1936 // grow the array
1937 void wxArrayString::Grow(size_t nIncrement)
1938 {
1939 // only do it if no more place
1940 if ( (m_nSize - m_nCount) < nIncrement ) {
1941 // if ARRAY_DEFAULT_INITIAL_SIZE were set to 0, the initially empty would
1942 // be never resized!
1943 #if ARRAY_DEFAULT_INITIAL_SIZE == 0
1944 #error "ARRAY_DEFAULT_INITIAL_SIZE must be > 0!"
1945 #endif
1946
1947 if ( m_nSize == 0 ) {
1948 // was empty, alloc some memory
1949 m_nSize = ARRAY_DEFAULT_INITIAL_SIZE;
1950 if (m_nSize < nIncrement)
1951 m_nSize = nIncrement;
1952 m_pItems = new wxChar *[m_nSize];
1953 }
1954 else {
1955 // otherwise when it's called for the first time, nIncrement would be 0
1956 // and the array would never be expanded
1957 // add 50% but not too much
1958 size_t ndefIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE
1959 ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1;
1960 if ( ndefIncrement > ARRAY_MAXSIZE_INCREMENT )
1961 ndefIncrement = ARRAY_MAXSIZE_INCREMENT;
1962 if ( nIncrement < ndefIncrement )
1963 nIncrement = ndefIncrement;
1964 m_nSize += nIncrement;
1965 wxChar **pNew = new wxChar *[m_nSize];
1966
1967 // copy data to new location
1968 memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *));
1969
1970 // delete old memory (but do not release the strings!)
1971 wxDELETEA(m_pItems);
1972
1973 m_pItems = pNew;
1974 }
1975 }
1976 }
1977
1978 void wxArrayString::Free()
1979 {
1980 for ( size_t n = 0; n < m_nCount; n++ ) {
1981 STRING(m_pItems[n])->GetStringData()->Unlock();
1982 }
1983 }
1984
1985 // deletes all the strings from the list
1986 void wxArrayString::Empty()
1987 {
1988 Free();
1989
1990 m_nCount = 0;
1991 }
1992
1993 // as Empty, but also frees memory
1994 void wxArrayString::Clear()
1995 {
1996 Free();
1997
1998 m_nSize =
1999 m_nCount = 0;
2000
2001 wxDELETEA(m_pItems);
2002 }
2003
2004 // dtor
2005 wxArrayString::~wxArrayString()
2006 {
2007 Free();
2008
2009 wxDELETEA(m_pItems);
2010 }
2011
2012 void wxArrayString::reserve(size_t nSize)
2013 {
2014 Alloc(nSize);
2015 }
2016
2017 // pre-allocates memory (frees the previous data!)
2018 void wxArrayString::Alloc(size_t nSize)
2019 {
2020 // only if old buffer was not big enough
2021 if ( nSize > m_nSize ) {
2022 Free();
2023 wxDELETEA(m_pItems);
2024 m_pItems = new wxChar *[nSize];
2025 m_nSize = nSize;
2026 }
2027
2028 m_nCount = 0;
2029 }
2030
2031 // minimizes the memory usage by freeing unused memory
2032 void wxArrayString::Shrink()
2033 {
2034 // only do it if we have some memory to free
2035 if( m_nCount < m_nSize ) {
2036 // allocates exactly as much memory as we need
2037 wxChar **pNew = new wxChar *[m_nCount];
2038
2039 // copy data to new location
2040 memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *));
2041 delete [] m_pItems;
2042 m_pItems = pNew;
2043 }
2044 }
2045
2046 #if WXWIN_COMPATIBILITY_2_4
2047
2048 // return a wxString[] as required for some control ctors.
2049 wxString* wxArrayString::GetStringArray() const
2050 {
2051 wxString *array = 0;
2052
2053 if( m_nCount > 0 )
2054 {
2055 array = new wxString[m_nCount];
2056 for( size_t i = 0; i < m_nCount; i++ )
2057 array[i] = m_pItems[i];
2058 }
2059
2060 return array;
2061 }
2062
2063 #endif // WXWIN_COMPATIBILITY_2_4
2064
2065 // searches the array for an item (forward or backwards)
2066 int wxArrayString::Index(const wxChar *sz, bool bCase, bool bFromEnd) const
2067 {
2068 if ( m_autoSort ) {
2069 // use binary search in the sorted array
2070 wxASSERT_MSG( bCase && !bFromEnd,
2071 wxT("search parameters ignored for auto sorted array") );
2072
2073 size_t i,
2074 lo = 0,
2075 hi = m_nCount;
2076 int res;
2077 while ( lo < hi ) {
2078 i = (lo + hi)/2;
2079
2080 res = wxStrcmp(sz, m_pItems[i]);
2081 if ( res < 0 )
2082 hi = i;
2083 else if ( res > 0 )
2084 lo = i + 1;
2085 else
2086 return i;
2087 }
2088
2089 return wxNOT_FOUND;
2090 }
2091 else {
2092 // use linear search in unsorted array
2093 if ( bFromEnd ) {
2094 if ( m_nCount > 0 ) {
2095 size_t ui = m_nCount;
2096 do {
2097 if ( STRING(m_pItems[--ui])->IsSameAs(sz, bCase) )
2098 return ui;
2099 }
2100 while ( ui != 0 );
2101 }
2102 }
2103 else {
2104 for( size_t ui = 0; ui < m_nCount; ui++ ) {
2105 if( STRING(m_pItems[ui])->IsSameAs(sz, bCase) )
2106 return ui;
2107 }
2108 }
2109 }
2110
2111 return wxNOT_FOUND;
2112 }
2113
2114 // add item at the end
2115 size_t wxArrayString::Add(const wxString& str, size_t nInsert)
2116 {
2117 if ( m_autoSort ) {
2118 // insert the string at the correct position to keep the array sorted
2119 size_t i,
2120 lo = 0,
2121 hi = m_nCount;
2122 int res;
2123 while ( lo < hi ) {
2124 i = (lo + hi)/2;
2125
2126 res = wxStrcmp(str, m_pItems[i]);
2127 if ( res < 0 )
2128 hi = i;
2129 else if ( res > 0 )
2130 lo = i + 1;
2131 else {
2132 lo = hi = i;
2133 break;
2134 }
2135 }
2136
2137 wxASSERT_MSG( lo == hi, wxT("binary search broken") );
2138
2139 Insert(str, lo, nInsert);
2140
2141 return (size_t)lo;
2142 }
2143 else {
2144 wxASSERT( str.GetStringData()->IsValid() );
2145
2146 Grow(nInsert);
2147
2148 for (size_t i = 0; i < nInsert; i++)
2149 {
2150 // the string data must not be deleted!
2151 str.GetStringData()->Lock();
2152
2153 // just append
2154 m_pItems[m_nCount + i] = (wxChar *)str.c_str(); // const_cast
2155 }
2156 size_t ret = m_nCount;
2157 m_nCount += nInsert;
2158 return ret;
2159 }
2160 }
2161
2162 // add item at the given position
2163 void wxArrayString::Insert(const wxString& str, size_t nIndex, size_t nInsert)
2164 {
2165 wxASSERT( str.GetStringData()->IsValid() );
2166
2167 wxCHECK_RET( nIndex <= m_nCount, wxT("bad index in wxArrayString::Insert") );
2168 wxCHECK_RET( m_nCount <= m_nCount + nInsert,
2169 wxT("array size overflow in wxArrayString::Insert") );
2170
2171 Grow(nInsert);
2172
2173 memmove(&m_pItems[nIndex + nInsert], &m_pItems[nIndex],
2174 (m_nCount - nIndex)*sizeof(wxChar *));
2175
2176 for (size_t i = 0; i < nInsert; i++)
2177 {
2178 str.GetStringData()->Lock();
2179 m_pItems[nIndex + i] = (wxChar *)str.c_str();
2180 }
2181 m_nCount += nInsert;
2182 }
2183
2184 // expand the array
2185 void wxArrayString::SetCount(size_t count)
2186 {
2187 Alloc(count);
2188
2189 wxString s;
2190 while ( m_nCount < count )
2191 m_pItems[m_nCount++] = (wxChar *)s.c_str();
2192 }
2193
2194 // removes item from array (by index)
2195 void wxArrayString::RemoveAt(size_t nIndex, size_t nRemove)
2196 {
2197 wxCHECK_RET( nIndex < m_nCount, wxT("bad index in wxArrayString::Remove") );
2198 wxCHECK_RET( nIndex + nRemove <= m_nCount,
2199 wxT("removing too many elements in wxArrayString::Remove") );
2200
2201 // release our lock
2202 for (size_t i = 0; i < nRemove; i++)
2203 Item(nIndex + i).GetStringData()->Unlock();
2204
2205 memmove(&m_pItems[nIndex], &m_pItems[nIndex + nRemove],
2206 (m_nCount - nIndex - nRemove)*sizeof(wxChar *));
2207 m_nCount -= nRemove;
2208 }
2209
2210 // removes item from array (by value)
2211 void wxArrayString::Remove(const wxChar *sz)
2212 {
2213 int iIndex = Index(sz);
2214
2215 wxCHECK_RET( iIndex != wxNOT_FOUND,
2216 wxT("removing inexistent element in wxArrayString::Remove") );
2217
2218 RemoveAt(iIndex);
2219 }
2220
2221 void wxArrayString::assign(const_iterator first, const_iterator last)
2222 {
2223 reserve(last - first);
2224 for(; first != last; ++first)
2225 push_back(*first);
2226 }
2227
2228 // ----------------------------------------------------------------------------
2229 // sorting
2230 // ----------------------------------------------------------------------------
2231
2232 // we can only sort one array at a time with the quick-sort based
2233 // implementation
2234 #if wxUSE_THREADS
2235 // need a critical section to protect access to gs_compareFunction and
2236 // gs_sortAscending variables
2237 static wxCriticalSection *gs_critsectStringSort = NULL;
2238
2239 // call this before the value of the global sort vars is changed/after
2240 // you're finished with them
2241 #define START_SORT() wxASSERT( !gs_critsectStringSort ); \
2242 gs_critsectStringSort = new wxCriticalSection; \
2243 gs_critsectStringSort->Enter()
2244 #define END_SORT() gs_critsectStringSort->Leave(); \
2245 delete gs_critsectStringSort; \
2246 gs_critsectStringSort = NULL
2247 #else // !threads
2248 #define START_SORT()
2249 #define END_SORT()
2250 #endif // wxUSE_THREADS
2251
2252 // function to use for string comparaison
2253 static wxArrayString::CompareFunction gs_compareFunction = NULL;
2254
2255 // if we don't use the compare function, this flag tells us if we sort the
2256 // array in ascending or descending order
2257 static bool gs_sortAscending = TRUE;
2258
2259 // function which is called by quick sort
2260 extern "C" int wxC_CALLING_CONV // LINKAGEMODE
2261 wxStringCompareFunction(const void *first, const void *second)
2262 {
2263 wxString *strFirst = (wxString *)first;
2264 wxString *strSecond = (wxString *)second;
2265
2266 if ( gs_compareFunction ) {
2267 return gs_compareFunction(*strFirst, *strSecond);
2268 }
2269 else {
2270 // maybe we should use wxStrcoll
2271 int result = wxStrcmp(strFirst->c_str(), strSecond->c_str());
2272
2273 return gs_sortAscending ? result : -result;
2274 }
2275 }
2276
2277 // sort array elements using passed comparaison function
2278 void wxArrayString::Sort(CompareFunction compareFunction)
2279 {
2280 START_SORT();
2281
2282 wxASSERT( !gs_compareFunction ); // must have been reset to NULL
2283 gs_compareFunction = compareFunction;
2284
2285 DoSort();
2286
2287 // reset it to NULL so that Sort(bool) will work the next time
2288 gs_compareFunction = NULL;
2289
2290 END_SORT();
2291 }
2292
2293 typedef int (wxC_CALLING_CONV * wxStringCompareFn)(const void *first, const void *second);
2294
2295 void wxArrayString::Sort(CompareFunction2 compareFunction)
2296 {
2297 qsort(m_pItems, m_nCount, sizeof(wxChar *), (wxStringCompareFn)compareFunction);
2298 }
2299
2300 void wxArrayString::Sort(bool reverseOrder)
2301 {
2302 Sort(reverseOrder ? wxStringSortDescending : wxStringSortAscending);
2303 }
2304
2305 void wxArrayString::DoSort()
2306 {
2307 wxCHECK_RET( !m_autoSort, wxT("can't use this method with sorted arrays") );
2308
2309 // just sort the pointers using qsort() - of course it only works because
2310 // wxString() *is* a pointer to its data
2311 qsort(m_pItems, m_nCount, sizeof(wxChar *), wxStringCompareFunction);
2312 }
2313
2314 bool wxArrayString::operator==(const wxArrayString& a) const
2315 {
2316 if ( m_nCount != a.m_nCount )
2317 return FALSE;
2318
2319 for ( size_t n = 0; n < m_nCount; n++ )
2320 {
2321 if ( Item(n) != a[n] )
2322 return FALSE;
2323 }
2324
2325 return TRUE;
2326 }
2327
2328 #endif // !wxUSE_STL
2329
2330 int wxCMPFUNC_CONV wxStringSortAscending(wxString* s1, wxString* s2)
2331 {
2332 return wxStrcmp(s1->c_str(), s2->c_str());
2333 }
2334
2335 int wxCMPFUNC_CONV wxStringSortDescending(wxString* s1, wxString* s2)
2336 {
2337 return -wxStrcmp(s1->c_str(), s2->c_str());
2338 }