define wxString::iterator::iterator_category correctly if wxUSE_STD_STRING and not...
[wxWidgets.git] / include / wx / string.h
1 ///////////////////////////////////////////////////////////////////////////////
2 // Name: wx/string.h
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 /*
13 Efficient string class [more or less] compatible with MFC CString,
14 wxWidgets version 1 wxString and std::string and some handy functions
15 missing from string.h.
16 */
17
18 #ifndef _WX_WXSTRING_H__
19 #define _WX_WXSTRING_H__
20
21 // ----------------------------------------------------------------------------
22 // headers
23 // ----------------------------------------------------------------------------
24
25 #include "wx/defs.h" // everybody should include this
26
27 #ifndef __WXPALMOS5__
28 #if defined(__WXMAC__) || defined(__VISAGECPP__)
29 #include <ctype.h>
30 #endif
31
32 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
33 // problem in VACPP V4 with including stdlib.h multiple times
34 // strconv includes it anyway
35 # include <stdio.h>
36 # include <string.h>
37 # include <stdarg.h>
38 # include <limits.h>
39 #else
40 # include <string.h>
41 # include <stdio.h>
42 # include <stdarg.h>
43 # include <limits.h>
44 # include <stdlib.h>
45 #endif
46
47 #ifdef HAVE_STRCASECMP_IN_STRINGS_H
48 #include <strings.h> // for strcasecmp()
49 #endif // HAVE_STRCASECMP_IN_STRINGS_H
50 #endif // ! __WXPALMOS5__
51
52 #include "wx/wxcrtbase.h" // for wxChar, wxStrlen() etc.
53 #include "wx/strvararg.h"
54 #include "wx/buffer.h" // for wxCharBuffer
55 #include "wx/strconv.h" // for wxConvertXXX() macros and wxMBConv classes
56 #include "wx/stringimpl.h"
57 #include "wx/stringops.h"
58 #include "wx/unichar.h"
59
60 // by default we cache the mapping of the positions in UTF-8 string to the byte
61 // offset as this results in noticeable performance improvements for loops over
62 // strings using indices; comment out this line to disable this
63 //
64 // notice that this optimization is well worth using even in debug builds as it
65 // changes asymptotic complexity of algorithms using indices to iterate over
66 // wxString back to expected linear from quadratic
67 //
68 // also notice that wxTLS_TYPE() (__declspec(thread) in this case) is unsafe to
69 // use in DLL build under pre-Vista Windows so we disable this code for now, if
70 // anybody really needs to use UTF-8 build under Windows with this optimization
71 // it would have to be re-tested and probably corrected
72 #if wxUSE_UNICODE_UTF8 && !defined(__WXMSW__)
73 #define wxUSE_STRING_POS_CACHE 1
74 #else
75 #define wxUSE_STRING_POS_CACHE 0
76 #endif
77
78 #if wxUSE_STRING_POS_CACHE
79 #include "wx/tls.h"
80
81 // change this 0 to 1 to enable additional (very expensive) asserts
82 // verifying that string caching logic works as expected
83 #if 0
84 #define wxSTRING_CACHE_ASSERT(cond) wxASSERT(cond)
85 #else
86 #define wxSTRING_CACHE_ASSERT(cond)
87 #endif
88 #endif // wxUSE_STRING_POS_CACHE
89
90 class WXDLLIMPEXP_FWD_BASE wxString;
91
92 // unless this symbol is predefined to disable the compatibility functions, do
93 // use them
94 #ifndef WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
95 #define WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER 1
96 #endif
97
98 namespace wxPrivate
99 {
100 template <typename T> struct wxStringAsBufHelper;
101 }
102
103 // ---------------------------------------------------------------------------
104 // macros
105 // ---------------------------------------------------------------------------
106
107 // casts [unfortunately!] needed to call some broken functions which require
108 // "char *" instead of "const char *"
109 #define WXSTRINGCAST (wxChar *)(const wxChar *)
110 #define wxCSTRINGCAST (wxChar *)(const wxChar *)
111 #define wxMBSTRINGCAST (char *)(const char *)
112 #define wxWCSTRINGCAST (wchar_t *)(const wchar_t *)
113
114 // ----------------------------------------------------------------------------
115 // constants
116 // ----------------------------------------------------------------------------
117
118 #if WXWIN_COMPATIBILITY_2_6
119
120 // deprecated in favour of wxString::npos, don't use in new code
121 //
122 // maximum possible length for a string means "take all string" everywhere
123 #define wxSTRING_MAXLEN wxString::npos
124
125 #endif // WXWIN_COMPATIBILITY_2_6
126
127 // ---------------------------------------------------------------------------
128 // global functions complementing standard C string library replacements for
129 // strlen() and portable strcasecmp()
130 //---------------------------------------------------------------------------
131
132 #if WXWIN_COMPATIBILITY_2_8
133 // Use wxXXX() functions from wxcrt.h instead! These functions are for
134 // backwards compatibility only.
135
136 // checks whether the passed in pointer is NULL and if the string is empty
137 wxDEPRECATED( inline bool IsEmpty(const char *p) );
138 inline bool IsEmpty(const char *p) { return (!p || !*p); }
139
140 // safe version of strlen() (returns 0 if passed NULL pointer)
141 wxDEPRECATED( inline size_t Strlen(const char *psz) );
142 inline size_t Strlen(const char *psz)
143 { return psz ? strlen(psz) : 0; }
144
145 // portable strcasecmp/_stricmp
146 wxDEPRECATED( inline int Stricmp(const char *psz1, const char *psz2) );
147 inline int Stricmp(const char *psz1, const char *psz2)
148 {
149 #if defined(__VISUALC__) && defined(__WXWINCE__)
150 register char c1, c2;
151 do {
152 c1 = tolower(*psz1++);
153 c2 = tolower(*psz2++);
154 } while ( c1 && (c1 == c2) );
155
156 return c1 - c2;
157 #elif defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
158 return _stricmp(psz1, psz2);
159 #elif defined(__SC__)
160 return _stricmp(psz1, psz2);
161 #elif defined(__BORLANDC__)
162 return stricmp(psz1, psz2);
163 #elif defined(__WATCOMC__)
164 return stricmp(psz1, psz2);
165 #elif defined(__DJGPP__)
166 return stricmp(psz1, psz2);
167 #elif defined(__EMX__)
168 return stricmp(psz1, psz2);
169 #elif defined(__WXPM__)
170 return stricmp(psz1, psz2);
171 #elif defined(__WXPALMOS__) || \
172 defined(HAVE_STRCASECMP_IN_STRING_H) || \
173 defined(HAVE_STRCASECMP_IN_STRINGS_H) || \
174 defined(__GNUWIN32__)
175 return strcasecmp(psz1, psz2);
176 #elif defined(__MWERKS__) && !defined(__INTEL__)
177 register char c1, c2;
178 do {
179 c1 = tolower(*psz1++);
180 c2 = tolower(*psz2++);
181 } while ( c1 && (c1 == c2) );
182
183 return c1 - c2;
184 #else
185 // almost all compilers/libraries provide this function (unfortunately under
186 // different names), that's why we don't implement our own which will surely
187 // be more efficient than this code (uncomment to use):
188 /*
189 register char c1, c2;
190 do {
191 c1 = tolower(*psz1++);
192 c2 = tolower(*psz2++);
193 } while ( c1 && (c1 == c2) );
194
195 return c1 - c2;
196 */
197
198 #error "Please define string case-insensitive compare for your OS/compiler"
199 #endif // OS/compiler
200 }
201
202 #endif // WXWIN_COMPATIBILITY_2_8
203
204 // ----------------------------------------------------------------------------
205 // wxCStrData
206 // ----------------------------------------------------------------------------
207
208 // Lightweight object returned by wxString::c_str() and implicitly convertible
209 // to either const char* or const wchar_t*.
210 class WXDLLIMPEXP_BASE wxCStrData
211 {
212 private:
213 // Ctors; for internal use by wxString and wxCStrData only
214 wxCStrData(const wxString *str, size_t offset = 0, bool owned = false)
215 : m_str(str), m_offset(offset), m_owned(owned) {}
216
217 public:
218 // Ctor constructs the object from char literal; they are needed to make
219 // operator?: compile and they intentionally take char*, not const char*
220 inline wxCStrData(char *buf);
221 inline wxCStrData(wchar_t *buf);
222 inline wxCStrData(const wxCStrData& data);
223
224 inline ~wxCStrData();
225
226 // methods defined inline below must be declared inline or mingw32 3.4.5
227 // warns about "<symbol> defined locally after being referenced with
228 // dllimport linkage"
229 #if wxUSE_UNICODE_WCHAR
230 inline
231 #endif
232 const wchar_t* AsWChar() const;
233 operator const wchar_t*() const { return AsWChar(); }
234
235 #if !wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY
236 inline
237 #endif
238 const char* AsChar() const;
239 const unsigned char* AsUnsignedChar() const
240 { return (const unsigned char *) AsChar(); }
241 operator const char*() const { return AsChar(); }
242 operator const unsigned char*() const { return AsUnsignedChar(); }
243
244 operator const void*() const { return AsChar(); }
245
246 inline const wxCharBuffer AsCharBuf() const;
247 inline const wxWCharBuffer AsWCharBuf() const;
248
249 inline wxString AsString() const;
250
251 // returns the value as C string in internal representation (equivalent
252 // to AsString().wx_str(), but more efficient)
253 const wxStringCharType *AsInternal() const;
254
255 // allow expressions like "c_str()[0]":
256 inline wxUniChar operator[](size_t n) const;
257 wxUniChar operator[](int n) const { return operator[](size_t(n)); }
258 wxUniChar operator[](long n) const { return operator[](size_t(n)); }
259 #ifndef wxSIZE_T_IS_UINT
260 wxUniChar operator[](unsigned int n) const { return operator[](size_t(n)); }
261 #endif // size_t != unsigned int
262
263 // these operators are needed to emulate the pointer semantics of c_str():
264 // expressions like "wxChar *p = str.c_str() + 1;" should continue to work
265 // (we need both versions to resolve ambiguities):
266 wxCStrData operator+(int n) const
267 { return wxCStrData(m_str, m_offset + n, m_owned); }
268 wxCStrData operator+(long n) const
269 { return wxCStrData(m_str, m_offset + n, m_owned); }
270 wxCStrData operator+(size_t n) const
271 { return wxCStrData(m_str, m_offset + n, m_owned); }
272
273 // and these for "str.c_str() + (p2 - p1)" (it also works for any integer
274 // expression but it must be ptrdiff_t and not e.g. int to work in this
275 // example):
276 wxCStrData operator-(ptrdiff_t n) const
277 {
278 wxASSERT_MSG( n <= (ptrdiff_t)m_offset,
279 _T("attempt to construct address before the beginning of the string") );
280 return wxCStrData(m_str, m_offset - n, m_owned);
281 }
282
283 // this operator is needed to make expressions like "*c_str()" or
284 // "*(c_str() + 2)" work
285 inline wxUniChar operator*() const;
286
287 private:
288 const wxString *m_str;
289 size_t m_offset;
290 bool m_owned;
291
292 friend class WXDLLIMPEXP_FWD_BASE wxString;
293 };
294
295 // ----------------------------------------------------------------------------
296 // wxStringPrintfMixin
297 // ---------------------------------------------------------------------------
298
299 // NB: VC6 has a bug that causes linker errors if you have template methods
300 // in a class using __declspec(dllimport). The solution is to split such
301 // class into two classes, one that contains the template methods and does
302 // *not* use WXDLLIMPEXP_BASE and another class that contains the rest
303 // (with DLL linkage).
304 //
305 // We only do this for VC6 here, because the code is less efficient
306 // (Printf() has to use dynamic_cast<>) and because OpenWatcom compiler
307 // cannot compile this code.
308
309 #if defined(__VISUALC__) && __VISUALC__ < 1300
310 #define wxNEEDS_WXSTRING_PRINTF_MIXIN
311 #endif
312
313 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
314 // this class contains implementation of wxString's vararg methods, it's
315 // exported from wxBase DLL
316 class WXDLLIMPEXP_BASE wxStringPrintfMixinBase
317 {
318 protected:
319 wxStringPrintfMixinBase() {}
320
321 #if !wxUSE_UTF8_LOCALE_ONLY
322 int DoPrintfWchar(const wxChar *format, ...);
323 static wxString DoFormatWchar(const wxChar *format, ...);
324 #endif
325 #if wxUSE_UNICODE_UTF8
326 int DoPrintfUtf8(const char *format, ...);
327 static wxString DoFormatUtf8(const char *format, ...);
328 #endif
329 };
330
331 // this class contains template wrappers for wxString's vararg methods, it's
332 // intentionally *not* exported from the DLL in order to fix the VC6 bug
333 // described above
334 class wxStringPrintfMixin : public wxStringPrintfMixinBase
335 {
336 private:
337 // to further complicate things, we can't return wxString from
338 // wxStringPrintfMixin::Format() because wxString is not yet declared at
339 // this point; the solution is to use this fake type trait template - this
340 // way the compiler won't know the return type until Format() is used
341 // (this doesn't compile with Watcom, but VC6 compiles it just fine):
342 template<typename T> struct StringReturnType
343 {
344 typedef wxString type;
345 };
346
347 public:
348 // these are duplicated wxString methods, they're also declared below
349 // if !wxNEEDS_WXSTRING_PRINTF_MIXIN:
350
351 // static wxString Format(const wString& format, ...) ATTRIBUTE_PRINTF_1;
352 WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType<T1>::type,
353 Format, 1, (const wxFormatString&),
354 DoFormatWchar, DoFormatUtf8)
355 // We have to implement the version without template arguments manually
356 // because of the StringReturnType<> hack, although WX_DEFINE_VARARG_FUNC
357 // normally does it itself. It has to be a template so that we can use
358 // the hack, even though there's no real template parameter. We can't move
359 // it to wxStrig, because it would shadow these versions of Format() then.
360 template<typename T>
361 inline static typename StringReturnType<T>::type
362 Format(const T& fmt)
363 {
364 // NB: this doesn't compile if T is not (some form of) a string;
365 // this makes Format's prototype equivalent to
366 // Format(const wxFormatString& fmt)
367 return DoFormatWchar(wxFormatString(fmt));
368 }
369
370 // int Printf(const wxString& format, ...);
371 WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),
372 DoPrintfWchar, DoPrintfUtf8)
373 // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;
374 WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),
375 DoPrintfWchar, DoPrintfUtf8)
376
377 protected:
378 wxStringPrintfMixin() : wxStringPrintfMixinBase() {}
379 };
380 #endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
381
382
383 // ----------------------------------------------------------------------------
384 // wxString: string class trying to be compatible with std::string, MFC
385 // CString and wxWindows 1.x wxString all at once
386 // ---------------------------------------------------------------------------
387
388 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
389 // "non dll-interface class 'wxStringPrintfMixin' used as base interface
390 // for dll-interface class 'wxString'" -- this is OK in our case
391 #pragma warning (disable:4275)
392 #endif
393
394 #if wxUSE_UNICODE_UTF8
395 // see the comment near wxString::iterator for why we need this
396 class WXDLLIMPEXP_BASE wxStringIteratorNode
397 {
398 public:
399 wxStringIteratorNode()
400 : m_str(NULL), m_citer(NULL), m_iter(NULL), m_prev(NULL), m_next(NULL) {}
401 wxStringIteratorNode(const wxString *str,
402 wxStringImpl::const_iterator *citer)
403 { DoSet(str, citer, NULL); }
404 wxStringIteratorNode(const wxString *str, wxStringImpl::iterator *iter)
405 { DoSet(str, NULL, iter); }
406 ~wxStringIteratorNode()
407 { clear(); }
408
409 inline void set(const wxString *str, wxStringImpl::const_iterator *citer)
410 { clear(); DoSet(str, citer, NULL); }
411 inline void set(const wxString *str, wxStringImpl::iterator *iter)
412 { clear(); DoSet(str, NULL, iter); }
413
414 const wxString *m_str;
415 wxStringImpl::const_iterator *m_citer;
416 wxStringImpl::iterator *m_iter;
417 wxStringIteratorNode *m_prev, *m_next;
418
419 private:
420 inline void clear();
421 inline void DoSet(const wxString *str,
422 wxStringImpl::const_iterator *citer,
423 wxStringImpl::iterator *iter);
424
425 // the node belongs to a particular iterator instance, it's not copied
426 // when a copy of the iterator is made
427 DECLARE_NO_COPY_CLASS(wxStringIteratorNode)
428 };
429 #endif // wxUSE_UNICODE_UTF8
430
431 class WXDLLIMPEXP_BASE wxString
432 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
433 : public wxStringPrintfMixin
434 #endif
435 {
436 // NB: special care was taken in arranging the member functions in such order
437 // that all inline functions can be effectively inlined, verify that all
438 // performance critical functions are still inlined if you change order!
439 public:
440 // an 'invalid' value for string index, moved to this place due to a CW bug
441 static const size_t npos;
442
443 private:
444 // if we hadn't made these operators private, it would be possible to
445 // compile "wxString s; s = 17;" without any warnings as 17 is implicitly
446 // converted to char in C and we do have operator=(char)
447 //
448 // NB: we don't need other versions (short/long and unsigned) as attempt
449 // to assign another numeric type to wxString will now result in
450 // ambiguity between operator=(char) and operator=(int)
451 wxString& operator=(int);
452
453 // these methods are not implemented - there is _no_ conversion from int to
454 // string, you're doing something wrong if the compiler wants to call it!
455 //
456 // try `s << i' or `s.Printf("%d", i)' instead
457 wxString(int);
458
459
460 // buffer for holding temporary substring when using any of the methods
461 // that take (char*,size_t) or (wchar_t*,size_t) arguments:
462 template<typename T>
463 struct SubstrBufFromType
464 {
465 T data;
466 size_t len;
467
468 SubstrBufFromType(const T& data_, size_t len_)
469 : data(data_), len(len_)
470 {
471 wxASSERT_MSG( len != npos, "must have real length" );
472 }
473 };
474
475 #if wxUSE_UNICODE_UTF8
476 // even char* -> char* needs conversion, from locale charset to UTF-8
477 typedef SubstrBufFromType<wxCharBuffer> SubstrBufFromWC;
478 typedef SubstrBufFromType<wxCharBuffer> SubstrBufFromMB;
479 #elif wxUSE_UNICODE_WCHAR
480 typedef SubstrBufFromType<const wchar_t*> SubstrBufFromWC;
481 typedef SubstrBufFromType<wxWCharBuffer> SubstrBufFromMB;
482 #else
483 typedef SubstrBufFromType<const char*> SubstrBufFromMB;
484 typedef SubstrBufFromType<wxCharBuffer> SubstrBufFromWC;
485 #endif
486
487
488 // Functions implementing primitive operations on string data; wxString
489 // methods and iterators are implemented in terms of it. The differences
490 // between UTF-8 and wchar_t* representations of the string are mostly
491 // contained here.
492
493 #if wxUSE_UNICODE_UTF8
494 static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength,
495 const wxMBConv& conv);
496 static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength,
497 const wxMBConv& conv);
498 #elif wxUSE_UNICODE_WCHAR
499 static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength,
500 const wxMBConv& conv);
501 #else
502 static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength,
503 const wxMBConv& conv);
504 #endif
505
506 #if !wxUSE_UNICODE_UTF8 // wxUSE_UNICODE_WCHAR or !wxUSE_UNICODE
507 // returns C string encoded as the implementation expects:
508 #if wxUSE_UNICODE
509 static const wchar_t* ImplStr(const wchar_t* str)
510 { return str ? str : wxT(""); }
511 static const SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
512 { return SubstrBufFromWC(str, (str && n == npos) ? wxWcslen(str) : n); }
513 static wxWCharBuffer ImplStr(const char* str,
514 const wxMBConv& conv = wxConvLibc)
515 { return ConvertStr(str, npos, conv).data; }
516 static SubstrBufFromMB ImplStr(const char* str, size_t n,
517 const wxMBConv& conv = wxConvLibc)
518 { return ConvertStr(str, n, conv); }
519 #else
520 static const char* ImplStr(const char* str,
521 const wxMBConv& WXUNUSED(conv) = wxConvLibc)
522 { return str ? str : ""; }
523 static const SubstrBufFromMB ImplStr(const char* str, size_t n,
524 const wxMBConv& WXUNUSED(conv) = wxConvLibc)
525 { return SubstrBufFromMB(str, (str && n == npos) ? wxStrlen(str) : n); }
526 static wxCharBuffer ImplStr(const wchar_t* str)
527 { return ConvertStr(str, npos, wxConvLibc).data; }
528 static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
529 { return ConvertStr(str, n, wxConvLibc); }
530 #endif
531
532 // translates position index in wxString to/from index in underlying
533 // wxStringImpl:
534 static size_t PosToImpl(size_t pos) { return pos; }
535 static void PosLenToImpl(size_t pos, size_t len,
536 size_t *implPos, size_t *implLen)
537 { *implPos = pos; *implLen = len; }
538 static size_t LenToImpl(size_t len) { return len; }
539 static size_t PosFromImpl(size_t pos) { return pos; }
540
541 // we don't want to define these as empty inline functions as it could
542 // result in noticeable (and quite unnecessary in non-UTF-8 build) slowdown
543 // in debug build where the inline functions are not effectively inlined
544 #define wxSTRING_INVALIDATE_CACHE()
545 #define wxSTRING_INVALIDATE_CACHED_LENGTH()
546 #define wxSTRING_UPDATE_CACHED_LENGTH(n)
547 #define wxSTRING_SET_CACHED_LENGTH(n)
548
549 #else // wxUSE_UNICODE_UTF8
550
551 static wxCharBuffer ImplStr(const char* str,
552 const wxMBConv& conv = wxConvLibc)
553 { return ConvertStr(str, npos, conv).data; }
554 static SubstrBufFromMB ImplStr(const char* str, size_t n,
555 const wxMBConv& conv = wxConvLibc)
556 { return ConvertStr(str, n, conv); }
557
558 static wxCharBuffer ImplStr(const wchar_t* str)
559 { return ConvertStr(str, npos, wxMBConvUTF8()).data; }
560 static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
561 { return ConvertStr(str, n, wxMBConvUTF8()); }
562
563 #if wxUSE_STRING_POS_CACHE
564 // this is an extremely simple cache used by PosToImpl(): each cache element
565 // contains the string it applies to and the index corresponding to the last
566 // used position in this wxString in its m_impl string
567 //
568 // NB: notice that this struct (and nested Element one) must be a POD or we
569 // wouldn't be able to use a thread-local variable of this type, in
570 // particular it should have no ctor -- we rely on statics being
571 // initialized to 0 instead
572 struct Cache
573 {
574 enum { SIZE = 8 };
575
576 struct Element
577 {
578 const wxString *str; // the string to which this element applies
579 size_t pos, // the cached index in this string
580 impl, // the corresponding position in its m_impl
581 len; // cached length or npos if unknown
582
583 // reset cached index to 0
584 void ResetPos() { pos = impl = 0; }
585
586 // reset position and length
587 void Reset() { ResetPos(); len = npos; }
588 };
589
590 // cache the indices mapping for the last few string used
591 Element cached[SIZE];
592
593 // the last used index
594 unsigned lastUsed;
595 };
596
597 // notice that we must use an accessor function and not a static variable
598 // because when the TLS variables support is implemented in the library (and
599 // not by the compiler), the global s_cache variable could be not yet
600 // initialized when a ctor of another global object is executed and if that
601 // ctor uses any wxString methods, bad things happen
602 //
603 // also note that for the same reason this function _is_ MT-safe: we know
604 // it's going to be called during the program startup (currently during
605 // globals initialization but even if they ever stop using wxString, it would
606 // still be called by wxInitialize()), i.e. before any threads are created
607 static Cache& GetCache()
608 {
609 static wxTLS_TYPE(Cache) s_cache;
610
611 return wxTLS_VALUE(s_cache);
612 }
613
614 static Cache::Element *GetCacheBegin() { return GetCache().cached; }
615 static Cache::Element *GetCacheEnd() { return GetCacheBegin() + Cache::SIZE; }
616 static unsigned& LastUsedCacheElement() { return GetCache().lastUsed; }
617
618 friend struct wxStrCacheDumper;
619
620 // uncomment this to have access to some profiling statistics on program
621 // termination
622 //#define wxPROFILE_STRING_CACHE
623
624 #ifdef wxPROFILE_STRING_CACHE
625 static struct PosToImplCacheStats
626 {
627 unsigned postot, // total non-trivial calls to PosToImpl
628 poshits, // cache hits from PosToImpl()
629 mishits, // cached position beyond the needed one
630 sumpos, // sum of all positions, used to compute the
631 // average position after dividing by postot
632 sumofs, // sum of all offsets after using the cache, used to
633 // compute the average after dividing by hits
634 lentot, // number of total calls to length()
635 lenhits; // number of cache hits in length()
636 } ms_cacheStats;
637
638 friend struct ShowCacheStats;
639
640 #define wxCACHE_PROFILE_FIELD_INC(field) ms_cacheStats.field++
641 #define wxCACHE_PROFILE_FIELD_ADD(field, val) ms_cacheStats.field += (val)
642 #else // !wxPROFILE_STRING_CACHE
643 #define wxCACHE_PROFILE_FIELD_INC(field)
644 #define wxCACHE_PROFILE_FIELD_ADD(field, val)
645 #endif // wxPROFILE_STRING_CACHE/!wxPROFILE_STRING_CACHE
646
647 // note: it could seem that the functions below shouldn't be inline because
648 // they are big, contain loops and so the compiler shouldn't be able to
649 // inline them anyhow, however moving them into string.cpp does decrease the
650 // code performance by ~5%, at least when using g++ 4.1 so do keep them here
651 // unless tests show that it's not advantageous any more
652
653 // return the pointer to the cache element for this string or NULL if not
654 // cached
655 Cache::Element *FindCacheElement() const
656 {
657 // profiling seems to show a small but consistent gain if we use this
658 // simple loop instead of starting from the last used element (there are
659 // a lot of misses in this function...)
660 for ( Cache::Element *c = GetCacheBegin(); c != GetCacheEnd(); c++ )
661 {
662 if ( c->str == this )
663 return c;
664 }
665
666 return NULL;
667 }
668
669 // unlike FindCacheElement(), this one always returns a valid pointer to the
670 // cache element for this string, it may have valid last cached position and
671 // its corresponding index in the byte string or not
672 Cache::Element *GetCacheElement() const
673 {
674 Cache::Element * const cacheBegin = GetCacheBegin();
675 Cache::Element * const cacheEnd = GetCacheEnd();
676 Cache::Element * const cacheStart = cacheBegin + LastUsedCacheElement();
677
678 // check the last used first, this does no (measurable) harm for a miss
679 // but does help for simple loops addressing the same string all the time
680 if ( cacheStart->str == this )
681 return cacheStart;
682
683 // notice that we're going to check cacheStart again inside this call but
684 // profiling shows that it's still faster to use a simple loop like
685 // inside FindCacheElement() than manually looping with wrapping starting
686 // from the cache entry after the start one
687 Cache::Element *c = FindCacheElement();
688 if ( !c )
689 {
690 // claim the next cache entry for this string
691 c = cacheStart;
692 if ( ++c == cacheEnd )
693 c = cacheBegin;
694
695 c->str = this;
696 c->Reset();
697
698 // and remember the last used element
699 LastUsedCacheElement() = c - cacheBegin;
700 }
701
702 return c;
703 }
704
705 size_t DoPosToImpl(size_t pos) const
706 {
707 wxCACHE_PROFILE_FIELD_INC(postot);
708
709 // NB: although the case of pos == 1 (and offset from cached position
710 // equal to 1) are common, nothing is gained by writing special code
711 // for handling them, the compiler (at least g++ 4.1 used) seems to
712 // optimize the code well enough on its own
713
714 wxCACHE_PROFILE_FIELD_ADD(sumpos, pos);
715
716 Cache::Element * const cache = GetCacheElement();
717
718 // cached position can't be 0 so if it is, it means that this entry was
719 // used for length caching only so far, i.e. it doesn't count as a hit
720 // from our point of view
721 if ( cache->pos )
722 wxCACHE_PROFILE_FIELD_INC(poshits);
723
724 if ( pos == cache->pos )
725 return cache->impl;
726
727 // this seems to happen only rarely so just reset the cache in this case
728 // instead of complicating code even further by seeking backwards in this
729 // case
730 if ( cache->pos > pos )
731 {
732 wxCACHE_PROFILE_FIELD_INC(mishits);
733
734 cache->ResetPos();
735 }
736
737 wxCACHE_PROFILE_FIELD_ADD(sumofs, pos - cache->pos);
738
739
740 wxStringImpl::const_iterator i(m_impl.begin() + cache->impl);
741 for ( size_t n = cache->pos; n < pos; n++ )
742 wxStringOperations::IncIter(i);
743
744 cache->pos = pos;
745 cache->impl = i - m_impl.begin();
746
747 wxSTRING_CACHE_ASSERT(
748 (int)cache->impl == (begin() + pos).impl() - m_impl.begin() );
749
750 return cache->impl;
751 }
752
753 void InvalidateCache()
754 {
755 Cache::Element * const cache = FindCacheElement();
756 if ( cache )
757 cache->Reset();
758 }
759
760 void InvalidateCachedLength()
761 {
762 Cache::Element * const cache = FindCacheElement();
763 if ( cache )
764 cache->len = npos;
765 }
766
767 void SetCachedLength(size_t len)
768 {
769 // we optimistically cache the length here even if the string wasn't
770 // present in the cache before, this seems to do no harm and the
771 // potential for avoiding length recomputation for long strings looks
772 // interesting
773 GetCacheElement()->len = len;
774 }
775
776 void UpdateCachedLength(ptrdiff_t delta)
777 {
778 Cache::Element * const cache = FindCacheElement();
779 if ( cache && cache->len != npos )
780 {
781 wxSTRING_CACHE_ASSERT( (ptrdiff_t)cache->len + delta >= 0 );
782
783 cache->len += delta;
784 }
785 }
786
787 #define wxSTRING_INVALIDATE_CACHE() InvalidateCache()
788 #define wxSTRING_INVALIDATE_CACHED_LENGTH() InvalidateCachedLength()
789 #define wxSTRING_UPDATE_CACHED_LENGTH(n) UpdateCachedLength(n)
790 #define wxSTRING_SET_CACHED_LENGTH(n) SetCachedLength(n)
791 #else // !wxUSE_STRING_POS_CACHE
792 size_t DoPosToImpl(size_t pos) const
793 {
794 return (begin() + pos).impl() - m_impl.begin();
795 }
796
797 #define wxSTRING_INVALIDATE_CACHE()
798 #define wxSTRING_INVALIDATE_CACHED_LENGTH()
799 #define wxSTRING_UPDATE_CACHED_LENGTH(n)
800 #define wxSTRING_SET_CACHED_LENGTH(n)
801 #endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
802
803 size_t PosToImpl(size_t pos) const
804 {
805 return pos == 0 || pos == npos ? pos : DoPosToImpl(pos);
806 }
807
808 void PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const;
809
810 size_t LenToImpl(size_t len) const
811 {
812 size_t pos, len2;
813 PosLenToImpl(0, len, &pos, &len2);
814 return len2;
815 }
816
817 size_t PosFromImpl(size_t pos) const
818 {
819 if ( pos == 0 || pos == npos )
820 return pos;
821 else
822 return const_iterator(this, m_impl.begin() + pos) - begin();
823 }
824 #endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8
825
826 public:
827 // standard types
828 typedef wxUniChar value_type;
829 typedef wxUniChar char_type;
830 typedef wxUniCharRef reference;
831 typedef wxChar* pointer;
832 typedef const wxChar* const_pointer;
833
834 typedef size_t size_type;
835 typedef wxUniChar const_reference;
836
837 #if wxUSE_STD_STRING
838 #if wxUSE_UNICODE_UTF8
839 // random access is not O(1), as required by Random Access Iterator
840 #define WX_STR_ITERATOR_TAG std::bidirectional_iterator_tag
841 #else
842 #define WX_STR_ITERATOR_TAG std::random_access_iterator_tag
843 #endif
844 #define WX_STR_ITERATOR_CATEGORY typedef WX_STR_ITERATOR_TAG iterator_category;
845 #else
846 // not defining iterator_category at all in this case is better than defining
847 // it as some dummy type -- at least it results in more intelligible error
848 // messages
849 #define WX_STR_ITERATOR_CATEGORY
850 #endif
851
852 #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \
853 private: \
854 typedef wxStringImpl::iterator_name underlying_iterator; \
855 public: \
856 WX_STR_ITERATOR_CATEGORY \
857 typedef wxUniChar value_type; \
858 typedef int difference_type; \
859 typedef reference_type reference; \
860 typedef pointer_type pointer; \
861 \
862 reference operator[](size_t n) const { return *(*this + n); } \
863 \
864 iterator_name& operator++() \
865 { wxStringOperations::IncIter(m_cur); return *this; } \
866 iterator_name& operator--() \
867 { wxStringOperations::DecIter(m_cur); return *this; } \
868 iterator_name operator++(int) \
869 { \
870 iterator_name tmp = *this; \
871 wxStringOperations::IncIter(m_cur); \
872 return tmp; \
873 } \
874 iterator_name operator--(int) \
875 { \
876 iterator_name tmp = *this; \
877 wxStringOperations::DecIter(m_cur); \
878 return tmp; \
879 } \
880 \
881 iterator_name& operator+=(ptrdiff_t n) \
882 { \
883 m_cur = wxStringOperations::AddToIter(m_cur, n); \
884 return *this; \
885 } \
886 iterator_name& operator-=(ptrdiff_t n) \
887 { \
888 m_cur = wxStringOperations::AddToIter(m_cur, -n); \
889 return *this; \
890 } \
891 \
892 difference_type operator-(const iterator_name& i) const \
893 { return wxStringOperations::DiffIters(m_cur, i.m_cur); } \
894 \
895 bool operator==(const iterator_name& i) const \
896 { return m_cur == i.m_cur; } \
897 bool operator!=(const iterator_name& i) const \
898 { return m_cur != i.m_cur; } \
899 \
900 bool operator<(const iterator_name& i) const \
901 { return m_cur < i.m_cur; } \
902 bool operator>(const iterator_name& i) const \
903 { return m_cur > i.m_cur; } \
904 bool operator<=(const iterator_name& i) const \
905 { return m_cur <= i.m_cur; } \
906 bool operator>=(const iterator_name& i) const \
907 { return m_cur >= i.m_cur; } \
908 \
909 private: \
910 /* for internal wxString use only: */ \
911 underlying_iterator impl() const { return m_cur; } \
912 \
913 friend class wxString; \
914 friend class wxCStrData; \
915 \
916 private: \
917 underlying_iterator m_cur
918
919 class WXDLLIMPEXP_FWD_BASE const_iterator;
920
921 #if wxUSE_UNICODE_UTF8
922 // NB: In UTF-8 build, (non-const) iterator needs to keep reference
923 // to the underlying wxStringImpl, because UTF-8 is variable-length
924 // encoding and changing the value pointer to by an iterator (using
925 // its operator*) requires calling wxStringImpl::replace() if the old
926 // and new values differ in their encoding's length.
927 //
928 // Furthermore, the replace() call may invalid all iterators for the
929 // string, so we have to keep track of outstanding iterators and update
930 // them if replace() happens.
931 //
932 // This is implemented by maintaining linked list of iterators for every
933 // string and traversing it in wxUniCharRef::operator=(). Head of the
934 // list is stored in wxString. (FIXME-UTF8)
935
936 class WXDLLIMPEXP_BASE iterator
937 {
938 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
939
940 public:
941 iterator() {}
942 iterator(const iterator& i)
943 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
944 iterator& operator=(const iterator& i)
945 {
946 if (&i != this)
947 {
948 m_cur = i.m_cur;
949 m_node.set(i.str(), &m_cur);
950 }
951 return *this;
952 }
953
954 reference operator*()
955 { return wxUniCharRef::CreateForString(*str(), m_cur); }
956
957 iterator operator+(ptrdiff_t n) const
958 { return iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
959 iterator operator-(ptrdiff_t n) const
960 { return iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
961
962 private:
963 iterator(wxString *str, underlying_iterator ptr)
964 : m_cur(ptr), m_node(str, &m_cur) {}
965
966 wxString* str() const { return wx_const_cast(wxString*, m_node.m_str); }
967
968 wxStringIteratorNode m_node;
969
970 friend class const_iterator;
971 };
972
973 class WXDLLIMPEXP_BASE const_iterator
974 {
975 // NB: reference_type is intentionally value, not reference, the character
976 // may be encoded differently in wxString data:
977 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
978
979 public:
980 const_iterator() {}
981 const_iterator(const const_iterator& i)
982 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
983 const_iterator(const iterator& i)
984 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
985
986 const_iterator& operator=(const const_iterator& i)
987 {
988 if (&i != this)
989 {
990 m_cur = i.m_cur;
991 m_node.set(i.str(), &m_cur);
992 }
993 return *this;
994 }
995 const_iterator& operator=(const iterator& i)
996 { m_cur = i.m_cur; m_node.set(i.str(), &m_cur); return *this; }
997
998 reference operator*() const
999 { return wxStringOperations::DecodeChar(m_cur); }
1000
1001 const_iterator operator+(ptrdiff_t n) const
1002 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
1003 const_iterator operator-(ptrdiff_t n) const
1004 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
1005
1006 private:
1007 // for internal wxString use only:
1008 const_iterator(const wxString *str, underlying_iterator ptr)
1009 : m_cur(ptr), m_node(str, &m_cur) {}
1010
1011 const wxString* str() const { return m_node.m_str; }
1012
1013 wxStringIteratorNode m_node;
1014 };
1015
1016 size_t IterToImplPos(wxString::iterator i) const
1017 { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); }
1018
1019 iterator GetIterForNthChar(size_t n)
1020 { return iterator(this, m_impl.begin() + PosToImpl(n)); }
1021 const_iterator GetIterForNthChar(size_t n) const
1022 { return const_iterator(this, m_impl.begin() + PosToImpl(n)); }
1023 #else // !wxUSE_UNICODE_UTF8
1024
1025 class WXDLLIMPEXP_BASE iterator
1026 {
1027 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
1028
1029 public:
1030 iterator() {}
1031 iterator(const iterator& i) : m_cur(i.m_cur) {}
1032
1033 reference operator*()
1034 { return wxUniCharRef::CreateForString(m_cur); }
1035
1036 iterator operator+(ptrdiff_t n) const
1037 { return iterator(wxStringOperations::AddToIter(m_cur, n)); }
1038 iterator operator-(ptrdiff_t n) const
1039 { return iterator(wxStringOperations::AddToIter(m_cur, -n)); }
1040
1041 private:
1042 // for internal wxString use only:
1043 iterator(underlying_iterator ptr) : m_cur(ptr) {}
1044 iterator(wxString *WXUNUSED(str), underlying_iterator ptr) : m_cur(ptr) {}
1045
1046 friend class const_iterator;
1047 };
1048
1049 class WXDLLIMPEXP_BASE const_iterator
1050 {
1051 // NB: reference_type is intentionally value, not reference, the character
1052 // may be encoded differently in wxString data:
1053 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
1054
1055 public:
1056 const_iterator() {}
1057 const_iterator(const const_iterator& i) : m_cur(i.m_cur) {}
1058 const_iterator(const iterator& i) : m_cur(i.m_cur) {}
1059
1060 reference operator*() const
1061 { return wxStringOperations::DecodeChar(m_cur); }
1062
1063 const_iterator operator+(ptrdiff_t n) const
1064 { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); }
1065 const_iterator operator-(ptrdiff_t n) const
1066 { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); }
1067
1068 private:
1069 // for internal wxString use only:
1070 const_iterator(underlying_iterator ptr) : m_cur(ptr) {}
1071 const_iterator(const wxString *WXUNUSED(str), underlying_iterator ptr)
1072 : m_cur(ptr) {}
1073 };
1074
1075 iterator GetIterForNthChar(size_t n) { return begin() + n; }
1076 const_iterator GetIterForNthChar(size_t n) const { return begin() + n; }
1077 #endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8
1078
1079 #undef WX_STR_ITERATOR_TAG
1080 #undef WX_STR_ITERATOR_IMPL
1081
1082 friend class iterator;
1083 friend class const_iterator;
1084
1085 template <typename T>
1086 class reverse_iterator_impl
1087 {
1088 public:
1089 typedef T iterator_type;
1090
1091 typedef typename T::iterator_category iterator_category;
1092 typedef typename T::value_type value_type;
1093 typedef typename T::difference_type difference_type;
1094 typedef typename T::reference reference;
1095 typedef typename T::pointer *pointer;
1096
1097 reverse_iterator_impl() {}
1098 reverse_iterator_impl(iterator_type i) : m_cur(i) {}
1099 reverse_iterator_impl(const reverse_iterator_impl& ri)
1100 : m_cur(ri.m_cur) {}
1101
1102 iterator_type base() const { return m_cur; }
1103
1104 reference operator*() const { return *(m_cur-1); }
1105 reference operator[](size_t n) const { return *(*this + n); }
1106
1107 reverse_iterator_impl& operator++()
1108 { --m_cur; return *this; }
1109 reverse_iterator_impl operator++(int)
1110 { reverse_iterator_impl tmp = *this; --m_cur; return tmp; }
1111 reverse_iterator_impl& operator--()
1112 { ++m_cur; return *this; }
1113 reverse_iterator_impl operator--(int)
1114 { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; }
1115
1116 // NB: explicit <T> in the functions below is to keep BCC 5.5 happy
1117 reverse_iterator_impl operator+(ptrdiff_t n) const
1118 { return reverse_iterator_impl<T>(m_cur - n); }
1119 reverse_iterator_impl operator-(ptrdiff_t n) const
1120 { return reverse_iterator_impl<T>(m_cur + n); }
1121 reverse_iterator_impl operator+=(ptrdiff_t n)
1122 { m_cur -= n; return *this; }
1123 reverse_iterator_impl operator-=(ptrdiff_t n)
1124 { m_cur += n; return *this; }
1125
1126 unsigned operator-(const reverse_iterator_impl& i) const
1127 { return i.m_cur - m_cur; }
1128
1129 bool operator==(const reverse_iterator_impl& ri) const
1130 { return m_cur == ri.m_cur; }
1131 bool operator!=(const reverse_iterator_impl& ri) const
1132 { return !(*this == ri); }
1133
1134 bool operator<(const reverse_iterator_impl& i) const
1135 { return m_cur > i.m_cur; }
1136 bool operator>(const reverse_iterator_impl& i) const
1137 { return m_cur < i.m_cur; }
1138 bool operator<=(const reverse_iterator_impl& i) const
1139 { return m_cur >= i.m_cur; }
1140 bool operator>=(const reverse_iterator_impl& i) const
1141 { return m_cur <= i.m_cur; }
1142
1143 private:
1144 iterator_type m_cur;
1145 };
1146
1147 typedef reverse_iterator_impl<iterator> reverse_iterator;
1148 typedef reverse_iterator_impl<const_iterator> const_reverse_iterator;
1149
1150 private:
1151 // used to transform an expression built using c_str() (and hence of type
1152 // wxCStrData) to an iterator into the string
1153 static const_iterator CreateConstIterator(const wxCStrData& data)
1154 {
1155 return const_iterator(data.m_str,
1156 (data.m_str->begin() + data.m_offset).impl());
1157 }
1158
1159 // in UTF-8 STL build, creation from std::string requires conversion under
1160 // non-UTF8 locales, so we can't have and use wxString(wxStringImpl) ctor;
1161 // instead we define dummy type that lets us have wxString ctor for creation
1162 // from wxStringImpl that couldn't be used by user code (in all other builds,
1163 // "standard" ctors can be used):
1164 #if wxUSE_UNICODE_UTF8 && wxUSE_STL_BASED_WXSTRING
1165 struct CtorFromStringImplTag {};
1166
1167 wxString(CtorFromStringImplTag* WXUNUSED(dummy), const wxStringImpl& src)
1168 : m_impl(src) {}
1169
1170 static wxString FromImpl(const wxStringImpl& src)
1171 { return wxString((CtorFromStringImplTag*)NULL, src); }
1172 #else
1173 #if !wxUSE_STL_BASED_WXSTRING
1174 wxString(const wxStringImpl& src) : m_impl(src) { }
1175 // else: already defined as wxString(wxStdString) below
1176 #endif
1177 static wxString FromImpl(const wxStringImpl& src) { return wxString(src); }
1178 #endif
1179
1180 public:
1181 // constructors and destructor
1182 // ctor for an empty string
1183 wxString() {}
1184
1185 // copy ctor
1186 wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { }
1187
1188 // string containing nRepeat copies of ch
1189 wxString(wxUniChar ch, size_t nRepeat = 1 )
1190 { assign(nRepeat, ch); }
1191 wxString(size_t nRepeat, wxUniChar ch)
1192 { assign(nRepeat, ch); }
1193 wxString(wxUniCharRef ch, size_t nRepeat = 1)
1194 { assign(nRepeat, ch); }
1195 wxString(size_t nRepeat, wxUniCharRef ch)
1196 { assign(nRepeat, ch); }
1197 wxString(char ch, size_t nRepeat = 1)
1198 { assign(nRepeat, ch); }
1199 wxString(size_t nRepeat, char ch)
1200 { assign(nRepeat, ch); }
1201 wxString(wchar_t ch, size_t nRepeat = 1)
1202 { assign(nRepeat, ch); }
1203 wxString(size_t nRepeat, wchar_t ch)
1204 { assign(nRepeat, ch); }
1205
1206 // ctors from char* strings:
1207 wxString(const char *psz)
1208 : m_impl(ImplStr(psz)) {}
1209 wxString(const char *psz, const wxMBConv& conv)
1210 : m_impl(ImplStr(psz, conv)) {}
1211 wxString(const char *psz, size_t nLength)
1212 { assign(psz, nLength); }
1213 wxString(const char *psz, const wxMBConv& conv, size_t nLength)
1214 {
1215 SubstrBufFromMB str(ImplStr(psz, nLength, conv));
1216 m_impl.assign(str.data, str.len);
1217 }
1218
1219 // and unsigned char*:
1220 wxString(const unsigned char *psz)
1221 : m_impl(ImplStr((const char*)psz)) {}
1222 wxString(const unsigned char *psz, const wxMBConv& conv)
1223 : m_impl(ImplStr((const char*)psz, conv)) {}
1224 wxString(const unsigned char *psz, size_t nLength)
1225 { assign((const char*)psz, nLength); }
1226 wxString(const unsigned char *psz, const wxMBConv& conv, size_t nLength)
1227 {
1228 SubstrBufFromMB str(ImplStr((const char*)psz, nLength, conv));
1229 m_impl.assign(str.data, str.len);
1230 }
1231
1232 // ctors from wchar_t* strings:
1233 wxString(const wchar_t *pwz)
1234 : m_impl(ImplStr(pwz)) {}
1235 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv))
1236 : m_impl(ImplStr(pwz)) {}
1237 wxString(const wchar_t *pwz, size_t nLength)
1238 { assign(pwz, nLength); }
1239 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv), size_t nLength)
1240 { assign(pwz, nLength); }
1241
1242 wxString(const wxCharBuffer& buf)
1243 { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
1244 wxString(const wxWCharBuffer& buf)
1245 { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
1246
1247 // NB: this version uses m_impl.c_str() to force making a copy of the
1248 // string, so that "wxString(str.c_str())" idiom for passing strings
1249 // between threads works
1250 wxString(const wxCStrData& cstr)
1251 : m_impl(cstr.AsString().m_impl.c_str()) { }
1252
1253 // as we provide both ctors with this signature for both char and unsigned
1254 // char string, we need to provide one for wxCStrData to resolve ambiguity
1255 wxString(const wxCStrData& cstr, size_t nLength)
1256 : m_impl(cstr.AsString().Mid(0, nLength).m_impl) {}
1257
1258 // and because wxString is convertible to wxCStrData and const wxChar *
1259 // we also need to provide this one
1260 wxString(const wxString& str, size_t nLength)
1261 { assign(str, nLength); }
1262
1263
1264 #if wxUSE_STRING_POS_CACHE
1265 ~wxString()
1266 {
1267 // we need to invalidate our cache entry as another string could be
1268 // recreated at the same address (unlikely, but still possible, with the
1269 // heap-allocated strings but perfectly common with stack-allocated ones)
1270 InvalidateCache();
1271 }
1272 #endif // wxUSE_STRING_POS_CACHE
1273
1274 // even if we're not built with wxUSE_STL == 1 it is very convenient to allow
1275 // implicit conversions from std::string to wxString and vice verse as this
1276 // allows to use the same strings in non-GUI and GUI code, however we don't
1277 // want to unconditionally add this ctor as it would make wx lib dependent on
1278 // libstdc++ on some Linux versions which is bad, so instead we ask the
1279 // client code to define this wxUSE_STD_STRING symbol if they need it
1280 #if wxUSE_STD_STRING
1281 #if wxUSE_UNICODE_WCHAR
1282 wxString(const wxStdWideString& str) : m_impl(str) {}
1283 #else // UTF-8 or ANSI
1284 wxString(const wxStdWideString& str)
1285 { assign(str.c_str(), str.length()); }
1286 #endif
1287
1288 #if !wxUSE_UNICODE // ANSI build
1289 // FIXME-UTF8: do this in UTF8 build #if wxUSE_UTF8_LOCALE_ONLY, too
1290 wxString(const std::string& str) : m_impl(str) {}
1291 #else // Unicode
1292 wxString(const std::string& str)
1293 { assign(str.c_str(), str.length()); }
1294 #endif
1295 #endif // wxUSE_STD_STRING
1296
1297 // Unlike ctor from std::string, we provide conversion to std::string only
1298 // if wxUSE_STL and not merely wxUSE_STD_STRING (which is on by default),
1299 // because it conflicts with operator const char/wchar_t*:
1300 #if wxUSE_STL
1301 #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING
1302 // wxStringImpl is std::string in the encoding we want
1303 operator const wxStdWideString&() const { return m_impl; }
1304 #else
1305 // wxStringImpl is either not std::string or needs conversion
1306 operator wxStdWideString() const
1307 // FIXME-UTF8: broken for embedded NULs
1308 { return wxStdWideString(wc_str()); }
1309 #endif
1310
1311 #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING
1312 // wxStringImpl is std::string in the encoding we want
1313 operator const std::string&() const { return m_impl; }
1314 #else
1315 // wxStringImpl is either not std::string or needs conversion
1316 operator std::string() const
1317 // FIXME-UTF8: broken for embedded NULs
1318 { return std::string(mb_str()); }
1319 #endif
1320 #endif // wxUSE_STL
1321
1322 wxString Clone() const
1323 {
1324 // make a deep copy of the string, i.e. the returned string will have
1325 // ref count = 1 with refcounted implementation
1326 return wxString::FromImpl(wxStringImpl(m_impl.c_str(), m_impl.length()));
1327 }
1328
1329 // first valid index position
1330 const_iterator begin() const { return const_iterator(this, m_impl.begin()); }
1331 iterator begin() { return iterator(this, m_impl.begin()); }
1332 // position one after the last valid one
1333 const_iterator end() const { return const_iterator(this, m_impl.end()); }
1334 iterator end() { return iterator(this, m_impl.end()); }
1335
1336 // first element of the reversed string
1337 const_reverse_iterator rbegin() const
1338 { return const_reverse_iterator(end()); }
1339 reverse_iterator rbegin()
1340 { return reverse_iterator(end()); }
1341 // one beyond the end of the reversed string
1342 const_reverse_iterator rend() const
1343 { return const_reverse_iterator(begin()); }
1344 reverse_iterator rend()
1345 { return reverse_iterator(begin()); }
1346
1347 // std::string methods:
1348 #if wxUSE_UNICODE_UTF8
1349 size_t length() const
1350 {
1351 #if wxUSE_STRING_POS_CACHE
1352 wxCACHE_PROFILE_FIELD_INC(lentot);
1353
1354 Cache::Element * const cache = GetCacheElement();
1355
1356 if ( cache->len == npos )
1357 {
1358 // it's probably not worth trying to be clever and using cache->pos
1359 // here as it's probably 0 anyhow -- you usually call length() before
1360 // starting to index the string
1361 cache->len = end() - begin();
1362 }
1363 else
1364 {
1365 wxCACHE_PROFILE_FIELD_INC(lenhits);
1366
1367 wxSTRING_CACHE_ASSERT( (int)cache->len == end() - begin() );
1368 }
1369
1370 return cache->len;
1371 #else // !wxUSE_STRING_POS_CACHE
1372 return end() - begin();
1373 #endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
1374 }
1375 #else
1376 size_t length() const { return m_impl.length(); }
1377 #endif
1378
1379 size_type size() const { return length(); }
1380 size_type max_size() const { return npos; }
1381
1382 bool empty() const { return m_impl.empty(); }
1383
1384 // NB: these methods don't have a well-defined meaning in UTF-8 case
1385 size_type capacity() const { return m_impl.capacity(); }
1386 void reserve(size_t sz) { m_impl.reserve(sz); }
1387
1388 void resize(size_t nSize, wxUniChar ch = wxT('\0'))
1389 {
1390 const size_t len = length();
1391 if ( nSize == len)
1392 return;
1393
1394 #if wxUSE_UNICODE_UTF8
1395 if ( nSize < len )
1396 {
1397 wxSTRING_INVALIDATE_CACHE();
1398
1399 // we can't use wxStringImpl::resize() for truncating the string as it
1400 // counts in bytes, not characters
1401 erase(nSize);
1402 return;
1403 }
1404
1405 // we also can't use (presumably more efficient) resize() if we have to
1406 // append characters taking more than one byte
1407 if ( !ch.IsAscii() )
1408 {
1409 append(nSize - len, ch);
1410 }
1411 else // can use (presumably faster) resize() version
1412 #endif // wxUSE_UNICODE_UTF8
1413 {
1414 wxSTRING_INVALIDATE_CACHED_LENGTH();
1415
1416 m_impl.resize(nSize, (wxStringCharType)ch);
1417 }
1418 }
1419
1420 wxString substr(size_t nStart = 0, size_t nLen = npos) const
1421 {
1422 size_t pos, len;
1423 PosLenToImpl(nStart, nLen, &pos, &len);
1424 return FromImpl(m_impl.substr(pos, len));
1425 }
1426
1427 // generic attributes & operations
1428 // as standard strlen()
1429 size_t Len() const { return length(); }
1430 // string contains any characters?
1431 bool IsEmpty() const { return empty(); }
1432 // empty string is "false", so !str will return true
1433 bool operator!() const { return empty(); }
1434 // truncate the string to given length
1435 wxString& Truncate(size_t uiLen);
1436 // empty string contents
1437 void Empty()
1438 {
1439 Truncate(0);
1440
1441 wxASSERT_MSG( empty(), _T("string not empty after call to Empty()?") );
1442 }
1443 // empty the string and free memory
1444 void Clear() { clear(); }
1445
1446 // contents test
1447 // Is an ascii value
1448 bool IsAscii() const;
1449 // Is a number
1450 bool IsNumber() const;
1451 // Is a word
1452 bool IsWord() const;
1453
1454 // data access (all indexes are 0 based)
1455 // read access
1456 wxUniChar at(size_t n) const
1457 { return wxStringOperations::DecodeChar(m_impl.begin() + PosToImpl(n)); }
1458 wxUniChar GetChar(size_t n) const
1459 { return at(n); }
1460 // read/write access
1461 wxUniCharRef at(size_t n)
1462 { return *GetIterForNthChar(n); }
1463 wxUniCharRef GetWritableChar(size_t n)
1464 { return at(n); }
1465 // write access
1466 void SetChar(size_t n, wxUniChar ch)
1467 { at(n) = ch; }
1468
1469 // get last character
1470 wxUniChar Last() const
1471 {
1472 wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
1473 return *rbegin();
1474 }
1475
1476 // get writable last character
1477 wxUniCharRef Last()
1478 {
1479 wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
1480 return *rbegin();
1481 }
1482
1483 /*
1484 Note that we we must define all of the overloads below to avoid
1485 ambiguity when using str[0].
1486 */
1487 wxUniChar operator[](int n) const
1488 { return at(n); }
1489 wxUniChar operator[](long n) const
1490 { return at(n); }
1491 wxUniChar operator[](size_t n) const
1492 { return at(n); }
1493 #ifndef wxSIZE_T_IS_UINT
1494 wxUniChar operator[](unsigned int n) const
1495 { return at(n); }
1496 #endif // size_t != unsigned int
1497
1498 // operator versions of GetWriteableChar()
1499 wxUniCharRef operator[](int n)
1500 { return at(n); }
1501 wxUniCharRef operator[](long n)
1502 { return at(n); }
1503 wxUniCharRef operator[](size_t n)
1504 { return at(n); }
1505 #ifndef wxSIZE_T_IS_UINT
1506 wxUniCharRef operator[](unsigned int n)
1507 { return at(n); }
1508 #endif // size_t != unsigned int
1509
1510 // explicit conversion to C string (use this with printf()!)
1511 wxCStrData c_str() const { return wxCStrData(this); }
1512 wxCStrData data() const { return c_str(); }
1513
1514 // implicit conversion to C string
1515 operator wxCStrData() const { return c_str(); }
1516
1517 // the first two operators conflict with operators for conversion to
1518 // std::string and they must be disabled in STL build; the next one only
1519 // makes sense if conversions to char* are also defined and not defining it
1520 // in STL build also helps us to get more clear error messages for the code
1521 // which relies on implicit conversion to char* in STL build
1522 #if !wxUSE_STL
1523 operator const char*() const { return c_str(); }
1524 operator const wchar_t*() const { return c_str(); }
1525
1526 // implicit conversion to untyped pointer for compatibility with previous
1527 // wxWidgets versions: this is the same as conversion to const char * so it
1528 // may fail!
1529 operator const void*() const { return c_str(); }
1530 #endif // wxUSE_STL
1531
1532 // identical to c_str(), for MFC compatibility
1533 const wxCStrData GetData() const { return c_str(); }
1534
1535 // explicit conversion to C string in internal representation (char*,
1536 // wchar_t*, UTF-8-encoded char*, depending on the build):
1537 const wxStringCharType *wx_str() const { return m_impl.c_str(); }
1538
1539 // conversion to *non-const* multibyte or widestring buffer; modifying
1540 // returned buffer won't affect the string, these methods are only useful
1541 // for passing values to const-incorrect functions
1542 wxWritableCharBuffer char_str(const wxMBConv& conv = wxConvLibc) const
1543 { return mb_str(conv); }
1544 wxWritableWCharBuffer wchar_str() const { return wc_str(); }
1545
1546 // conversion to the buffer of the given type T (= char or wchar_t) and
1547 // also optionally return the buffer length
1548 //
1549 // this is mostly/only useful for the template functions
1550 //
1551 // FIXME-VC6: the second argument only exists for VC6 which doesn't support
1552 // explicit template function selection, do not use it unless
1553 // you must support VC6!
1554 template <typename T>
1555 wxCharTypeBuffer<T> tchar_str(size_t *len = NULL,
1556 T * WXUNUSED(dummy) = NULL) const
1557 {
1558 #if wxUSE_UNICODE
1559 // we need a helper dispatcher depending on type
1560 return wxPrivate::wxStringAsBufHelper<T>::Get(*this, len);
1561 #else // ANSI
1562 // T can only be char in ANSI build
1563 if ( len )
1564 *len = length();
1565
1566 return wxCharTypeBuffer<T>::CreateNonOwned(wx_str());
1567 #endif // Unicode build kind
1568 }
1569
1570 // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for
1571 // converting numbers or strings which are certain not to contain special
1572 // chars (typically system functions, X atoms, environment variables etc.)
1573 //
1574 // the behaviour of these functions with the strings containing anything
1575 // else than 7 bit ASCII characters is undefined, use at your own risk.
1576 #if wxUSE_UNICODE
1577 static wxString FromAscii(const char *ascii, size_t len);
1578 static wxString FromAscii(const char *ascii);
1579 static wxString FromAscii(char ascii);
1580 const wxCharBuffer ToAscii() const;
1581 #else // ANSI
1582 static wxString FromAscii(const char *ascii) { return wxString( ascii ); }
1583 static wxString FromAscii(const char *ascii, size_t len)
1584 { return wxString( ascii, len ); }
1585 static wxString FromAscii(char ascii) { return wxString( ascii ); }
1586 const char *ToAscii() const { return c_str(); }
1587 #endif // Unicode/!Unicode
1588
1589 // also provide unsigned char overloads as signed/unsigned doesn't matter
1590 // for 7 bit ASCII characters
1591 static wxString FromAscii(const unsigned char *ascii)
1592 { return FromAscii((const char *)ascii); }
1593 static wxString FromAscii(const unsigned char *ascii, size_t len)
1594 { return FromAscii((const char *)ascii, len); }
1595
1596 // conversion to/from UTF-8:
1597 #if wxUSE_UNICODE_UTF8
1598 static wxString FromUTF8Unchecked(const char *utf8)
1599 {
1600 if ( !utf8 )
1601 return wxEmptyString;
1602
1603 wxASSERT( wxStringOperations::IsValidUtf8String(utf8) );
1604 return FromImpl(wxStringImpl(utf8));
1605 }
1606 static wxString FromUTF8Unchecked(const char *utf8, size_t len)
1607 {
1608 if ( !utf8 )
1609 return wxEmptyString;
1610 if ( len == npos )
1611 return FromUTF8Unchecked(utf8);
1612
1613 wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) );
1614 return FromImpl(wxStringImpl(utf8, len));
1615 }
1616
1617 static wxString FromUTF8(const char *utf8)
1618 {
1619 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8) )
1620 return "";
1621
1622 return FromImpl(wxStringImpl(utf8));
1623 }
1624 static wxString FromUTF8(const char *utf8, size_t len)
1625 {
1626 if ( len == npos )
1627 return FromUTF8(utf8);
1628
1629 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8, len) )
1630 return "";
1631
1632 return FromImpl(wxStringImpl(utf8, len));
1633 }
1634
1635 const char* utf8_str() const { return wx_str(); }
1636 const char* ToUTF8() const { return wx_str(); }
1637
1638 // this function exists in UTF-8 build only and returns the length of the
1639 // internal UTF-8 representation
1640 size_t utf8_length() const { return m_impl.length(); }
1641 #elif wxUSE_UNICODE_WCHAR
1642 static wxString FromUTF8(const char *utf8, size_t len = npos)
1643 { return wxString(utf8, wxMBConvUTF8(), len); }
1644 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1645 {
1646 const wxString s(utf8, wxMBConvUTF8(), len);
1647 wxASSERT_MSG( !utf8 || !*utf8 || !s.empty(),
1648 "string must be valid UTF-8" );
1649 return s;
1650 }
1651 const wxCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); }
1652 const wxCharBuffer ToUTF8() const { return utf8_str(); }
1653 #else // ANSI
1654 static wxString FromUTF8(const char *utf8)
1655 { return wxString(wxMBConvUTF8().cMB2WC(utf8)); }
1656 static wxString FromUTF8(const char *utf8, size_t len)
1657 {
1658 size_t wlen;
1659 wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen));
1660 return wxString(buf.data(), wlen);
1661 }
1662 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1663 {
1664 size_t wlen;
1665 wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8,
1666 len == npos ? wxNO_LEN : len,
1667 &wlen));
1668 wxASSERT_MSG( !utf8 || !*utf8 || wlen,
1669 "string must be valid UTF-8" );
1670
1671 return wxString(buf.data(), wlen);
1672 }
1673 const wxCharBuffer utf8_str() const
1674 { return wxMBConvUTF8().cWC2MB(wc_str()); }
1675 const wxCharBuffer ToUTF8() const { return utf8_str(); }
1676 #endif
1677
1678 // functions for storing binary data in wxString:
1679 #if wxUSE_UNICODE
1680 static wxString From8BitData(const char *data, size_t len)
1681 { return wxString(data, wxConvISO8859_1, len); }
1682 // version for NUL-terminated data:
1683 static wxString From8BitData(const char *data)
1684 { return wxString(data, wxConvISO8859_1); }
1685 const wxCharBuffer To8BitData() const { return mb_str(wxConvISO8859_1); }
1686 #else // ANSI
1687 static wxString From8BitData(const char *data, size_t len)
1688 { return wxString(data, len); }
1689 // version for NUL-terminated data:
1690 static wxString From8BitData(const char *data)
1691 { return wxString(data); }
1692 const char *To8BitData() const { return c_str(); }
1693 #endif // Unicode/ANSI
1694
1695 // conversions with (possible) format conversions: have to return a
1696 // buffer with temporary data
1697 //
1698 // the functions defined (in either Unicode or ANSI) mode are mb_str() to
1699 // return an ANSI (multibyte) string, wc_str() to return a wide string and
1700 // fn_str() to return a string which should be used with the OS APIs
1701 // accepting the file names. The return value is always the same, but the
1702 // type differs because a function may either return pointer to the buffer
1703 // directly or have to use intermediate buffer for translation.
1704 #if wxUSE_UNICODE
1705
1706 #if wxUSE_UTF8_LOCALE_ONLY
1707 const char* mb_str() const { return wx_str(); }
1708 const wxCharBuffer mb_str(const wxMBConv& conv) const;
1709 #else
1710 const wxCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const;
1711 #endif
1712
1713 const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
1714
1715 #if wxUSE_UNICODE_WCHAR
1716 const wchar_t* wc_str() const { return wx_str(); }
1717 #elif wxUSE_UNICODE_UTF8
1718 const wxWCharBuffer wc_str() const;
1719 #endif
1720 // for compatibility with !wxUSE_UNICODE version
1721 const wxWX2WCbuf wc_str(const wxMBConv& WXUNUSED(conv)) const
1722 { return wc_str(); }
1723
1724 #if wxMBFILES
1725 const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
1726 #else // !wxMBFILES
1727 const wxWX2WCbuf fn_str() const { return wc_str(); }
1728 #endif // wxMBFILES/!wxMBFILES
1729
1730 #else // ANSI
1731 const wxChar* mb_str() const { return wx_str(); }
1732
1733 // for compatibility with wxUSE_UNICODE version
1734 const char* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); }
1735
1736 const wxWX2MBbuf mbc_str() const { return mb_str(); }
1737
1738 #if wxUSE_WCHAR_T
1739 const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const;
1740 #endif // wxUSE_WCHAR_T
1741 const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); }
1742 #endif // Unicode/ANSI
1743
1744 #if wxUSE_UNICODE_UTF8
1745 const wxWCharBuffer t_str() const { return wc_str(); }
1746 #elif wxUSE_UNICODE_WCHAR
1747 const wchar_t* t_str() const { return wx_str(); }
1748 #else
1749 const char* t_str() const { return wx_str(); }
1750 #endif
1751
1752
1753 // overloaded assignment
1754 // from another wxString
1755 wxString& operator=(const wxString& stringSrc)
1756 {
1757 if ( this != &stringSrc )
1758 {
1759 wxSTRING_INVALIDATE_CACHE();
1760
1761 m_impl = stringSrc.m_impl;
1762 }
1763
1764 return *this;
1765 }
1766
1767 wxString& operator=(const wxCStrData& cstr)
1768 { return *this = cstr.AsString(); }
1769 // from a character
1770 wxString& operator=(wxUniChar ch)
1771 {
1772 wxSTRING_INVALIDATE_CACHE();
1773
1774 #if wxUSE_UNICODE_UTF8
1775 if ( !ch.IsAscii() )
1776 m_impl = wxStringOperations::EncodeChar(ch);
1777 else
1778 #endif // wxUSE_UNICODE_UTF8
1779 m_impl = (wxStringCharType)ch;
1780 return *this;
1781 }
1782
1783 wxString& operator=(wxUniCharRef ch)
1784 { return operator=((wxUniChar)ch); }
1785 wxString& operator=(char ch)
1786 { return operator=(wxUniChar(ch)); }
1787 wxString& operator=(unsigned char ch)
1788 { return operator=(wxUniChar(ch)); }
1789 wxString& operator=(wchar_t ch)
1790 { return operator=(wxUniChar(ch)); }
1791 // from a C string - STL probably will crash on NULL,
1792 // so we need to compensate in that case
1793 #if wxUSE_STL_BASED_WXSTRING
1794 wxString& operator=(const char *psz)
1795 {
1796 wxSTRING_INVALIDATE_CACHE();
1797
1798 if ( psz )
1799 m_impl = ImplStr(psz);
1800 else
1801 clear();
1802
1803 return *this;
1804 }
1805
1806 wxString& operator=(const wchar_t *pwz)
1807 {
1808 wxSTRING_INVALIDATE_CACHE();
1809
1810 if ( pwz )
1811 m_impl = ImplStr(pwz);
1812 else
1813 clear();
1814
1815 return *this;
1816 }
1817 #else // !wxUSE_STL_BASED_WXSTRING
1818 wxString& operator=(const char *psz)
1819 {
1820 wxSTRING_INVALIDATE_CACHE();
1821
1822 m_impl = ImplStr(psz);
1823
1824 return *this;
1825 }
1826
1827 wxString& operator=(const wchar_t *pwz)
1828 {
1829 wxSTRING_INVALIDATE_CACHE();
1830
1831 m_impl = ImplStr(pwz);
1832
1833 return *this;
1834 }
1835 #endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING
1836
1837 wxString& operator=(const unsigned char *psz)
1838 { return operator=((const char*)psz); }
1839
1840 // from wxWCharBuffer
1841 wxString& operator=(const wxWCharBuffer& s)
1842 { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
1843 // from wxCharBuffer
1844 wxString& operator=(const wxCharBuffer& s)
1845 { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
1846
1847 // string concatenation
1848 // in place concatenation
1849 /*
1850 Concatenate and return the result. Note that the left to right
1851 associativity of << allows to write things like "str << str1 << str2
1852 << ..." (unlike with +=)
1853 */
1854 // string += string
1855 wxString& operator<<(const wxString& s)
1856 {
1857 #if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
1858 wxASSERT_MSG( s.IsValid(),
1859 _T("did you forget to call UngetWriteBuf()?") );
1860 #endif
1861
1862 append(s);
1863 return *this;
1864 }
1865 // string += C string
1866 wxString& operator<<(const char *psz)
1867 { append(psz); return *this; }
1868 wxString& operator<<(const wchar_t *pwz)
1869 { append(pwz); return *this; }
1870 wxString& operator<<(const wxCStrData& psz)
1871 { append(psz.AsString()); return *this; }
1872 // string += char
1873 wxString& operator<<(wxUniChar ch) { append(1, ch); return *this; }
1874 wxString& operator<<(wxUniCharRef ch) { append(1, ch); return *this; }
1875 wxString& operator<<(char ch) { append(1, ch); return *this; }
1876 wxString& operator<<(unsigned char ch) { append(1, ch); return *this; }
1877 wxString& operator<<(wchar_t ch) { append(1, ch); return *this; }
1878
1879 // string += buffer (i.e. from wxGetString)
1880 wxString& operator<<(const wxWCharBuffer& s)
1881 { return operator<<((const wchar_t *)s); }
1882 wxString& operator<<(const wxCharBuffer& s)
1883 { return operator<<((const char *)s); }
1884
1885 // string += C string
1886 wxString& Append(const wxString& s)
1887 {
1888 // test for empty() to share the string if possible
1889 if ( empty() )
1890 *this = s;
1891 else
1892 append(s);
1893 return *this;
1894 }
1895 wxString& Append(const char* psz)
1896 { append(psz); return *this; }
1897 wxString& Append(const wchar_t* pwz)
1898 { append(pwz); return *this; }
1899 wxString& Append(const wxCStrData& psz)
1900 { append(psz); return *this; }
1901 wxString& Append(const wxCharBuffer& psz)
1902 { append(psz); return *this; }
1903 wxString& Append(const wxWCharBuffer& psz)
1904 { append(psz); return *this; }
1905 wxString& Append(const char* psz, size_t nLen)
1906 { append(psz, nLen); return *this; }
1907 wxString& Append(const wchar_t* pwz, size_t nLen)
1908 { append(pwz, nLen); return *this; }
1909 wxString& Append(const wxCStrData& psz, size_t nLen)
1910 { append(psz, nLen); return *this; }
1911 wxString& Append(const wxCharBuffer& psz, size_t nLen)
1912 { append(psz, nLen); return *this; }
1913 wxString& Append(const wxWCharBuffer& psz, size_t nLen)
1914 { append(psz, nLen); return *this; }
1915 // append count copies of given character
1916 wxString& Append(wxUniChar ch, size_t count = 1u)
1917 { append(count, ch); return *this; }
1918 wxString& Append(wxUniCharRef ch, size_t count = 1u)
1919 { append(count, ch); return *this; }
1920 wxString& Append(char ch, size_t count = 1u)
1921 { append(count, ch); return *this; }
1922 wxString& Append(unsigned char ch, size_t count = 1u)
1923 { append(count, ch); return *this; }
1924 wxString& Append(wchar_t ch, size_t count = 1u)
1925 { append(count, ch); return *this; }
1926
1927 // prepend a string, return the string itself
1928 wxString& Prepend(const wxString& str)
1929 { *this = str + *this; return *this; }
1930
1931 // non-destructive concatenation
1932 // two strings
1933 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string1,
1934 const wxString& string2);
1935 // string with a single char
1936 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
1937 // char with a string
1938 friend wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
1939 // string with C string
1940 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
1941 const char *psz);
1942 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
1943 const wchar_t *pwz);
1944 // C string with string
1945 friend wxString WXDLLIMPEXP_BASE operator+(const char *psz,
1946 const wxString& string);
1947 friend wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz,
1948 const wxString& string);
1949
1950 // stream-like functions
1951 // insert an int into string
1952 wxString& operator<<(int i)
1953 { return (*this) << Format(_T("%d"), i); }
1954 // insert an unsigned int into string
1955 wxString& operator<<(unsigned int ui)
1956 { return (*this) << Format(_T("%u"), ui); }
1957 // insert a long into string
1958 wxString& operator<<(long l)
1959 { return (*this) << Format(_T("%ld"), l); }
1960 // insert an unsigned long into string
1961 wxString& operator<<(unsigned long ul)
1962 { return (*this) << Format(_T("%lu"), ul); }
1963 #if defined wxLongLong_t && !defined wxLongLongIsLong
1964 // insert a long long if they exist and aren't longs
1965 wxString& operator<<(wxLongLong_t ll)
1966 {
1967 const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("d");
1968 return (*this) << Format(fmt, ll);
1969 }
1970 // insert an unsigned long long
1971 wxString& operator<<(wxULongLong_t ull)
1972 {
1973 const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("u");
1974 return (*this) << Format(fmt , ull);
1975 }
1976 #endif // wxLongLong_t && !wxLongLongIsLong
1977 // insert a float into string
1978 wxString& operator<<(float f)
1979 { return (*this) << Format(_T("%f"), f); }
1980 // insert a double into string
1981 wxString& operator<<(double d)
1982 { return (*this) << Format(_T("%g"), d); }
1983
1984 // string comparison
1985 // case-sensitive comparison (returns a value < 0, = 0 or > 0)
1986 int Cmp(const char *psz) const
1987 { return compare(psz); }
1988 int Cmp(const wchar_t *pwz) const
1989 { return compare(pwz); }
1990 int Cmp(const wxString& s) const
1991 { return compare(s); }
1992 int Cmp(const wxCStrData& s) const
1993 { return compare(s); }
1994 int Cmp(const wxCharBuffer& s) const
1995 { return compare(s); }
1996 int Cmp(const wxWCharBuffer& s) const
1997 { return compare(s); }
1998 // same as Cmp() but not case-sensitive
1999 int CmpNoCase(const wxString& s) const;
2000
2001 // test for the string equality, either considering case or not
2002 // (if compareWithCase then the case matters)
2003 bool IsSameAs(const wxString& str, bool compareWithCase = true) const
2004 {
2005 #if !wxUSE_UNICODE_UTF8
2006 // in UTF-8 build, length() is O(n) and doing this would be _slower_
2007 if ( length() != str.length() )
2008 return false;
2009 #endif
2010 return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0;
2011 }
2012 bool IsSameAs(const char *str, bool compareWithCase = true) const
2013 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2014 bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const
2015 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2016
2017 bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const
2018 { return IsSameAs(str.AsString(), compareWithCase); }
2019 bool IsSameAs(const wxCharBuffer& str, bool compareWithCase = true) const
2020 { return IsSameAs(str.data(), compareWithCase); }
2021 bool IsSameAs(const wxWCharBuffer& str, bool compareWithCase = true) const
2022 { return IsSameAs(str.data(), compareWithCase); }
2023 // comparison with a single character: returns true if equal
2024 bool IsSameAs(wxUniChar c, bool compareWithCase = true) const;
2025 // FIXME-UTF8: remove these overloads
2026 bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const
2027 { return IsSameAs(wxUniChar(c), compareWithCase); }
2028 bool IsSameAs(char c, bool compareWithCase = true) const
2029 { return IsSameAs(wxUniChar(c), compareWithCase); }
2030 bool IsSameAs(unsigned char c, bool compareWithCase = true) const
2031 { return IsSameAs(wxUniChar(c), compareWithCase); }
2032 bool IsSameAs(wchar_t c, bool compareWithCase = true) const
2033 { return IsSameAs(wxUniChar(c), compareWithCase); }
2034 bool IsSameAs(int c, bool compareWithCase = true) const
2035 { return IsSameAs(wxUniChar(c), compareWithCase); }
2036
2037 // simple sub-string extraction
2038 // return substring starting at nFirst of length nCount (or till the end
2039 // if nCount = default value)
2040 wxString Mid(size_t nFirst, size_t nCount = npos) const;
2041
2042 // operator version of Mid()
2043 wxString operator()(size_t start, size_t len) const
2044 { return Mid(start, len); }
2045
2046 // check if the string starts with the given prefix and return the rest
2047 // of the string in the provided pointer if it is not NULL; otherwise
2048 // return false
2049 bool StartsWith(const wxString& prefix, wxString *rest = NULL) const;
2050 // check if the string ends with the given suffix and return the
2051 // beginning of the string before the suffix in the provided pointer if
2052 // it is not NULL; otherwise return false
2053 bool EndsWith(const wxString& suffix, wxString *rest = NULL) const;
2054
2055 // get first nCount characters
2056 wxString Left(size_t nCount) const;
2057 // get last nCount characters
2058 wxString Right(size_t nCount) const;
2059 // get all characters before the first occurance of ch
2060 // (returns the whole string if ch not found)
2061 wxString BeforeFirst(wxUniChar ch) const;
2062 // get all characters before the last occurence of ch
2063 // (returns empty string if ch not found)
2064 wxString BeforeLast(wxUniChar ch) const;
2065 // get all characters after the first occurence of ch
2066 // (returns empty string if ch not found)
2067 wxString AfterFirst(wxUniChar ch) const;
2068 // get all characters after the last occurence of ch
2069 // (returns the whole string if ch not found)
2070 wxString AfterLast(wxUniChar ch) const;
2071
2072 // for compatibility only, use more explicitly named functions above
2073 wxString Before(wxUniChar ch) const { return BeforeLast(ch); }
2074 wxString After(wxUniChar ch) const { return AfterFirst(ch); }
2075
2076 // case conversion
2077 // convert to upper case in place, return the string itself
2078 wxString& MakeUpper();
2079 // convert to upper case, return the copy of the string
2080 wxString Upper() const { return wxString(*this).MakeUpper(); }
2081 // convert to lower case in place, return the string itself
2082 wxString& MakeLower();
2083 // convert to lower case, return the copy of the string
2084 wxString Lower() const { return wxString(*this).MakeLower(); }
2085 // convert the first character to the upper case and the rest to the
2086 // lower one, return the modified string itself
2087 wxString& MakeCapitalized();
2088 // convert the first character to the upper case and the rest to the
2089 // lower one, return the copy of the string
2090 wxString Capitalize() const { return wxString(*this).MakeCapitalized(); }
2091
2092 // trimming/padding whitespace (either side) and truncating
2093 // remove spaces from left or from right (default) side
2094 wxString& Trim(bool bFromRight = true);
2095 // add nCount copies chPad in the beginning or at the end (default)
2096 wxString& Pad(size_t nCount, wxUniChar chPad = wxT(' '), bool bFromRight = true);
2097
2098 // searching and replacing
2099 // searching (return starting index, or -1 if not found)
2100 int Find(wxUniChar ch, bool bFromEnd = false) const; // like strchr/strrchr
2101 int Find(wxUniCharRef ch, bool bFromEnd = false) const
2102 { return Find(wxUniChar(ch), bFromEnd); }
2103 int Find(char ch, bool bFromEnd = false) const
2104 { return Find(wxUniChar(ch), bFromEnd); }
2105 int Find(unsigned char ch, bool bFromEnd = false) const
2106 { return Find(wxUniChar(ch), bFromEnd); }
2107 int Find(wchar_t ch, bool bFromEnd = false) const
2108 { return Find(wxUniChar(ch), bFromEnd); }
2109 // searching (return starting index, or -1 if not found)
2110 int Find(const wxString& sub) const // like strstr
2111 {
2112 size_type idx = find(sub);
2113 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2114 }
2115 int Find(const char *sub) const // like strstr
2116 {
2117 size_type idx = find(sub);
2118 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2119 }
2120 int Find(const wchar_t *sub) const // like strstr
2121 {
2122 size_type idx = find(sub);
2123 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2124 }
2125
2126 int Find(const wxCStrData& sub) const
2127 { return Find(sub.AsString()); }
2128 int Find(const wxCharBuffer& sub) const
2129 { return Find(sub.data()); }
2130 int Find(const wxWCharBuffer& sub) const
2131 { return Find(sub.data()); }
2132
2133 // replace first (or all of bReplaceAll) occurences of substring with
2134 // another string, returns the number of replacements made
2135 size_t Replace(const wxString& strOld,
2136 const wxString& strNew,
2137 bool bReplaceAll = true);
2138
2139 // check if the string contents matches a mask containing '*' and '?'
2140 bool Matches(const wxString& mask) const;
2141
2142 // conversion to numbers: all functions return true only if the whole
2143 // string is a number and put the value of this number into the pointer
2144 // provided, the base is the numeric base in which the conversion should be
2145 // done and must be comprised between 2 and 36 or be 0 in which case the
2146 // standard C rules apply (leading '0' => octal, "0x" => hex)
2147 // convert to a signed integer
2148 bool ToLong(long *val, int base = 10) const;
2149 // convert to an unsigned integer
2150 bool ToULong(unsigned long *val, int base = 10) const;
2151 // convert to wxLongLong
2152 #if defined(wxLongLong_t)
2153 bool ToLongLong(wxLongLong_t *val, int base = 10) const;
2154 // convert to wxULongLong
2155 bool ToULongLong(wxULongLong_t *val, int base = 10) const;
2156 #endif // wxLongLong_t
2157 // convert to a double
2158 bool ToDouble(double *val) const;
2159
2160
2161 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2162 // formatted input/output
2163 // as sprintf(), returns the number of characters written or < 0 on error
2164 // (take 'this' into account in attribute parameter count)
2165 // int Printf(const wxString& format, ...);
2166 WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),
2167 DoPrintfWchar, DoPrintfUtf8)
2168 #ifdef __WATCOMC__
2169 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2170 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxString&),
2171 (wxFormatString(f1)));
2172 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxCStrData&),
2173 (wxFormatString(f1)));
2174 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const char*),
2175 (wxFormatString(f1)));
2176 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wchar_t*),
2177 (wxFormatString(f1)));
2178 #endif
2179 #endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN
2180 // as vprintf(), returns the number of characters written or < 0 on error
2181 int PrintfV(const wxString& format, va_list argptr);
2182
2183 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2184 // returns the string containing the result of Printf() to it
2185 // static wxString Format(const wxString& format, ...) ATTRIBUTE_PRINTF_1;
2186 WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxFormatString&),
2187 DoFormatWchar, DoFormatUtf8)
2188 #ifdef __WATCOMC__
2189 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2190 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxString&),
2191 (wxFormatString(f1)));
2192 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxCStrData&),
2193 (wxFormatString(f1)));
2194 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const char*),
2195 (wxFormatString(f1)));
2196 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wchar_t*),
2197 (wxFormatString(f1)));
2198 #endif
2199 #endif
2200 // the same as above, but takes a va_list
2201 static wxString FormatV(const wxString& format, va_list argptr);
2202
2203 // raw access to string memory
2204 // ensure that string has space for at least nLen characters
2205 // only works if the data of this string is not shared
2206 bool Alloc(size_t nLen) { reserve(nLen); return capacity() >= nLen; }
2207 // minimize the string's memory
2208 // only works if the data of this string is not shared
2209 bool Shrink();
2210 #if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
2211 // These are deprecated, use wxStringBuffer or wxStringBufferLength instead
2212 //
2213 // get writable buffer of at least nLen bytes. Unget() *must* be called
2214 // a.s.a.p. to put string back in a reasonable state!
2215 wxDEPRECATED( wxStringCharType *GetWriteBuf(size_t nLen) );
2216 // call this immediately after GetWriteBuf() has been used
2217 wxDEPRECATED( void UngetWriteBuf() );
2218 wxDEPRECATED( void UngetWriteBuf(size_t nLen) );
2219 #endif // WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && wxUSE_UNICODE_UTF8
2220
2221 // wxWidgets version 1 compatibility functions
2222
2223 // use Mid()
2224 wxString SubString(size_t from, size_t to) const
2225 { return Mid(from, (to - from + 1)); }
2226 // values for second parameter of CompareTo function
2227 enum caseCompare {exact, ignoreCase};
2228 // values for first parameter of Strip function
2229 enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
2230
2231 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2232 // use Printf()
2233 // (take 'this' into account in attribute parameter count)
2234 // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;
2235 WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),
2236 DoPrintfWchar, DoPrintfUtf8)
2237 #ifdef __WATCOMC__
2238 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2239 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxString&),
2240 (wxFormatString(f1)));
2241 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxCStrData&),
2242 (wxFormatString(f1)));
2243 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const char*),
2244 (wxFormatString(f1)));
2245 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wchar_t*),
2246 (wxFormatString(f1)));
2247 #endif
2248 #endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
2249
2250 // use Cmp()
2251 int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
2252 { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
2253
2254 // use length()
2255 size_t Length() const { return length(); }
2256 // Count the number of characters
2257 int Freq(wxUniChar ch) const;
2258 // use MakeLower
2259 void LowerCase() { MakeLower(); }
2260 // use MakeUpper
2261 void UpperCase() { MakeUpper(); }
2262 // use Trim except that it doesn't change this string
2263 wxString Strip(stripType w = trailing) const;
2264
2265 // use Find (more general variants not yet supported)
2266 size_t Index(const wxChar* psz) const { return Find(psz); }
2267 size_t Index(wxUniChar ch) const { return Find(ch); }
2268 // use Truncate
2269 wxString& Remove(size_t pos) { return Truncate(pos); }
2270 wxString& RemoveLast(size_t n = 1) { return Truncate(length() - n); }
2271
2272 wxString& Remove(size_t nStart, size_t nLen)
2273 { return (wxString&)erase( nStart, nLen ); }
2274
2275 // use Find()
2276 int First( wxUniChar ch ) const { return Find(ch); }
2277 int First( wxUniCharRef ch ) const { return Find(ch); }
2278 int First( char ch ) const { return Find(ch); }
2279 int First( unsigned char ch ) const { return Find(ch); }
2280 int First( wchar_t ch ) const { return Find(ch); }
2281 int First( const wxString& str ) const { return Find(str); }
2282 int Last( wxUniChar ch ) const { return Find(ch, true); }
2283 bool Contains(const wxString& str) const { return Find(str) != wxNOT_FOUND; }
2284
2285 // use empty()
2286 bool IsNull() const { return empty(); }
2287
2288 // std::string compatibility functions
2289
2290 // take nLen chars starting at nPos
2291 wxString(const wxString& str, size_t nPos, size_t nLen)
2292 { assign(str, nPos, nLen); }
2293 // take all characters from first to last
2294 wxString(const_iterator first, const_iterator last)
2295 : m_impl(first.impl(), last.impl()) { }
2296 #if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2297 // the 2 overloads below are for compatibility with the existing code using
2298 // pointers instead of iterators
2299 wxString(const char *first, const char *last)
2300 {
2301 SubstrBufFromMB str(ImplStr(first, last - first));
2302 m_impl.assign(str.data, str.len);
2303 }
2304 wxString(const wchar_t *first, const wchar_t *last)
2305 {
2306 SubstrBufFromWC str(ImplStr(first, last - first));
2307 m_impl.assign(str.data, str.len);
2308 }
2309 // and this one is needed to compile code adding offsets to c_str() result
2310 wxString(const wxCStrData& first, const wxCStrData& last)
2311 : m_impl(CreateConstIterator(first).impl(),
2312 CreateConstIterator(last).impl())
2313 {
2314 wxASSERT_MSG( first.m_str == last.m_str,
2315 _T("pointers must be into the same string") );
2316 }
2317 #endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2318
2319 // lib.string.modifiers
2320 // append elements str[pos], ..., str[pos+n]
2321 wxString& append(const wxString& str, size_t pos, size_t n)
2322 {
2323 wxSTRING_UPDATE_CACHED_LENGTH(n);
2324
2325 size_t from, len;
2326 str.PosLenToImpl(pos, n, &from, &len);
2327 m_impl.append(str.m_impl, from, len);
2328 return *this;
2329 }
2330 // append a string
2331 wxString& append(const wxString& str)
2332 {
2333 wxSTRING_UPDATE_CACHED_LENGTH(str.length());
2334
2335 m_impl.append(str.m_impl);
2336 return *this;
2337 }
2338
2339 // append first n (or all if n == npos) characters of sz
2340 wxString& append(const char *sz)
2341 {
2342 wxSTRING_INVALIDATE_CACHED_LENGTH();
2343
2344 m_impl.append(ImplStr(sz));
2345 return *this;
2346 }
2347
2348 wxString& append(const wchar_t *sz)
2349 {
2350 wxSTRING_INVALIDATE_CACHED_LENGTH();
2351
2352 m_impl.append(ImplStr(sz));
2353 return *this;
2354 }
2355
2356 wxString& append(const char *sz, size_t n)
2357 {
2358 wxSTRING_INVALIDATE_CACHED_LENGTH();
2359
2360 SubstrBufFromMB str(ImplStr(sz, n));
2361 m_impl.append(str.data, str.len);
2362 return *this;
2363 }
2364 wxString& append(const wchar_t *sz, size_t n)
2365 {
2366 wxSTRING_UPDATE_CACHED_LENGTH(n);
2367
2368 SubstrBufFromWC str(ImplStr(sz, n));
2369 m_impl.append(str.data, str.len);
2370 return *this;
2371 }
2372
2373 wxString& append(const wxCStrData& str)
2374 { return append(str.AsString()); }
2375 wxString& append(const wxCharBuffer& str)
2376 { return append(str.data()); }
2377 wxString& append(const wxWCharBuffer& str)
2378 { return append(str.data()); }
2379 wxString& append(const wxCStrData& str, size_t n)
2380 { return append(str.AsString(), 0, n); }
2381 wxString& append(const wxCharBuffer& str, size_t n)
2382 { return append(str.data(), n); }
2383 wxString& append(const wxWCharBuffer& str, size_t n)
2384 { return append(str.data(), n); }
2385
2386 // append n copies of ch
2387 wxString& append(size_t n, wxUniChar ch)
2388 {
2389 #if wxUSE_UNICODE_UTF8
2390 if ( !ch.IsAscii() )
2391 {
2392 wxSTRING_INVALIDATE_CACHED_LENGTH();
2393
2394 m_impl.append(wxStringOperations::EncodeNChars(n, ch));
2395 }
2396 else // ASCII
2397 #endif
2398 {
2399 wxSTRING_UPDATE_CACHED_LENGTH(n);
2400
2401 m_impl.append(n, (wxStringCharType)ch);
2402 }
2403
2404 return *this;
2405 }
2406
2407 wxString& append(size_t n, wxUniCharRef ch)
2408 { return append(n, wxUniChar(ch)); }
2409 wxString& append(size_t n, char ch)
2410 { return append(n, wxUniChar(ch)); }
2411 wxString& append(size_t n, unsigned char ch)
2412 { return append(n, wxUniChar(ch)); }
2413 wxString& append(size_t n, wchar_t ch)
2414 { return append(n, wxUniChar(ch)); }
2415
2416 // append from first to last
2417 wxString& append(const_iterator first, const_iterator last)
2418 {
2419 wxSTRING_INVALIDATE_CACHED_LENGTH();
2420
2421 m_impl.append(first.impl(), last.impl());
2422 return *this;
2423 }
2424 #if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2425 wxString& append(const char *first, const char *last)
2426 { return append(first, last - first); }
2427 wxString& append(const wchar_t *first, const wchar_t *last)
2428 { return append(first, last - first); }
2429 wxString& append(const wxCStrData& first, const wxCStrData& last)
2430 { return append(CreateConstIterator(first), CreateConstIterator(last)); }
2431 #endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2432
2433 // same as `this_string = str'
2434 wxString& assign(const wxString& str)
2435 {
2436 wxSTRING_SET_CACHED_LENGTH(str.length());
2437
2438 m_impl = str.m_impl;
2439
2440 return *this;
2441 }
2442
2443 wxString& assign(const wxString& str, size_t len)
2444 {
2445 wxSTRING_SET_CACHED_LENGTH(len);
2446
2447 m_impl.assign(str.m_impl, 0, str.LenToImpl(len));
2448
2449 return *this;
2450 }
2451
2452 // same as ` = str[pos..pos + n]
2453 wxString& assign(const wxString& str, size_t pos, size_t n)
2454 {
2455 size_t from, len;
2456 str.PosLenToImpl(pos, n, &from, &len);
2457 m_impl.assign(str.m_impl, from, len);
2458
2459 // it's important to call this after PosLenToImpl() above in case str is
2460 // the same string as this one
2461 wxSTRING_SET_CACHED_LENGTH(n);
2462
2463 return *this;
2464 }
2465
2466 // same as `= first n (or all if n == npos) characters of sz'
2467 wxString& assign(const char *sz)
2468 {
2469 wxSTRING_INVALIDATE_CACHE();
2470
2471 m_impl.assign(ImplStr(sz));
2472
2473 return *this;
2474 }
2475
2476 wxString& assign(const wchar_t *sz)
2477 {
2478 wxSTRING_INVALIDATE_CACHE();
2479
2480 m_impl.assign(ImplStr(sz));
2481
2482 return *this;
2483 }
2484
2485 wxString& assign(const char *sz, size_t n)
2486 {
2487 wxSTRING_SET_CACHED_LENGTH(n);
2488
2489 SubstrBufFromMB str(ImplStr(sz, n));
2490 m_impl.assign(str.data, str.len);
2491
2492 return *this;
2493 }
2494
2495 wxString& assign(const wchar_t *sz, size_t n)
2496 {
2497 wxSTRING_SET_CACHED_LENGTH(n);
2498
2499 SubstrBufFromWC str(ImplStr(sz, n));
2500 m_impl.assign(str.data, str.len);
2501
2502 return *this;
2503 }
2504
2505 wxString& assign(const wxCStrData& str)
2506 { return assign(str.AsString()); }
2507 wxString& assign(const wxCharBuffer& str)
2508 { return assign(str.data()); }
2509 wxString& assign(const wxWCharBuffer& str)
2510 { return assign(str.data()); }
2511 wxString& assign(const wxCStrData& str, size_t len)
2512 { return assign(str.AsString(), len); }
2513 wxString& assign(const wxCharBuffer& str, size_t len)
2514 { return assign(str.data(), len); }
2515 wxString& assign(const wxWCharBuffer& str, size_t len)
2516 { return assign(str.data(), len); }
2517
2518 // same as `= n copies of ch'
2519 wxString& assign(size_t n, wxUniChar ch)
2520 {
2521 wxSTRING_SET_CACHED_LENGTH(n);
2522
2523 #if wxUSE_UNICODE_UTF8
2524 if ( !ch.IsAscii() )
2525 m_impl.assign(wxStringOperations::EncodeNChars(n, ch));
2526 else
2527 #endif
2528 m_impl.assign(n, (wxStringCharType)ch);
2529
2530 return *this;
2531 }
2532
2533 wxString& assign(size_t n, wxUniCharRef ch)
2534 { return assign(n, wxUniChar(ch)); }
2535 wxString& assign(size_t n, char ch)
2536 { return assign(n, wxUniChar(ch)); }
2537 wxString& assign(size_t n, unsigned char ch)
2538 { return assign(n, wxUniChar(ch)); }
2539 wxString& assign(size_t n, wchar_t ch)
2540 { return assign(n, wxUniChar(ch)); }
2541
2542 // assign from first to last
2543 wxString& assign(const_iterator first, const_iterator last)
2544 {
2545 wxSTRING_INVALIDATE_CACHE();
2546
2547 m_impl.assign(first.impl(), last.impl());
2548
2549 return *this;
2550 }
2551 #if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2552 wxString& assign(const char *first, const char *last)
2553 { return assign(first, last - first); }
2554 wxString& assign(const wchar_t *first, const wchar_t *last)
2555 { return assign(first, last - first); }
2556 wxString& assign(const wxCStrData& first, const wxCStrData& last)
2557 { return assign(CreateConstIterator(first), CreateConstIterator(last)); }
2558 #endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2559
2560 // string comparison
2561 int compare(const wxString& str) const;
2562 int compare(const char* sz) const;
2563 int compare(const wchar_t* sz) const;
2564 int compare(const wxCStrData& str) const
2565 { return compare(str.AsString()); }
2566 int compare(const wxCharBuffer& str) const
2567 { return compare(str.data()); }
2568 int compare(const wxWCharBuffer& str) const
2569 { return compare(str.data()); }
2570 // comparison with a substring
2571 int compare(size_t nStart, size_t nLen, const wxString& str) const;
2572 // comparison of 2 substrings
2573 int compare(size_t nStart, size_t nLen,
2574 const wxString& str, size_t nStart2, size_t nLen2) const;
2575 // substring comparison with first nCount characters of sz
2576 int compare(size_t nStart, size_t nLen,
2577 const char* sz, size_t nCount = npos) const;
2578 int compare(size_t nStart, size_t nLen,
2579 const wchar_t* sz, size_t nCount = npos) const;
2580
2581 // insert another string
2582 wxString& insert(size_t nPos, const wxString& str)
2583 { insert(GetIterForNthChar(nPos), str.begin(), str.end()); return *this; }
2584 // insert n chars of str starting at nStart (in str)
2585 wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
2586 {
2587 wxSTRING_UPDATE_CACHED_LENGTH(n);
2588
2589 size_t from, len;
2590 str.PosLenToImpl(nStart, n, &from, &len);
2591 m_impl.insert(PosToImpl(nPos), str.m_impl, from, len);
2592
2593 return *this;
2594 }
2595
2596 // insert first n (or all if n == npos) characters of sz
2597 wxString& insert(size_t nPos, const char *sz)
2598 {
2599 wxSTRING_INVALIDATE_CACHE();
2600
2601 m_impl.insert(PosToImpl(nPos), ImplStr(sz));
2602
2603 return *this;
2604 }
2605
2606 wxString& insert(size_t nPos, const wchar_t *sz)
2607 {
2608 wxSTRING_INVALIDATE_CACHE();
2609
2610 m_impl.insert(PosToImpl(nPos), ImplStr(sz)); return *this;
2611 }
2612
2613 wxString& insert(size_t nPos, const char *sz, size_t n)
2614 {
2615 wxSTRING_UPDATE_CACHED_LENGTH(n);
2616
2617 SubstrBufFromMB str(ImplStr(sz, n));
2618 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2619
2620 return *this;
2621 }
2622
2623 wxString& insert(size_t nPos, const wchar_t *sz, size_t n)
2624 {
2625 wxSTRING_UPDATE_CACHED_LENGTH(n);
2626
2627 SubstrBufFromWC str(ImplStr(sz, n));
2628 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2629
2630 return *this;
2631 }
2632
2633 // insert n copies of ch
2634 wxString& insert(size_t nPos, size_t n, wxUniChar ch)
2635 {
2636 wxSTRING_UPDATE_CACHED_LENGTH(n);
2637
2638 #if wxUSE_UNICODE_UTF8
2639 if ( !ch.IsAscii() )
2640 m_impl.insert(PosToImpl(nPos), wxStringOperations::EncodeNChars(n, ch));
2641 else
2642 #endif
2643 m_impl.insert(PosToImpl(nPos), n, (wxStringCharType)ch);
2644 return *this;
2645 }
2646
2647 iterator insert(iterator it, wxUniChar ch)
2648 {
2649 wxSTRING_UPDATE_CACHED_LENGTH(1);
2650
2651 #if wxUSE_UNICODE_UTF8
2652 if ( !ch.IsAscii() )
2653 {
2654 size_t pos = IterToImplPos(it);
2655 m_impl.insert(pos, wxStringOperations::EncodeChar(ch));
2656 return iterator(this, m_impl.begin() + pos);
2657 }
2658 else
2659 #endif
2660 return iterator(this, m_impl.insert(it.impl(), (wxStringCharType)ch));
2661 }
2662
2663 void insert(iterator it, const_iterator first, const_iterator last)
2664 {
2665 wxSTRING_INVALIDATE_CACHE();
2666
2667 m_impl.insert(it.impl(), first.impl(), last.impl());
2668 }
2669
2670 #if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2671 void insert(iterator it, const char *first, const char *last)
2672 { insert(it - begin(), first, last - first); }
2673 void insert(iterator it, const wchar_t *first, const wchar_t *last)
2674 { insert(it - begin(), first, last - first); }
2675 void insert(iterator it, const wxCStrData& first, const wxCStrData& last)
2676 { insert(it, CreateConstIterator(first), CreateConstIterator(last)); }
2677 #endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2678
2679 void insert(iterator it, size_type n, wxUniChar ch)
2680 {
2681 wxSTRING_UPDATE_CACHED_LENGTH(n);
2682
2683 #if wxUSE_UNICODE_UTF8
2684 if ( !ch.IsAscii() )
2685 m_impl.insert(IterToImplPos(it), wxStringOperations::EncodeNChars(n, ch));
2686 else
2687 #endif
2688 m_impl.insert(it.impl(), n, (wxStringCharType)ch);
2689 }
2690
2691 // delete characters from nStart to nStart + nLen
2692 wxString& erase(size_type pos = 0, size_type n = npos)
2693 {
2694 wxSTRING_INVALIDATE_CACHE();
2695
2696 size_t from, len;
2697 PosLenToImpl(pos, n, &from, &len);
2698 m_impl.erase(from, len);
2699
2700 return *this;
2701 }
2702
2703 // delete characters from first up to last
2704 iterator erase(iterator first, iterator last)
2705 {
2706 wxSTRING_INVALIDATE_CACHE();
2707
2708 return iterator(this, m_impl.erase(first.impl(), last.impl()));
2709 }
2710
2711 iterator erase(iterator first)
2712 {
2713 wxSTRING_UPDATE_CACHED_LENGTH(-1);
2714
2715 return iterator(this, m_impl.erase(first.impl()));
2716 }
2717
2718 #ifdef wxSTRING_BASE_HASNT_CLEAR
2719 void clear() { erase(); }
2720 #else
2721 void clear()
2722 {
2723 wxSTRING_SET_CACHED_LENGTH(0);
2724
2725 m_impl.clear();
2726 }
2727 #endif
2728
2729 // replaces the substring of length nLen starting at nStart
2730 wxString& replace(size_t nStart, size_t nLen, const char* sz)
2731 {
2732 wxSTRING_INVALIDATE_CACHE();
2733
2734 size_t from, len;
2735 PosLenToImpl(nStart, nLen, &from, &len);
2736 m_impl.replace(from, len, ImplStr(sz));
2737
2738 return *this;
2739 }
2740
2741 wxString& replace(size_t nStart, size_t nLen, const wchar_t* sz)
2742 {
2743 wxSTRING_INVALIDATE_CACHE();
2744
2745 size_t from, len;
2746 PosLenToImpl(nStart, nLen, &from, &len);
2747 m_impl.replace(from, len, ImplStr(sz));
2748
2749 return *this;
2750 }
2751
2752 // replaces the substring of length nLen starting at nStart
2753 wxString& replace(size_t nStart, size_t nLen, const wxString& str)
2754 {
2755 wxSTRING_INVALIDATE_CACHE();
2756
2757 size_t from, len;
2758 PosLenToImpl(nStart, nLen, &from, &len);
2759 m_impl.replace(from, len, str.m_impl);
2760
2761 return *this;
2762 }
2763
2764 // replaces the substring with nCount copies of ch
2765 wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxUniChar ch)
2766 {
2767 wxSTRING_INVALIDATE_CACHE();
2768
2769 size_t from, len;
2770 PosLenToImpl(nStart, nLen, &from, &len);
2771 #if wxUSE_UNICODE_UTF8
2772 if ( !ch.IsAscii() )
2773 m_impl.replace(from, len, wxStringOperations::EncodeNChars(nCount, ch));
2774 else
2775 #endif
2776 m_impl.replace(from, len, nCount, (wxStringCharType)ch);
2777
2778 return *this;
2779 }
2780
2781 // replaces a substring with another substring
2782 wxString& replace(size_t nStart, size_t nLen,
2783 const wxString& str, size_t nStart2, size_t nLen2)
2784 {
2785 wxSTRING_INVALIDATE_CACHE();
2786
2787 size_t from, len;
2788 PosLenToImpl(nStart, nLen, &from, &len);
2789
2790 size_t from2, len2;
2791 str.PosLenToImpl(nStart2, nLen2, &from2, &len2);
2792
2793 m_impl.replace(from, len, str.m_impl, from2, len2);
2794
2795 return *this;
2796 }
2797
2798 // replaces the substring with first nCount chars of sz
2799 wxString& replace(size_t nStart, size_t nLen,
2800 const char* sz, size_t nCount)
2801 {
2802 wxSTRING_INVALIDATE_CACHE();
2803
2804 size_t from, len;
2805 PosLenToImpl(nStart, nLen, &from, &len);
2806
2807 SubstrBufFromMB str(ImplStr(sz, nCount));
2808
2809 m_impl.replace(from, len, str.data, str.len);
2810
2811 return *this;
2812 }
2813
2814 wxString& replace(size_t nStart, size_t nLen,
2815 const wchar_t* sz, size_t nCount)
2816 {
2817 wxSTRING_INVALIDATE_CACHE();
2818
2819 size_t from, len;
2820 PosLenToImpl(nStart, nLen, &from, &len);
2821
2822 SubstrBufFromWC str(ImplStr(sz, nCount));
2823
2824 m_impl.replace(from, len, str.data, str.len);
2825
2826 return *this;
2827 }
2828
2829 wxString& replace(size_t nStart, size_t nLen,
2830 const wxString& s, size_t nCount)
2831 {
2832 wxSTRING_INVALIDATE_CACHE();
2833
2834 size_t from, len;
2835 PosLenToImpl(nStart, nLen, &from, &len);
2836 m_impl.replace(from, len, s.m_impl.c_str(), s.LenToImpl(nCount));
2837
2838 return *this;
2839 }
2840
2841 wxString& replace(iterator first, iterator last, const char* s)
2842 {
2843 wxSTRING_INVALIDATE_CACHE();
2844
2845 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2846
2847 return *this;
2848 }
2849
2850 wxString& replace(iterator first, iterator last, const wchar_t* s)
2851 {
2852 wxSTRING_INVALIDATE_CACHE();
2853
2854 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2855
2856 return *this;
2857 }
2858
2859 wxString& replace(iterator first, iterator last, const char* s, size_type n)
2860 {
2861 wxSTRING_INVALIDATE_CACHE();
2862
2863 SubstrBufFromMB str(ImplStr(s, n));
2864 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2865
2866 return *this;
2867 }
2868
2869 wxString& replace(iterator first, iterator last, const wchar_t* s, size_type n)
2870 {
2871 wxSTRING_INVALIDATE_CACHE();
2872
2873 SubstrBufFromWC str(ImplStr(s, n));
2874 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2875
2876 return *this;
2877 }
2878
2879 wxString& replace(iterator first, iterator last, const wxString& s)
2880 {
2881 wxSTRING_INVALIDATE_CACHE();
2882
2883 m_impl.replace(first.impl(), last.impl(), s.m_impl);
2884
2885 return *this;
2886 }
2887
2888 wxString& replace(iterator first, iterator last, size_type n, wxUniChar ch)
2889 {
2890 wxSTRING_INVALIDATE_CACHE();
2891
2892 #if wxUSE_UNICODE_UTF8
2893 if ( !ch.IsAscii() )
2894 m_impl.replace(first.impl(), last.impl(),
2895 wxStringOperations::EncodeNChars(n, ch));
2896 else
2897 #endif
2898 m_impl.replace(first.impl(), last.impl(), n, (wxStringCharType)ch);
2899
2900 return *this;
2901 }
2902
2903 wxString& replace(iterator first, iterator last,
2904 const_iterator first1, const_iterator last1)
2905 {
2906 wxSTRING_INVALIDATE_CACHE();
2907
2908 m_impl.replace(first.impl(), last.impl(), first1.impl(), last1.impl());
2909
2910 return *this;
2911 }
2912
2913 wxString& replace(iterator first, iterator last,
2914 const char *first1, const char *last1)
2915 { replace(first, last, first1, last1 - first1); return *this; }
2916 wxString& replace(iterator first, iterator last,
2917 const wchar_t *first1, const wchar_t *last1)
2918 { replace(first, last, first1, last1 - first1); return *this; }
2919
2920 // swap two strings
2921 void swap(wxString& str)
2922 {
2923 #if wxUSE_STRING_POS_CACHE
2924 // we modify not only this string but also the other one directly so we
2925 // need to invalidate cache for both of them (we could also try to
2926 // exchange their cache entries but it seems unlikely to be worth it)
2927 InvalidateCache();
2928 str.InvalidateCache();
2929 #endif // wxUSE_STRING_POS_CACHE
2930
2931 m_impl.swap(str.m_impl);
2932 }
2933
2934 // find a substring
2935 size_t find(const wxString& str, size_t nStart = 0) const
2936 { return PosFromImpl(m_impl.find(str.m_impl, PosToImpl(nStart))); }
2937
2938 // find first n characters of sz
2939 size_t find(const char* sz, size_t nStart = 0, size_t n = npos) const
2940 {
2941 SubstrBufFromMB str(ImplStr(sz, n));
2942 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
2943 }
2944 size_t find(const wchar_t* sz, size_t nStart = 0, size_t n = npos) const
2945 {
2946 SubstrBufFromWC str(ImplStr(sz, n));
2947 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
2948 }
2949 size_t find(const wxCharBuffer& s, size_t nStart = 0, size_t n = npos) const
2950 { return find(s.data(), nStart, n); }
2951 size_t find(const wxWCharBuffer& s, size_t nStart = 0, size_t n = npos) const
2952 { return find(s.data(), nStart, n); }
2953 size_t find(const wxCStrData& s, size_t nStart = 0, size_t n = npos) const
2954 { return find(s.AsWChar(), nStart, n); }
2955
2956 // find the first occurence of character ch after nStart
2957 size_t find(wxUniChar ch, size_t nStart = 0) const
2958 {
2959 #if wxUSE_UNICODE_UTF8
2960 if ( !ch.IsAscii() )
2961 return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch),
2962 PosToImpl(nStart)));
2963 else
2964 #endif
2965 return PosFromImpl(m_impl.find((wxStringCharType)ch,
2966 PosToImpl(nStart)));
2967
2968 }
2969 size_t find(wxUniCharRef ch, size_t nStart = 0) const
2970 { return find(wxUniChar(ch), nStart); }
2971 size_t find(char ch, size_t nStart = 0) const
2972 { return find(wxUniChar(ch), nStart); }
2973 size_t find(unsigned char ch, size_t nStart = 0) const
2974 { return find(wxUniChar(ch), nStart); }
2975 size_t find(wchar_t ch, size_t nStart = 0) const
2976 { return find(wxUniChar(ch), nStart); }
2977
2978 // rfind() family is exactly like find() but works right to left
2979
2980 // as find, but from the end
2981 size_t rfind(const wxString& str, size_t nStart = npos) const
2982 { return PosFromImpl(m_impl.rfind(str.m_impl, PosToImpl(nStart))); }
2983
2984 // as find, but from the end
2985 size_t rfind(const char* sz, size_t nStart = npos, size_t n = npos) const
2986 {
2987 SubstrBufFromMB str(ImplStr(sz, n));
2988 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
2989 }
2990 size_t rfind(const wchar_t* sz, size_t nStart = npos, size_t n = npos) const
2991 {
2992 SubstrBufFromWC str(ImplStr(sz, n));
2993 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
2994 }
2995 size_t rfind(const wxCharBuffer& s, size_t nStart = npos, size_t n = npos) const
2996 { return rfind(s.data(), nStart, n); }
2997 size_t rfind(const wxWCharBuffer& s, size_t nStart = npos, size_t n = npos) const
2998 { return rfind(s.data(), nStart, n); }
2999 size_t rfind(const wxCStrData& s, size_t nStart = npos, size_t n = npos) const
3000 { return rfind(s.AsWChar(), nStart, n); }
3001 // as find, but from the end
3002 size_t rfind(wxUniChar ch, size_t nStart = npos) const
3003 {
3004 #if wxUSE_UNICODE_UTF8
3005 if ( !ch.IsAscii() )
3006 return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch),
3007 PosToImpl(nStart)));
3008 else
3009 #endif
3010 return PosFromImpl(m_impl.rfind((wxStringCharType)ch,
3011 PosToImpl(nStart)));
3012 }
3013 size_t rfind(wxUniCharRef ch, size_t nStart = npos) const
3014 { return rfind(wxUniChar(ch), nStart); }
3015 size_t rfind(char ch, size_t nStart = npos) const
3016 { return rfind(wxUniChar(ch), nStart); }
3017 size_t rfind(unsigned char ch, size_t nStart = npos) const
3018 { return rfind(wxUniChar(ch), nStart); }
3019 size_t rfind(wchar_t ch, size_t nStart = npos) const
3020 { return rfind(wxUniChar(ch), nStart); }
3021
3022 // find first/last occurence of any character (not) in the set:
3023 #if wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3024 // FIXME-UTF8: this is not entirely correct, because it doesn't work if
3025 // sizeof(wchar_t)==2 and surrogates are present in the string;
3026 // should we care? Probably not.
3027 size_t find_first_of(const wxString& str, size_t nStart = 0) const
3028 { return m_impl.find_first_of(str.m_impl, nStart); }
3029 size_t find_first_of(const char* sz, size_t nStart = 0) const
3030 { return m_impl.find_first_of(ImplStr(sz), nStart); }
3031 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const
3032 { return m_impl.find_first_of(ImplStr(sz), nStart); }
3033 size_t find_first_of(const char* sz, size_t nStart, size_t n) const
3034 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
3035 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const
3036 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
3037 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3038 { return m_impl.find_first_of((wxChar)c, nStart); }
3039
3040 size_t find_last_of(const wxString& str, size_t nStart = npos) const
3041 { return m_impl.find_last_of(str.m_impl, nStart); }
3042 size_t find_last_of(const char* sz, size_t nStart = npos) const
3043 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3044 size_t find_last_of(const wchar_t* sz, size_t nStart = npos) const
3045 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3046 size_t find_last_of(const char* sz, size_t nStart, size_t n) const
3047 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3048 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const
3049 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3050 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3051 { return m_impl.find_last_of((wxChar)c, nStart); }
3052
3053 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
3054 { return m_impl.find_first_not_of(str.m_impl, nStart); }
3055 size_t find_first_not_of(const char* sz, size_t nStart = 0) const
3056 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
3057 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const
3058 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
3059 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const
3060 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
3061 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const
3062 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
3063 size_t find_first_not_of(wxUniChar c, size_t nStart = 0) const
3064 { return m_impl.find_first_not_of((wxChar)c, nStart); }
3065
3066 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
3067 { return m_impl.find_last_not_of(str.m_impl, nStart); }
3068 size_t find_last_not_of(const char* sz, size_t nStart = npos) const
3069 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
3070 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const
3071 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
3072 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const
3073 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
3074 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const
3075 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
3076 size_t find_last_not_of(wxUniChar c, size_t nStart = npos) const
3077 { return m_impl.find_last_not_of((wxChar)c, nStart); }
3078 #else
3079 // we can't use std::string implementation in UTF-8 build, because the
3080 // character sets would be interpreted wrongly:
3081
3082 // as strpbrk() but starts at nStart, returns npos if not found
3083 size_t find_first_of(const wxString& str, size_t nStart = 0) const
3084 #if wxUSE_UNICODE // FIXME-UTF8: temporary
3085 { return find_first_of(str.wc_str(), nStart); }
3086 #else
3087 { return find_first_of(str.mb_str(), nStart); }
3088 #endif
3089 // same as above
3090 size_t find_first_of(const char* sz, size_t nStart = 0) const;
3091 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const;
3092 size_t find_first_of(const char* sz, size_t nStart, size_t n) const;
3093 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const;
3094 // same as find(char, size_t)
3095 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3096 { return find(c, nStart); }
3097 // find the last (starting from nStart) char from str in this string
3098 size_t find_last_of (const wxString& str, size_t nStart = npos) const
3099 #if wxUSE_UNICODE // FIXME-UTF8: temporary
3100 { return find_last_of(str.wc_str(), nStart); }
3101 #else
3102 { return find_last_of(str.mb_str(), nStart); }
3103 #endif
3104 // same as above
3105 size_t find_last_of (const char* sz, size_t nStart = npos) const;
3106 size_t find_last_of (const wchar_t* sz, size_t nStart = npos) const;
3107 size_t find_last_of(const char* sz, size_t nStart, size_t n) const;
3108 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const;
3109 // same as above
3110 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3111 { return rfind(c, nStart); }
3112
3113 // find first/last occurence of any character not in the set
3114
3115 // as strspn() (starting from nStart), returns npos on failure
3116 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
3117 #if wxUSE_UNICODE // FIXME-UTF8: temporary
3118 { return find_first_not_of(str.wc_str(), nStart); }
3119 #else
3120 { return find_first_not_of(str.mb_str(), nStart); }
3121 #endif
3122 // same as above
3123 size_t find_first_not_of(const char* sz, size_t nStart = 0) const;
3124 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const;
3125 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const;
3126 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3127 // same as above
3128 size_t find_first_not_of(wxUniChar ch, size_t nStart = 0) const;
3129 // as strcspn()
3130 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
3131 #if wxUSE_UNICODE // FIXME-UTF8: temporary
3132 { return find_last_not_of(str.wc_str(), nStart); }
3133 #else
3134 { return find_last_not_of(str.mb_str(), nStart); }
3135 #endif
3136 // same as above
3137 size_t find_last_not_of(const char* sz, size_t nStart = npos) const;
3138 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const;
3139 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const;
3140 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3141 // same as above
3142 size_t find_last_not_of(wxUniChar ch, size_t nStart = npos) const;
3143 #endif // wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 or not
3144
3145 // provide char/wchar_t/wxUniCharRef overloads for char-finding functions
3146 // above to resolve ambiguities:
3147 size_t find_first_of(wxUniCharRef ch, size_t nStart = 0) const
3148 { return find_first_of(wxUniChar(ch), nStart); }
3149 size_t find_first_of(char ch, size_t nStart = 0) const
3150 { return find_first_of(wxUniChar(ch), nStart); }
3151 size_t find_first_of(unsigned char ch, size_t nStart = 0) const
3152 { return find_first_of(wxUniChar(ch), nStart); }
3153 size_t find_first_of(wchar_t ch, size_t nStart = 0) const
3154 { return find_first_of(wxUniChar(ch), nStart); }
3155 size_t find_last_of(wxUniCharRef ch, size_t nStart = npos) const
3156 { return find_last_of(wxUniChar(ch), nStart); }
3157 size_t find_last_of(char ch, size_t nStart = npos) const
3158 { return find_last_of(wxUniChar(ch), nStart); }
3159 size_t find_last_of(unsigned char ch, size_t nStart = npos) const
3160 { return find_last_of(wxUniChar(ch), nStart); }
3161 size_t find_last_of(wchar_t ch, size_t nStart = npos) const
3162 { return find_last_of(wxUniChar(ch), nStart); }
3163 size_t find_first_not_of(wxUniCharRef ch, size_t nStart = 0) const
3164 { return find_first_not_of(wxUniChar(ch), nStart); }
3165 size_t find_first_not_of(char ch, size_t nStart = 0) const
3166 { return find_first_not_of(wxUniChar(ch), nStart); }
3167 size_t find_first_not_of(unsigned char ch, size_t nStart = 0) const
3168 { return find_first_not_of(wxUniChar(ch), nStart); }
3169 size_t find_first_not_of(wchar_t ch, size_t nStart = 0) const
3170 { return find_first_not_of(wxUniChar(ch), nStart); }
3171 size_t find_last_not_of(wxUniCharRef ch, size_t nStart = npos) const
3172 { return find_last_not_of(wxUniChar(ch), nStart); }
3173 size_t find_last_not_of(char ch, size_t nStart = npos) const
3174 { return find_last_not_of(wxUniChar(ch), nStart); }
3175 size_t find_last_not_of(unsigned char ch, size_t nStart = npos) const
3176 { return find_last_not_of(wxUniChar(ch), nStart); }
3177 size_t find_last_not_of(wchar_t ch, size_t nStart = npos) const
3178 { return find_last_not_of(wxUniChar(ch), nStart); }
3179
3180 // and additional overloads for the versions taking strings:
3181 size_t find_first_of(const wxCStrData& sz, size_t nStart = 0) const
3182 { return find_first_of(sz.AsString(), nStart); }
3183 size_t find_first_of(const wxCharBuffer& sz, size_t nStart = 0) const
3184 { return find_first_of(sz.data(), nStart); }
3185 size_t find_first_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3186 { return find_first_of(sz.data(), nStart); }
3187 size_t find_first_of(const wxCStrData& sz, size_t nStart, size_t n) const
3188 { return find_first_of(sz.AsWChar(), nStart, n); }
3189 size_t find_first_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3190 { return find_first_of(sz.data(), nStart, n); }
3191 size_t find_first_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3192 { return find_first_of(sz.data(), nStart, n); }
3193
3194 size_t find_last_of(const wxCStrData& sz, size_t nStart = 0) const
3195 { return find_last_of(sz.AsString(), nStart); }
3196 size_t find_last_of(const wxCharBuffer& sz, size_t nStart = 0) const
3197 { return find_last_of(sz.data(), nStart); }
3198 size_t find_last_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3199 { return find_last_of(sz.data(), nStart); }
3200 size_t find_last_of(const wxCStrData& sz, size_t nStart, size_t n) const
3201 { return find_last_of(sz.AsWChar(), nStart, n); }
3202 size_t find_last_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3203 { return find_last_of(sz.data(), nStart, n); }
3204 size_t find_last_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3205 { return find_last_of(sz.data(), nStart, n); }
3206
3207 size_t find_first_not_of(const wxCStrData& sz, size_t nStart = 0) const
3208 { return find_first_not_of(sz.AsString(), nStart); }
3209 size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart = 0) const
3210 { return find_first_not_of(sz.data(), nStart); }
3211 size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3212 { return find_first_not_of(sz.data(), nStart); }
3213 size_t find_first_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3214 { return find_first_not_of(sz.AsWChar(), nStart, n); }
3215 size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3216 { return find_first_not_of(sz.data(), nStart, n); }
3217 size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3218 { return find_first_not_of(sz.data(), nStart, n); }
3219
3220 size_t find_last_not_of(const wxCStrData& sz, size_t nStart = 0) const
3221 { return find_last_not_of(sz.AsString(), nStart); }
3222 size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart = 0) const
3223 { return find_last_not_of(sz.data(), nStart); }
3224 size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3225 { return find_last_not_of(sz.data(), nStart); }
3226 size_t find_last_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3227 { return find_last_not_of(sz.AsWChar(), nStart, n); }
3228 size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3229 { return find_last_not_of(sz.data(), nStart, n); }
3230 size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3231 { return find_last_not_of(sz.data(), nStart, n); }
3232
3233 // string += string
3234 wxString& operator+=(const wxString& s)
3235 {
3236 wxSTRING_INVALIDATE_CACHED_LENGTH();
3237
3238 m_impl += s.m_impl;
3239 return *this;
3240 }
3241 // string += C string
3242 wxString& operator+=(const char *psz)
3243 {
3244 wxSTRING_INVALIDATE_CACHED_LENGTH();
3245
3246 m_impl += ImplStr(psz);
3247 return *this;
3248 }
3249 wxString& operator+=(const wchar_t *pwz)
3250 {
3251 wxSTRING_INVALIDATE_CACHED_LENGTH();
3252
3253 m_impl += ImplStr(pwz);
3254 return *this;
3255 }
3256 wxString& operator+=(const wxCStrData& s)
3257 {
3258 wxSTRING_INVALIDATE_CACHED_LENGTH();
3259
3260 m_impl += s.AsString().m_impl;
3261 return *this;
3262 }
3263 wxString& operator+=(const wxCharBuffer& s)
3264 { return operator+=(s.data()); }
3265 wxString& operator+=(const wxWCharBuffer& s)
3266 { return operator+=(s.data()); }
3267 // string += char
3268 wxString& operator+=(wxUniChar ch)
3269 {
3270 wxSTRING_UPDATE_CACHED_LENGTH(1);
3271
3272 #if wxUSE_UNICODE_UTF8
3273 if ( !ch.IsAscii() )
3274 m_impl += wxStringOperations::EncodeChar(ch);
3275 else
3276 #endif
3277 m_impl += (wxStringCharType)ch;
3278 return *this;
3279 }
3280 wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); }
3281 wxString& operator+=(int ch) { return *this += wxUniChar(ch); }
3282 wxString& operator+=(char ch) { return *this += wxUniChar(ch); }
3283 wxString& operator+=(unsigned char ch) { return *this += wxUniChar(ch); }
3284 wxString& operator+=(wchar_t ch) { return *this += wxUniChar(ch); }
3285
3286 private:
3287 #if !wxUSE_STL_BASED_WXSTRING
3288 // helpers for wxStringBuffer and wxStringBufferLength
3289 wxStringCharType *DoGetWriteBuf(size_t nLen)
3290 {
3291 return m_impl.DoGetWriteBuf(nLen);
3292 }
3293
3294 void DoUngetWriteBuf()
3295 {
3296 wxSTRING_INVALIDATE_CACHE();
3297
3298 m_impl.DoUngetWriteBuf();
3299 }
3300
3301 void DoUngetWriteBuf(size_t nLen)
3302 {
3303 wxSTRING_SET_CACHED_LENGTH(nLen);
3304
3305 m_impl.DoUngetWriteBuf(nLen);
3306 }
3307 #endif // !wxUSE_STL_BASED_WXSTRING
3308
3309 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
3310 #if !wxUSE_UTF8_LOCALE_ONLY
3311 int DoPrintfWchar(const wxChar *format, ...);
3312 static wxString DoFormatWchar(const wxChar *format, ...);
3313 #endif
3314 #if wxUSE_UNICODE_UTF8
3315 int DoPrintfUtf8(const char *format, ...);
3316 static wxString DoFormatUtf8(const char *format, ...);
3317 #endif
3318 #endif
3319
3320 #if !wxUSE_STL_BASED_WXSTRING
3321 // check string's data validity
3322 bool IsValid() const { return m_impl.GetStringData()->IsValid(); }
3323 #endif
3324
3325 private:
3326 wxStringImpl m_impl;
3327
3328 // buffers for compatibility conversion from (char*)c_str() and
3329 // (wchar_t*)c_str():
3330 // FIXME-UTF8: bechmark various approaches to keeping compatibility buffers
3331 template<typename T>
3332 struct ConvertedBuffer
3333 {
3334 ConvertedBuffer() : m_buf(NULL) {}
3335 ~ConvertedBuffer()
3336 { free(m_buf); }
3337
3338 operator T*() const { return m_buf; }
3339
3340 ConvertedBuffer& operator=(T *str)
3341 {
3342 free(m_buf);
3343 m_buf = str;
3344 return *this;
3345 }
3346
3347 T *m_buf;
3348 };
3349 #if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY
3350 ConvertedBuffer<char> m_convertedToChar;
3351 #endif
3352 #if !wxUSE_UNICODE_WCHAR
3353 ConvertedBuffer<wchar_t> m_convertedToWChar;
3354 #endif
3355
3356 #if wxUSE_UNICODE_UTF8
3357 // FIXME-UTF8: (try to) move this elsewhere (TLS) or solve differently
3358 // assigning to character pointer to by wxString::interator may
3359 // change the underlying wxStringImpl iterator, so we have to
3360 // keep track of all iterators and update them as necessary:
3361 struct wxStringIteratorNodeHead
3362 {
3363 wxStringIteratorNodeHead() : ptr(NULL) {}
3364 wxStringIteratorNode *ptr;
3365
3366 // copying is disallowed as it would result in more than one pointer into
3367 // the same linked list
3368 DECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead)
3369 };
3370
3371 wxStringIteratorNodeHead m_iterators;
3372
3373 friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode;
3374 friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef;
3375 #endif // wxUSE_UNICODE_UTF8
3376
3377 friend class WXDLLIMPEXP_FWD_BASE wxCStrData;
3378 friend class wxStringInternalBuffer;
3379 friend class wxStringInternalBufferLength;
3380 };
3381
3382 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
3383 #pragma warning (default:4275)
3384 #endif
3385
3386 // string iterator operators that satisfy STL Random Access Iterator
3387 // requirements:
3388 inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i)
3389 { return i + n; }
3390 inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i)
3391 { return i + n; }
3392 inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i)
3393 { return i + n; }
3394 inline wxString::const_reverse_iterator operator+(ptrdiff_t n, wxString::const_reverse_iterator i)
3395 { return i + n; }
3396
3397 // notice that even though for many compilers the friend declarations above are
3398 // enough, from the point of view of C++ standard we must have the declarations
3399 // here as friend ones are not injected in the enclosing namespace and without
3400 // them the code fails to compile with conforming compilers such as xlC or g++4
3401 wxString WXDLLIMPEXP_BASE operator+(const wxString& string1, const wxString& string2);
3402 wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const char *psz);
3403 wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const wchar_t *pwz);
3404 wxString WXDLLIMPEXP_BASE operator+(const char *psz, const wxString& string);
3405 wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz, const wxString& string);
3406
3407 wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
3408 wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
3409
3410 inline wxString operator+(const wxString& string, wxUniCharRef ch)
3411 { return string + (wxUniChar)ch; }
3412 inline wxString operator+(const wxString& string, char ch)
3413 { return string + wxUniChar(ch); }
3414 inline wxString operator+(const wxString& string, wchar_t ch)
3415 { return string + wxUniChar(ch); }
3416 inline wxString operator+(wxUniCharRef ch, const wxString& string)
3417 { return (wxUniChar)ch + string; }
3418 inline wxString operator+(char ch, const wxString& string)
3419 { return wxUniChar(ch) + string; }
3420 inline wxString operator+(wchar_t ch, const wxString& string)
3421 { return wxUniChar(ch) + string; }
3422
3423
3424 #define wxGetEmptyString() wxString()
3425
3426 // ----------------------------------------------------------------------------
3427 // helper functions which couldn't be defined inline
3428 // ----------------------------------------------------------------------------
3429
3430 namespace wxPrivate
3431 {
3432
3433 #if wxUSE_UNICODE_WCHAR
3434
3435 template <>
3436 struct wxStringAsBufHelper<char>
3437 {
3438 static wxCharBuffer Get(const wxString& s, size_t *len)
3439 {
3440 wxCharBuffer buf(s.mb_str());
3441 if ( len )
3442 *len = buf ? strlen(buf) : 0;
3443 return buf;
3444 }
3445 };
3446
3447 template <>
3448 struct wxStringAsBufHelper<wchar_t>
3449 {
3450 static wxWCharBuffer Get(const wxString& s, size_t *len)
3451 {
3452 if ( len )
3453 *len = s.length();
3454 return wxWCharBuffer::CreateNonOwned(s.wx_str());
3455 }
3456 };
3457
3458 #elif wxUSE_UNICODE_UTF8
3459
3460 template <>
3461 struct wxStringAsBufHelper<char>
3462 {
3463 static wxCharBuffer Get(const wxString& s, size_t *len)
3464 {
3465 if ( len )
3466 *len = s.utf8_length();
3467 return wxCharBuffer::CreateNonOwned(s.wx_str());
3468 }
3469 };
3470
3471 template <>
3472 struct wxStringAsBufHelper<wchar_t>
3473 {
3474 static wxWCharBuffer Get(const wxString& s, size_t *len)
3475 {
3476 wxWCharBuffer wbuf(s.wc_str());
3477 if ( len )
3478 *len = wxWcslen(wbuf);
3479 return wbuf;
3480 }
3481 };
3482
3483 #endif // Unicode build kind
3484
3485 } // namespace wxPrivate
3486
3487 // ----------------------------------------------------------------------------
3488 // wxStringBuffer: a tiny class allowing to get a writable pointer into string
3489 // ----------------------------------------------------------------------------
3490
3491 #if !wxUSE_STL_BASED_WXSTRING
3492 // string buffer for direct access to string data in their native
3493 // representation:
3494 class wxStringInternalBuffer
3495 {
3496 public:
3497 typedef wxStringCharType CharType;
3498
3499 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
3500 : m_str(str), m_buf(NULL)
3501 { m_buf = m_str.DoGetWriteBuf(lenWanted); }
3502
3503 ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); }
3504
3505 operator wxStringCharType*() const { return m_buf; }
3506
3507 private:
3508 wxString& m_str;
3509 wxStringCharType *m_buf;
3510
3511 DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)
3512 };
3513
3514 class wxStringInternalBufferLength
3515 {
3516 public:
3517 typedef wxStringCharType CharType;
3518
3519 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
3520 : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false)
3521 {
3522 m_buf = m_str.DoGetWriteBuf(lenWanted);
3523 wxASSERT(m_buf != NULL);
3524 }
3525
3526 ~wxStringInternalBufferLength()
3527 {
3528 wxASSERT(m_lenSet);
3529 m_str.DoUngetWriteBuf(m_len);
3530 }
3531
3532 operator wxStringCharType*() const { return m_buf; }
3533 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3534
3535 private:
3536 wxString& m_str;
3537 wxStringCharType *m_buf;
3538 size_t m_len;
3539 bool m_lenSet;
3540
3541 DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)
3542 };
3543
3544 #endif // !wxUSE_STL_BASED_WXSTRING
3545
3546 template<typename T>
3547 class WXDLLIMPEXP_BASE wxStringTypeBufferBase
3548 {
3549 public:
3550 typedef T CharType;
3551
3552 wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024)
3553 : m_str(str), m_buf(lenWanted)
3554 {
3555 // for compatibility with old wxStringBuffer which provided direct
3556 // access to wxString internal buffer, initialize ourselves with the
3557 // string initial contents
3558
3559 // FIXME-VC6: remove the ugly (CharType *)NULL and use normal
3560 // tchar_str<CharType>
3561 size_t len;
3562 const wxCharTypeBuffer<CharType> buf(str.tchar_str(&len, (CharType *)NULL));
3563 if ( buf )
3564 {
3565 if ( len > lenWanted )
3566 {
3567 // in this case there is not enough space for terminating NUL,
3568 // ensure that we still put it there
3569 m_buf.data()[lenWanted] = 0;
3570 len = lenWanted - 1;
3571 }
3572
3573 memcpy(m_buf.data(), buf, (len + 1)*sizeof(CharType));
3574 }
3575 //else: conversion failed, this can happen when trying to get Unicode
3576 // string contents into a char string
3577 }
3578
3579 operator CharType*() { return m_buf.data(); }
3580
3581 protected:
3582 wxString& m_str;
3583 wxCharTypeBuffer<CharType> m_buf;
3584 };
3585
3586 template<typename T>
3587 class WXDLLIMPEXP_BASE wxStringTypeBufferLengthBase
3588 : public wxStringTypeBufferBase<T>
3589 {
3590 public:
3591 wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024)
3592 : wxStringTypeBufferBase<T>(str, lenWanted),
3593 m_len(0),
3594 m_lenSet(false)
3595 { }
3596
3597 ~wxStringTypeBufferLengthBase()
3598 {
3599 wxASSERT_MSG( this->m_lenSet, "forgot to call SetLength()" );
3600 }
3601
3602 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3603
3604 protected:
3605 size_t m_len;
3606 bool m_lenSet;
3607 };
3608
3609 template<typename T>
3610 class wxStringTypeBuffer : public wxStringTypeBufferBase<T>
3611 {
3612 public:
3613 wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024)
3614 : wxStringTypeBufferBase<T>(str, lenWanted)
3615 { }
3616
3617 ~wxStringTypeBuffer()
3618 {
3619 this->m_str.assign(this->m_buf.data());
3620 }
3621
3622 DECLARE_NO_COPY_CLASS(wxStringTypeBuffer)
3623 };
3624
3625 template<typename T>
3626 class wxStringTypeBufferLength : public wxStringTypeBufferLengthBase<T>
3627 {
3628 public:
3629 wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024)
3630 : wxStringTypeBufferLengthBase<T>(str, lenWanted)
3631 { }
3632
3633 ~wxStringTypeBufferLength()
3634 {
3635 this->m_str.assign(this->m_buf.data(), this->m_len);
3636 }
3637
3638 DECLARE_NO_COPY_CLASS(wxStringTypeBufferLength)
3639 };
3640
3641 #if wxUSE_STL_BASED_WXSTRING
3642
3643 WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<wxStringCharType> )
3644
3645 class wxStringInternalBuffer : public wxStringTypeBufferBase<wxStringCharType>
3646 {
3647 public:
3648 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
3649 : wxStringTypeBufferBase<wxStringCharType>(str, lenWanted) {}
3650 ~wxStringInternalBuffer()
3651 { m_str.m_impl.assign(m_buf.data()); }
3652
3653 DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)
3654 };
3655
3656 WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE(
3657 wxStringTypeBufferLengthBase<wxStringCharType> )
3658
3659 class wxStringInternalBufferLength
3660 : public wxStringTypeBufferLengthBase<wxStringCharType>
3661 {
3662 public:
3663 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
3664 : wxStringTypeBufferLengthBase<wxStringCharType>(str, lenWanted) {}
3665
3666 ~wxStringInternalBufferLength()
3667 {
3668 m_str.m_impl.assign(m_buf.data(), m_len);
3669 }
3670
3671 DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)
3672 };
3673
3674 #endif // wxUSE_STL_BASED_WXSTRING
3675
3676
3677 #if wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8
3678 typedef wxStringTypeBuffer<wxChar> wxStringBuffer;
3679 typedef wxStringTypeBufferLength<wxChar> wxStringBufferLength;
3680 #else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3681 typedef wxStringInternalBuffer wxStringBuffer;
3682 typedef wxStringInternalBufferLength wxStringBufferLength;
3683 #endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3684
3685 #if wxUSE_UNICODE_UTF8
3686 typedef wxStringInternalBuffer wxUTF8StringBuffer;
3687 typedef wxStringInternalBufferLength wxUTF8StringBufferLength;
3688 #elif wxUSE_UNICODE_WCHAR
3689
3690 WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<char> )
3691
3692 class WXDLLIMPEXP_BASE wxUTF8StringBuffer : public wxStringTypeBufferBase<char>
3693 {
3694 public:
3695 wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024)
3696 : wxStringTypeBufferBase<char>(str, lenWanted) {}
3697 ~wxUTF8StringBuffer();
3698
3699 DECLARE_NO_COPY_CLASS(wxUTF8StringBuffer)
3700 };
3701
3702 WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase<char> )
3703
3704 class WXDLLIMPEXP_BASE wxUTF8StringBufferLength
3705 : public wxStringTypeBufferLengthBase<char>
3706 {
3707 public:
3708 wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024)
3709 : wxStringTypeBufferLengthBase<char>(str, lenWanted) {}
3710 ~wxUTF8StringBufferLength();
3711
3712 DECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength)
3713 };
3714 #endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR
3715
3716
3717 // ---------------------------------------------------------------------------
3718 // wxString comparison functions: operator versions are always case sensitive
3719 // ---------------------------------------------------------------------------
3720
3721 #define wxCMP_WXCHAR_STRING(p, s, op) 0 op s.Cmp(p)
3722
3723 wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING)
3724
3725 #undef wxCMP_WXCHAR_STRING
3726
3727 inline bool operator==(const wxString& s1, const wxString& s2)
3728 { return s1.IsSameAs(s2); }
3729 inline bool operator!=(const wxString& s1, const wxString& s2)
3730 { return !s1.IsSameAs(s2); }
3731 inline bool operator< (const wxString& s1, const wxString& s2)
3732 { return s1.Cmp(s2) < 0; }
3733 inline bool operator> (const wxString& s1, const wxString& s2)
3734 { return s1.Cmp(s2) > 0; }
3735 inline bool operator<=(const wxString& s1, const wxString& s2)
3736 { return s1.Cmp(s2) <= 0; }
3737 inline bool operator>=(const wxString& s1, const wxString& s2)
3738 { return s1.Cmp(s2) >= 0; }
3739
3740 inline bool operator==(const wxString& s1, const wxCStrData& s2)
3741 { return s1 == s2.AsString(); }
3742 inline bool operator==(const wxCStrData& s1, const wxString& s2)
3743 { return s1.AsString() == s2; }
3744 inline bool operator!=(const wxString& s1, const wxCStrData& s2)
3745 { return s1 != s2.AsString(); }
3746 inline bool operator!=(const wxCStrData& s1, const wxString& s2)
3747 { return s1.AsString() != s2; }
3748
3749 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
3750 { return (s1.Cmp((const wchar_t *)s2) == 0); }
3751 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
3752 { return (s2.Cmp((const wchar_t *)s1) == 0); }
3753 inline bool operator!=(const wxString& s1, const wxWCharBuffer& s2)
3754 { return (s1.Cmp((const wchar_t *)s2) != 0); }
3755 inline bool operator!=(const wxWCharBuffer& s1, const wxString& s2)
3756 { return (s2.Cmp((const wchar_t *)s1) != 0); }
3757
3758 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
3759 { return (s1.Cmp((const char *)s2) == 0); }
3760 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
3761 { return (s2.Cmp((const char *)s1) == 0); }
3762 inline bool operator!=(const wxString& s1, const wxCharBuffer& s2)
3763 { return (s1.Cmp((const char *)s2) != 0); }
3764 inline bool operator!=(const wxCharBuffer& s1, const wxString& s2)
3765 { return (s2.Cmp((const char *)s1) != 0); }
3766
3767 inline wxString operator+(const wxString& string, const wxWCharBuffer& buf)
3768 { return string + (const wchar_t *)buf; }
3769 inline wxString operator+(const wxWCharBuffer& buf, const wxString& string)
3770 { return (const wchar_t *)buf + string; }
3771
3772 inline wxString operator+(const wxString& string, const wxCharBuffer& buf)
3773 { return string + (const char *)buf; }
3774 inline wxString operator+(const wxCharBuffer& buf, const wxString& string)
3775 { return (const char *)buf + string; }
3776
3777 // comparison with char
3778 inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); }
3779 inline bool operator==(const wxUniCharRef& c, const wxString& s) { return s.IsSameAs(c); }
3780 inline bool operator==(char c, const wxString& s) { return s.IsSameAs(c); }
3781 inline bool operator==(wchar_t c, const wxString& s) { return s.IsSameAs(c); }
3782 inline bool operator==(int c, const wxString& s) { return s.IsSameAs(c); }
3783 inline bool operator==(const wxString& s, const wxUniChar& c) { return s.IsSameAs(c); }
3784 inline bool operator==(const wxString& s, const wxUniCharRef& c) { return s.IsSameAs(c); }
3785 inline bool operator==(const wxString& s, char c) { return s.IsSameAs(c); }
3786 inline bool operator==(const wxString& s, wchar_t c) { return s.IsSameAs(c); }
3787 inline bool operator!=(const wxUniChar& c, const wxString& s) { return !s.IsSameAs(c); }
3788 inline bool operator!=(const wxUniCharRef& c, const wxString& s) { return !s.IsSameAs(c); }
3789 inline bool operator!=(char c, const wxString& s) { return !s.IsSameAs(c); }
3790 inline bool operator!=(wchar_t c, const wxString& s) { return !s.IsSameAs(c); }
3791 inline bool operator!=(int c, const wxString& s) { return !s.IsSameAs(c); }
3792 inline bool operator!=(const wxString& s, const wxUniChar& c) { return !s.IsSameAs(c); }
3793 inline bool operator!=(const wxString& s, const wxUniCharRef& c) { return !s.IsSameAs(c); }
3794 inline bool operator!=(const wxString& s, char c) { return !s.IsSameAs(c); }
3795 inline bool operator!=(const wxString& s, wchar_t c) { return !s.IsSameAs(c); }
3796
3797 // comparison with C string in Unicode build
3798 #if wxUSE_UNICODE
3799
3800 #define wxCMP_CHAR_STRING(p, s, op) wxString(p) op s
3801
3802 wxDEFINE_ALL_COMPARISONS(const char *, const wxString&, wxCMP_CHAR_STRING)
3803
3804 #undef wxCMP_CHAR_STRING
3805
3806 #endif // wxUSE_UNICODE
3807
3808 // we also need to provide the operators for comparison with wxCStrData to
3809 // resolve ambiguity between operator(const wxChar *,const wxString &) and
3810 // operator(const wxChar *, const wxChar *) for "p == s.c_str()"
3811 //
3812 // notice that these are (shallow) pointer comparisons, not (deep) string ones
3813 #define wxCMP_CHAR_CSTRDATA(p, s, op) p op s.AsChar()
3814 #define wxCMP_WCHAR_CSTRDATA(p, s, op) p op s.AsWChar()
3815
3816 wxDEFINE_ALL_COMPARISONS(const wchar_t *, const wxCStrData&, wxCMP_WCHAR_CSTRDATA)
3817 wxDEFINE_ALL_COMPARISONS(const char *, const wxCStrData&, wxCMP_CHAR_CSTRDATA)
3818
3819 #undef wxCMP_CHAR_CSTRDATA
3820 #undef wxCMP_WCHAR_CSTRDATA
3821
3822 // ---------------------------------------------------------------------------
3823 // Implementation only from here until the end of file
3824 // ---------------------------------------------------------------------------
3825
3826 #if wxUSE_STD_IOSTREAM
3827
3828 #include "wx/iosfwrap.h"
3829
3830 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxString&);
3831 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCStrData&);
3832 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCharBuffer&);
3833 #ifndef __BORLANDC__
3834 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&);
3835 #endif
3836
3837 #if wxUSE_UNICODE && defined(HAVE_WOSTREAM)
3838
3839 WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxString&);
3840 WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxCStrData&);
3841 WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxWCharBuffer&);
3842
3843 #endif // wxUSE_UNICODE && defined(HAVE_WOSTREAM)
3844
3845 #endif // wxUSE_STD_IOSTREAM
3846
3847 // ---------------------------------------------------------------------------
3848 // wxCStrData implementation
3849 // ---------------------------------------------------------------------------
3850
3851 inline wxCStrData::wxCStrData(char *buf)
3852 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
3853 inline wxCStrData::wxCStrData(wchar_t *buf)
3854 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
3855
3856 inline wxCStrData::wxCStrData(const wxCStrData& data)
3857 : m_str(data.m_owned ? new wxString(*data.m_str) : data.m_str),
3858 m_offset(data.m_offset),
3859 m_owned(data.m_owned)
3860 {
3861 }
3862
3863 inline wxCStrData::~wxCStrData()
3864 {
3865 if ( m_owned )
3866 delete wx_const_cast(wxString*, m_str); // cast to silence warnings
3867 }
3868
3869 // simple cases for AsChar() and AsWChar(), the complicated ones are
3870 // in string.cpp
3871 #if wxUSE_UNICODE_WCHAR
3872 inline const wchar_t* wxCStrData::AsWChar() const
3873 {
3874 return m_str->wx_str() + m_offset;
3875 }
3876 #endif // wxUSE_UNICODE_WCHAR
3877
3878 #if !wxUSE_UNICODE
3879 inline const char* wxCStrData::AsChar() const
3880 {
3881 return m_str->wx_str() + m_offset;
3882 }
3883 #endif // !wxUSE_UNICODE
3884
3885 #if wxUSE_UTF8_LOCALE_ONLY
3886 inline const char* wxCStrData::AsChar() const
3887 {
3888 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
3889 }
3890 #endif // wxUSE_UTF8_LOCALE_ONLY
3891
3892 inline const wxCharBuffer wxCStrData::AsCharBuf() const
3893 {
3894 #if !wxUSE_UNICODE
3895 return wxCharBuffer::CreateNonOwned(AsChar());
3896 #else
3897 return AsString().mb_str();
3898 #endif
3899 }
3900
3901 inline const wxWCharBuffer wxCStrData::AsWCharBuf() const
3902 {
3903 #if wxUSE_UNICODE_WCHAR
3904 return wxWCharBuffer::CreateNonOwned(AsWChar());
3905 #else
3906 return AsString().wc_str();
3907 #endif
3908 }
3909
3910 inline wxString wxCStrData::AsString() const
3911 {
3912 if ( m_offset == 0 )
3913 return *m_str;
3914 else
3915 return m_str->Mid(m_offset);
3916 }
3917
3918 inline const wxStringCharType *wxCStrData::AsInternal() const
3919 {
3920 #if wxUSE_UNICODE_UTF8
3921 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
3922 #else
3923 return m_str->wx_str() + m_offset;
3924 #endif
3925 }
3926
3927 inline wxUniChar wxCStrData::operator*() const
3928 {
3929 if ( m_str->empty() )
3930 return wxUniChar(_T('\0'));
3931 else
3932 return (*m_str)[m_offset];
3933 }
3934
3935 inline wxUniChar wxCStrData::operator[](size_t n) const
3936 {
3937 // NB: we intentionally use operator[] and not at() here because the former
3938 // works for the terminating NUL while the latter does not
3939 return (*m_str)[m_offset + n];
3940 }
3941
3942 // ----------------------------------------------------------------------------
3943 // more wxCStrData operators
3944 // ----------------------------------------------------------------------------
3945
3946 // we need to define those to allow "size_t pos = p - s.c_str()" where p is
3947 // some pointer into the string
3948 inline size_t operator-(const char *p, const wxCStrData& cs)
3949 {
3950 return p - cs.AsChar();
3951 }
3952
3953 inline size_t operator-(const wchar_t *p, const wxCStrData& cs)
3954 {
3955 return p - cs.AsWChar();
3956 }
3957
3958 // ----------------------------------------------------------------------------
3959 // implementation of wx[W]CharBuffer inline methods using wxCStrData
3960 // ----------------------------------------------------------------------------
3961
3962 // FIXME-UTF8: move this to buffer.h
3963 inline wxCharBuffer::wxCharBuffer(const wxCStrData& cstr)
3964 : wxCharTypeBufferBase(cstr.AsCharBuf())
3965 {
3966 }
3967
3968 inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr)
3969 : wxCharTypeBufferBase(cstr.AsWCharBuf())
3970 {
3971 }
3972
3973 #if wxUSE_UNICODE_UTF8
3974 // ----------------------------------------------------------------------------
3975 // implementation of wxStringIteratorNode inline methods
3976 // ----------------------------------------------------------------------------
3977
3978 void wxStringIteratorNode::DoSet(const wxString *str,
3979 wxStringImpl::const_iterator *citer,
3980 wxStringImpl::iterator *iter)
3981 {
3982 m_prev = NULL;
3983 m_iter = iter;
3984 m_citer = citer;
3985 m_str = str;
3986 if ( str )
3987 {
3988 m_next = str->m_iterators.ptr;
3989 wx_const_cast(wxString*, m_str)->m_iterators.ptr = this;
3990 if ( m_next )
3991 m_next->m_prev = this;
3992 }
3993 else
3994 {
3995 m_next = NULL;
3996 }
3997 }
3998
3999 void wxStringIteratorNode::clear()
4000 {
4001 if ( m_next )
4002 m_next->m_prev = m_prev;
4003 if ( m_prev )
4004 m_prev->m_next = m_next;
4005 else if ( m_str ) // first in the list
4006 wx_const_cast(wxString*, m_str)->m_iterators.ptr = m_next;
4007
4008 m_next = m_prev = NULL;
4009 m_citer = NULL;
4010 m_iter = NULL;
4011 m_str = NULL;
4012 }
4013 #endif // wxUSE_UNICODE_UTF8
4014
4015 #if WXWIN_COMPATIBILITY_2_8
4016 // lot of code out there doesn't explicitly include wx/crt.h, but uses
4017 // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h,
4018 // so let's include this header now that wxString is defined and it's safe
4019 // to do it:
4020 #include "wx/crt.h"
4021 #endif
4022
4023 #endif // _WX_WXSTRING_H_