]> git.saurik.com Git - wxWidgets.git/blame_incremental - include/wx/string.h
fixing override, fixes #9985 (second error)
[wxWidgets.git] / include / wx / string.h
... / ...
CommitLineData
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
90class 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
98namespace 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
137wxDEPRECATED( inline bool IsEmpty(const char *p) );
138inline bool IsEmpty(const char *p) { return (!p || !*p); }
139
140// safe version of strlen() (returns 0 if passed NULL pointer)
141wxDEPRECATED( inline size_t Strlen(const char *psz) );
142inline size_t Strlen(const char *psz)
143 { return psz ? strlen(psz) : 0; }
144
145// portable strcasecmp/_stricmp
146wxDEPRECATED( inline int Stricmp(const char *psz1, const char *psz2) );
147inline 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*.
210class WXDLLIMPEXP_BASE wxCStrData
211{
212private:
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
217public:
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
287private:
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
316class WXDLLIMPEXP_BASE wxStringPrintfMixinBase
317{
318protected:
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
334class wxStringPrintfMixin : public wxStringPrintfMixinBase
335{
336private:
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
347public:
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
377protected:
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
396class WXDLLIMPEXP_BASE wxStringIteratorNode
397{
398public:
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
419private:
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
431class 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!
439public:
440 // an 'invalid' value for string index, moved to this place due to a CW bug
441 static const size_t npos;
442
443private:
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#ifndef wxHAS_COMPILER_TLS
598 // we must use an accessor function and not a static variable when the TLS
599 // variables support is implemented in the library (and not by the compiler)
600 // because the global s_cache variable could be not yet initialized when a
601 // ctor of another global object is executed and if that ctor uses any
602 // wxString methods, bad things happen
603 //
604 // however notice that this approach does not work when compiler TLS is used,
605 // at least not with g++ 4.1.2 under amd64 as it apparently compiles code
606 // using this accessor incorrectly when optimizations are enabled (-O2 is
607 // enough) -- luckily we don't need it then neither as static __thread
608 // variables are initialized by 0 anyhow then and so we can use the variable
609 // directly
610 WXEXPORT static Cache& GetCache()
611 {
612 static wxTLS_TYPE(Cache) s_cache;
613
614 return wxTLS_VALUE(s_cache);
615 }
616
617 // this helper struct is used to ensure that GetCache() is called during
618 // static initialization time, i.e. before any threads creation, as otherwise
619 // the static s_cache construction inside GetCache() wouldn't be MT-safe
620 friend struct wxStrCacheInitializer;
621#else // wxHAS_COMPILER_TLS
622 static wxTLS_TYPE(Cache) ms_cache;
623 static Cache& GetCache() { return wxTLS_VALUE(ms_cache); }
624#endif // !wxHAS_COMPILER_TLS/wxHAS_COMPILER_TLS
625
626 static Cache::Element *GetCacheBegin() { return GetCache().cached; }
627 static Cache::Element *GetCacheEnd() { return GetCacheBegin() + Cache::SIZE; }
628 static unsigned& LastUsedCacheElement() { return GetCache().lastUsed; }
629
630 // this is used in debug builds only to provide a convenient function,
631 // callable from a debugger, to show the cache contents
632 friend struct wxStrCacheDumper;
633
634 // uncomment this to have access to some profiling statistics on program
635 // termination
636 //#define wxPROFILE_STRING_CACHE
637
638#ifdef wxPROFILE_STRING_CACHE
639 static struct PosToImplCacheStats
640 {
641 unsigned postot, // total non-trivial calls to PosToImpl
642 poshits, // cache hits from PosToImpl()
643 mishits, // cached position beyond the needed one
644 sumpos, // sum of all positions, used to compute the
645 // average position after dividing by postot
646 sumofs, // sum of all offsets after using the cache, used to
647 // compute the average after dividing by hits
648 lentot, // number of total calls to length()
649 lenhits; // number of cache hits in length()
650 } ms_cacheStats;
651
652 friend struct wxStrCacheStatsDumper;
653
654 #define wxCACHE_PROFILE_FIELD_INC(field) ms_cacheStats.field++
655 #define wxCACHE_PROFILE_FIELD_ADD(field, val) ms_cacheStats.field += (val)
656#else // !wxPROFILE_STRING_CACHE
657 #define wxCACHE_PROFILE_FIELD_INC(field)
658 #define wxCACHE_PROFILE_FIELD_ADD(field, val)
659#endif // wxPROFILE_STRING_CACHE/!wxPROFILE_STRING_CACHE
660
661 // note: it could seem that the functions below shouldn't be inline because
662 // they are big, contain loops and so the compiler shouldn't be able to
663 // inline them anyhow, however moving them into string.cpp does decrease the
664 // code performance by ~5%, at least when using g++ 4.1 so do keep them here
665 // unless tests show that it's not advantageous any more
666
667 // return the pointer to the cache element for this string or NULL if not
668 // cached
669 Cache::Element *FindCacheElement() const
670 {
671 // profiling seems to show a small but consistent gain if we use this
672 // simple loop instead of starting from the last used element (there are
673 // a lot of misses in this function...)
674 Cache::Element * const cacheBegin = GetCacheBegin();
675#ifndef wxHAS_COMPILER_TLS
676 // during destruction tls calls may return NULL, in this case return NULL
677 // immediately without accessing anything else
678 if ( cacheBegin == NULL )
679 return NULL;
680#endif
681 Cache::Element * const cacheEnd = GetCacheEnd();
682 for ( Cache::Element *c = cacheBegin; c != cacheEnd; c++ )
683 {
684 if ( c->str == this )
685 return c;
686 }
687
688 return NULL;
689 }
690
691 // unlike FindCacheElement(), this one always returns a valid pointer to the
692 // cache element for this string, it may have valid last cached position and
693 // its corresponding index in the byte string or not
694 Cache::Element *GetCacheElement() const
695 {
696 Cache::Element * const cacheBegin = GetCacheBegin();
697 Cache::Element * const cacheEnd = GetCacheEnd();
698 Cache::Element * const cacheStart = cacheBegin + LastUsedCacheElement();
699
700 // check the last used first, this does no (measurable) harm for a miss
701 // but does help for simple loops addressing the same string all the time
702 if ( cacheStart->str == this )
703 return cacheStart;
704
705 // notice that we're going to check cacheStart again inside this call but
706 // profiling shows that it's still faster to use a simple loop like
707 // inside FindCacheElement() than manually looping with wrapping starting
708 // from the cache entry after the start one
709 Cache::Element *c = FindCacheElement();
710 if ( !c )
711 {
712 // claim the next cache entry for this string
713 c = cacheStart;
714 if ( ++c == cacheEnd )
715 c = cacheBegin;
716
717 c->str = this;
718 c->Reset();
719
720 // and remember the last used element
721 LastUsedCacheElement() = c - cacheBegin;
722 }
723
724 return c;
725 }
726
727 size_t DoPosToImpl(size_t pos) const
728 {
729 wxCACHE_PROFILE_FIELD_INC(postot);
730
731 // NB: although the case of pos == 1 (and offset from cached position
732 // equal to 1) are common, nothing is gained by writing special code
733 // for handling them, the compiler (at least g++ 4.1 used) seems to
734 // optimize the code well enough on its own
735
736 wxCACHE_PROFILE_FIELD_ADD(sumpos, pos);
737
738 Cache::Element * const cache = GetCacheElement();
739
740 // cached position can't be 0 so if it is, it means that this entry was
741 // used for length caching only so far, i.e. it doesn't count as a hit
742 // from our point of view
743 if ( cache->pos )
744 {
745 wxCACHE_PROFILE_FIELD_INC(poshits);
746 }
747
748 if ( pos == cache->pos )
749 return cache->impl;
750
751 // this seems to happen only rarely so just reset the cache in this case
752 // instead of complicating code even further by seeking backwards in this
753 // case
754 if ( cache->pos > pos )
755 {
756 wxCACHE_PROFILE_FIELD_INC(mishits);
757
758 cache->ResetPos();
759 }
760
761 wxCACHE_PROFILE_FIELD_ADD(sumofs, pos - cache->pos);
762
763
764 wxStringImpl::const_iterator i(m_impl.begin() + cache->impl);
765 for ( size_t n = cache->pos; n < pos; n++ )
766 wxStringOperations::IncIter(i);
767
768 cache->pos = pos;
769 cache->impl = i - m_impl.begin();
770
771 wxSTRING_CACHE_ASSERT(
772 (int)cache->impl == (begin() + pos).impl() - m_impl.begin() );
773
774 return cache->impl;
775 }
776
777 void InvalidateCache()
778 {
779 Cache::Element * const cache = FindCacheElement();
780 if ( cache )
781 cache->Reset();
782 }
783
784 void InvalidateCachedLength()
785 {
786 Cache::Element * const cache = FindCacheElement();
787 if ( cache )
788 cache->len = npos;
789 }
790
791 void SetCachedLength(size_t len)
792 {
793 // we optimistically cache the length here even if the string wasn't
794 // present in the cache before, this seems to do no harm and the
795 // potential for avoiding length recomputation for long strings looks
796 // interesting
797 GetCacheElement()->len = len;
798 }
799
800 void UpdateCachedLength(ptrdiff_t delta)
801 {
802 Cache::Element * const cache = FindCacheElement();
803 if ( cache && cache->len != npos )
804 {
805 wxSTRING_CACHE_ASSERT( (ptrdiff_t)cache->len + delta >= 0 );
806
807 cache->len += delta;
808 }
809 }
810
811 #define wxSTRING_INVALIDATE_CACHE() InvalidateCache()
812 #define wxSTRING_INVALIDATE_CACHED_LENGTH() InvalidateCachedLength()
813 #define wxSTRING_UPDATE_CACHED_LENGTH(n) UpdateCachedLength(n)
814 #define wxSTRING_SET_CACHED_LENGTH(n) SetCachedLength(n)
815#else // !wxUSE_STRING_POS_CACHE
816 size_t DoPosToImpl(size_t pos) const
817 {
818 return (begin() + pos).impl() - m_impl.begin();
819 }
820
821 #define wxSTRING_INVALIDATE_CACHE()
822 #define wxSTRING_INVALIDATE_CACHED_LENGTH()
823 #define wxSTRING_UPDATE_CACHED_LENGTH(n)
824 #define wxSTRING_SET_CACHED_LENGTH(n)
825#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
826
827 size_t PosToImpl(size_t pos) const
828 {
829 return pos == 0 || pos == npos ? pos : DoPosToImpl(pos);
830 }
831
832 void PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const;
833
834 size_t LenToImpl(size_t len) const
835 {
836 size_t pos, len2;
837 PosLenToImpl(0, len, &pos, &len2);
838 return len2;
839 }
840
841 size_t PosFromImpl(size_t pos) const
842 {
843 if ( pos == 0 || pos == npos )
844 return pos;
845 else
846 return const_iterator(this, m_impl.begin() + pos) - begin();
847 }
848#endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8
849
850public:
851 // standard types
852 typedef wxUniChar value_type;
853 typedef wxUniChar char_type;
854 typedef wxUniCharRef reference;
855 typedef wxChar* pointer;
856 typedef const wxChar* const_pointer;
857
858 typedef size_t size_type;
859 typedef wxUniChar const_reference;
860
861#if wxUSE_STD_STRING
862 #if wxUSE_UNICODE_UTF8
863 // random access is not O(1), as required by Random Access Iterator
864 #define WX_STR_ITERATOR_TAG std::bidirectional_iterator_tag
865 #else
866 #define WX_STR_ITERATOR_TAG std::random_access_iterator_tag
867 #endif
868 #define WX_DEFINE_ITERATOR_CATEGORY(cat) typedef cat iterator_category;
869#else
870 // not defining iterator_category at all in this case is better than defining
871 // it as some dummy type -- at least it results in more intelligible error
872 // messages
873 #define WX_DEFINE_ITERATOR_CATEGORY(cat)
874#endif
875
876 #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \
877 private: \
878 typedef wxStringImpl::iterator_name underlying_iterator; \
879 public: \
880 WX_DEFINE_ITERATOR_CATEGORY(WX_STR_ITERATOR_TAG) \
881 typedef wxUniChar value_type; \
882 typedef int difference_type; \
883 typedef reference_type reference; \
884 typedef pointer_type pointer; \
885 \
886 reference operator[](size_t n) const { return *(*this + n); } \
887 \
888 iterator_name& operator++() \
889 { wxStringOperations::IncIter(m_cur); return *this; } \
890 iterator_name& operator--() \
891 { wxStringOperations::DecIter(m_cur); return *this; } \
892 iterator_name operator++(int) \
893 { \
894 iterator_name tmp = *this; \
895 wxStringOperations::IncIter(m_cur); \
896 return tmp; \
897 } \
898 iterator_name operator--(int) \
899 { \
900 iterator_name tmp = *this; \
901 wxStringOperations::DecIter(m_cur); \
902 return tmp; \
903 } \
904 \
905 iterator_name& operator+=(ptrdiff_t n) \
906 { \
907 m_cur = wxStringOperations::AddToIter(m_cur, n); \
908 return *this; \
909 } \
910 iterator_name& operator-=(ptrdiff_t n) \
911 { \
912 m_cur = wxStringOperations::AddToIter(m_cur, -n); \
913 return *this; \
914 } \
915 \
916 difference_type operator-(const iterator_name& i) const \
917 { return wxStringOperations::DiffIters(m_cur, i.m_cur); } \
918 \
919 bool operator==(const iterator_name& i) const \
920 { return m_cur == i.m_cur; } \
921 bool operator!=(const iterator_name& i) const \
922 { return m_cur != i.m_cur; } \
923 \
924 bool operator<(const iterator_name& i) const \
925 { return m_cur < i.m_cur; } \
926 bool operator>(const iterator_name& i) const \
927 { return m_cur > i.m_cur; } \
928 bool operator<=(const iterator_name& i) const \
929 { return m_cur <= i.m_cur; } \
930 bool operator>=(const iterator_name& i) const \
931 { return m_cur >= i.m_cur; } \
932 \
933 private: \
934 /* for internal wxString use only: */ \
935 underlying_iterator impl() const { return m_cur; } \
936 \
937 friend class wxString; \
938 friend class wxCStrData; \
939 \
940 private: \
941 underlying_iterator m_cur
942
943 class WXDLLIMPEXP_FWD_BASE const_iterator;
944
945#if wxUSE_UNICODE_UTF8
946 // NB: In UTF-8 build, (non-const) iterator needs to keep reference
947 // to the underlying wxStringImpl, because UTF-8 is variable-length
948 // encoding and changing the value pointer to by an iterator (using
949 // its operator*) requires calling wxStringImpl::replace() if the old
950 // and new values differ in their encoding's length.
951 //
952 // Furthermore, the replace() call may invalid all iterators for the
953 // string, so we have to keep track of outstanding iterators and update
954 // them if replace() happens.
955 //
956 // This is implemented by maintaining linked list of iterators for every
957 // string and traversing it in wxUniCharRef::operator=(). Head of the
958 // list is stored in wxString. (FIXME-UTF8)
959
960 class WXDLLIMPEXP_BASE iterator
961 {
962 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
963
964 public:
965 iterator() {}
966 iterator(const iterator& i)
967 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
968 iterator& operator=(const iterator& i)
969 {
970 if (&i != this)
971 {
972 m_cur = i.m_cur;
973 m_node.set(i.str(), &m_cur);
974 }
975 return *this;
976 }
977
978 reference operator*()
979 { return wxUniCharRef::CreateForString(*str(), m_cur); }
980
981 iterator operator+(ptrdiff_t n) const
982 { return iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
983 iterator operator-(ptrdiff_t n) const
984 { return iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
985
986 private:
987 iterator(wxString *str, underlying_iterator ptr)
988 : m_cur(ptr), m_node(str, &m_cur) {}
989
990 wxString* str() const { return wx_const_cast(wxString*, m_node.m_str); }
991
992 wxStringIteratorNode m_node;
993
994 friend class const_iterator;
995 };
996
997 class WXDLLIMPEXP_BASE const_iterator
998 {
999 // NB: reference_type is intentionally value, not reference, the character
1000 // may be encoded differently in wxString data:
1001 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
1002
1003 public:
1004 const_iterator() {}
1005 const_iterator(const const_iterator& i)
1006 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
1007 const_iterator(const iterator& i)
1008 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
1009
1010 const_iterator& operator=(const const_iterator& i)
1011 {
1012 if (&i != this)
1013 {
1014 m_cur = i.m_cur;
1015 m_node.set(i.str(), &m_cur);
1016 }
1017 return *this;
1018 }
1019 const_iterator& operator=(const iterator& i)
1020 { m_cur = i.m_cur; m_node.set(i.str(), &m_cur); return *this; }
1021
1022 reference operator*() const
1023 { return wxStringOperations::DecodeChar(m_cur); }
1024
1025 const_iterator operator+(ptrdiff_t n) const
1026 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
1027 const_iterator operator-(ptrdiff_t n) const
1028 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
1029
1030 private:
1031 // for internal wxString use only:
1032 const_iterator(const wxString *str, underlying_iterator ptr)
1033 : m_cur(ptr), m_node(str, &m_cur) {}
1034
1035 const wxString* str() const { return m_node.m_str; }
1036
1037 wxStringIteratorNode m_node;
1038 };
1039
1040 size_t IterToImplPos(wxString::iterator i) const
1041 { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); }
1042
1043 iterator GetIterForNthChar(size_t n)
1044 { return iterator(this, m_impl.begin() + PosToImpl(n)); }
1045 const_iterator GetIterForNthChar(size_t n) const
1046 { return const_iterator(this, m_impl.begin() + PosToImpl(n)); }
1047#else // !wxUSE_UNICODE_UTF8
1048
1049 class WXDLLIMPEXP_BASE iterator
1050 {
1051 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
1052
1053 public:
1054 iterator() {}
1055 iterator(const iterator& i) : m_cur(i.m_cur) {}
1056
1057 reference operator*()
1058 { return wxUniCharRef::CreateForString(m_cur); }
1059
1060 iterator operator+(ptrdiff_t n) const
1061 { return iterator(wxStringOperations::AddToIter(m_cur, n)); }
1062 iterator operator-(ptrdiff_t n) const
1063 { return iterator(wxStringOperations::AddToIter(m_cur, -n)); }
1064
1065 private:
1066 // for internal wxString use only:
1067 iterator(underlying_iterator ptr) : m_cur(ptr) {}
1068 iterator(wxString *WXUNUSED(str), underlying_iterator ptr) : m_cur(ptr) {}
1069
1070 friend class const_iterator;
1071 };
1072
1073 class WXDLLIMPEXP_BASE const_iterator
1074 {
1075 // NB: reference_type is intentionally value, not reference, the character
1076 // may be encoded differently in wxString data:
1077 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
1078
1079 public:
1080 const_iterator() {}
1081 const_iterator(const const_iterator& i) : m_cur(i.m_cur) {}
1082 const_iterator(const iterator& i) : m_cur(i.m_cur) {}
1083
1084 reference operator*() const
1085 { return wxStringOperations::DecodeChar(m_cur); }
1086
1087 const_iterator operator+(ptrdiff_t n) const
1088 { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); }
1089 const_iterator operator-(ptrdiff_t n) const
1090 { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); }
1091
1092 private:
1093 // for internal wxString use only:
1094 const_iterator(underlying_iterator ptr) : m_cur(ptr) {}
1095 const_iterator(const wxString *WXUNUSED(str), underlying_iterator ptr)
1096 : m_cur(ptr) {}
1097 };
1098
1099 iterator GetIterForNthChar(size_t n) { return begin() + n; }
1100 const_iterator GetIterForNthChar(size_t n) const { return begin() + n; }
1101#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8
1102
1103 #undef WX_STR_ITERATOR_TAG
1104 #undef WX_STR_ITERATOR_IMPL
1105
1106 friend class iterator;
1107 friend class const_iterator;
1108
1109 template <typename T>
1110 class reverse_iterator_impl
1111 {
1112 public:
1113 typedef T iterator_type;
1114
1115 WX_DEFINE_ITERATOR_CATEGORY(typename T::iterator_category)
1116 typedef typename T::value_type value_type;
1117 typedef typename T::difference_type difference_type;
1118 typedef typename T::reference reference;
1119 typedef typename T::pointer *pointer;
1120
1121 reverse_iterator_impl() {}
1122 reverse_iterator_impl(iterator_type i) : m_cur(i) {}
1123 reverse_iterator_impl(const reverse_iterator_impl& ri)
1124 : m_cur(ri.m_cur) {}
1125
1126 iterator_type base() const { return m_cur; }
1127
1128 reference operator*() const { return *(m_cur-1); }
1129 reference operator[](size_t n) const { return *(*this + n); }
1130
1131 reverse_iterator_impl& operator++()
1132 { --m_cur; return *this; }
1133 reverse_iterator_impl operator++(int)
1134 { reverse_iterator_impl tmp = *this; --m_cur; return tmp; }
1135 reverse_iterator_impl& operator--()
1136 { ++m_cur; return *this; }
1137 reverse_iterator_impl operator--(int)
1138 { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; }
1139
1140 // NB: explicit <T> in the functions below is to keep BCC 5.5 happy
1141 reverse_iterator_impl operator+(ptrdiff_t n) const
1142 { return reverse_iterator_impl<T>(m_cur - n); }
1143 reverse_iterator_impl operator-(ptrdiff_t n) const
1144 { return reverse_iterator_impl<T>(m_cur + n); }
1145 reverse_iterator_impl operator+=(ptrdiff_t n)
1146 { m_cur -= n; return *this; }
1147 reverse_iterator_impl operator-=(ptrdiff_t n)
1148 { m_cur += n; return *this; }
1149
1150 unsigned operator-(const reverse_iterator_impl& i) const
1151 { return i.m_cur - m_cur; }
1152
1153 bool operator==(const reverse_iterator_impl& ri) const
1154 { return m_cur == ri.m_cur; }
1155 bool operator!=(const reverse_iterator_impl& ri) const
1156 { return !(*this == ri); }
1157
1158 bool operator<(const reverse_iterator_impl& i) const
1159 { return m_cur > i.m_cur; }
1160 bool operator>(const reverse_iterator_impl& i) const
1161 { return m_cur < i.m_cur; }
1162 bool operator<=(const reverse_iterator_impl& i) const
1163 { return m_cur >= i.m_cur; }
1164 bool operator>=(const reverse_iterator_impl& i) const
1165 { return m_cur <= i.m_cur; }
1166
1167 private:
1168 iterator_type m_cur;
1169 };
1170
1171 typedef reverse_iterator_impl<iterator> reverse_iterator;
1172 typedef reverse_iterator_impl<const_iterator> const_reverse_iterator;
1173
1174private:
1175 // used to transform an expression built using c_str() (and hence of type
1176 // wxCStrData) to an iterator into the string
1177 static const_iterator CreateConstIterator(const wxCStrData& data)
1178 {
1179 return const_iterator(data.m_str,
1180 (data.m_str->begin() + data.m_offset).impl());
1181 }
1182
1183 // in UTF-8 STL build, creation from std::string requires conversion under
1184 // non-UTF8 locales, so we can't have and use wxString(wxStringImpl) ctor;
1185 // instead we define dummy type that lets us have wxString ctor for creation
1186 // from wxStringImpl that couldn't be used by user code (in all other builds,
1187 // "standard" ctors can be used):
1188#if wxUSE_UNICODE_UTF8 && wxUSE_STL_BASED_WXSTRING
1189 struct CtorFromStringImplTag {};
1190
1191 wxString(CtorFromStringImplTag* WXUNUSED(dummy), const wxStringImpl& src)
1192 : m_impl(src) {}
1193
1194 static wxString FromImpl(const wxStringImpl& src)
1195 { return wxString((CtorFromStringImplTag*)NULL, src); }
1196#else
1197 #if !wxUSE_STL_BASED_WXSTRING
1198 wxString(const wxStringImpl& src) : m_impl(src) { }
1199 // else: already defined as wxString(wxStdString) below
1200 #endif
1201 static wxString FromImpl(const wxStringImpl& src) { return wxString(src); }
1202#endif
1203
1204public:
1205 // constructors and destructor
1206 // ctor for an empty string
1207 wxString() {}
1208
1209 // copy ctor
1210 wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { }
1211
1212 // string containing nRepeat copies of ch
1213 wxString(wxUniChar ch, size_t nRepeat = 1 )
1214 { assign(nRepeat, ch); }
1215 wxString(size_t nRepeat, wxUniChar ch)
1216 { assign(nRepeat, ch); }
1217 wxString(wxUniCharRef ch, size_t nRepeat = 1)
1218 { assign(nRepeat, ch); }
1219 wxString(size_t nRepeat, wxUniCharRef ch)
1220 { assign(nRepeat, ch); }
1221 wxString(char ch, size_t nRepeat = 1)
1222 { assign(nRepeat, ch); }
1223 wxString(size_t nRepeat, char ch)
1224 { assign(nRepeat, ch); }
1225 wxString(wchar_t ch, size_t nRepeat = 1)
1226 { assign(nRepeat, ch); }
1227 wxString(size_t nRepeat, wchar_t ch)
1228 { assign(nRepeat, ch); }
1229
1230 // ctors from char* strings:
1231 wxString(const char *psz)
1232 : m_impl(ImplStr(psz)) {}
1233 wxString(const char *psz, const wxMBConv& conv)
1234 : m_impl(ImplStr(psz, conv)) {}
1235 wxString(const char *psz, size_t nLength)
1236 { assign(psz, nLength); }
1237 wxString(const char *psz, const wxMBConv& conv, size_t nLength)
1238 {
1239 SubstrBufFromMB str(ImplStr(psz, nLength, conv));
1240 m_impl.assign(str.data, str.len);
1241 }
1242
1243 // and unsigned char*:
1244 wxString(const unsigned char *psz)
1245 : m_impl(ImplStr((const char*)psz)) {}
1246 wxString(const unsigned char *psz, const wxMBConv& conv)
1247 : m_impl(ImplStr((const char*)psz, conv)) {}
1248 wxString(const unsigned char *psz, size_t nLength)
1249 { assign((const char*)psz, nLength); }
1250 wxString(const unsigned char *psz, const wxMBConv& conv, size_t nLength)
1251 {
1252 SubstrBufFromMB str(ImplStr((const char*)psz, nLength, conv));
1253 m_impl.assign(str.data, str.len);
1254 }
1255
1256 // ctors from wchar_t* strings:
1257 wxString(const wchar_t *pwz)
1258 : m_impl(ImplStr(pwz)) {}
1259 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv))
1260 : m_impl(ImplStr(pwz)) {}
1261 wxString(const wchar_t *pwz, size_t nLength)
1262 { assign(pwz, nLength); }
1263 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv), size_t nLength)
1264 { assign(pwz, nLength); }
1265
1266 wxString(const wxCharBuffer& buf)
1267 { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
1268 wxString(const wxWCharBuffer& buf)
1269 { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
1270
1271 // NB: this version uses m_impl.c_str() to force making a copy of the
1272 // string, so that "wxString(str.c_str())" idiom for passing strings
1273 // between threads works
1274 wxString(const wxCStrData& cstr)
1275 : m_impl(cstr.AsString().m_impl.c_str()) { }
1276
1277 // as we provide both ctors with this signature for both char and unsigned
1278 // char string, we need to provide one for wxCStrData to resolve ambiguity
1279 wxString(const wxCStrData& cstr, size_t nLength)
1280 : m_impl(cstr.AsString().Mid(0, nLength).m_impl) {}
1281
1282 // and because wxString is convertible to wxCStrData and const wxChar *
1283 // we also need to provide this one
1284 wxString(const wxString& str, size_t nLength)
1285 { assign(str, nLength); }
1286
1287
1288#if wxUSE_STRING_POS_CACHE
1289 ~wxString()
1290 {
1291 // we need to invalidate our cache entry as another string could be
1292 // recreated at the same address (unlikely, but still possible, with the
1293 // heap-allocated strings but perfectly common with stack-allocated ones)
1294 InvalidateCache();
1295 }
1296#endif // wxUSE_STRING_POS_CACHE
1297
1298 // even if we're not built with wxUSE_STL == 1 it is very convenient to allow
1299 // implicit conversions from std::string to wxString and vice verse as this
1300 // allows to use the same strings in non-GUI and GUI code, however we don't
1301 // want to unconditionally add this ctor as it would make wx lib dependent on
1302 // libstdc++ on some Linux versions which is bad, so instead we ask the
1303 // client code to define this wxUSE_STD_STRING symbol if they need it
1304#if wxUSE_STD_STRING
1305 #if wxUSE_UNICODE_WCHAR
1306 wxString(const wxStdWideString& str) : m_impl(str) {}
1307 #else // UTF-8 or ANSI
1308 wxString(const wxStdWideString& str)
1309 { assign(str.c_str(), str.length()); }
1310 #endif
1311
1312 #if !wxUSE_UNICODE // ANSI build
1313 // FIXME-UTF8: do this in UTF8 build #if wxUSE_UTF8_LOCALE_ONLY, too
1314 wxString(const std::string& str) : m_impl(str) {}
1315 #else // Unicode
1316 wxString(const std::string& str)
1317 { assign(str.c_str(), str.length()); }
1318 #endif
1319#endif // wxUSE_STD_STRING
1320
1321 // Unlike ctor from std::string, we provide conversion to std::string only
1322 // if wxUSE_STL and not merely wxUSE_STD_STRING (which is on by default),
1323 // because it conflicts with operator const char/wchar_t*:
1324#if wxUSE_STL
1325 #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING
1326 // wxStringImpl is std::string in the encoding we want
1327 operator const wxStdWideString&() const { return m_impl; }
1328 #else
1329 // wxStringImpl is either not std::string or needs conversion
1330 operator wxStdWideString() const
1331 // FIXME-UTF8: broken for embedded NULs
1332 { return wxStdWideString(wc_str()); }
1333 #endif
1334
1335 #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING
1336 // wxStringImpl is std::string in the encoding we want
1337 operator const std::string&() const { return m_impl; }
1338 #else
1339 // wxStringImpl is either not std::string or needs conversion
1340 operator std::string() const
1341 // FIXME-UTF8: broken for embedded NULs
1342 { return std::string(mb_str()); }
1343 #endif
1344#endif // wxUSE_STL
1345
1346 wxString Clone() const
1347 {
1348 // make a deep copy of the string, i.e. the returned string will have
1349 // ref count = 1 with refcounted implementation
1350 return wxString::FromImpl(wxStringImpl(m_impl.c_str(), m_impl.length()));
1351 }
1352
1353 // first valid index position
1354 const_iterator begin() const { return const_iterator(this, m_impl.begin()); }
1355 iterator begin() { return iterator(this, m_impl.begin()); }
1356 // position one after the last valid one
1357 const_iterator end() const { return const_iterator(this, m_impl.end()); }
1358 iterator end() { return iterator(this, m_impl.end()); }
1359
1360 // first element of the reversed string
1361 const_reverse_iterator rbegin() const
1362 { return const_reverse_iterator(end()); }
1363 reverse_iterator rbegin()
1364 { return reverse_iterator(end()); }
1365 // one beyond the end of the reversed string
1366 const_reverse_iterator rend() const
1367 { return const_reverse_iterator(begin()); }
1368 reverse_iterator rend()
1369 { return reverse_iterator(begin()); }
1370
1371 // std::string methods:
1372#if wxUSE_UNICODE_UTF8
1373 size_t length() const
1374 {
1375#if wxUSE_STRING_POS_CACHE
1376 wxCACHE_PROFILE_FIELD_INC(lentot);
1377
1378 Cache::Element * const cache = GetCacheElement();
1379
1380 if ( cache->len == npos )
1381 {
1382 // it's probably not worth trying to be clever and using cache->pos
1383 // here as it's probably 0 anyhow -- you usually call length() before
1384 // starting to index the string
1385 cache->len = end() - begin();
1386 }
1387 else
1388 {
1389 wxCACHE_PROFILE_FIELD_INC(lenhits);
1390
1391 wxSTRING_CACHE_ASSERT( (int)cache->len == end() - begin() );
1392 }
1393
1394 return cache->len;
1395#else // !wxUSE_STRING_POS_CACHE
1396 return end() - begin();
1397#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
1398 }
1399#else
1400 size_t length() const { return m_impl.length(); }
1401#endif
1402
1403 size_type size() const { return length(); }
1404 size_type max_size() const { return npos; }
1405
1406 bool empty() const { return m_impl.empty(); }
1407
1408 // NB: these methods don't have a well-defined meaning in UTF-8 case
1409 size_type capacity() const { return m_impl.capacity(); }
1410 void reserve(size_t sz) { m_impl.reserve(sz); }
1411
1412 void resize(size_t nSize, wxUniChar ch = wxT('\0'))
1413 {
1414 const size_t len = length();
1415 if ( nSize == len)
1416 return;
1417
1418#if wxUSE_UNICODE_UTF8
1419 if ( nSize < len )
1420 {
1421 wxSTRING_INVALIDATE_CACHE();
1422
1423 // we can't use wxStringImpl::resize() for truncating the string as it
1424 // counts in bytes, not characters
1425 erase(nSize);
1426 return;
1427 }
1428
1429 // we also can't use (presumably more efficient) resize() if we have to
1430 // append characters taking more than one byte
1431 if ( !ch.IsAscii() )
1432 {
1433 append(nSize - len, ch);
1434 }
1435 else // can use (presumably faster) resize() version
1436#endif // wxUSE_UNICODE_UTF8
1437 {
1438 wxSTRING_INVALIDATE_CACHED_LENGTH();
1439
1440 m_impl.resize(nSize, (wxStringCharType)ch);
1441 }
1442 }
1443
1444 wxString substr(size_t nStart = 0, size_t nLen = npos) const
1445 {
1446 size_t pos, len;
1447 PosLenToImpl(nStart, nLen, &pos, &len);
1448 return FromImpl(m_impl.substr(pos, len));
1449 }
1450
1451 // generic attributes & operations
1452 // as standard strlen()
1453 size_t Len() const { return length(); }
1454 // string contains any characters?
1455 bool IsEmpty() const { return empty(); }
1456 // empty string is "false", so !str will return true
1457 bool operator!() const { return empty(); }
1458 // truncate the string to given length
1459 wxString& Truncate(size_t uiLen);
1460 // empty string contents
1461 void Empty() { clear(); }
1462 // empty the string and free memory
1463 void Clear() { clear(); }
1464
1465 // contents test
1466 // Is an ascii value
1467 bool IsAscii() const;
1468 // Is a number
1469 bool IsNumber() const;
1470 // Is a word
1471 bool IsWord() const;
1472
1473 // data access (all indexes are 0 based)
1474 // read access
1475 wxUniChar at(size_t n) const
1476 { return wxStringOperations::DecodeChar(m_impl.begin() + PosToImpl(n)); }
1477 wxUniChar GetChar(size_t n) const
1478 { return at(n); }
1479 // read/write access
1480 wxUniCharRef at(size_t n)
1481 { return *GetIterForNthChar(n); }
1482 wxUniCharRef GetWritableChar(size_t n)
1483 { return at(n); }
1484 // write access
1485 void SetChar(size_t n, wxUniChar ch)
1486 { at(n) = ch; }
1487
1488 // get last character
1489 wxUniChar Last() const
1490 {
1491 wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
1492 return *rbegin();
1493 }
1494
1495 // get writable last character
1496 wxUniCharRef Last()
1497 {
1498 wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
1499 return *rbegin();
1500 }
1501
1502 /*
1503 Note that we we must define all of the overloads below to avoid
1504 ambiguity when using str[0].
1505 */
1506 wxUniChar operator[](int n) const
1507 { return at(n); }
1508 wxUniChar operator[](long n) const
1509 { return at(n); }
1510 wxUniChar operator[](size_t n) const
1511 { return at(n); }
1512#ifndef wxSIZE_T_IS_UINT
1513 wxUniChar operator[](unsigned int n) const
1514 { return at(n); }
1515#endif // size_t != unsigned int
1516
1517 // operator versions of GetWriteableChar()
1518 wxUniCharRef operator[](int n)
1519 { return at(n); }
1520 wxUniCharRef operator[](long n)
1521 { return at(n); }
1522 wxUniCharRef operator[](size_t n)
1523 { return at(n); }
1524#ifndef wxSIZE_T_IS_UINT
1525 wxUniCharRef operator[](unsigned int n)
1526 { return at(n); }
1527#endif // size_t != unsigned int
1528
1529 // explicit conversion to C string (use this with printf()!)
1530 wxCStrData c_str() const { return wxCStrData(this); }
1531 wxCStrData data() const { return c_str(); }
1532
1533 // implicit conversion to C string
1534 operator wxCStrData() const { return c_str(); }
1535
1536 // the first two operators conflict with operators for conversion to
1537 // std::string and they must be disabled in STL build; the next one only
1538 // makes sense if conversions to char* are also defined and not defining it
1539 // in STL build also helps us to get more clear error messages for the code
1540 // which relies on implicit conversion to char* in STL build
1541#if !wxUSE_STL
1542 operator const char*() const { return c_str(); }
1543 operator const wchar_t*() const { return c_str(); }
1544
1545 // implicit conversion to untyped pointer for compatibility with previous
1546 // wxWidgets versions: this is the same as conversion to const char * so it
1547 // may fail!
1548 operator const void*() const { return c_str(); }
1549#endif // wxUSE_STL
1550
1551 // identical to c_str(), for MFC compatibility
1552 const wxCStrData GetData() const { return c_str(); }
1553
1554 // explicit conversion to C string in internal representation (char*,
1555 // wchar_t*, UTF-8-encoded char*, depending on the build):
1556 const wxStringCharType *wx_str() const { return m_impl.c_str(); }
1557
1558 // conversion to *non-const* multibyte or widestring buffer; modifying
1559 // returned buffer won't affect the string, these methods are only useful
1560 // for passing values to const-incorrect functions
1561 wxWritableCharBuffer char_str(const wxMBConv& conv = wxConvLibc) const
1562 { return mb_str(conv); }
1563 wxWritableWCharBuffer wchar_str() const { return wc_str(); }
1564
1565 // conversion to the buffer of the given type T (= char or wchar_t) and
1566 // also optionally return the buffer length
1567 //
1568 // this is mostly/only useful for the template functions
1569 //
1570 // FIXME-VC6: the second argument only exists for VC6 which doesn't support
1571 // explicit template function selection, do not use it unless
1572 // you must support VC6!
1573 template <typename T>
1574 wxCharTypeBuffer<T> tchar_str(size_t *len = NULL,
1575 T * WXUNUSED(dummy) = NULL) const
1576 {
1577#if wxUSE_UNICODE
1578 // we need a helper dispatcher depending on type
1579 return wxPrivate::wxStringAsBufHelper<T>::Get(*this, len);
1580#else // ANSI
1581 // T can only be char in ANSI build
1582 if ( len )
1583 *len = length();
1584
1585 return wxCharTypeBuffer<T>::CreateNonOwned(wx_str());
1586#endif // Unicode build kind
1587 }
1588
1589 // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for
1590 // converting numbers or strings which are certain not to contain special
1591 // chars (typically system functions, X atoms, environment variables etc.)
1592 //
1593 // the behaviour of these functions with the strings containing anything
1594 // else than 7 bit ASCII characters is undefined, use at your own risk.
1595#if wxUSE_UNICODE
1596 static wxString FromAscii(const char *ascii, size_t len);
1597 static wxString FromAscii(const char *ascii);
1598 static wxString FromAscii(char ascii);
1599 const wxCharBuffer ToAscii() const;
1600#else // ANSI
1601 static wxString FromAscii(const char *ascii) { return wxString( ascii ); }
1602 static wxString FromAscii(const char *ascii, size_t len)
1603 { return wxString( ascii, len ); }
1604 static wxString FromAscii(char ascii) { return wxString( ascii ); }
1605 const char *ToAscii() const { return c_str(); }
1606#endif // Unicode/!Unicode
1607
1608 // also provide unsigned char overloads as signed/unsigned doesn't matter
1609 // for 7 bit ASCII characters
1610 static wxString FromAscii(const unsigned char *ascii)
1611 { return FromAscii((const char *)ascii); }
1612 static wxString FromAscii(const unsigned char *ascii, size_t len)
1613 { return FromAscii((const char *)ascii, len); }
1614
1615 // conversion to/from UTF-8:
1616#if wxUSE_UNICODE_UTF8
1617 static wxString FromUTF8Unchecked(const char *utf8)
1618 {
1619 if ( !utf8 )
1620 return wxEmptyString;
1621
1622 wxASSERT( wxStringOperations::IsValidUtf8String(utf8) );
1623 return FromImpl(wxStringImpl(utf8));
1624 }
1625 static wxString FromUTF8Unchecked(const char *utf8, size_t len)
1626 {
1627 if ( !utf8 )
1628 return wxEmptyString;
1629 if ( len == npos )
1630 return FromUTF8Unchecked(utf8);
1631
1632 wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) );
1633 return FromImpl(wxStringImpl(utf8, len));
1634 }
1635
1636 static wxString FromUTF8(const char *utf8)
1637 {
1638 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8) )
1639 return "";
1640
1641 return FromImpl(wxStringImpl(utf8));
1642 }
1643 static wxString FromUTF8(const char *utf8, size_t len)
1644 {
1645 if ( len == npos )
1646 return FromUTF8(utf8);
1647
1648 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8, len) )
1649 return "";
1650
1651 return FromImpl(wxStringImpl(utf8, len));
1652 }
1653
1654 const char* utf8_str() const { return wx_str(); }
1655 const char* ToUTF8() const { return wx_str(); }
1656
1657 // this function exists in UTF-8 build only and returns the length of the
1658 // internal UTF-8 representation
1659 size_t utf8_length() const { return m_impl.length(); }
1660#elif wxUSE_UNICODE_WCHAR
1661 static wxString FromUTF8(const char *utf8, size_t len = npos)
1662 { return wxString(utf8, wxMBConvUTF8(), len); }
1663 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1664 {
1665 const wxString s(utf8, wxMBConvUTF8(), len);
1666 wxASSERT_MSG( !utf8 || !*utf8 || !s.empty(),
1667 "string must be valid UTF-8" );
1668 return s;
1669 }
1670 const wxCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); }
1671 const wxCharBuffer ToUTF8() const { return utf8_str(); }
1672#else // ANSI
1673 static wxString FromUTF8(const char *utf8)
1674 { return wxString(wxMBConvUTF8().cMB2WC(utf8)); }
1675 static wxString FromUTF8(const char *utf8, size_t len)
1676 {
1677 size_t wlen;
1678 wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen));
1679 return wxString(buf.data(), wlen);
1680 }
1681 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1682 {
1683 size_t wlen;
1684 wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8,
1685 len == npos ? wxNO_LEN : len,
1686 &wlen));
1687 wxASSERT_MSG( !utf8 || !*utf8 || wlen,
1688 "string must be valid UTF-8" );
1689
1690 return wxString(buf.data(), wlen);
1691 }
1692 const wxCharBuffer utf8_str() const
1693 { return wxMBConvUTF8().cWC2MB(wc_str()); }
1694 const wxCharBuffer ToUTF8() const { return utf8_str(); }
1695#endif
1696
1697 // functions for storing binary data in wxString:
1698#if wxUSE_UNICODE
1699 static wxString From8BitData(const char *data, size_t len)
1700 { return wxString(data, wxConvISO8859_1, len); }
1701 // version for NUL-terminated data:
1702 static wxString From8BitData(const char *data)
1703 { return wxString(data, wxConvISO8859_1); }
1704 const wxCharBuffer To8BitData() const { return mb_str(wxConvISO8859_1); }
1705#else // ANSI
1706 static wxString From8BitData(const char *data, size_t len)
1707 { return wxString(data, len); }
1708 // version for NUL-terminated data:
1709 static wxString From8BitData(const char *data)
1710 { return wxString(data); }
1711 const char *To8BitData() const { return c_str(); }
1712#endif // Unicode/ANSI
1713
1714 // conversions with (possible) format conversions: have to return a
1715 // buffer with temporary data
1716 //
1717 // the functions defined (in either Unicode or ANSI) mode are mb_str() to
1718 // return an ANSI (multibyte) string, wc_str() to return a wide string and
1719 // fn_str() to return a string which should be used with the OS APIs
1720 // accepting the file names. The return value is always the same, but the
1721 // type differs because a function may either return pointer to the buffer
1722 // directly or have to use intermediate buffer for translation.
1723#if wxUSE_UNICODE
1724
1725#if wxUSE_UTF8_LOCALE_ONLY
1726 const char* mb_str() const { return wx_str(); }
1727 const wxCharBuffer mb_str(const wxMBConv& conv) const;
1728#else
1729 const wxCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const;
1730#endif
1731
1732 const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
1733
1734#if wxUSE_UNICODE_WCHAR
1735 const wchar_t* wc_str() const { return wx_str(); }
1736#elif wxUSE_UNICODE_UTF8
1737 const wxWCharBuffer wc_str() const;
1738#endif
1739 // for compatibility with !wxUSE_UNICODE version
1740 const wxWX2WCbuf wc_str(const wxMBConv& WXUNUSED(conv)) const
1741 { return wc_str(); }
1742
1743#if wxMBFILES
1744 const wxCharBuffer fn_str() const { return mb_str(wxConvFile); }
1745#else // !wxMBFILES
1746 const wxWX2WCbuf fn_str() const { return wc_str(); }
1747#endif // wxMBFILES/!wxMBFILES
1748
1749#else // ANSI
1750 const wxChar* mb_str() const { return wx_str(); }
1751
1752 // for compatibility with wxUSE_UNICODE version
1753 const char* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); }
1754
1755 const wxWX2MBbuf mbc_str() const { return mb_str(); }
1756
1757#if wxUSE_WCHAR_T
1758 const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const;
1759#endif // wxUSE_WCHAR_T
1760 const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); }
1761#endif // Unicode/ANSI
1762
1763#if wxUSE_UNICODE_UTF8
1764 const wxWCharBuffer t_str() const { return wc_str(); }
1765#elif wxUSE_UNICODE_WCHAR
1766 const wchar_t* t_str() const { return wx_str(); }
1767#else
1768 const char* t_str() const { return wx_str(); }
1769#endif
1770
1771
1772 // overloaded assignment
1773 // from another wxString
1774 wxString& operator=(const wxString& stringSrc)
1775 {
1776 if ( this != &stringSrc )
1777 {
1778 wxSTRING_INVALIDATE_CACHE();
1779
1780 m_impl = stringSrc.m_impl;
1781 }
1782
1783 return *this;
1784 }
1785
1786 wxString& operator=(const wxCStrData& cstr)
1787 { return *this = cstr.AsString(); }
1788 // from a character
1789 wxString& operator=(wxUniChar ch)
1790 {
1791 wxSTRING_INVALIDATE_CACHE();
1792
1793#if wxUSE_UNICODE_UTF8
1794 if ( !ch.IsAscii() )
1795 m_impl = wxStringOperations::EncodeChar(ch);
1796 else
1797#endif // wxUSE_UNICODE_UTF8
1798 m_impl = (wxStringCharType)ch;
1799 return *this;
1800 }
1801
1802 wxString& operator=(wxUniCharRef ch)
1803 { return operator=((wxUniChar)ch); }
1804 wxString& operator=(char ch)
1805 { return operator=(wxUniChar(ch)); }
1806 wxString& operator=(unsigned char ch)
1807 { return operator=(wxUniChar(ch)); }
1808 wxString& operator=(wchar_t ch)
1809 { return operator=(wxUniChar(ch)); }
1810 // from a C string - STL probably will crash on NULL,
1811 // so we need to compensate in that case
1812#if wxUSE_STL_BASED_WXSTRING
1813 wxString& operator=(const char *psz)
1814 {
1815 wxSTRING_INVALIDATE_CACHE();
1816
1817 if ( psz )
1818 m_impl = ImplStr(psz);
1819 else
1820 clear();
1821
1822 return *this;
1823 }
1824
1825 wxString& operator=(const wchar_t *pwz)
1826 {
1827 wxSTRING_INVALIDATE_CACHE();
1828
1829 if ( pwz )
1830 m_impl = ImplStr(pwz);
1831 else
1832 clear();
1833
1834 return *this;
1835 }
1836#else // !wxUSE_STL_BASED_WXSTRING
1837 wxString& operator=(const char *psz)
1838 {
1839 wxSTRING_INVALIDATE_CACHE();
1840
1841 m_impl = ImplStr(psz);
1842
1843 return *this;
1844 }
1845
1846 wxString& operator=(const wchar_t *pwz)
1847 {
1848 wxSTRING_INVALIDATE_CACHE();
1849
1850 m_impl = ImplStr(pwz);
1851
1852 return *this;
1853 }
1854#endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING
1855
1856 wxString& operator=(const unsigned char *psz)
1857 { return operator=((const char*)psz); }
1858
1859 // from wxWCharBuffer
1860 wxString& operator=(const wxWCharBuffer& s)
1861 { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
1862 // from wxCharBuffer
1863 wxString& operator=(const wxCharBuffer& s)
1864 { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
1865
1866 // string concatenation
1867 // in place concatenation
1868 /*
1869 Concatenate and return the result. Note that the left to right
1870 associativity of << allows to write things like "str << str1 << str2
1871 << ..." (unlike with +=)
1872 */
1873 // string += string
1874 wxString& operator<<(const wxString& s)
1875 {
1876#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
1877 wxASSERT_MSG( s.IsValid(),
1878 _T("did you forget to call UngetWriteBuf()?") );
1879#endif
1880
1881 append(s);
1882 return *this;
1883 }
1884 // string += C string
1885 wxString& operator<<(const char *psz)
1886 { append(psz); return *this; }
1887 wxString& operator<<(const wchar_t *pwz)
1888 { append(pwz); return *this; }
1889 wxString& operator<<(const wxCStrData& psz)
1890 { append(psz.AsString()); return *this; }
1891 // string += char
1892 wxString& operator<<(wxUniChar ch) { append(1, ch); return *this; }
1893 wxString& operator<<(wxUniCharRef ch) { append(1, ch); return *this; }
1894 wxString& operator<<(char ch) { append(1, ch); return *this; }
1895 wxString& operator<<(unsigned char ch) { append(1, ch); return *this; }
1896 wxString& operator<<(wchar_t ch) { append(1, ch); return *this; }
1897
1898 // string += buffer (i.e. from wxGetString)
1899 wxString& operator<<(const wxWCharBuffer& s)
1900 { return operator<<((const wchar_t *)s); }
1901 wxString& operator<<(const wxCharBuffer& s)
1902 { return operator<<((const char *)s); }
1903
1904 // string += C string
1905 wxString& Append(const wxString& s)
1906 {
1907 // test for empty() to share the string if possible
1908 if ( empty() )
1909 *this = s;
1910 else
1911 append(s);
1912 return *this;
1913 }
1914 wxString& Append(const char* psz)
1915 { append(psz); return *this; }
1916 wxString& Append(const wchar_t* pwz)
1917 { append(pwz); return *this; }
1918 wxString& Append(const wxCStrData& psz)
1919 { append(psz); return *this; }
1920 wxString& Append(const wxCharBuffer& psz)
1921 { append(psz); return *this; }
1922 wxString& Append(const wxWCharBuffer& psz)
1923 { append(psz); return *this; }
1924 wxString& Append(const char* psz, size_t nLen)
1925 { append(psz, nLen); return *this; }
1926 wxString& Append(const wchar_t* pwz, size_t nLen)
1927 { append(pwz, nLen); return *this; }
1928 wxString& Append(const wxCStrData& psz, size_t nLen)
1929 { append(psz, nLen); return *this; }
1930 wxString& Append(const wxCharBuffer& psz, size_t nLen)
1931 { append(psz, nLen); return *this; }
1932 wxString& Append(const wxWCharBuffer& psz, size_t nLen)
1933 { append(psz, nLen); return *this; }
1934 // append count copies of given character
1935 wxString& Append(wxUniChar ch, size_t count = 1u)
1936 { append(count, ch); return *this; }
1937 wxString& Append(wxUniCharRef ch, size_t count = 1u)
1938 { append(count, ch); return *this; }
1939 wxString& Append(char ch, size_t count = 1u)
1940 { append(count, ch); return *this; }
1941 wxString& Append(unsigned char ch, size_t count = 1u)
1942 { append(count, ch); return *this; }
1943 wxString& Append(wchar_t ch, size_t count = 1u)
1944 { append(count, ch); return *this; }
1945
1946 // prepend a string, return the string itself
1947 wxString& Prepend(const wxString& str)
1948 { *this = str + *this; return *this; }
1949
1950 // non-destructive concatenation
1951 // two strings
1952 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string1,
1953 const wxString& string2);
1954 // string with a single char
1955 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
1956 // char with a string
1957 friend wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
1958 // string with C string
1959 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
1960 const char *psz);
1961 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
1962 const wchar_t *pwz);
1963 // C string with string
1964 friend wxString WXDLLIMPEXP_BASE operator+(const char *psz,
1965 const wxString& string);
1966 friend wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz,
1967 const wxString& string);
1968
1969 // stream-like functions
1970 // insert an int into string
1971 wxString& operator<<(int i)
1972 { return (*this) << Format(_T("%d"), i); }
1973 // insert an unsigned int into string
1974 wxString& operator<<(unsigned int ui)
1975 { return (*this) << Format(_T("%u"), ui); }
1976 // insert a long into string
1977 wxString& operator<<(long l)
1978 { return (*this) << Format(_T("%ld"), l); }
1979 // insert an unsigned long into string
1980 wxString& operator<<(unsigned long ul)
1981 { return (*this) << Format(_T("%lu"), ul); }
1982#if defined wxLongLong_t && !defined wxLongLongIsLong
1983 // insert a long long if they exist and aren't longs
1984 wxString& operator<<(wxLongLong_t ll)
1985 {
1986 const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("d");
1987 return (*this) << Format(fmt, ll);
1988 }
1989 // insert an unsigned long long
1990 wxString& operator<<(wxULongLong_t ull)
1991 {
1992 const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("u");
1993 return (*this) << Format(fmt , ull);
1994 }
1995#endif // wxLongLong_t && !wxLongLongIsLong
1996 // insert a float into string
1997 wxString& operator<<(float f)
1998 { return (*this) << Format(_T("%f"), f); }
1999 // insert a double into string
2000 wxString& operator<<(double d)
2001 { return (*this) << Format(_T("%g"), d); }
2002
2003 // string comparison
2004 // case-sensitive comparison (returns a value < 0, = 0 or > 0)
2005 int Cmp(const char *psz) const
2006 { return compare(psz); }
2007 int Cmp(const wchar_t *pwz) const
2008 { return compare(pwz); }
2009 int Cmp(const wxString& s) const
2010 { return compare(s); }
2011 int Cmp(const wxCStrData& s) const
2012 { return compare(s); }
2013 int Cmp(const wxCharBuffer& s) const
2014 { return compare(s); }
2015 int Cmp(const wxWCharBuffer& s) const
2016 { return compare(s); }
2017 // same as Cmp() but not case-sensitive
2018 int CmpNoCase(const wxString& s) const;
2019
2020 // test for the string equality, either considering case or not
2021 // (if compareWithCase then the case matters)
2022 bool IsSameAs(const wxString& str, bool compareWithCase = true) const
2023 {
2024#if !wxUSE_UNICODE_UTF8
2025 // in UTF-8 build, length() is O(n) and doing this would be _slower_
2026 if ( length() != str.length() )
2027 return false;
2028#endif
2029 return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0;
2030 }
2031 bool IsSameAs(const char *str, bool compareWithCase = true) const
2032 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2033 bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const
2034 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2035
2036 bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const
2037 { return IsSameAs(str.AsString(), compareWithCase); }
2038 bool IsSameAs(const wxCharBuffer& str, bool compareWithCase = true) const
2039 { return IsSameAs(str.data(), compareWithCase); }
2040 bool IsSameAs(const wxWCharBuffer& str, bool compareWithCase = true) const
2041 { return IsSameAs(str.data(), compareWithCase); }
2042 // comparison with a single character: returns true if equal
2043 bool IsSameAs(wxUniChar c, bool compareWithCase = true) const;
2044 // FIXME-UTF8: remove these overloads
2045 bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const
2046 { return IsSameAs(wxUniChar(c), compareWithCase); }
2047 bool IsSameAs(char c, bool compareWithCase = true) const
2048 { return IsSameAs(wxUniChar(c), compareWithCase); }
2049 bool IsSameAs(unsigned char c, bool compareWithCase = true) const
2050 { return IsSameAs(wxUniChar(c), compareWithCase); }
2051 bool IsSameAs(wchar_t c, bool compareWithCase = true) const
2052 { return IsSameAs(wxUniChar(c), compareWithCase); }
2053 bool IsSameAs(int c, bool compareWithCase = true) const
2054 { return IsSameAs(wxUniChar(c), compareWithCase); }
2055
2056 // simple sub-string extraction
2057 // return substring starting at nFirst of length nCount (or till the end
2058 // if nCount = default value)
2059 wxString Mid(size_t nFirst, size_t nCount = npos) const;
2060
2061 // operator version of Mid()
2062 wxString operator()(size_t start, size_t len) const
2063 { return Mid(start, len); }
2064
2065 // check if the string starts with the given prefix and return the rest
2066 // of the string in the provided pointer if it is not NULL; otherwise
2067 // return false
2068 bool StartsWith(const wxString& prefix, wxString *rest = NULL) const;
2069 // check if the string ends with the given suffix and return the
2070 // beginning of the string before the suffix in the provided pointer if
2071 // it is not NULL; otherwise return false
2072 bool EndsWith(const wxString& suffix, wxString *rest = NULL) const;
2073
2074 // get first nCount characters
2075 wxString Left(size_t nCount) const;
2076 // get last nCount characters
2077 wxString Right(size_t nCount) const;
2078 // get all characters before the first occurance of ch
2079 // (returns the whole string if ch not found)
2080 wxString BeforeFirst(wxUniChar ch) const;
2081 // get all characters before the last occurence of ch
2082 // (returns empty string if ch not found)
2083 wxString BeforeLast(wxUniChar ch) const;
2084 // get all characters after the first occurence of ch
2085 // (returns empty string if ch not found)
2086 wxString AfterFirst(wxUniChar ch) const;
2087 // get all characters after the last occurence of ch
2088 // (returns the whole string if ch not found)
2089 wxString AfterLast(wxUniChar ch) const;
2090
2091 // for compatibility only, use more explicitly named functions above
2092 wxString Before(wxUniChar ch) const { return BeforeLast(ch); }
2093 wxString After(wxUniChar ch) const { return AfterFirst(ch); }
2094
2095 // case conversion
2096 // convert to upper case in place, return the string itself
2097 wxString& MakeUpper();
2098 // convert to upper case, return the copy of the string
2099 wxString Upper() const { return wxString(*this).MakeUpper(); }
2100 // convert to lower case in place, return the string itself
2101 wxString& MakeLower();
2102 // convert to lower case, return the copy of the string
2103 wxString Lower() const { return wxString(*this).MakeLower(); }
2104 // convert the first character to the upper case and the rest to the
2105 // lower one, return the modified string itself
2106 wxString& MakeCapitalized();
2107 // convert the first character to the upper case and the rest to the
2108 // lower one, return the copy of the string
2109 wxString Capitalize() const { return wxString(*this).MakeCapitalized(); }
2110
2111 // trimming/padding whitespace (either side) and truncating
2112 // remove spaces from left or from right (default) side
2113 wxString& Trim(bool bFromRight = true);
2114 // add nCount copies chPad in the beginning or at the end (default)
2115 wxString& Pad(size_t nCount, wxUniChar chPad = wxT(' '), bool bFromRight = true);
2116
2117 // searching and replacing
2118 // searching (return starting index, or -1 if not found)
2119 int Find(wxUniChar ch, bool bFromEnd = false) const; // like strchr/strrchr
2120 int Find(wxUniCharRef ch, bool bFromEnd = false) const
2121 { return Find(wxUniChar(ch), bFromEnd); }
2122 int Find(char ch, bool bFromEnd = false) const
2123 { return Find(wxUniChar(ch), bFromEnd); }
2124 int Find(unsigned char ch, bool bFromEnd = false) const
2125 { return Find(wxUniChar(ch), bFromEnd); }
2126 int Find(wchar_t ch, bool bFromEnd = false) const
2127 { return Find(wxUniChar(ch), bFromEnd); }
2128 // searching (return starting index, or -1 if not found)
2129 int Find(const wxString& sub) const // like strstr
2130 {
2131 size_type idx = find(sub);
2132 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2133 }
2134 int Find(const char *sub) const // like strstr
2135 {
2136 size_type idx = find(sub);
2137 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2138 }
2139 int Find(const wchar_t *sub) const // like strstr
2140 {
2141 size_type idx = find(sub);
2142 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2143 }
2144
2145 int Find(const wxCStrData& sub) const
2146 { return Find(sub.AsString()); }
2147 int Find(const wxCharBuffer& sub) const
2148 { return Find(sub.data()); }
2149 int Find(const wxWCharBuffer& sub) const
2150 { return Find(sub.data()); }
2151
2152 // replace first (or all of bReplaceAll) occurences of substring with
2153 // another string, returns the number of replacements made
2154 size_t Replace(const wxString& strOld,
2155 const wxString& strNew,
2156 bool bReplaceAll = true);
2157
2158 // check if the string contents matches a mask containing '*' and '?'
2159 bool Matches(const wxString& mask) const;
2160
2161 // conversion to numbers: all functions return true only if the whole
2162 // string is a number and put the value of this number into the pointer
2163 // provided, the base is the numeric base in which the conversion should be
2164 // done and must be comprised between 2 and 36 or be 0 in which case the
2165 // standard C rules apply (leading '0' => octal, "0x" => hex)
2166 // convert to a signed integer
2167 bool ToLong(long *val, int base = 10) const;
2168 // convert to an unsigned integer
2169 bool ToULong(unsigned long *val, int base = 10) const;
2170 // convert to wxLongLong
2171#if defined(wxLongLong_t)
2172 bool ToLongLong(wxLongLong_t *val, int base = 10) const;
2173 // convert to wxULongLong
2174 bool ToULongLong(wxULongLong_t *val, int base = 10) const;
2175#endif // wxLongLong_t
2176 // convert to a double
2177 bool ToDouble(double *val) const;
2178
2179
2180#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2181 // formatted input/output
2182 // as sprintf(), returns the number of characters written or < 0 on error
2183 // (take 'this' into account in attribute parameter count)
2184 // int Printf(const wxString& format, ...);
2185 WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),
2186 DoPrintfWchar, DoPrintfUtf8)
2187#ifdef __WATCOMC__
2188 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2189 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxString&),
2190 (wxFormatString(f1)));
2191 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxCStrData&),
2192 (wxFormatString(f1)));
2193 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const char*),
2194 (wxFormatString(f1)));
2195 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wchar_t*),
2196 (wxFormatString(f1)));
2197#endif
2198#endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN
2199 // as vprintf(), returns the number of characters written or < 0 on error
2200 int PrintfV(const wxString& format, va_list argptr);
2201
2202#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2203 // returns the string containing the result of Printf() to it
2204 // static wxString Format(const wxString& format, ...) ATTRIBUTE_PRINTF_1;
2205 WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxFormatString&),
2206 DoFormatWchar, DoFormatUtf8)
2207#ifdef __WATCOMC__
2208 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2209 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxString&),
2210 (wxFormatString(f1)));
2211 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxCStrData&),
2212 (wxFormatString(f1)));
2213 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const char*),
2214 (wxFormatString(f1)));
2215 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wchar_t*),
2216 (wxFormatString(f1)));
2217#endif
2218#endif
2219 // the same as above, but takes a va_list
2220 static wxString FormatV(const wxString& format, va_list argptr);
2221
2222 // raw access to string memory
2223 // ensure that string has space for at least nLen characters
2224 // only works if the data of this string is not shared
2225 bool Alloc(size_t nLen) { reserve(nLen); return capacity() >= nLen; }
2226 // minimize the string's memory
2227 // only works if the data of this string is not shared
2228 bool Shrink();
2229#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
2230 // These are deprecated, use wxStringBuffer or wxStringBufferLength instead
2231 //
2232 // get writable buffer of at least nLen bytes. Unget() *must* be called
2233 // a.s.a.p. to put string back in a reasonable state!
2234 wxDEPRECATED( wxStringCharType *GetWriteBuf(size_t nLen) );
2235 // call this immediately after GetWriteBuf() has been used
2236 wxDEPRECATED( void UngetWriteBuf() );
2237 wxDEPRECATED( void UngetWriteBuf(size_t nLen) );
2238#endif // WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && wxUSE_UNICODE_UTF8
2239
2240 // wxWidgets version 1 compatibility functions
2241
2242 // use Mid()
2243 wxString SubString(size_t from, size_t to) const
2244 { return Mid(from, (to - from + 1)); }
2245 // values for second parameter of CompareTo function
2246 enum caseCompare {exact, ignoreCase};
2247 // values for first parameter of Strip function
2248 enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
2249
2250#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
2251 // use Printf()
2252 // (take 'this' into account in attribute parameter count)
2253 // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;
2254 WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),
2255 DoPrintfWchar, DoPrintfUtf8)
2256#ifdef __WATCOMC__
2257 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2258 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxString&),
2259 (wxFormatString(f1)));
2260 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxCStrData&),
2261 (wxFormatString(f1)));
2262 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const char*),
2263 (wxFormatString(f1)));
2264 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wchar_t*),
2265 (wxFormatString(f1)));
2266#endif
2267#endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
2268
2269 // use Cmp()
2270 int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
2271 { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
2272
2273 // use length()
2274 size_t Length() const { return length(); }
2275 // Count the number of characters
2276 int Freq(wxUniChar ch) const;
2277 // use MakeLower
2278 void LowerCase() { MakeLower(); }
2279 // use MakeUpper
2280 void UpperCase() { MakeUpper(); }
2281 // use Trim except that it doesn't change this string
2282 wxString Strip(stripType w = trailing) const;
2283
2284 // use Find (more general variants not yet supported)
2285 size_t Index(const wxChar* psz) const { return Find(psz); }
2286 size_t Index(wxUniChar ch) const { return Find(ch); }
2287 // use Truncate
2288 wxString& Remove(size_t pos) { return Truncate(pos); }
2289 wxString& RemoveLast(size_t n = 1) { return Truncate(length() - n); }
2290
2291 wxString& Remove(size_t nStart, size_t nLen)
2292 { return (wxString&)erase( nStart, nLen ); }
2293
2294 // use Find()
2295 int First( wxUniChar ch ) const { return Find(ch); }
2296 int First( wxUniCharRef ch ) const { return Find(ch); }
2297 int First( char ch ) const { return Find(ch); }
2298 int First( unsigned char ch ) const { return Find(ch); }
2299 int First( wchar_t ch ) const { return Find(ch); }
2300 int First( const wxString& str ) const { return Find(str); }
2301 int Last( wxUniChar ch ) const { return Find(ch, true); }
2302 bool Contains(const wxString& str) const { return Find(str) != wxNOT_FOUND; }
2303
2304 // use empty()
2305 bool IsNull() const { return empty(); }
2306
2307 // std::string compatibility functions
2308
2309 // take nLen chars starting at nPos
2310 wxString(const wxString& str, size_t nPos, size_t nLen)
2311 { assign(str, nPos, nLen); }
2312 // take all characters from first to last
2313 wxString(const_iterator first, const_iterator last)
2314 : m_impl(first.impl(), last.impl()) { }
2315#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2316 // the 2 overloads below are for compatibility with the existing code using
2317 // pointers instead of iterators
2318 wxString(const char *first, const char *last)
2319 {
2320 SubstrBufFromMB str(ImplStr(first, last - first));
2321 m_impl.assign(str.data, str.len);
2322 }
2323 wxString(const wchar_t *first, const wchar_t *last)
2324 {
2325 SubstrBufFromWC str(ImplStr(first, last - first));
2326 m_impl.assign(str.data, str.len);
2327 }
2328 // and this one is needed to compile code adding offsets to c_str() result
2329 wxString(const wxCStrData& first, const wxCStrData& last)
2330 : m_impl(CreateConstIterator(first).impl(),
2331 CreateConstIterator(last).impl())
2332 {
2333 wxASSERT_MSG( first.m_str == last.m_str,
2334 _T("pointers must be into the same string") );
2335 }
2336#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2337
2338 // lib.string.modifiers
2339 // append elements str[pos], ..., str[pos+n]
2340 wxString& append(const wxString& str, size_t pos, size_t n)
2341 {
2342 wxSTRING_UPDATE_CACHED_LENGTH(n);
2343
2344 size_t from, len;
2345 str.PosLenToImpl(pos, n, &from, &len);
2346 m_impl.append(str.m_impl, from, len);
2347 return *this;
2348 }
2349 // append a string
2350 wxString& append(const wxString& str)
2351 {
2352 wxSTRING_UPDATE_CACHED_LENGTH(str.length());
2353
2354 m_impl.append(str.m_impl);
2355 return *this;
2356 }
2357
2358 // append first n (or all if n == npos) characters of sz
2359 wxString& append(const char *sz)
2360 {
2361 wxSTRING_INVALIDATE_CACHED_LENGTH();
2362
2363 m_impl.append(ImplStr(sz));
2364 return *this;
2365 }
2366
2367 wxString& append(const wchar_t *sz)
2368 {
2369 wxSTRING_INVALIDATE_CACHED_LENGTH();
2370
2371 m_impl.append(ImplStr(sz));
2372 return *this;
2373 }
2374
2375 wxString& append(const char *sz, size_t n)
2376 {
2377 wxSTRING_INVALIDATE_CACHED_LENGTH();
2378
2379 SubstrBufFromMB str(ImplStr(sz, n));
2380 m_impl.append(str.data, str.len);
2381 return *this;
2382 }
2383 wxString& append(const wchar_t *sz, size_t n)
2384 {
2385 wxSTRING_UPDATE_CACHED_LENGTH(n);
2386
2387 SubstrBufFromWC str(ImplStr(sz, n));
2388 m_impl.append(str.data, str.len);
2389 return *this;
2390 }
2391
2392 wxString& append(const wxCStrData& str)
2393 { return append(str.AsString()); }
2394 wxString& append(const wxCharBuffer& str)
2395 { return append(str.data()); }
2396 wxString& append(const wxWCharBuffer& str)
2397 { return append(str.data()); }
2398 wxString& append(const wxCStrData& str, size_t n)
2399 { return append(str.AsString(), 0, n); }
2400 wxString& append(const wxCharBuffer& str, size_t n)
2401 { return append(str.data(), n); }
2402 wxString& append(const wxWCharBuffer& str, size_t n)
2403 { return append(str.data(), n); }
2404
2405 // append n copies of ch
2406 wxString& append(size_t n, wxUniChar ch)
2407 {
2408#if wxUSE_UNICODE_UTF8
2409 if ( !ch.IsAscii() )
2410 {
2411 wxSTRING_INVALIDATE_CACHED_LENGTH();
2412
2413 m_impl.append(wxStringOperations::EncodeNChars(n, ch));
2414 }
2415 else // ASCII
2416#endif
2417 {
2418 wxSTRING_UPDATE_CACHED_LENGTH(n);
2419
2420 m_impl.append(n, (wxStringCharType)ch);
2421 }
2422
2423 return *this;
2424 }
2425
2426 wxString& append(size_t n, wxUniCharRef ch)
2427 { return append(n, wxUniChar(ch)); }
2428 wxString& append(size_t n, char ch)
2429 { return append(n, wxUniChar(ch)); }
2430 wxString& append(size_t n, unsigned char ch)
2431 { return append(n, wxUniChar(ch)); }
2432 wxString& append(size_t n, wchar_t ch)
2433 { return append(n, wxUniChar(ch)); }
2434
2435 // append from first to last
2436 wxString& append(const_iterator first, const_iterator last)
2437 {
2438 wxSTRING_INVALIDATE_CACHED_LENGTH();
2439
2440 m_impl.append(first.impl(), last.impl());
2441 return *this;
2442 }
2443#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2444 wxString& append(const char *first, const char *last)
2445 { return append(first, last - first); }
2446 wxString& append(const wchar_t *first, const wchar_t *last)
2447 { return append(first, last - first); }
2448 wxString& append(const wxCStrData& first, const wxCStrData& last)
2449 { return append(CreateConstIterator(first), CreateConstIterator(last)); }
2450#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2451
2452 // same as `this_string = str'
2453 wxString& assign(const wxString& str)
2454 {
2455 wxSTRING_SET_CACHED_LENGTH(str.length());
2456
2457 m_impl = str.m_impl;
2458
2459 return *this;
2460 }
2461
2462 wxString& assign(const wxString& str, size_t len)
2463 {
2464 wxSTRING_SET_CACHED_LENGTH(len);
2465
2466 m_impl.assign(str.m_impl, 0, str.LenToImpl(len));
2467
2468 return *this;
2469 }
2470
2471 // same as ` = str[pos..pos + n]
2472 wxString& assign(const wxString& str, size_t pos, size_t n)
2473 {
2474 size_t from, len;
2475 str.PosLenToImpl(pos, n, &from, &len);
2476 m_impl.assign(str.m_impl, from, len);
2477
2478 // it's important to call this after PosLenToImpl() above in case str is
2479 // the same string as this one
2480 wxSTRING_SET_CACHED_LENGTH(n);
2481
2482 return *this;
2483 }
2484
2485 // same as `= first n (or all if n == npos) characters of sz'
2486 wxString& assign(const char *sz)
2487 {
2488 wxSTRING_INVALIDATE_CACHE();
2489
2490 m_impl.assign(ImplStr(sz));
2491
2492 return *this;
2493 }
2494
2495 wxString& assign(const wchar_t *sz)
2496 {
2497 wxSTRING_INVALIDATE_CACHE();
2498
2499 m_impl.assign(ImplStr(sz));
2500
2501 return *this;
2502 }
2503
2504 wxString& assign(const char *sz, size_t n)
2505 {
2506 wxSTRING_SET_CACHED_LENGTH(n);
2507
2508 SubstrBufFromMB str(ImplStr(sz, n));
2509 m_impl.assign(str.data, str.len);
2510
2511 return *this;
2512 }
2513
2514 wxString& assign(const wchar_t *sz, size_t n)
2515 {
2516 wxSTRING_SET_CACHED_LENGTH(n);
2517
2518 SubstrBufFromWC str(ImplStr(sz, n));
2519 m_impl.assign(str.data, str.len);
2520
2521 return *this;
2522 }
2523
2524 wxString& assign(const wxCStrData& str)
2525 { return assign(str.AsString()); }
2526 wxString& assign(const wxCharBuffer& str)
2527 { return assign(str.data()); }
2528 wxString& assign(const wxWCharBuffer& str)
2529 { return assign(str.data()); }
2530 wxString& assign(const wxCStrData& str, size_t len)
2531 { return assign(str.AsString(), len); }
2532 wxString& assign(const wxCharBuffer& str, size_t len)
2533 { return assign(str.data(), len); }
2534 wxString& assign(const wxWCharBuffer& str, size_t len)
2535 { return assign(str.data(), len); }
2536
2537 // same as `= n copies of ch'
2538 wxString& assign(size_t n, wxUniChar ch)
2539 {
2540 wxSTRING_SET_CACHED_LENGTH(n);
2541
2542#if wxUSE_UNICODE_UTF8
2543 if ( !ch.IsAscii() )
2544 m_impl.assign(wxStringOperations::EncodeNChars(n, ch));
2545 else
2546#endif
2547 m_impl.assign(n, (wxStringCharType)ch);
2548
2549 return *this;
2550 }
2551
2552 wxString& assign(size_t n, wxUniCharRef ch)
2553 { return assign(n, wxUniChar(ch)); }
2554 wxString& assign(size_t n, char ch)
2555 { return assign(n, wxUniChar(ch)); }
2556 wxString& assign(size_t n, unsigned char ch)
2557 { return assign(n, wxUniChar(ch)); }
2558 wxString& assign(size_t n, wchar_t ch)
2559 { return assign(n, wxUniChar(ch)); }
2560
2561 // assign from first to last
2562 wxString& assign(const_iterator first, const_iterator last)
2563 {
2564 wxSTRING_INVALIDATE_CACHE();
2565
2566 m_impl.assign(first.impl(), last.impl());
2567
2568 return *this;
2569 }
2570#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2571 wxString& assign(const char *first, const char *last)
2572 { return assign(first, last - first); }
2573 wxString& assign(const wchar_t *first, const wchar_t *last)
2574 { return assign(first, last - first); }
2575 wxString& assign(const wxCStrData& first, const wxCStrData& last)
2576 { return assign(CreateConstIterator(first), CreateConstIterator(last)); }
2577#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2578
2579 // string comparison
2580 int compare(const wxString& str) const;
2581 int compare(const char* sz) const;
2582 int compare(const wchar_t* sz) const;
2583 int compare(const wxCStrData& str) const
2584 { return compare(str.AsString()); }
2585 int compare(const wxCharBuffer& str) const
2586 { return compare(str.data()); }
2587 int compare(const wxWCharBuffer& str) const
2588 { return compare(str.data()); }
2589 // comparison with a substring
2590 int compare(size_t nStart, size_t nLen, const wxString& str) const;
2591 // comparison of 2 substrings
2592 int compare(size_t nStart, size_t nLen,
2593 const wxString& str, size_t nStart2, size_t nLen2) const;
2594 // substring comparison with first nCount characters of sz
2595 int compare(size_t nStart, size_t nLen,
2596 const char* sz, size_t nCount = npos) const;
2597 int compare(size_t nStart, size_t nLen,
2598 const wchar_t* sz, size_t nCount = npos) const;
2599
2600 // insert another string
2601 wxString& insert(size_t nPos, const wxString& str)
2602 { insert(GetIterForNthChar(nPos), str.begin(), str.end()); return *this; }
2603 // insert n chars of str starting at nStart (in str)
2604 wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
2605 {
2606 wxSTRING_UPDATE_CACHED_LENGTH(n);
2607
2608 size_t from, len;
2609 str.PosLenToImpl(nStart, n, &from, &len);
2610 m_impl.insert(PosToImpl(nPos), str.m_impl, from, len);
2611
2612 return *this;
2613 }
2614
2615 // insert first n (or all if n == npos) characters of sz
2616 wxString& insert(size_t nPos, const char *sz)
2617 {
2618 wxSTRING_INVALIDATE_CACHE();
2619
2620 m_impl.insert(PosToImpl(nPos), ImplStr(sz));
2621
2622 return *this;
2623 }
2624
2625 wxString& insert(size_t nPos, const wchar_t *sz)
2626 {
2627 wxSTRING_INVALIDATE_CACHE();
2628
2629 m_impl.insert(PosToImpl(nPos), ImplStr(sz)); return *this;
2630 }
2631
2632 wxString& insert(size_t nPos, const char *sz, size_t n)
2633 {
2634 wxSTRING_UPDATE_CACHED_LENGTH(n);
2635
2636 SubstrBufFromMB str(ImplStr(sz, n));
2637 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2638
2639 return *this;
2640 }
2641
2642 wxString& insert(size_t nPos, const wchar_t *sz, size_t n)
2643 {
2644 wxSTRING_UPDATE_CACHED_LENGTH(n);
2645
2646 SubstrBufFromWC str(ImplStr(sz, n));
2647 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2648
2649 return *this;
2650 }
2651
2652 // insert n copies of ch
2653 wxString& insert(size_t nPos, size_t n, wxUniChar ch)
2654 {
2655 wxSTRING_UPDATE_CACHED_LENGTH(n);
2656
2657#if wxUSE_UNICODE_UTF8
2658 if ( !ch.IsAscii() )
2659 m_impl.insert(PosToImpl(nPos), wxStringOperations::EncodeNChars(n, ch));
2660 else
2661#endif
2662 m_impl.insert(PosToImpl(nPos), n, (wxStringCharType)ch);
2663 return *this;
2664 }
2665
2666 iterator insert(iterator it, wxUniChar ch)
2667 {
2668 wxSTRING_UPDATE_CACHED_LENGTH(1);
2669
2670#if wxUSE_UNICODE_UTF8
2671 if ( !ch.IsAscii() )
2672 {
2673 size_t pos = IterToImplPos(it);
2674 m_impl.insert(pos, wxStringOperations::EncodeChar(ch));
2675 return iterator(this, m_impl.begin() + pos);
2676 }
2677 else
2678#endif
2679 return iterator(this, m_impl.insert(it.impl(), (wxStringCharType)ch));
2680 }
2681
2682 void insert(iterator it, const_iterator first, const_iterator last)
2683 {
2684 wxSTRING_INVALIDATE_CACHE();
2685
2686 m_impl.insert(it.impl(), first.impl(), last.impl());
2687 }
2688
2689#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2690 void insert(iterator it, const char *first, const char *last)
2691 { insert(it - begin(), first, last - first); }
2692 void insert(iterator it, const wchar_t *first, const wchar_t *last)
2693 { insert(it - begin(), first, last - first); }
2694 void insert(iterator it, const wxCStrData& first, const wxCStrData& last)
2695 { insert(it, CreateConstIterator(first), CreateConstIterator(last)); }
2696#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2697
2698 void insert(iterator it, size_type n, wxUniChar ch)
2699 {
2700 wxSTRING_UPDATE_CACHED_LENGTH(n);
2701
2702#if wxUSE_UNICODE_UTF8
2703 if ( !ch.IsAscii() )
2704 m_impl.insert(IterToImplPos(it), wxStringOperations::EncodeNChars(n, ch));
2705 else
2706#endif
2707 m_impl.insert(it.impl(), n, (wxStringCharType)ch);
2708 }
2709
2710 // delete characters from nStart to nStart + nLen
2711 wxString& erase(size_type pos = 0, size_type n = npos)
2712 {
2713 wxSTRING_INVALIDATE_CACHE();
2714
2715 size_t from, len;
2716 PosLenToImpl(pos, n, &from, &len);
2717 m_impl.erase(from, len);
2718
2719 return *this;
2720 }
2721
2722 // delete characters from first up to last
2723 iterator erase(iterator first, iterator last)
2724 {
2725 wxSTRING_INVALIDATE_CACHE();
2726
2727 return iterator(this, m_impl.erase(first.impl(), last.impl()));
2728 }
2729
2730 iterator erase(iterator first)
2731 {
2732 wxSTRING_UPDATE_CACHED_LENGTH(-1);
2733
2734 return iterator(this, m_impl.erase(first.impl()));
2735 }
2736
2737#ifdef wxSTRING_BASE_HASNT_CLEAR
2738 void clear() { erase(); }
2739#else
2740 void clear()
2741 {
2742 wxSTRING_SET_CACHED_LENGTH(0);
2743
2744 m_impl.clear();
2745 }
2746#endif
2747
2748 // replaces the substring of length nLen starting at nStart
2749 wxString& replace(size_t nStart, size_t nLen, const char* sz)
2750 {
2751 wxSTRING_INVALIDATE_CACHE();
2752
2753 size_t from, len;
2754 PosLenToImpl(nStart, nLen, &from, &len);
2755 m_impl.replace(from, len, ImplStr(sz));
2756
2757 return *this;
2758 }
2759
2760 wxString& replace(size_t nStart, size_t nLen, const wchar_t* sz)
2761 {
2762 wxSTRING_INVALIDATE_CACHE();
2763
2764 size_t from, len;
2765 PosLenToImpl(nStart, nLen, &from, &len);
2766 m_impl.replace(from, len, ImplStr(sz));
2767
2768 return *this;
2769 }
2770
2771 // replaces the substring of length nLen starting at nStart
2772 wxString& replace(size_t nStart, size_t nLen, const wxString& str)
2773 {
2774 wxSTRING_INVALIDATE_CACHE();
2775
2776 size_t from, len;
2777 PosLenToImpl(nStart, nLen, &from, &len);
2778 m_impl.replace(from, len, str.m_impl);
2779
2780 return *this;
2781 }
2782
2783 // replaces the substring with nCount copies of ch
2784 wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxUniChar ch)
2785 {
2786 wxSTRING_INVALIDATE_CACHE();
2787
2788 size_t from, len;
2789 PosLenToImpl(nStart, nLen, &from, &len);
2790#if wxUSE_UNICODE_UTF8
2791 if ( !ch.IsAscii() )
2792 m_impl.replace(from, len, wxStringOperations::EncodeNChars(nCount, ch));
2793 else
2794#endif
2795 m_impl.replace(from, len, nCount, (wxStringCharType)ch);
2796
2797 return *this;
2798 }
2799
2800 // replaces a substring with another substring
2801 wxString& replace(size_t nStart, size_t nLen,
2802 const wxString& str, size_t nStart2, size_t nLen2)
2803 {
2804 wxSTRING_INVALIDATE_CACHE();
2805
2806 size_t from, len;
2807 PosLenToImpl(nStart, nLen, &from, &len);
2808
2809 size_t from2, len2;
2810 str.PosLenToImpl(nStart2, nLen2, &from2, &len2);
2811
2812 m_impl.replace(from, len, str.m_impl, from2, len2);
2813
2814 return *this;
2815 }
2816
2817 // replaces the substring with first nCount chars of sz
2818 wxString& replace(size_t nStart, size_t nLen,
2819 const char* sz, size_t nCount)
2820 {
2821 wxSTRING_INVALIDATE_CACHE();
2822
2823 size_t from, len;
2824 PosLenToImpl(nStart, nLen, &from, &len);
2825
2826 SubstrBufFromMB str(ImplStr(sz, nCount));
2827
2828 m_impl.replace(from, len, str.data, str.len);
2829
2830 return *this;
2831 }
2832
2833 wxString& replace(size_t nStart, size_t nLen,
2834 const wchar_t* sz, size_t nCount)
2835 {
2836 wxSTRING_INVALIDATE_CACHE();
2837
2838 size_t from, len;
2839 PosLenToImpl(nStart, nLen, &from, &len);
2840
2841 SubstrBufFromWC str(ImplStr(sz, nCount));
2842
2843 m_impl.replace(from, len, str.data, str.len);
2844
2845 return *this;
2846 }
2847
2848 wxString& replace(size_t nStart, size_t nLen,
2849 const wxString& s, size_t nCount)
2850 {
2851 wxSTRING_INVALIDATE_CACHE();
2852
2853 size_t from, len;
2854 PosLenToImpl(nStart, nLen, &from, &len);
2855 m_impl.replace(from, len, s.m_impl.c_str(), s.LenToImpl(nCount));
2856
2857 return *this;
2858 }
2859
2860 wxString& replace(iterator first, iterator last, const char* s)
2861 {
2862 wxSTRING_INVALIDATE_CACHE();
2863
2864 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2865
2866 return *this;
2867 }
2868
2869 wxString& replace(iterator first, iterator last, const wchar_t* s)
2870 {
2871 wxSTRING_INVALIDATE_CACHE();
2872
2873 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2874
2875 return *this;
2876 }
2877
2878 wxString& replace(iterator first, iterator last, const char* s, size_type n)
2879 {
2880 wxSTRING_INVALIDATE_CACHE();
2881
2882 SubstrBufFromMB str(ImplStr(s, n));
2883 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2884
2885 return *this;
2886 }
2887
2888 wxString& replace(iterator first, iterator last, const wchar_t* s, size_type n)
2889 {
2890 wxSTRING_INVALIDATE_CACHE();
2891
2892 SubstrBufFromWC str(ImplStr(s, n));
2893 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2894
2895 return *this;
2896 }
2897
2898 wxString& replace(iterator first, iterator last, const wxString& s)
2899 {
2900 wxSTRING_INVALIDATE_CACHE();
2901
2902 m_impl.replace(first.impl(), last.impl(), s.m_impl);
2903
2904 return *this;
2905 }
2906
2907 wxString& replace(iterator first, iterator last, size_type n, wxUniChar ch)
2908 {
2909 wxSTRING_INVALIDATE_CACHE();
2910
2911#if wxUSE_UNICODE_UTF8
2912 if ( !ch.IsAscii() )
2913 m_impl.replace(first.impl(), last.impl(),
2914 wxStringOperations::EncodeNChars(n, ch));
2915 else
2916#endif
2917 m_impl.replace(first.impl(), last.impl(), n, (wxStringCharType)ch);
2918
2919 return *this;
2920 }
2921
2922 wxString& replace(iterator first, iterator last,
2923 const_iterator first1, const_iterator last1)
2924 {
2925 wxSTRING_INVALIDATE_CACHE();
2926
2927 m_impl.replace(first.impl(), last.impl(), first1.impl(), last1.impl());
2928
2929 return *this;
2930 }
2931
2932 wxString& replace(iterator first, iterator last,
2933 const char *first1, const char *last1)
2934 { replace(first, last, first1, last1 - first1); return *this; }
2935 wxString& replace(iterator first, iterator last,
2936 const wchar_t *first1, const wchar_t *last1)
2937 { replace(first, last, first1, last1 - first1); return *this; }
2938
2939 // swap two strings
2940 void swap(wxString& str)
2941 {
2942#if wxUSE_STRING_POS_CACHE
2943 // we modify not only this string but also the other one directly so we
2944 // need to invalidate cache for both of them (we could also try to
2945 // exchange their cache entries but it seems unlikely to be worth it)
2946 InvalidateCache();
2947 str.InvalidateCache();
2948#endif // wxUSE_STRING_POS_CACHE
2949
2950 m_impl.swap(str.m_impl);
2951 }
2952
2953 // find a substring
2954 size_t find(const wxString& str, size_t nStart = 0) const
2955 { return PosFromImpl(m_impl.find(str.m_impl, PosToImpl(nStart))); }
2956
2957 // find first n characters of sz
2958 size_t find(const char* sz, size_t nStart = 0, size_t n = npos) const
2959 {
2960 SubstrBufFromMB str(ImplStr(sz, n));
2961 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
2962 }
2963 size_t find(const wchar_t* sz, size_t nStart = 0, size_t n = npos) const
2964 {
2965 SubstrBufFromWC str(ImplStr(sz, n));
2966 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
2967 }
2968 size_t find(const wxCharBuffer& s, size_t nStart = 0, size_t n = npos) const
2969 { return find(s.data(), nStart, n); }
2970 size_t find(const wxWCharBuffer& s, size_t nStart = 0, size_t n = npos) const
2971 { return find(s.data(), nStart, n); }
2972 size_t find(const wxCStrData& s, size_t nStart = 0, size_t n = npos) const
2973 { return find(s.AsWChar(), nStart, n); }
2974
2975 // find the first occurence of character ch after nStart
2976 size_t find(wxUniChar ch, size_t nStart = 0) const
2977 {
2978#if wxUSE_UNICODE_UTF8
2979 if ( !ch.IsAscii() )
2980 return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch),
2981 PosToImpl(nStart)));
2982 else
2983#endif
2984 return PosFromImpl(m_impl.find((wxStringCharType)ch,
2985 PosToImpl(nStart)));
2986
2987 }
2988 size_t find(wxUniCharRef ch, size_t nStart = 0) const
2989 { return find(wxUniChar(ch), nStart); }
2990 size_t find(char ch, size_t nStart = 0) const
2991 { return find(wxUniChar(ch), nStart); }
2992 size_t find(unsigned char ch, size_t nStart = 0) const
2993 { return find(wxUniChar(ch), nStart); }
2994 size_t find(wchar_t ch, size_t nStart = 0) const
2995 { return find(wxUniChar(ch), nStart); }
2996
2997 // rfind() family is exactly like find() but works right to left
2998
2999 // as find, but from the end
3000 size_t rfind(const wxString& str, size_t nStart = npos) const
3001 { return PosFromImpl(m_impl.rfind(str.m_impl, PosToImpl(nStart))); }
3002
3003 // as find, but from the end
3004 size_t rfind(const char* sz, size_t nStart = npos, size_t n = npos) const
3005 {
3006 SubstrBufFromMB str(ImplStr(sz, n));
3007 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
3008 }
3009 size_t rfind(const wchar_t* sz, size_t nStart = npos, size_t n = npos) const
3010 {
3011 SubstrBufFromWC str(ImplStr(sz, n));
3012 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
3013 }
3014 size_t rfind(const wxCharBuffer& s, size_t nStart = npos, size_t n = npos) const
3015 { return rfind(s.data(), nStart, n); }
3016 size_t rfind(const wxWCharBuffer& s, size_t nStart = npos, size_t n = npos) const
3017 { return rfind(s.data(), nStart, n); }
3018 size_t rfind(const wxCStrData& s, size_t nStart = npos, size_t n = npos) const
3019 { return rfind(s.AsWChar(), nStart, n); }
3020 // as find, but from the end
3021 size_t rfind(wxUniChar ch, size_t nStart = npos) const
3022 {
3023#if wxUSE_UNICODE_UTF8
3024 if ( !ch.IsAscii() )
3025 return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch),
3026 PosToImpl(nStart)));
3027 else
3028#endif
3029 return PosFromImpl(m_impl.rfind((wxStringCharType)ch,
3030 PosToImpl(nStart)));
3031 }
3032 size_t rfind(wxUniCharRef ch, size_t nStart = npos) const
3033 { return rfind(wxUniChar(ch), nStart); }
3034 size_t rfind(char ch, size_t nStart = npos) const
3035 { return rfind(wxUniChar(ch), nStart); }
3036 size_t rfind(unsigned char ch, size_t nStart = npos) const
3037 { return rfind(wxUniChar(ch), nStart); }
3038 size_t rfind(wchar_t ch, size_t nStart = npos) const
3039 { return rfind(wxUniChar(ch), nStart); }
3040
3041 // find first/last occurence of any character (not) in the set:
3042#if wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3043 // FIXME-UTF8: this is not entirely correct, because it doesn't work if
3044 // sizeof(wchar_t)==2 and surrogates are present in the string;
3045 // should we care? Probably not.
3046 size_t find_first_of(const wxString& str, size_t nStart = 0) const
3047 { return m_impl.find_first_of(str.m_impl, nStart); }
3048 size_t find_first_of(const char* sz, size_t nStart = 0) const
3049 { return m_impl.find_first_of(ImplStr(sz), nStart); }
3050 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const
3051 { return m_impl.find_first_of(ImplStr(sz), nStart); }
3052 size_t find_first_of(const char* sz, size_t nStart, size_t n) const
3053 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
3054 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const
3055 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
3056 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3057 { return m_impl.find_first_of((wxChar)c, nStart); }
3058
3059 size_t find_last_of(const wxString& str, size_t nStart = npos) const
3060 { return m_impl.find_last_of(str.m_impl, nStart); }
3061 size_t find_last_of(const char* sz, size_t nStart = npos) const
3062 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3063 size_t find_last_of(const wchar_t* sz, size_t nStart = npos) const
3064 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3065 size_t find_last_of(const char* sz, size_t nStart, size_t n) const
3066 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3067 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const
3068 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3069 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3070 { return m_impl.find_last_of((wxChar)c, nStart); }
3071
3072 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
3073 { return m_impl.find_first_not_of(str.m_impl, nStart); }
3074 size_t find_first_not_of(const char* sz, size_t nStart = 0) const
3075 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
3076 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const
3077 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
3078 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const
3079 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
3080 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const
3081 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
3082 size_t find_first_not_of(wxUniChar c, size_t nStart = 0) const
3083 { return m_impl.find_first_not_of((wxChar)c, nStart); }
3084
3085 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
3086 { return m_impl.find_last_not_of(str.m_impl, nStart); }
3087 size_t find_last_not_of(const char* sz, size_t nStart = npos) const
3088 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
3089 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const
3090 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
3091 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const
3092 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
3093 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const
3094 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
3095 size_t find_last_not_of(wxUniChar c, size_t nStart = npos) const
3096 { return m_impl.find_last_not_of((wxChar)c, nStart); }
3097#else
3098 // we can't use std::string implementation in UTF-8 build, because the
3099 // character sets would be interpreted wrongly:
3100
3101 // as strpbrk() but starts at nStart, returns npos if not found
3102 size_t find_first_of(const wxString& str, size_t nStart = 0) const
3103#if wxUSE_UNICODE // FIXME-UTF8: temporary
3104 { return find_first_of(str.wc_str(), nStart); }
3105#else
3106 { return find_first_of(str.mb_str(), nStart); }
3107#endif
3108 // same as above
3109 size_t find_first_of(const char* sz, size_t nStart = 0) const;
3110 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const;
3111 size_t find_first_of(const char* sz, size_t nStart, size_t n) const;
3112 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const;
3113 // same as find(char, size_t)
3114 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3115 { return find(c, nStart); }
3116 // find the last (starting from nStart) char from str in this string
3117 size_t find_last_of (const wxString& str, size_t nStart = npos) const
3118#if wxUSE_UNICODE // FIXME-UTF8: temporary
3119 { return find_last_of(str.wc_str(), nStart); }
3120#else
3121 { return find_last_of(str.mb_str(), nStart); }
3122#endif
3123 // same as above
3124 size_t find_last_of (const char* sz, size_t nStart = npos) const;
3125 size_t find_last_of (const wchar_t* sz, size_t nStart = npos) const;
3126 size_t find_last_of(const char* sz, size_t nStart, size_t n) const;
3127 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const;
3128 // same as above
3129 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3130 { return rfind(c, nStart); }
3131
3132 // find first/last occurence of any character not in the set
3133
3134 // as strspn() (starting from nStart), returns npos on failure
3135 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
3136#if wxUSE_UNICODE // FIXME-UTF8: temporary
3137 { return find_first_not_of(str.wc_str(), nStart); }
3138#else
3139 { return find_first_not_of(str.mb_str(), nStart); }
3140#endif
3141 // same as above
3142 size_t find_first_not_of(const char* sz, size_t nStart = 0) const;
3143 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const;
3144 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const;
3145 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3146 // same as above
3147 size_t find_first_not_of(wxUniChar ch, size_t nStart = 0) const;
3148 // as strcspn()
3149 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
3150#if wxUSE_UNICODE // FIXME-UTF8: temporary
3151 { return find_last_not_of(str.wc_str(), nStart); }
3152#else
3153 { return find_last_not_of(str.mb_str(), nStart); }
3154#endif
3155 // same as above
3156 size_t find_last_not_of(const char* sz, size_t nStart = npos) const;
3157 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const;
3158 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const;
3159 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3160 // same as above
3161 size_t find_last_not_of(wxUniChar ch, size_t nStart = npos) const;
3162#endif // wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 or not
3163
3164 // provide char/wchar_t/wxUniCharRef overloads for char-finding functions
3165 // above to resolve ambiguities:
3166 size_t find_first_of(wxUniCharRef ch, size_t nStart = 0) const
3167 { return find_first_of(wxUniChar(ch), nStart); }
3168 size_t find_first_of(char ch, size_t nStart = 0) const
3169 { return find_first_of(wxUniChar(ch), nStart); }
3170 size_t find_first_of(unsigned char ch, size_t nStart = 0) const
3171 { return find_first_of(wxUniChar(ch), nStart); }
3172 size_t find_first_of(wchar_t ch, size_t nStart = 0) const
3173 { return find_first_of(wxUniChar(ch), nStart); }
3174 size_t find_last_of(wxUniCharRef ch, size_t nStart = npos) const
3175 { return find_last_of(wxUniChar(ch), nStart); }
3176 size_t find_last_of(char ch, size_t nStart = npos) const
3177 { return find_last_of(wxUniChar(ch), nStart); }
3178 size_t find_last_of(unsigned char ch, size_t nStart = npos) const
3179 { return find_last_of(wxUniChar(ch), nStart); }
3180 size_t find_last_of(wchar_t ch, size_t nStart = npos) const
3181 { return find_last_of(wxUniChar(ch), nStart); }
3182 size_t find_first_not_of(wxUniCharRef ch, size_t nStart = 0) const
3183 { return find_first_not_of(wxUniChar(ch), nStart); }
3184 size_t find_first_not_of(char ch, size_t nStart = 0) const
3185 { return find_first_not_of(wxUniChar(ch), nStart); }
3186 size_t find_first_not_of(unsigned char ch, size_t nStart = 0) const
3187 { return find_first_not_of(wxUniChar(ch), nStart); }
3188 size_t find_first_not_of(wchar_t ch, size_t nStart = 0) const
3189 { return find_first_not_of(wxUniChar(ch), nStart); }
3190 size_t find_last_not_of(wxUniCharRef ch, size_t nStart = npos) const
3191 { return find_last_not_of(wxUniChar(ch), nStart); }
3192 size_t find_last_not_of(char ch, size_t nStart = npos) const
3193 { return find_last_not_of(wxUniChar(ch), nStart); }
3194 size_t find_last_not_of(unsigned char ch, size_t nStart = npos) const
3195 { return find_last_not_of(wxUniChar(ch), nStart); }
3196 size_t find_last_not_of(wchar_t ch, size_t nStart = npos) const
3197 { return find_last_not_of(wxUniChar(ch), nStart); }
3198
3199 // and additional overloads for the versions taking strings:
3200 size_t find_first_of(const wxCStrData& sz, size_t nStart = 0) const
3201 { return find_first_of(sz.AsString(), nStart); }
3202 size_t find_first_of(const wxCharBuffer& sz, size_t nStart = 0) const
3203 { return find_first_of(sz.data(), nStart); }
3204 size_t find_first_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3205 { return find_first_of(sz.data(), nStart); }
3206 size_t find_first_of(const wxCStrData& sz, size_t nStart, size_t n) const
3207 { return find_first_of(sz.AsWChar(), nStart, n); }
3208 size_t find_first_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3209 { return find_first_of(sz.data(), nStart, n); }
3210 size_t find_first_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3211 { return find_first_of(sz.data(), nStart, n); }
3212
3213 size_t find_last_of(const wxCStrData& sz, size_t nStart = 0) const
3214 { return find_last_of(sz.AsString(), nStart); }
3215 size_t find_last_of(const wxCharBuffer& sz, size_t nStart = 0) const
3216 { return find_last_of(sz.data(), nStart); }
3217 size_t find_last_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3218 { return find_last_of(sz.data(), nStart); }
3219 size_t find_last_of(const wxCStrData& sz, size_t nStart, size_t n) const
3220 { return find_last_of(sz.AsWChar(), nStart, n); }
3221 size_t find_last_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3222 { return find_last_of(sz.data(), nStart, n); }
3223 size_t find_last_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3224 { return find_last_of(sz.data(), nStart, n); }
3225
3226 size_t find_first_not_of(const wxCStrData& sz, size_t nStart = 0) const
3227 { return find_first_not_of(sz.AsString(), nStart); }
3228 size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart = 0) const
3229 { return find_first_not_of(sz.data(), nStart); }
3230 size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3231 { return find_first_not_of(sz.data(), nStart); }
3232 size_t find_first_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3233 { return find_first_not_of(sz.AsWChar(), nStart, n); }
3234 size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3235 { return find_first_not_of(sz.data(), nStart, n); }
3236 size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3237 { return find_first_not_of(sz.data(), nStart, n); }
3238
3239 size_t find_last_not_of(const wxCStrData& sz, size_t nStart = 0) const
3240 { return find_last_not_of(sz.AsString(), nStart); }
3241 size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart = 0) const
3242 { return find_last_not_of(sz.data(), nStart); }
3243 size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const
3244 { return find_last_not_of(sz.data(), nStart); }
3245 size_t find_last_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3246 { return find_last_not_of(sz.AsWChar(), nStart, n); }
3247 size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const
3248 { return find_last_not_of(sz.data(), nStart, n); }
3249 size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const
3250 { return find_last_not_of(sz.data(), nStart, n); }
3251
3252 // string += string
3253 wxString& operator+=(const wxString& s)
3254 {
3255 wxSTRING_INVALIDATE_CACHED_LENGTH();
3256
3257 m_impl += s.m_impl;
3258 return *this;
3259 }
3260 // string += C string
3261 wxString& operator+=(const char *psz)
3262 {
3263 wxSTRING_INVALIDATE_CACHED_LENGTH();
3264
3265 m_impl += ImplStr(psz);
3266 return *this;
3267 }
3268 wxString& operator+=(const wchar_t *pwz)
3269 {
3270 wxSTRING_INVALIDATE_CACHED_LENGTH();
3271
3272 m_impl += ImplStr(pwz);
3273 return *this;
3274 }
3275 wxString& operator+=(const wxCStrData& s)
3276 {
3277 wxSTRING_INVALIDATE_CACHED_LENGTH();
3278
3279 m_impl += s.AsString().m_impl;
3280 return *this;
3281 }
3282 wxString& operator+=(const wxCharBuffer& s)
3283 { return operator+=(s.data()); }
3284 wxString& operator+=(const wxWCharBuffer& s)
3285 { return operator+=(s.data()); }
3286 // string += char
3287 wxString& operator+=(wxUniChar ch)
3288 {
3289 wxSTRING_UPDATE_CACHED_LENGTH(1);
3290
3291#if wxUSE_UNICODE_UTF8
3292 if ( !ch.IsAscii() )
3293 m_impl += wxStringOperations::EncodeChar(ch);
3294 else
3295#endif
3296 m_impl += (wxStringCharType)ch;
3297 return *this;
3298 }
3299 wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); }
3300 wxString& operator+=(int ch) { return *this += wxUniChar(ch); }
3301 wxString& operator+=(char ch) { return *this += wxUniChar(ch); }
3302 wxString& operator+=(unsigned char ch) { return *this += wxUniChar(ch); }
3303 wxString& operator+=(wchar_t ch) { return *this += wxUniChar(ch); }
3304
3305private:
3306#if !wxUSE_STL_BASED_WXSTRING
3307 // helpers for wxStringBuffer and wxStringBufferLength
3308 wxStringCharType *DoGetWriteBuf(size_t nLen)
3309 {
3310 return m_impl.DoGetWriteBuf(nLen);
3311 }
3312
3313 void DoUngetWriteBuf()
3314 {
3315 wxSTRING_INVALIDATE_CACHE();
3316
3317 m_impl.DoUngetWriteBuf();
3318 }
3319
3320 void DoUngetWriteBuf(size_t nLen)
3321 {
3322 wxSTRING_SET_CACHED_LENGTH(nLen);
3323
3324 m_impl.DoUngetWriteBuf(nLen);
3325 }
3326#endif // !wxUSE_STL_BASED_WXSTRING
3327
3328#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
3329 #if !wxUSE_UTF8_LOCALE_ONLY
3330 int DoPrintfWchar(const wxChar *format, ...);
3331 static wxString DoFormatWchar(const wxChar *format, ...);
3332 #endif
3333 #if wxUSE_UNICODE_UTF8
3334 int DoPrintfUtf8(const char *format, ...);
3335 static wxString DoFormatUtf8(const char *format, ...);
3336 #endif
3337#endif
3338
3339#if !wxUSE_STL_BASED_WXSTRING
3340 // check string's data validity
3341 bool IsValid() const { return m_impl.GetStringData()->IsValid(); }
3342#endif
3343
3344private:
3345 wxStringImpl m_impl;
3346
3347 // buffers for compatibility conversion from (char*)c_str() and
3348 // (wchar_t*)c_str():
3349 // FIXME-UTF8: bechmark various approaches to keeping compatibility buffers
3350 template<typename T>
3351 struct ConvertedBuffer
3352 {
3353 ConvertedBuffer() : m_buf(NULL) {}
3354 ~ConvertedBuffer()
3355 { free(m_buf); }
3356
3357 operator T*() const { return m_buf; }
3358
3359 ConvertedBuffer& operator=(T *str)
3360 {
3361 free(m_buf);
3362 m_buf = str;
3363 return *this;
3364 }
3365
3366 T *m_buf;
3367 };
3368#if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY
3369 ConvertedBuffer<char> m_convertedToChar;
3370#endif
3371#if !wxUSE_UNICODE_WCHAR
3372 ConvertedBuffer<wchar_t> m_convertedToWChar;
3373#endif
3374
3375#if wxUSE_UNICODE_UTF8
3376 // FIXME-UTF8: (try to) move this elsewhere (TLS) or solve differently
3377 // assigning to character pointer to by wxString::interator may
3378 // change the underlying wxStringImpl iterator, so we have to
3379 // keep track of all iterators and update them as necessary:
3380 struct wxStringIteratorNodeHead
3381 {
3382 wxStringIteratorNodeHead() : ptr(NULL) {}
3383 wxStringIteratorNode *ptr;
3384
3385 // copying is disallowed as it would result in more than one pointer into
3386 // the same linked list
3387 DECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead)
3388 };
3389
3390 wxStringIteratorNodeHead m_iterators;
3391
3392 friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode;
3393 friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef;
3394#endif // wxUSE_UNICODE_UTF8
3395
3396 friend class WXDLLIMPEXP_FWD_BASE wxCStrData;
3397 friend class wxStringInternalBuffer;
3398 friend class wxStringInternalBufferLength;
3399};
3400
3401#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
3402 #pragma warning (default:4275)
3403#endif
3404
3405// string iterator operators that satisfy STL Random Access Iterator
3406// requirements:
3407inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i)
3408 { return i + n; }
3409inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i)
3410 { return i + n; }
3411inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i)
3412 { return i + n; }
3413inline wxString::const_reverse_iterator operator+(ptrdiff_t n, wxString::const_reverse_iterator i)
3414 { return i + n; }
3415
3416// notice that even though for many compilers the friend declarations above are
3417// enough, from the point of view of C++ standard we must have the declarations
3418// here as friend ones are not injected in the enclosing namespace and without
3419// them the code fails to compile with conforming compilers such as xlC or g++4
3420wxString WXDLLIMPEXP_BASE operator+(const wxString& string1, const wxString& string2);
3421wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const char *psz);
3422wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const wchar_t *pwz);
3423wxString WXDLLIMPEXP_BASE operator+(const char *psz, const wxString& string);
3424wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz, const wxString& string);
3425
3426wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
3427wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
3428
3429inline wxString operator+(const wxString& string, wxUniCharRef ch)
3430 { return string + (wxUniChar)ch; }
3431inline wxString operator+(const wxString& string, char ch)
3432 { return string + wxUniChar(ch); }
3433inline wxString operator+(const wxString& string, wchar_t ch)
3434 { return string + wxUniChar(ch); }
3435inline wxString operator+(wxUniCharRef ch, const wxString& string)
3436 { return (wxUniChar)ch + string; }
3437inline wxString operator+(char ch, const wxString& string)
3438 { return wxUniChar(ch) + string; }
3439inline wxString operator+(wchar_t ch, const wxString& string)
3440 { return wxUniChar(ch) + string; }
3441
3442
3443#define wxGetEmptyString() wxString()
3444
3445// ----------------------------------------------------------------------------
3446// helper functions which couldn't be defined inline
3447// ----------------------------------------------------------------------------
3448
3449namespace wxPrivate
3450{
3451
3452#if wxUSE_UNICODE_WCHAR
3453
3454template <>
3455struct wxStringAsBufHelper<char>
3456{
3457 static wxCharBuffer Get(const wxString& s, size_t *len)
3458 {
3459 wxCharBuffer buf(s.mb_str());
3460 if ( len )
3461 *len = buf ? strlen(buf) : 0;
3462 return buf;
3463 }
3464};
3465
3466template <>
3467struct wxStringAsBufHelper<wchar_t>
3468{
3469 static wxWCharBuffer Get(const wxString& s, size_t *len)
3470 {
3471 if ( len )
3472 *len = s.length();
3473 return wxWCharBuffer::CreateNonOwned(s.wx_str());
3474 }
3475};
3476
3477#elif wxUSE_UNICODE_UTF8
3478
3479template <>
3480struct wxStringAsBufHelper<char>
3481{
3482 static wxCharBuffer Get(const wxString& s, size_t *len)
3483 {
3484 if ( len )
3485 *len = s.utf8_length();
3486 return wxCharBuffer::CreateNonOwned(s.wx_str());
3487 }
3488};
3489
3490template <>
3491struct wxStringAsBufHelper<wchar_t>
3492{
3493 static wxWCharBuffer Get(const wxString& s, size_t *len)
3494 {
3495 wxWCharBuffer wbuf(s.wc_str());
3496 if ( len )
3497 *len = wxWcslen(wbuf);
3498 return wbuf;
3499 }
3500};
3501
3502#endif // Unicode build kind
3503
3504} // namespace wxPrivate
3505
3506// ----------------------------------------------------------------------------
3507// wxStringBuffer: a tiny class allowing to get a writable pointer into string
3508// ----------------------------------------------------------------------------
3509
3510#if !wxUSE_STL_BASED_WXSTRING
3511// string buffer for direct access to string data in their native
3512// representation:
3513class wxStringInternalBuffer
3514{
3515public:
3516 typedef wxStringCharType CharType;
3517
3518 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
3519 : m_str(str), m_buf(NULL)
3520 { m_buf = m_str.DoGetWriteBuf(lenWanted); }
3521
3522 ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); }
3523
3524 operator wxStringCharType*() const { return m_buf; }
3525
3526private:
3527 wxString& m_str;
3528 wxStringCharType *m_buf;
3529
3530 DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)
3531};
3532
3533class wxStringInternalBufferLength
3534{
3535public:
3536 typedef wxStringCharType CharType;
3537
3538 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
3539 : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false)
3540 {
3541 m_buf = m_str.DoGetWriteBuf(lenWanted);
3542 wxASSERT(m_buf != NULL);
3543 }
3544
3545 ~wxStringInternalBufferLength()
3546 {
3547 wxASSERT(m_lenSet);
3548 m_str.DoUngetWriteBuf(m_len);
3549 }
3550
3551 operator wxStringCharType*() const { return m_buf; }
3552 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3553
3554private:
3555 wxString& m_str;
3556 wxStringCharType *m_buf;
3557 size_t m_len;
3558 bool m_lenSet;
3559
3560 DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)
3561};
3562
3563#endif // !wxUSE_STL_BASED_WXSTRING
3564
3565template<typename T>
3566class WXDLLIMPEXP_BASE wxStringTypeBufferBase
3567{
3568public:
3569 typedef T CharType;
3570
3571 wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024)
3572 : m_str(str), m_buf(lenWanted)
3573 {
3574 // for compatibility with old wxStringBuffer which provided direct
3575 // access to wxString internal buffer, initialize ourselves with the
3576 // string initial contents
3577
3578 // FIXME-VC6: remove the ugly (CharType *)NULL and use normal
3579 // tchar_str<CharType>
3580 size_t len;
3581 const wxCharTypeBuffer<CharType> buf(str.tchar_str(&len, (CharType *)NULL));
3582 if ( buf )
3583 {
3584 if ( len > lenWanted )
3585 {
3586 // in this case there is not enough space for terminating NUL,
3587 // ensure that we still put it there
3588 m_buf.data()[lenWanted] = 0;
3589 len = lenWanted - 1;
3590 }
3591
3592 memcpy(m_buf.data(), buf, (len + 1)*sizeof(CharType));
3593 }
3594 //else: conversion failed, this can happen when trying to get Unicode
3595 // string contents into a char string
3596 }
3597
3598 operator CharType*() { return m_buf.data(); }
3599
3600protected:
3601 wxString& m_str;
3602 wxCharTypeBuffer<CharType> m_buf;
3603};
3604
3605template<typename T>
3606class WXDLLIMPEXP_BASE wxStringTypeBufferLengthBase
3607 : public wxStringTypeBufferBase<T>
3608{
3609public:
3610 wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024)
3611 : wxStringTypeBufferBase<T>(str, lenWanted),
3612 m_len(0),
3613 m_lenSet(false)
3614 { }
3615
3616 ~wxStringTypeBufferLengthBase()
3617 {
3618 wxASSERT_MSG( this->m_lenSet, "forgot to call SetLength()" );
3619 }
3620
3621 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3622
3623protected:
3624 size_t m_len;
3625 bool m_lenSet;
3626};
3627
3628template<typename T>
3629class wxStringTypeBuffer : public wxStringTypeBufferBase<T>
3630{
3631public:
3632 wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024)
3633 : wxStringTypeBufferBase<T>(str, lenWanted)
3634 { }
3635
3636 ~wxStringTypeBuffer()
3637 {
3638 this->m_str.assign(this->m_buf.data());
3639 }
3640
3641 DECLARE_NO_COPY_CLASS(wxStringTypeBuffer)
3642};
3643
3644template<typename T>
3645class wxStringTypeBufferLength : public wxStringTypeBufferLengthBase<T>
3646{
3647public:
3648 wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024)
3649 : wxStringTypeBufferLengthBase<T>(str, lenWanted)
3650 { }
3651
3652 ~wxStringTypeBufferLength()
3653 {
3654 this->m_str.assign(this->m_buf.data(), this->m_len);
3655 }
3656
3657 DECLARE_NO_COPY_CLASS(wxStringTypeBufferLength)
3658};
3659
3660#if wxUSE_STL_BASED_WXSTRING
3661
3662WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<wxStringCharType> )
3663
3664class wxStringInternalBuffer : public wxStringTypeBufferBase<wxStringCharType>
3665{
3666public:
3667 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
3668 : wxStringTypeBufferBase<wxStringCharType>(str, lenWanted) {}
3669 ~wxStringInternalBuffer()
3670 { m_str.m_impl.assign(m_buf.data()); }
3671
3672 DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)
3673};
3674
3675WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE(
3676 wxStringTypeBufferLengthBase<wxStringCharType> )
3677
3678class wxStringInternalBufferLength
3679 : public wxStringTypeBufferLengthBase<wxStringCharType>
3680{
3681public:
3682 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
3683 : wxStringTypeBufferLengthBase<wxStringCharType>(str, lenWanted) {}
3684
3685 ~wxStringInternalBufferLength()
3686 {
3687 m_str.m_impl.assign(m_buf.data(), m_len);
3688 }
3689
3690 DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)
3691};
3692
3693#endif // wxUSE_STL_BASED_WXSTRING
3694
3695
3696#if wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8
3697typedef wxStringTypeBuffer<wxChar> wxStringBuffer;
3698typedef wxStringTypeBufferLength<wxChar> wxStringBufferLength;
3699#else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3700typedef wxStringInternalBuffer wxStringBuffer;
3701typedef wxStringInternalBufferLength wxStringBufferLength;
3702#endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3703
3704#if wxUSE_UNICODE_UTF8
3705typedef wxStringInternalBuffer wxUTF8StringBuffer;
3706typedef wxStringInternalBufferLength wxUTF8StringBufferLength;
3707#elif wxUSE_UNICODE_WCHAR
3708
3709WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<char> )
3710
3711class WXDLLIMPEXP_BASE wxUTF8StringBuffer : public wxStringTypeBufferBase<char>
3712{
3713public:
3714 wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024)
3715 : wxStringTypeBufferBase<char>(str, lenWanted) {}
3716 ~wxUTF8StringBuffer();
3717
3718 DECLARE_NO_COPY_CLASS(wxUTF8StringBuffer)
3719};
3720
3721WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase<char> )
3722
3723class WXDLLIMPEXP_BASE wxUTF8StringBufferLength
3724 : public wxStringTypeBufferLengthBase<char>
3725{
3726public:
3727 wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024)
3728 : wxStringTypeBufferLengthBase<char>(str, lenWanted) {}
3729 ~wxUTF8StringBufferLength();
3730
3731 DECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength)
3732};
3733#endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR
3734
3735
3736// ---------------------------------------------------------------------------
3737// wxString comparison functions: operator versions are always case sensitive
3738// ---------------------------------------------------------------------------
3739
3740#define wxCMP_WXCHAR_STRING(p, s, op) 0 op s.Cmp(p)
3741
3742wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING)
3743
3744#undef wxCMP_WXCHAR_STRING
3745
3746inline bool operator==(const wxString& s1, const wxString& s2)
3747 { return s1.IsSameAs(s2); }
3748inline bool operator!=(const wxString& s1, const wxString& s2)
3749 { return !s1.IsSameAs(s2); }
3750inline bool operator< (const wxString& s1, const wxString& s2)
3751 { return s1.Cmp(s2) < 0; }
3752inline bool operator> (const wxString& s1, const wxString& s2)
3753 { return s1.Cmp(s2) > 0; }
3754inline bool operator<=(const wxString& s1, const wxString& s2)
3755 { return s1.Cmp(s2) <= 0; }
3756inline bool operator>=(const wxString& s1, const wxString& s2)
3757 { return s1.Cmp(s2) >= 0; }
3758
3759inline bool operator==(const wxString& s1, const wxCStrData& s2)
3760 { return s1 == s2.AsString(); }
3761inline bool operator==(const wxCStrData& s1, const wxString& s2)
3762 { return s1.AsString() == s2; }
3763inline bool operator!=(const wxString& s1, const wxCStrData& s2)
3764 { return s1 != s2.AsString(); }
3765inline bool operator!=(const wxCStrData& s1, const wxString& s2)
3766 { return s1.AsString() != s2; }
3767
3768inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
3769 { return (s1.Cmp((const wchar_t *)s2) == 0); }
3770inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
3771 { return (s2.Cmp((const wchar_t *)s1) == 0); }
3772inline bool operator!=(const wxString& s1, const wxWCharBuffer& s2)
3773 { return (s1.Cmp((const wchar_t *)s2) != 0); }
3774inline bool operator!=(const wxWCharBuffer& s1, const wxString& s2)
3775 { return (s2.Cmp((const wchar_t *)s1) != 0); }
3776
3777inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
3778 { return (s1.Cmp((const char *)s2) == 0); }
3779inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
3780 { return (s2.Cmp((const char *)s1) == 0); }
3781inline bool operator!=(const wxString& s1, const wxCharBuffer& s2)
3782 { return (s1.Cmp((const char *)s2) != 0); }
3783inline bool operator!=(const wxCharBuffer& s1, const wxString& s2)
3784 { return (s2.Cmp((const char *)s1) != 0); }
3785
3786inline wxString operator+(const wxString& string, const wxWCharBuffer& buf)
3787 { return string + (const wchar_t *)buf; }
3788inline wxString operator+(const wxWCharBuffer& buf, const wxString& string)
3789 { return (const wchar_t *)buf + string; }
3790
3791inline wxString operator+(const wxString& string, const wxCharBuffer& buf)
3792 { return string + (const char *)buf; }
3793inline wxString operator+(const wxCharBuffer& buf, const wxString& string)
3794 { return (const char *)buf + string; }
3795
3796// comparison with char
3797inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); }
3798inline bool operator==(const wxUniCharRef& c, const wxString& s) { return s.IsSameAs(c); }
3799inline bool operator==(char c, const wxString& s) { return s.IsSameAs(c); }
3800inline bool operator==(wchar_t c, const wxString& s) { return s.IsSameAs(c); }
3801inline bool operator==(int c, const wxString& s) { return s.IsSameAs(c); }
3802inline bool operator==(const wxString& s, const wxUniChar& c) { return s.IsSameAs(c); }
3803inline bool operator==(const wxString& s, const wxUniCharRef& c) { return s.IsSameAs(c); }
3804inline bool operator==(const wxString& s, char c) { return s.IsSameAs(c); }
3805inline bool operator==(const wxString& s, wchar_t c) { return s.IsSameAs(c); }
3806inline bool operator!=(const wxUniChar& c, const wxString& s) { return !s.IsSameAs(c); }
3807inline bool operator!=(const wxUniCharRef& c, const wxString& s) { return !s.IsSameAs(c); }
3808inline bool operator!=(char c, const wxString& s) { return !s.IsSameAs(c); }
3809inline bool operator!=(wchar_t c, const wxString& s) { return !s.IsSameAs(c); }
3810inline bool operator!=(int c, const wxString& s) { return !s.IsSameAs(c); }
3811inline bool operator!=(const wxString& s, const wxUniChar& c) { return !s.IsSameAs(c); }
3812inline bool operator!=(const wxString& s, const wxUniCharRef& c) { return !s.IsSameAs(c); }
3813inline bool operator!=(const wxString& s, char c) { return !s.IsSameAs(c); }
3814inline bool operator!=(const wxString& s, wchar_t c) { return !s.IsSameAs(c); }
3815
3816// comparison with C string in Unicode build
3817#if wxUSE_UNICODE
3818
3819#define wxCMP_CHAR_STRING(p, s, op) wxString(p) op s
3820
3821wxDEFINE_ALL_COMPARISONS(const char *, const wxString&, wxCMP_CHAR_STRING)
3822
3823#undef wxCMP_CHAR_STRING
3824
3825#endif // wxUSE_UNICODE
3826
3827// we also need to provide the operators for comparison with wxCStrData to
3828// resolve ambiguity between operator(const wxChar *,const wxString &) and
3829// operator(const wxChar *, const wxChar *) for "p == s.c_str()"
3830//
3831// notice that these are (shallow) pointer comparisons, not (deep) string ones
3832#define wxCMP_CHAR_CSTRDATA(p, s, op) p op s.AsChar()
3833#define wxCMP_WCHAR_CSTRDATA(p, s, op) p op s.AsWChar()
3834
3835wxDEFINE_ALL_COMPARISONS(const wchar_t *, const wxCStrData&, wxCMP_WCHAR_CSTRDATA)
3836wxDEFINE_ALL_COMPARISONS(const char *, const wxCStrData&, wxCMP_CHAR_CSTRDATA)
3837
3838#undef wxCMP_CHAR_CSTRDATA
3839#undef wxCMP_WCHAR_CSTRDATA
3840
3841// ---------------------------------------------------------------------------
3842// Implementation only from here until the end of file
3843// ---------------------------------------------------------------------------
3844
3845#if wxUSE_STD_IOSTREAM
3846
3847#include "wx/iosfwrap.h"
3848
3849WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxString&);
3850WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCStrData&);
3851WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCharBuffer&);
3852#ifndef __BORLANDC__
3853WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&);
3854#endif
3855
3856#if wxUSE_UNICODE && defined(HAVE_WOSTREAM)
3857
3858WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxString&);
3859WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxCStrData&);
3860WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxWCharBuffer&);
3861
3862#endif // wxUSE_UNICODE && defined(HAVE_WOSTREAM)
3863
3864#endif // wxUSE_STD_IOSTREAM
3865
3866// ---------------------------------------------------------------------------
3867// wxCStrData implementation
3868// ---------------------------------------------------------------------------
3869
3870inline wxCStrData::wxCStrData(char *buf)
3871 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
3872inline wxCStrData::wxCStrData(wchar_t *buf)
3873 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
3874
3875inline wxCStrData::wxCStrData(const wxCStrData& data)
3876 : m_str(data.m_owned ? new wxString(*data.m_str) : data.m_str),
3877 m_offset(data.m_offset),
3878 m_owned(data.m_owned)
3879{
3880}
3881
3882inline wxCStrData::~wxCStrData()
3883{
3884 if ( m_owned )
3885 delete wx_const_cast(wxString*, m_str); // cast to silence warnings
3886}
3887
3888// simple cases for AsChar() and AsWChar(), the complicated ones are
3889// in string.cpp
3890#if wxUSE_UNICODE_WCHAR
3891inline const wchar_t* wxCStrData::AsWChar() const
3892{
3893 return m_str->wx_str() + m_offset;
3894}
3895#endif // wxUSE_UNICODE_WCHAR
3896
3897#if !wxUSE_UNICODE
3898inline const char* wxCStrData::AsChar() const
3899{
3900 return m_str->wx_str() + m_offset;
3901}
3902#endif // !wxUSE_UNICODE
3903
3904#if wxUSE_UTF8_LOCALE_ONLY
3905inline const char* wxCStrData::AsChar() const
3906{
3907 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
3908}
3909#endif // wxUSE_UTF8_LOCALE_ONLY
3910
3911inline const wxCharBuffer wxCStrData::AsCharBuf() const
3912{
3913#if !wxUSE_UNICODE
3914 return wxCharBuffer::CreateNonOwned(AsChar());
3915#else
3916 return AsString().mb_str();
3917#endif
3918}
3919
3920inline const wxWCharBuffer wxCStrData::AsWCharBuf() const
3921{
3922#if wxUSE_UNICODE_WCHAR
3923 return wxWCharBuffer::CreateNonOwned(AsWChar());
3924#else
3925 return AsString().wc_str();
3926#endif
3927}
3928
3929inline wxString wxCStrData::AsString() const
3930{
3931 if ( m_offset == 0 )
3932 return *m_str;
3933 else
3934 return m_str->Mid(m_offset);
3935}
3936
3937inline const wxStringCharType *wxCStrData::AsInternal() const
3938{
3939#if wxUSE_UNICODE_UTF8
3940 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
3941#else
3942 return m_str->wx_str() + m_offset;
3943#endif
3944}
3945
3946inline wxUniChar wxCStrData::operator*() const
3947{
3948 if ( m_str->empty() )
3949 return wxUniChar(_T('\0'));
3950 else
3951 return (*m_str)[m_offset];
3952}
3953
3954inline wxUniChar wxCStrData::operator[](size_t n) const
3955{
3956 // NB: we intentionally use operator[] and not at() here because the former
3957 // works for the terminating NUL while the latter does not
3958 return (*m_str)[m_offset + n];
3959}
3960
3961// ----------------------------------------------------------------------------
3962// more wxCStrData operators
3963// ----------------------------------------------------------------------------
3964
3965// we need to define those to allow "size_t pos = p - s.c_str()" where p is
3966// some pointer into the string
3967inline size_t operator-(const char *p, const wxCStrData& cs)
3968{
3969 return p - cs.AsChar();
3970}
3971
3972inline size_t operator-(const wchar_t *p, const wxCStrData& cs)
3973{
3974 return p - cs.AsWChar();
3975}
3976
3977// ----------------------------------------------------------------------------
3978// implementation of wx[W]CharBuffer inline methods using wxCStrData
3979// ----------------------------------------------------------------------------
3980
3981// FIXME-UTF8: move this to buffer.h
3982inline wxCharBuffer::wxCharBuffer(const wxCStrData& cstr)
3983 : wxCharTypeBufferBase(cstr.AsCharBuf())
3984{
3985}
3986
3987inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr)
3988 : wxCharTypeBufferBase(cstr.AsWCharBuf())
3989{
3990}
3991
3992#if wxUSE_UNICODE_UTF8
3993// ----------------------------------------------------------------------------
3994// implementation of wxStringIteratorNode inline methods
3995// ----------------------------------------------------------------------------
3996
3997void wxStringIteratorNode::DoSet(const wxString *str,
3998 wxStringImpl::const_iterator *citer,
3999 wxStringImpl::iterator *iter)
4000{
4001 m_prev = NULL;
4002 m_iter = iter;
4003 m_citer = citer;
4004 m_str = str;
4005 if ( str )
4006 {
4007 m_next = str->m_iterators.ptr;
4008 wx_const_cast(wxString*, m_str)->m_iterators.ptr = this;
4009 if ( m_next )
4010 m_next->m_prev = this;
4011 }
4012 else
4013 {
4014 m_next = NULL;
4015 }
4016}
4017
4018void wxStringIteratorNode::clear()
4019{
4020 if ( m_next )
4021 m_next->m_prev = m_prev;
4022 if ( m_prev )
4023 m_prev->m_next = m_next;
4024 else if ( m_str ) // first in the list
4025 wx_const_cast(wxString*, m_str)->m_iterators.ptr = m_next;
4026
4027 m_next = m_prev = NULL;
4028 m_citer = NULL;
4029 m_iter = NULL;
4030 m_str = NULL;
4031}
4032#endif // wxUSE_UNICODE_UTF8
4033
4034#if WXWIN_COMPATIBILITY_2_8
4035 // lot of code out there doesn't explicitly include wx/crt.h, but uses
4036 // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h,
4037 // so let's include this header now that wxString is defined and it's safe
4038 // to do it:
4039 #include "wx/crt.h"
4040#endif
4041
4042#endif // _WX_WXSTRING_H_