]> git.saurik.com Git - wxWidgets.git/blame - include/wx/string.h
Fix OpenVMS compile support for new version in src/stc
[wxWidgets.git] / include / wx / string.h
CommitLineData
3c67202d 1///////////////////////////////////////////////////////////////////////////////
1c2d1459 2// Name: wx/string.h
a7ea63e2 3// Purpose: wxString class
c801d85f
KB
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>
65571936 9// Licence: wxWindows licence
3c67202d 10///////////////////////////////////////////////////////////////////////////////
c801d85f 11
e90c1d2a
VZ
12/*
13 Efficient string class [more or less] compatible with MFC CString,
77ffb593 14 wxWidgets version 1 wxString and std::string and some handy functions
e90c1d2a
VZ
15 missing from string.h.
16*/
17
a7ea63e2
VS
18#ifndef _WX_WXSTRING_H__
19#define _WX_WXSTRING_H__
c801d85f 20
e90c1d2a
VZ
21// ----------------------------------------------------------------------------
22// headers
23// ----------------------------------------------------------------------------
24
5737d05f
VZ
25#include "wx/defs.h" // everybody should include this
26
9b4da627 27#ifndef __WXPALMOS5__
9dea36ef 28#if defined(__WXMAC__) || defined(__VISAGECPP__)
3f4a0c5b 29 #include <ctype.h>
17dff81c 30#endif
3f4a0c5b 31
9dea36ef
DW
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
c801d85f 46
1c2d1459 47#ifdef HAVE_STRCASECMP_IN_STRINGS_H
1bfcb0b6 48 #include <strings.h> // for strcasecmp()
1c2d1459 49#endif // HAVE_STRCASECMP_IN_STRINGS_H
9b4da627 50#endif // ! __WXPALMOS5__
ffecfa5a 51
52de37c7 52#include "wx/wxcrtbase.h" // for wxChar, wxStrlen() etc.
c9f78968 53#include "wx/strvararg.h"
e90c1d2a
VZ
54#include "wx/buffer.h" // for wxCharBuffer
55#include "wx/strconv.h" // for wxConvertXXX() macros and wxMBConv classes
a7ea63e2 56#include "wx/stringimpl.h"
467175ab 57#include "wx/stringops.h"
a7ea63e2 58#include "wx/unichar.h"
3f4a0c5b 59
68482dc5
VZ
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
b69470e4
SC
72// CS: under OSX release builds the string destructor/cache cleanup sometimes
73// crashes, disable until we find the true reason or a better workaround
74#if wxUSE_UNICODE_UTF8 && !defined(__WXMSW__) && !defined(__WXOSX__)
68482dc5
VZ
75 #define wxUSE_STRING_POS_CACHE 1
76#else
77 #define wxUSE_STRING_POS_CACHE 0
78#endif
79
80#if wxUSE_STRING_POS_CACHE
81 #include "wx/tls.h"
82
83 // change this 0 to 1 to enable additional (very expensive) asserts
84 // verifying that string caching logic works as expected
85 #if 0
86 #define wxSTRING_CACHE_ASSERT(cond) wxASSERT(cond)
87 #else
88 #define wxSTRING_CACHE_ASSERT(cond)
89 #endif
90#endif // wxUSE_STRING_POS_CACHE
91
b5dbe15d 92class WXDLLIMPEXP_FWD_BASE wxString;
fb3c9042 93
67cff9dc
VZ
94// unless this symbol is predefined to disable the compatibility functions, do
95// use them
96#ifndef WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
97 #define WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER 1
98#endif
99
062dc5fc
VZ
100namespace wxPrivate
101{
102 template <typename T> struct wxStringAsBufHelper;
103}
104
c801d85f
KB
105// ---------------------------------------------------------------------------
106// macros
107// ---------------------------------------------------------------------------
108
5737d05f
VZ
109// casts [unfortunately!] needed to call some broken functions which require
110// "char *" instead of "const char *"
2bb67b80 111#define WXSTRINGCAST (wxChar *)(const wxChar *)
e90c1d2a
VZ
112#define wxCSTRINGCAST (wxChar *)(const wxChar *)
113#define wxMBSTRINGCAST (char *)(const char *)
114#define wxWCSTRINGCAST (wchar_t *)(const wchar_t *)
c801d85f 115
e90c1d2a
VZ
116// ----------------------------------------------------------------------------
117// constants
118// ----------------------------------------------------------------------------
119
13874b5c
VZ
120#if WXWIN_COMPATIBILITY_2_6
121
122// deprecated in favour of wxString::npos, don't use in new code
123//
e90c1d2a 124// maximum possible length for a string means "take all string" everywhere
8f93a29f 125#define wxSTRING_MAXLEN wxString::npos
66b6b045 126
13874b5c
VZ
127#endif // WXWIN_COMPATIBILITY_2_6
128
c801d85f 129// ---------------------------------------------------------------------------
e90c1d2a 130// global functions complementing standard C string library replacements for
3c67202d
VZ
131// strlen() and portable strcasecmp()
132//---------------------------------------------------------------------------
e90c1d2a 133
c7554da8 134#if WXWIN_COMPATIBILITY_2_8
e3f6cbd9 135// Use wxXXX() functions from wxcrt.h instead! These functions are for
e90c1d2a 136// backwards compatibility only.
88150e60 137
3c67202d 138// checks whether the passed in pointer is NULL and if the string is empty
c7554da8 139wxDEPRECATED( inline bool IsEmpty(const char *p) );
6d56eb5c 140inline bool IsEmpty(const char *p) { return (!p || !*p); }
c801d85f 141
3c67202d 142// safe version of strlen() (returns 0 if passed NULL pointer)
c7554da8 143wxDEPRECATED( inline size_t Strlen(const char *psz) );
6d56eb5c 144inline size_t Strlen(const char *psz)
c801d85f
KB
145 { return psz ? strlen(psz) : 0; }
146
3c67202d 147// portable strcasecmp/_stricmp
c7554da8 148wxDEPRECATED( inline int Stricmp(const char *psz1, const char *psz2) );
6d56eb5c 149inline int Stricmp(const char *psz1, const char *psz2)
dd1eaa89 150{
7f0586ef
JS
151#if defined(__VISUALC__) && defined(__WXWINCE__)
152 register char c1, c2;
153 do {
154 c1 = tolower(*psz1++);
155 c2 = tolower(*psz2++);
156 } while ( c1 && (c1 == c2) );
157
158 return c1 - c2;
159#elif defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) )
dd1eaa89 160 return _stricmp(psz1, psz2);
91b8de8d 161#elif defined(__SC__)
2432b92d 162 return _stricmp(psz1, psz2);
dd1eaa89
VZ
163#elif defined(__BORLANDC__)
164 return stricmp(psz1, psz2);
7be1f0d9
JS
165#elif defined(__WATCOMC__)
166 return stricmp(psz1, psz2);
14704334
VS
167#elif defined(__DJGPP__)
168 return stricmp(psz1, psz2);
91b8de8d
RR
169#elif defined(__EMX__)
170 return stricmp(psz1, psz2);
e2c87f4c 171#elif defined(__WXPM__)
1777b9bb 172 return stricmp(psz1, psz2);
a0be6908
WS
173#elif defined(__WXPALMOS__) || \
174 defined(HAVE_STRCASECMP_IN_STRING_H) || \
1c2d1459
VZ
175 defined(HAVE_STRCASECMP_IN_STRINGS_H) || \
176 defined(__GNUWIN32__)
dd1eaa89 177 return strcasecmp(psz1, psz2);
8be97d65 178#elif defined(__MWERKS__) && !defined(__INTEL__)
17dff81c
SC
179 register char c1, c2;
180 do {
181 c1 = tolower(*psz1++);
182 c2 = tolower(*psz2++);
183 } while ( c1 && (c1 == c2) );
184
185 return c1 - c2;
dd1eaa89
VZ
186#else
187 // almost all compilers/libraries provide this function (unfortunately under
188 // different names), that's why we don't implement our own which will surely
189 // be more efficient than this code (uncomment to use):
190 /*
191 register char c1, c2;
192 do {
193 c1 = tolower(*psz1++);
194 c2 = tolower(*psz2++);
195 } while ( c1 && (c1 == c2) );
196
197 return c1 - c2;
198 */
199
200 #error "Please define string case-insensitive compare for your OS/compiler"
201#endif // OS/compiler
202}
c801d85f 203
c7554da8
VS
204#endif // WXWIN_COMPATIBILITY_2_8
205
c9f78968
VS
206// ----------------------------------------------------------------------------
207// wxCStrData
208// ----------------------------------------------------------------------------
209
210// Lightweight object returned by wxString::c_str() and implicitly convertible
211// to either const char* or const wchar_t*.
113f6233 212class wxCStrData
c9f78968
VS
213{
214private:
215 // Ctors; for internal use by wxString and wxCStrData only
216 wxCStrData(const wxString *str, size_t offset = 0, bool owned = false)
217 : m_str(str), m_offset(offset), m_owned(owned) {}
218
219public:
220 // Ctor constructs the object from char literal; they are needed to make
221 // operator?: compile and they intentionally take char*, not const char*
92a17abc
VS
222 inline wxCStrData(char *buf);
223 inline wxCStrData(wchar_t *buf);
224 inline wxCStrData(const wxCStrData& data);
c9f78968 225
9aee0061 226 inline ~wxCStrData();
c9f78968 227
05d676f5
VZ
228 // AsWChar() and AsChar() can't be defined here as they use wxString and so
229 // must come after it and because of this won't be inlined when called from
230 // wxString methods (without a lot of work to extract these wxString methods
231 // from inside the class itself). But we still define them being inline
232 // below to let compiler inline them from elsewhere. And because of this we
233 // must declare them as inline here because otherwise some compilers give
234 // warnings about them, e.g. mingw32 3.4.5 warns about "<symbol> defined
235 // locally after being referenced with dllimport linkage" while IRIX
236 // mipsPro 7.4 warns about "function declared inline after being called".
237 inline const wchar_t* AsWChar() const;
c9f78968 238 operator const wchar_t*() const { return AsWChar(); }
11aac4ba 239
05d676f5 240 inline const char* AsChar() const;
dbea442a
VZ
241 const unsigned char* AsUnsignedChar() const
242 { return (const unsigned char *) AsChar(); }
c9f78968 243 operator const char*() const { return AsChar(); }
dbea442a 244 operator const unsigned char*() const { return AsUnsignedChar(); }
11aac4ba
VS
245
246 operator const void*() const { return AsChar(); }
c9f78968 247
de4983f3 248 // returns buffers that are valid as long as the associated wxString exists
f54cb154
VZ
249 const wxScopedCharBuffer AsCharBuf() const
250 {
251 return wxScopedCharBuffer::CreateNonOwned(AsChar());
252 }
253
254 const wxScopedWCharBuffer AsWCharBuf() const
255 {
256 return wxScopedWCharBuffer::CreateNonOwned(AsWChar());
257 }
681e4412 258
9aee0061 259 inline wxString AsString() const;
c9f78968 260
2523e9b7
VS
261 // returns the value as C string in internal representation (equivalent
262 // to AsString().wx_str(), but more efficient)
263 const wxStringCharType *AsInternal() const;
264
c9f78968 265 // allow expressions like "c_str()[0]":
9aee0061 266 inline wxUniChar operator[](size_t n) const;
c9f78968 267 wxUniChar operator[](int n) const { return operator[](size_t(n)); }
c1df0a3b 268 wxUniChar operator[](long n) const { return operator[](size_t(n)); }
c9f78968
VS
269#ifndef wxSIZE_T_IS_UINT
270 wxUniChar operator[](unsigned int n) const { return operator[](size_t(n)); }
271#endif // size_t != unsigned int
272
52988b5f 273 // These operators are needed to emulate the pointer semantics of c_str():
c9f78968 274 // expressions like "wxChar *p = str.c_str() + 1;" should continue to work
52988b5f
VS
275 // (we need both versions to resolve ambiguities). Note that this means
276 // the 'n' value is interpreted as addition to char*/wchar_t* pointer, it
277 // is *not* number of Unicode characters in wxString.
c9f78968
VS
278 wxCStrData operator+(int n) const
279 { return wxCStrData(m_str, m_offset + n, m_owned); }
acf0c9ac
VZ
280 wxCStrData operator+(long n) const
281 { return wxCStrData(m_str, m_offset + n, m_owned); }
c9f78968
VS
282 wxCStrData operator+(size_t n) const
283 { return wxCStrData(m_str, m_offset + n, m_owned); }
284
b5343e06
VZ
285 // and these for "str.c_str() + (p2 - p1)" (it also works for any integer
286 // expression but it must be ptrdiff_t and not e.g. int to work in this
287 // example):
288 wxCStrData operator-(ptrdiff_t n) const
f4c90fdf 289 {
b5343e06 290 wxASSERT_MSG( n <= (ptrdiff_t)m_offset,
9a83f860 291 wxT("attempt to construct address before the beginning of the string") );
f4c90fdf
VS
292 return wxCStrData(m_str, m_offset - n, m_owned);
293 }
294
b77afb41 295 // this operator is needed to make expressions like "*c_str()" or
c9f78968 296 // "*(c_str() + 2)" work
9aee0061 297 inline wxUniChar operator*() const;
c9f78968
VS
298
299private:
52988b5f 300 // the wxString this object was returned for
c9f78968 301 const wxString *m_str;
52988b5f
VS
302 // Offset into c_str() return value. Note that this is *not* offset in
303 // m_str in Unicode characters. Instead, it is index into the
304 // char*/wchar_t* buffer returned by c_str(). It's interpretation depends
305 // on how is the wxCStrData instance used: if it is eventually cast to
306 // const char*, m_offset will be in bytes form string's start; if it is
307 // cast to const wchar_t*, it will be in wchar_t values.
c9f78968 308 size_t m_offset;
52988b5f 309 // should m_str be deleted, i.e. is it owned by us?
c9f78968
VS
310 bool m_owned;
311
b5dbe15d 312 friend class WXDLLIMPEXP_FWD_BASE wxString;
c9f78968
VS
313};
314
315// ----------------------------------------------------------------------------
316// wxStringPrintfMixin
317// ---------------------------------------------------------------------------
318
319// NB: VC6 has a bug that causes linker errors if you have template methods
320// in a class using __declspec(dllimport). The solution is to split such
321// class into two classes, one that contains the template methods and does
322// *not* use WXDLLIMPEXP_BASE and another class that contains the rest
323// (with DLL linkage).
324//
325// We only do this for VC6 here, because the code is less efficient
326// (Printf() has to use dynamic_cast<>) and because OpenWatcom compiler
327// cannot compile this code.
328
329#if defined(__VISUALC__) && __VISUALC__ < 1300
330 #define wxNEEDS_WXSTRING_PRINTF_MIXIN
331#endif
332
333#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
334// this class contains implementation of wxString's vararg methods, it's
335// exported from wxBase DLL
336class WXDLLIMPEXP_BASE wxStringPrintfMixinBase
337{
338protected:
339 wxStringPrintfMixinBase() {}
340
d1f6e2cf
VS
341#if !wxUSE_UTF8_LOCALE_ONLY
342 int DoPrintfWchar(const wxChar *format, ...);
343 static wxString DoFormatWchar(const wxChar *format, ...);
344#endif
345#if wxUSE_UNICODE_UTF8
346 int DoPrintfUtf8(const char *format, ...);
347 static wxString DoFormatUtf8(const char *format, ...);
348#endif
c9f78968
VS
349};
350
351// this class contains template wrappers for wxString's vararg methods, it's
352// intentionally *not* exported from the DLL in order to fix the VC6 bug
353// described above
354class wxStringPrintfMixin : public wxStringPrintfMixinBase
355{
356private:
357 // to further complicate things, we can't return wxString from
358 // wxStringPrintfMixin::Format() because wxString is not yet declared at
359 // this point; the solution is to use this fake type trait template - this
360 // way the compiler won't know the return type until Format() is used
361 // (this doesn't compile with Watcom, but VC6 compiles it just fine):
362 template<typename T> struct StringReturnType
363 {
364 typedef wxString type;
365 };
366
367public:
368 // these are duplicated wxString methods, they're also declared below
369 // if !wxNEEDS_WXSTRING_PRINTF_MIXIN:
370
862bb5c7 371 // static wxString Format(const wString& format, ...) WX_ATTRIBUTE_PRINTF_1;
d1f6e2cf 372 WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType<T1>::type,
1528e0b8 373 Format, 1, (const wxFormatString&),
d1f6e2cf 374 DoFormatWchar, DoFormatUtf8)
2523e9b7
VS
375 // We have to implement the version without template arguments manually
376 // because of the StringReturnType<> hack, although WX_DEFINE_VARARG_FUNC
377 // normally does it itself. It has to be a template so that we can use
81051829
VS
378 // the hack, even though there's no real template parameter. We can't move
379 // it to wxStrig, because it would shadow these versions of Format() then.
2523e9b7
VS
380 template<typename T>
381 inline static typename StringReturnType<T>::type
81051829 382 Format(const T& fmt)
2523e9b7 383 {
81051829
VS
384 // NB: this doesn't compile if T is not (some form of) a string;
385 // this makes Format's prototype equivalent to
386 // Format(const wxFormatString& fmt)
387 return DoFormatWchar(wxFormatString(fmt));
2523e9b7
VS
388 }
389
390 // int Printf(const wxString& format, ...);
1528e0b8 391 WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),
d1f6e2cf 392 DoPrintfWchar, DoPrintfUtf8)
862bb5c7 393 // int sprintf(const wxString& format, ...) WX_ATTRIBUTE_PRINTF_2;
1528e0b8 394 WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),
d1f6e2cf 395 DoPrintfWchar, DoPrintfUtf8)
c9f78968
VS
396
397protected:
398 wxStringPrintfMixin() : wxStringPrintfMixinBase() {}
399};
400#endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
401
402
b6339dde
VZ
403// ----------------------------------------------------------------------------
404// wxString: string class trying to be compatible with std::string, MFC
405// CString and wxWindows 1.x wxString all at once
e87b7833
MB
406// ---------------------------------------------------------------------------
407
c9f78968
VS
408#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
409 // "non dll-interface class 'wxStringPrintfMixin' used as base interface
410 // for dll-interface class 'wxString'" -- this is OK in our case
fb330c2e 411 #pragma warning (push)
c9f78968 412 #pragma warning (disable:4275)
e87b7833
MB
413#endif
414
b0c4d5d7
VS
415#if wxUSE_UNICODE_UTF8
416// see the comment near wxString::iterator for why we need this
326a863a 417class WXDLLIMPEXP_BASE wxStringIteratorNode
b0c4d5d7 418{
326a863a 419public:
39c20230
VS
420 wxStringIteratorNode()
421 : m_str(NULL), m_citer(NULL), m_iter(NULL), m_prev(NULL), m_next(NULL) {}
422 wxStringIteratorNode(const wxString *str,
423 wxStringImpl::const_iterator *citer)
424 { DoSet(str, citer, NULL); }
425 wxStringIteratorNode(const wxString *str, wxStringImpl::iterator *iter)
426 { DoSet(str, NULL, iter); }
427 ~wxStringIteratorNode()
428 { clear(); }
429
430 inline void set(const wxString *str, wxStringImpl::const_iterator *citer)
431 { clear(); DoSet(str, citer, NULL); }
432 inline void set(const wxString *str, wxStringImpl::iterator *iter)
433 { clear(); DoSet(str, NULL, iter); }
b0c4d5d7
VS
434
435 const wxString *m_str;
436 wxStringImpl::const_iterator *m_citer;
437 wxStringImpl::iterator *m_iter;
438 wxStringIteratorNode *m_prev, *m_next;
300b44a9 439
39c20230
VS
440private:
441 inline void clear();
442 inline void DoSet(const wxString *str,
443 wxStringImpl::const_iterator *citer,
444 wxStringImpl::iterator *iter);
445
300b44a9
VS
446 // the node belongs to a particular iterator instance, it's not copied
447 // when a copy of the iterator is made
c0c133e1 448 wxDECLARE_NO_COPY_CLASS(wxStringIteratorNode);
b0c4d5d7
VS
449};
450#endif // wxUSE_UNICODE_UTF8
451
8f93a29f 452class WXDLLIMPEXP_BASE wxString
c9f78968 453#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
8f93a29f 454 : public wxStringPrintfMixin
c9f78968
VS
455#endif
456{
e87b7833
MB
457 // NB: special care was taken in arranging the member functions in such order
458 // that all inline functions can be effectively inlined, verify that all
20aa779e 459 // performance critical functions are still inlined if you change order!
8f93a29f 460public:
8f93a29f
VS
461 // an 'invalid' value for string index, moved to this place due to a CW bug
462 static const size_t npos;
8f93a29f 463
e87b7833
MB
464private:
465 // if we hadn't made these operators private, it would be possible to
466 // compile "wxString s; s = 17;" without any warnings as 17 is implicitly
467 // converted to char in C and we do have operator=(char)
468 //
469 // NB: we don't need other versions (short/long and unsigned) as attempt
470 // to assign another numeric type to wxString will now result in
471 // ambiguity between operator=(char) and operator=(int)
472 wxString& operator=(int);
473
8f93a29f
VS
474 // these methods are not implemented - there is _no_ conversion from int to
475 // string, you're doing something wrong if the compiler wants to call it!
476 //
477 // try `s << i' or `s.Printf("%d", i)' instead
478 wxString(int);
479
480
481 // buffer for holding temporary substring when using any of the methods
482 // that take (char*,size_t) or (wchar_t*,size_t) arguments:
8f93a29f
VS
483 template<typename T>
484 struct SubstrBufFromType
485 {
486 T data;
487 size_t len;
488
8f93a29f 489 SubstrBufFromType(const T& data_, size_t len_)
9ef1ad0d
VZ
490 : data(data_), len(len_)
491 {
492 wxASSERT_MSG( len != npos, "must have real length" );
493 }
8f93a29f
VS
494 };
495
496#if wxUSE_UNICODE_UTF8
81727065 497 // even char* -> char* needs conversion, from locale charset to UTF-8
de4983f3
VS
498 typedef SubstrBufFromType<wxScopedCharBuffer> SubstrBufFromWC;
499 typedef SubstrBufFromType<wxScopedCharBuffer> SubstrBufFromMB;
8f93a29f 500#elif wxUSE_UNICODE_WCHAR
de4983f3
VS
501 typedef SubstrBufFromType<const wchar_t*> SubstrBufFromWC;
502 typedef SubstrBufFromType<wxScopedWCharBuffer> SubstrBufFromMB;
8f93a29f 503#else
de4983f3
VS
504 typedef SubstrBufFromType<const char*> SubstrBufFromMB;
505 typedef SubstrBufFromType<wxScopedCharBuffer> SubstrBufFromWC;
8f93a29f
VS
506#endif
507
508
509 // Functions implementing primitive operations on string data; wxString
510 // methods and iterators are implemented in terms of it. The differences
511 // between UTF-8 and wchar_t* representations of the string are mostly
512 // contained here.
513
81727065
VS
514#if wxUSE_UNICODE_UTF8
515 static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength,
516 const wxMBConv& conv);
517 static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength,
518 const wxMBConv& conv);
519#elif wxUSE_UNICODE_WCHAR
8f93a29f 520 static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength,
a962cdf4 521 const wxMBConv& conv);
8f93a29f
VS
522#else
523 static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength,
a962cdf4 524 const wxMBConv& conv);
8f93a29f
VS
525#endif
526
527#if !wxUSE_UNICODE_UTF8 // wxUSE_UNICODE_WCHAR or !wxUSE_UNICODE
a962cdf4 528 // returns C string encoded as the implementation expects:
8f93a29f 529 #if wxUSE_UNICODE
a962cdf4 530 static const wchar_t* ImplStr(const wchar_t* str)
e8f59039 531 { return str ? str : wxT(""); }
a962cdf4 532 static const SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
e8f59039 533 { return SubstrBufFromWC(str, (str && n == npos) ? wxWcslen(str) : n); }
de4983f3
VS
534 static wxScopedWCharBuffer ImplStr(const char* str,
535 const wxMBConv& conv = wxConvLibc)
e8f59039
VS
536 { return ConvertStr(str, npos, conv).data; }
537 static SubstrBufFromMB ImplStr(const char* str, size_t n,
538 const wxMBConv& conv = wxConvLibc)
539 { return ConvertStr(str, n, conv); }
8f93a29f 540 #else
e8f59039
VS
541 static const char* ImplStr(const char* str,
542 const wxMBConv& WXUNUSED(conv) = wxConvLibc)
543 { return str ? str : ""; }
544 static const SubstrBufFromMB ImplStr(const char* str, size_t n,
545 const wxMBConv& WXUNUSED(conv) = wxConvLibc)
546 { return SubstrBufFromMB(str, (str && n == npos) ? wxStrlen(str) : n); }
de4983f3 547 static wxScopedCharBuffer ImplStr(const wchar_t* str)
8f93a29f
VS
548 { return ConvertStr(str, npos, wxConvLibc).data; }
549 static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
550 { return ConvertStr(str, n, wxConvLibc); }
551 #endif
552
8f93a29f
VS
553 // translates position index in wxString to/from index in underlying
554 // wxStringImpl:
555 static size_t PosToImpl(size_t pos) { return pos; }
556 static void PosLenToImpl(size_t pos, size_t len,
557 size_t *implPos, size_t *implLen)
558 { *implPos = pos; *implLen = len; }
11aac4ba 559 static size_t LenToImpl(size_t len) { return len; }
8f93a29f
VS
560 static size_t PosFromImpl(size_t pos) { return pos; }
561
68482dc5
VZ
562 // we don't want to define these as empty inline functions as it could
563 // result in noticeable (and quite unnecessary in non-UTF-8 build) slowdown
564 // in debug build where the inline functions are not effectively inlined
565 #define wxSTRING_INVALIDATE_CACHE()
566 #define wxSTRING_INVALIDATE_CACHED_LENGTH()
567 #define wxSTRING_UPDATE_CACHED_LENGTH(n)
568 #define wxSTRING_SET_CACHED_LENGTH(n)
569
8f93a29f
VS
570#else // wxUSE_UNICODE_UTF8
571
de4983f3
VS
572 static wxScopedCharBuffer ImplStr(const char* str,
573 const wxMBConv& conv = wxConvLibc)
467175ab
VS
574 { return ConvertStr(str, npos, conv).data; }
575 static SubstrBufFromMB ImplStr(const char* str, size_t n,
576 const wxMBConv& conv = wxConvLibc)
577 { return ConvertStr(str, n, conv); }
81727065 578
de4983f3 579 static wxScopedCharBuffer ImplStr(const wchar_t* str)
5487ff0f 580 { return ConvertStr(str, npos, wxMBConvUTF8()).data; }
467175ab 581 static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
5487ff0f 582 { return ConvertStr(str, n, wxMBConvUTF8()); }
81727065 583
68482dc5
VZ
584#if wxUSE_STRING_POS_CACHE
585 // this is an extremely simple cache used by PosToImpl(): each cache element
586 // contains the string it applies to and the index corresponding to the last
587 // used position in this wxString in its m_impl string
588 //
589 // NB: notice that this struct (and nested Element one) must be a POD or we
590 // wouldn't be able to use a thread-local variable of this type, in
591 // particular it should have no ctor -- we rely on statics being
592 // initialized to 0 instead
593 struct Cache
594 {
595 enum { SIZE = 8 };
596
597 struct Element
598 {
599 const wxString *str; // the string to which this element applies
600 size_t pos, // the cached index in this string
601 impl, // the corresponding position in its m_impl
602 len; // cached length or npos if unknown
603
604 // reset cached index to 0
605 void ResetPos() { pos = impl = 0; }
606
607 // reset position and length
608 void Reset() { ResetPos(); len = npos; }
609 };
610
611 // cache the indices mapping for the last few string used
612 Element cached[SIZE];
613
614 // the last used index
615 unsigned lastUsed;
616 };
617
e810df36
VZ
618#ifndef wxHAS_COMPILER_TLS
619 // we must use an accessor function and not a static variable when the TLS
620 // variables support is implemented in the library (and not by the compiler)
621 // because the global s_cache variable could be not yet initialized when a
622 // ctor of another global object is executed and if that ctor uses any
623 // wxString methods, bad things happen
624 //
625 // however notice that this approach does not work when compiler TLS is used,
626 // at least not with g++ 4.1.2 under amd64 as it apparently compiles code
627 // using this accessor incorrectly when optimizations are enabled (-O2 is
628 // enough) -- luckily we don't need it then neither as static __thread
629 // variables are initialized by 0 anyhow then and so we can use the variable
630 // directly
e317bd3f 631 WXEXPORT static Cache& GetCache()
8b73c531
VZ
632 {
633 static wxTLS_TYPE(Cache) s_cache;
634
635 return wxTLS_VALUE(s_cache);
636 }
cbec0f40 637
ad8ae788
VZ
638 // this helper struct is used to ensure that GetCache() is called during
639 // static initialization time, i.e. before any threads creation, as otherwise
640 // the static s_cache construction inside GetCache() wouldn't be MT-safe
641 friend struct wxStrCacheInitializer;
e810df36
VZ
642#else // wxHAS_COMPILER_TLS
643 static wxTLS_TYPE(Cache) ms_cache;
644 static Cache& GetCache() { return wxTLS_VALUE(ms_cache); }
645#endif // !wxHAS_COMPILER_TLS/wxHAS_COMPILER_TLS
646
647 static Cache::Element *GetCacheBegin() { return GetCache().cached; }
648 static Cache::Element *GetCacheEnd() { return GetCacheBegin() + Cache::SIZE; }
649 static unsigned& LastUsedCacheElement() { return GetCache().lastUsed; }
ad8ae788
VZ
650
651 // this is used in debug builds only to provide a convenient function,
652 // callable from a debugger, to show the cache contents
68482dc5
VZ
653 friend struct wxStrCacheDumper;
654
655 // uncomment this to have access to some profiling statistics on program
656 // termination
657 //#define wxPROFILE_STRING_CACHE
658
659#ifdef wxPROFILE_STRING_CACHE
660 static struct PosToImplCacheStats
661 {
662 unsigned postot, // total non-trivial calls to PosToImpl
663 poshits, // cache hits from PosToImpl()
664 mishits, // cached position beyond the needed one
665 sumpos, // sum of all positions, used to compute the
666 // average position after dividing by postot
667 sumofs, // sum of all offsets after using the cache, used to
668 // compute the average after dividing by hits
669 lentot, // number of total calls to length()
670 lenhits; // number of cache hits in length()
671 } ms_cacheStats;
672
8c3b65d9 673 friend struct wxStrCacheStatsDumper;
68482dc5
VZ
674
675 #define wxCACHE_PROFILE_FIELD_INC(field) ms_cacheStats.field++
676 #define wxCACHE_PROFILE_FIELD_ADD(field, val) ms_cacheStats.field += (val)
677#else // !wxPROFILE_STRING_CACHE
678 #define wxCACHE_PROFILE_FIELD_INC(field)
679 #define wxCACHE_PROFILE_FIELD_ADD(field, val)
680#endif // wxPROFILE_STRING_CACHE/!wxPROFILE_STRING_CACHE
681
682 // note: it could seem that the functions below shouldn't be inline because
683 // they are big, contain loops and so the compiler shouldn't be able to
684 // inline them anyhow, however moving them into string.cpp does decrease the
685 // code performance by ~5%, at least when using g++ 4.1 so do keep them here
686 // unless tests show that it's not advantageous any more
687
688 // return the pointer to the cache element for this string or NULL if not
689 // cached
690 Cache::Element *FindCacheElement() const
691 {
692 // profiling seems to show a small but consistent gain if we use this
693 // simple loop instead of starting from the last used element (there are
694 // a lot of misses in this function...)
e317bd3f 695 Cache::Element * const cacheBegin = GetCacheBegin();
6e42b980 696#ifndef wxHAS_COMPILER_TLS
cbec0f40 697 // during destruction tls calls may return NULL, in this case return NULL
e317bd3f
SC
698 // immediately without accessing anything else
699 if ( cacheBegin == NULL )
c1158f7b 700 return NULL;
e317bd3f
SC
701#endif
702 Cache::Element * const cacheEnd = GetCacheEnd();
703 for ( Cache::Element *c = cacheBegin; c != cacheEnd; c++ )
68482dc5
VZ
704 {
705 if ( c->str == this )
706 return c;
707 }
708
709 return NULL;
710 }
711
712 // unlike FindCacheElement(), this one always returns a valid pointer to the
713 // cache element for this string, it may have valid last cached position and
714 // its corresponding index in the byte string or not
715 Cache::Element *GetCacheElement() const
716 {
8b73c531
VZ
717 Cache::Element * const cacheBegin = GetCacheBegin();
718 Cache::Element * const cacheEnd = GetCacheEnd();
719 Cache::Element * const cacheStart = cacheBegin + LastUsedCacheElement();
68482dc5
VZ
720
721 // check the last used first, this does no (measurable) harm for a miss
722 // but does help for simple loops addressing the same string all the time
723 if ( cacheStart->str == this )
724 return cacheStart;
725
726 // notice that we're going to check cacheStart again inside this call but
727 // profiling shows that it's still faster to use a simple loop like
728 // inside FindCacheElement() than manually looping with wrapping starting
729 // from the cache entry after the start one
730 Cache::Element *c = FindCacheElement();
731 if ( !c )
732 {
733 // claim the next cache entry for this string
734 c = cacheStart;
735 if ( ++c == cacheEnd )
736 c = cacheBegin;
737
738 c->str = this;
739 c->Reset();
740
741 // and remember the last used element
8b73c531 742 LastUsedCacheElement() = c - cacheBegin;
68482dc5
VZ
743 }
744
745 return c;
746 }
747
748 size_t DoPosToImpl(size_t pos) const
749 {
750 wxCACHE_PROFILE_FIELD_INC(postot);
751
752 // NB: although the case of pos == 1 (and offset from cached position
753 // equal to 1) are common, nothing is gained by writing special code
754 // for handling them, the compiler (at least g++ 4.1 used) seems to
755 // optimize the code well enough on its own
756
757 wxCACHE_PROFILE_FIELD_ADD(sumpos, pos);
758
759 Cache::Element * const cache = GetCacheElement();
760
761 // cached position can't be 0 so if it is, it means that this entry was
762 // used for length caching only so far, i.e. it doesn't count as a hit
763 // from our point of view
764 if ( cache->pos )
1224b8a3 765 {
68482dc5 766 wxCACHE_PROFILE_FIELD_INC(poshits);
1224b8a3 767 }
68482dc5
VZ
768
769 if ( pos == cache->pos )
770 return cache->impl;
771
772 // this seems to happen only rarely so just reset the cache in this case
773 // instead of complicating code even further by seeking backwards in this
774 // case
775 if ( cache->pos > pos )
776 {
777 wxCACHE_PROFILE_FIELD_INC(mishits);
778
779 cache->ResetPos();
780 }
781
782 wxCACHE_PROFILE_FIELD_ADD(sumofs, pos - cache->pos);
783
784
785 wxStringImpl::const_iterator i(m_impl.begin() + cache->impl);
786 for ( size_t n = cache->pos; n < pos; n++ )
787 wxStringOperations::IncIter(i);
788
789 cache->pos = pos;
790 cache->impl = i - m_impl.begin();
791
792 wxSTRING_CACHE_ASSERT(
793 (int)cache->impl == (begin() + pos).impl() - m_impl.begin() );
794
795 return cache->impl;
796 }
797
798 void InvalidateCache()
799 {
800 Cache::Element * const cache = FindCacheElement();
801 if ( cache )
802 cache->Reset();
803 }
804
805 void InvalidateCachedLength()
806 {
807 Cache::Element * const cache = FindCacheElement();
808 if ( cache )
809 cache->len = npos;
810 }
811
812 void SetCachedLength(size_t len)
813 {
814 // we optimistically cache the length here even if the string wasn't
815 // present in the cache before, this seems to do no harm and the
816 // potential for avoiding length recomputation for long strings looks
817 // interesting
818 GetCacheElement()->len = len;
819 }
820
821 void UpdateCachedLength(ptrdiff_t delta)
822 {
823 Cache::Element * const cache = FindCacheElement();
824 if ( cache && cache->len != npos )
825 {
826 wxSTRING_CACHE_ASSERT( (ptrdiff_t)cache->len + delta >= 0 );
827
828 cache->len += delta;
829 }
830 }
831
832 #define wxSTRING_INVALIDATE_CACHE() InvalidateCache()
833 #define wxSTRING_INVALIDATE_CACHED_LENGTH() InvalidateCachedLength()
834 #define wxSTRING_UPDATE_CACHED_LENGTH(n) UpdateCachedLength(n)
835 #define wxSTRING_SET_CACHED_LENGTH(n) SetCachedLength(n)
836#else // !wxUSE_STRING_POS_CACHE
837 size_t DoPosToImpl(size_t pos) const
838 {
839 return (begin() + pos).impl() - m_impl.begin();
840 }
841
842 #define wxSTRING_INVALIDATE_CACHE()
843 #define wxSTRING_INVALIDATE_CACHED_LENGTH()
844 #define wxSTRING_UPDATE_CACHED_LENGTH(n)
845 #define wxSTRING_SET_CACHED_LENGTH(n)
846#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
847
8f93a29f
VS
848 size_t PosToImpl(size_t pos) const
849 {
68482dc5 850 return pos == 0 || pos == npos ? pos : DoPosToImpl(pos);
64a044d5
VZ
851 }
852
81727065
VS
853 void PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const;
854
855 size_t LenToImpl(size_t len) const
856 {
857 size_t pos, len2;
858 PosLenToImpl(0, len, &pos, &len2);
859 return len2;
860 }
861
8f93a29f
VS
862 size_t PosFromImpl(size_t pos) const
863 {
864 if ( pos == 0 || pos == npos )
865 return pos;
866 else
b0c4d5d7 867 return const_iterator(this, m_impl.begin() + pos) - begin();
8f93a29f 868 }
8f93a29f
VS
869#endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8
870
8f93a29f
VS
871public:
872 // standard types
873 typedef wxUniChar value_type;
874 typedef wxUniChar char_type;
875 typedef wxUniCharRef reference;
876 typedef wxChar* pointer;
877 typedef const wxChar* const_pointer;
878
879 typedef size_t size_type;
880 typedef wxUniChar const_reference;
881
302acc45 882#if wxUSE_STD_STRING
81727065
VS
883 #if wxUSE_UNICODE_UTF8
884 // random access is not O(1), as required by Random Access Iterator
885 #define WX_STR_ITERATOR_TAG std::bidirectional_iterator_tag
886 #else
887 #define WX_STR_ITERATOR_TAG std::random_access_iterator_tag
888 #endif
0514adaf 889 #define WX_DEFINE_ITERATOR_CATEGORY(cat) typedef cat iterator_category;
a962cdf4 890#else
302acc45
VZ
891 // not defining iterator_category at all in this case is better than defining
892 // it as some dummy type -- at least it results in more intelligible error
893 // messages
0514adaf 894 #define WX_DEFINE_ITERATOR_CATEGORY(cat)
a962cdf4
VS
895#endif
896
b0c4d5d7 897 #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \
8f93a29f
VS
898 private: \
899 typedef wxStringImpl::iterator_name underlying_iterator; \
900 public: \
0514adaf 901 WX_DEFINE_ITERATOR_CATEGORY(WX_STR_ITERATOR_TAG) \
8f93a29f 902 typedef wxUniChar value_type; \
a962cdf4 903 typedef int difference_type; \
8f93a29f
VS
904 typedef reference_type reference; \
905 typedef pointer_type pointer; \
906 \
a962cdf4 907 reference operator[](size_t n) const { return *(*this + n); } \
8f93a29f
VS
908 \
909 iterator_name& operator++() \
467175ab 910 { wxStringOperations::IncIter(m_cur); return *this; } \
8f93a29f 911 iterator_name& operator--() \
467175ab 912 { wxStringOperations::DecIter(m_cur); return *this; } \
8f93a29f
VS
913 iterator_name operator++(int) \
914 { \
915 iterator_name tmp = *this; \
467175ab 916 wxStringOperations::IncIter(m_cur); \
8f93a29f
VS
917 return tmp; \
918 } \
919 iterator_name operator--(int) \
920 { \
921 iterator_name tmp = *this; \
467175ab 922 wxStringOperations::DecIter(m_cur); \
8f93a29f
VS
923 return tmp; \
924 } \
925 \
b5343e06 926 iterator_name& operator+=(ptrdiff_t n) \
467175ab
VS
927 { \
928 m_cur = wxStringOperations::AddToIter(m_cur, n); \
929 return *this; \
930 } \
b5343e06 931 iterator_name& operator-=(ptrdiff_t n) \
467175ab
VS
932 { \
933 m_cur = wxStringOperations::AddToIter(m_cur, -n); \
934 return *this; \
467175ab 935 } \
8f93a29f 936 \
89360a8c 937 difference_type operator-(const iterator_name& i) const \
467175ab 938 { return wxStringOperations::DiffIters(m_cur, i.m_cur); } \
8f93a29f 939 \
f2a1b1bd 940 bool operator==(const iterator_name& i) const \
8f93a29f
VS
941 { return m_cur == i.m_cur; } \
942 bool operator!=(const iterator_name& i) const \
943 { return m_cur != i.m_cur; } \
944 \
945 bool operator<(const iterator_name& i) const \
946 { return m_cur < i.m_cur; } \
947 bool operator>(const iterator_name& i) const \
948 { return m_cur > i.m_cur; } \
949 bool operator<=(const iterator_name& i) const \
950 { return m_cur <= i.m_cur; } \
951 bool operator>=(const iterator_name& i) const \
952 { return m_cur >= i.m_cur; } \
953 \
954 private: \
955 /* for internal wxString use only: */ \
cf9a878b 956 underlying_iterator impl() const { return m_cur; } \
8f93a29f 957 \
b5dbe15d
VS
958 friend class wxString; \
959 friend class wxCStrData; \
8f93a29f
VS
960 \
961 private: \
89360a8c 962 underlying_iterator m_cur
8f93a29f 963
b5dbe15d 964 class WXDLLIMPEXP_FWD_BASE const_iterator;
8f93a29f 965
81727065 966#if wxUSE_UNICODE_UTF8
b0c4d5d7
VS
967 // NB: In UTF-8 build, (non-const) iterator needs to keep reference
968 // to the underlying wxStringImpl, because UTF-8 is variable-length
969 // encoding and changing the value pointer to by an iterator (using
970 // its operator*) requires calling wxStringImpl::replace() if the old
971 // and new values differ in their encoding's length.
972 //
973 // Furthermore, the replace() call may invalid all iterators for the
974 // string, so we have to keep track of outstanding iterators and update
975 // them if replace() happens.
976 //
977 // This is implemented by maintaining linked list of iterators for every
978 // string and traversing it in wxUniCharRef::operator=(). Head of the
979 // list is stored in wxString. (FIXME-UTF8)
980
541aa821 981 class WXDLLIMPEXP_BASE iterator
81727065 982 {
b0c4d5d7 983 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
c7dc0057 984
c67b782d 985 public:
39c20230 986 iterator() {}
b0c4d5d7
VS
987 iterator(const iterator& i)
988 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
300b44a9 989 iterator& operator=(const iterator& i)
162e998c
PC
990 {
991 if (&i != this)
992 {
993 m_cur = i.m_cur;
994 m_node.set(i.str(), &m_cur);
995 }
996 return *this;
997 }
b0c4d5d7
VS
998
999 reference operator*()
6bd4f281 1000 { return wxUniCharRef::CreateForString(*str(), m_cur); }
81727065 1001
b5343e06 1002 iterator operator+(ptrdiff_t n) const
b0c4d5d7 1003 { return iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
b5343e06 1004 iterator operator-(ptrdiff_t n) const
b0c4d5d7 1005 { return iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
81727065
VS
1006
1007 private:
f6363458
VZ
1008 iterator(wxString *wxstr, underlying_iterator ptr)
1009 : m_cur(ptr), m_node(wxstr, &m_cur) {}
c7dc0057 1010
5c33522f 1011 wxString* str() const { return const_cast<wxString*>(m_node.m_str); }
b0c4d5d7
VS
1012
1013 wxStringIteratorNode m_node;
81727065
VS
1014
1015 friend class const_iterator;
1016 };
cf9a878b 1017
541aa821 1018 class WXDLLIMPEXP_BASE const_iterator
b0c4d5d7
VS
1019 {
1020 // NB: reference_type is intentionally value, not reference, the character
1021 // may be encoded differently in wxString data:
1022 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
1023
1024 public:
39c20230 1025 const_iterator() {}
b0c4d5d7
VS
1026 const_iterator(const const_iterator& i)
1027 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
1028 const_iterator(const iterator& i)
1029 : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
1030
300b44a9 1031 const_iterator& operator=(const const_iterator& i)
162e998c
PC
1032 {
1033 if (&i != this)
1034 {
1035 m_cur = i.m_cur;
1036 m_node.set(i.str(), &m_cur);
1037 }
1038 return *this;
1039 }
300b44a9 1040 const_iterator& operator=(const iterator& i)
39c20230 1041 { m_cur = i.m_cur; m_node.set(i.str(), &m_cur); return *this; }
300b44a9 1042
b0c4d5d7
VS
1043 reference operator*() const
1044 { return wxStringOperations::DecodeChar(m_cur); }
1045
b5343e06 1046 const_iterator operator+(ptrdiff_t n) const
b0c4d5d7 1047 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
b5343e06 1048 const_iterator operator-(ptrdiff_t n) const
b0c4d5d7 1049 { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
b0c4d5d7
VS
1050
1051 private:
1052 // for internal wxString use only:
f6363458
VZ
1053 const_iterator(const wxString *wxstr, underlying_iterator ptr)
1054 : m_cur(ptr), m_node(wxstr, &m_cur) {}
b0c4d5d7
VS
1055
1056 const wxString* str() const { return m_node.m_str; }
1057
1058 wxStringIteratorNode m_node;
1059 };
1060
cf9a878b
VS
1061 size_t IterToImplPos(wxString::iterator i) const
1062 { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); }
1063
68482dc5
VZ
1064 iterator GetIterForNthChar(size_t n)
1065 { return iterator(this, m_impl.begin() + PosToImpl(n)); }
1066 const_iterator GetIterForNthChar(size_t n) const
1067 { return const_iterator(this, m_impl.begin() + PosToImpl(n)); }
81727065 1068#else // !wxUSE_UNICODE_UTF8
cf9a878b 1069
541aa821 1070 class WXDLLIMPEXP_BASE iterator
8f93a29f 1071 {
b0c4d5d7 1072 WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);
8f93a29f 1073
81727065 1074 public:
39c20230 1075 iterator() {}
81727065
VS
1076 iterator(const iterator& i) : m_cur(i.m_cur) {}
1077
b0c4d5d7
VS
1078 reference operator*()
1079 { return wxUniCharRef::CreateForString(m_cur); }
1080
b5343e06 1081 iterator operator+(ptrdiff_t n) const
467175ab 1082 { return iterator(wxStringOperations::AddToIter(m_cur, n)); }
b5343e06 1083 iterator operator-(ptrdiff_t n) const
467175ab 1084 { return iterator(wxStringOperations::AddToIter(m_cur, -n)); }
81727065
VS
1085
1086 private:
1087 // for internal wxString use only:
1088 iterator(underlying_iterator ptr) : m_cur(ptr) {}
1089 iterator(wxString *WXUNUSED(str), underlying_iterator ptr) : m_cur(ptr) {}
1090
8f93a29f
VS
1091 friend class const_iterator;
1092 };
1093
541aa821 1094 class WXDLLIMPEXP_BASE const_iterator
8f93a29f
VS
1095 {
1096 // NB: reference_type is intentionally value, not reference, the character
1097 // may be encoded differently in wxString data:
b0c4d5d7 1098 WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
8f93a29f
VS
1099
1100 public:
39c20230 1101 const_iterator() {}
81727065 1102 const_iterator(const const_iterator& i) : m_cur(i.m_cur) {}
8f93a29f 1103 const_iterator(const iterator& i) : m_cur(i.m_cur) {}
81727065 1104
b0c4d5d7
VS
1105 reference operator*() const
1106 { return wxStringOperations::DecodeChar(m_cur); }
1107
b5343e06 1108 const_iterator operator+(ptrdiff_t n) const
467175ab 1109 { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); }
b5343e06 1110 const_iterator operator-(ptrdiff_t n) const
467175ab 1111 { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); }
81727065
VS
1112
1113 private:
1114 // for internal wxString use only:
1115 const_iterator(underlying_iterator ptr) : m_cur(ptr) {}
b0c4d5d7
VS
1116 const_iterator(const wxString *WXUNUSED(str), underlying_iterator ptr)
1117 : m_cur(ptr) {}
8f93a29f 1118 };
68482dc5
VZ
1119
1120 iterator GetIterForNthChar(size_t n) { return begin() + n; }
1121 const_iterator GetIterForNthChar(size_t n) const { return begin() + n; }
b0c4d5d7 1122#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8
8f93a29f 1123
a962cdf4 1124 #undef WX_STR_ITERATOR_TAG
8f93a29f
VS
1125 #undef WX_STR_ITERATOR_IMPL
1126
1127 friend class iterator;
1128 friend class const_iterator;
1129
1130 template <typename T>
1131 class reverse_iterator_impl
1132 {
1133 public:
1134 typedef T iterator_type;
a962cdf4 1135
0514adaf 1136 WX_DEFINE_ITERATOR_CATEGORY(typename T::iterator_category)
8f93a29f 1137 typedef typename T::value_type value_type;
a962cdf4 1138 typedef typename T::difference_type difference_type;
8f93a29f
VS
1139 typedef typename T::reference reference;
1140 typedef typename T::pointer *pointer;
1141
39c20230 1142 reverse_iterator_impl() {}
8f93a29f
VS
1143 reverse_iterator_impl(iterator_type i) : m_cur(i) {}
1144 reverse_iterator_impl(const reverse_iterator_impl& ri)
1145 : m_cur(ri.m_cur) {}
1146
1147 iterator_type base() const { return m_cur; }
1148
1149 reference operator*() const { return *(m_cur-1); }
a962cdf4 1150 reference operator[](size_t n) const { return *(*this + n); }
8f93a29f
VS
1151
1152 reverse_iterator_impl& operator++()
1153 { --m_cur; return *this; }
1154 reverse_iterator_impl operator++(int)
1155 { reverse_iterator_impl tmp = *this; --m_cur; return tmp; }
1156 reverse_iterator_impl& operator--()
1157 { ++m_cur; return *this; }
1158 reverse_iterator_impl operator--(int)
1159 { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; }
1160
50eeb96f 1161 // NB: explicit <T> in the functions below is to keep BCC 5.5 happy
b5343e06 1162 reverse_iterator_impl operator+(ptrdiff_t n) const
50eeb96f 1163 { return reverse_iterator_impl<T>(m_cur - n); }
b5343e06 1164 reverse_iterator_impl operator-(ptrdiff_t n) const
50eeb96f 1165 { return reverse_iterator_impl<T>(m_cur + n); }
b5343e06 1166 reverse_iterator_impl operator+=(ptrdiff_t n)
8f93a29f 1167 { m_cur -= n; return *this; }
b5343e06 1168 reverse_iterator_impl operator-=(ptrdiff_t n)
8f93a29f
VS
1169 { m_cur += n; return *this; }
1170
1171 unsigned operator-(const reverse_iterator_impl& i) const
1172 { return i.m_cur - m_cur; }
1173
1174 bool operator==(const reverse_iterator_impl& ri) const
1175 { return m_cur == ri.m_cur; }
1176 bool operator!=(const reverse_iterator_impl& ri) const
1177 { return !(*this == ri); }
1178
1179 bool operator<(const reverse_iterator_impl& i) const
1180 { return m_cur > i.m_cur; }
1181 bool operator>(const reverse_iterator_impl& i) const
1182 { return m_cur < i.m_cur; }
1183 bool operator<=(const reverse_iterator_impl& i) const
1184 { return m_cur >= i.m_cur; }
1185 bool operator>=(const reverse_iterator_impl& i) const
1186 { return m_cur <= i.m_cur; }
1187
1188 private:
1189 iterator_type m_cur;
1190 };
e87b7833 1191
8f93a29f
VS
1192 typedef reverse_iterator_impl<iterator> reverse_iterator;
1193 typedef reverse_iterator_impl<const_iterator> const_reverse_iterator;
e90c1d2a 1194
67cff9dc
VZ
1195private:
1196 // used to transform an expression built using c_str() (and hence of type
1197 // wxCStrData) to an iterator into the string
1198 static const_iterator CreateConstIterator(const wxCStrData& data)
1199 {
b0c4d5d7
VS
1200 return const_iterator(data.m_str,
1201 (data.m_str->begin() + data.m_offset).impl());
67cff9dc
VZ
1202 }
1203
59953bf4
VS
1204 // in UTF-8 STL build, creation from std::string requires conversion under
1205 // non-UTF8 locales, so we can't have and use wxString(wxStringImpl) ctor;
1206 // instead we define dummy type that lets us have wxString ctor for creation
1207 // from wxStringImpl that couldn't be used by user code (in all other builds,
1208 // "standard" ctors can be used):
1209#if wxUSE_UNICODE_UTF8 && wxUSE_STL_BASED_WXSTRING
1210 struct CtorFromStringImplTag {};
1211
1212 wxString(CtorFromStringImplTag* WXUNUSED(dummy), const wxStringImpl& src)
1213 : m_impl(src) {}
1214
1215 static wxString FromImpl(const wxStringImpl& src)
1216 { return wxString((CtorFromStringImplTag*)NULL, src); }
1217#else
82a99c69 1218 #if !wxUSE_STL_BASED_WXSTRING
59953bf4 1219 wxString(const wxStringImpl& src) : m_impl(src) { }
82a99c69
VS
1220 // else: already defined as wxString(wxStdString) below
1221 #endif
59953bf4
VS
1222 static wxString FromImpl(const wxStringImpl& src) { return wxString(src); }
1223#endif
1224
67cff9dc
VZ
1225public:
1226 // constructors and destructor
1227 // ctor for an empty string
1228 wxString() {}
1229
1230 // copy ctor
67cff9dc
VZ
1231 wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { }
1232
1233 // string containing nRepeat copies of ch
623633ee 1234 wxString(wxUniChar ch, size_t nRepeat = 1 )
67cff9dc
VZ
1235 { assign(nRepeat, ch); }
1236 wxString(size_t nRepeat, wxUniChar ch)
1237 { assign(nRepeat, ch); }
1238 wxString(wxUniCharRef ch, size_t nRepeat = 1)
1239 { assign(nRepeat, ch); }
1240 wxString(size_t nRepeat, wxUniCharRef ch)
1241 { assign(nRepeat, ch); }
1242 wxString(char ch, size_t nRepeat = 1)
1243 { assign(nRepeat, ch); }
1244 wxString(size_t nRepeat, char ch)
1245 { assign(nRepeat, ch); }
1246 wxString(wchar_t ch, size_t nRepeat = 1)
1247 { assign(nRepeat, ch); }
1248 wxString(size_t nRepeat, wchar_t ch)
1249 { assign(nRepeat, ch); }
1250
1251 // ctors from char* strings:
1252 wxString(const char *psz)
1253 : m_impl(ImplStr(psz)) {}
1254 wxString(const char *psz, const wxMBConv& conv)
1255 : m_impl(ImplStr(psz, conv)) {}
1256 wxString(const char *psz, size_t nLength)
1257 { assign(psz, nLength); }
1258 wxString(const char *psz, const wxMBConv& conv, size_t nLength)
1259 {
1260 SubstrBufFromMB str(ImplStr(psz, nLength, conv));
1261 m_impl.assign(str.data, str.len);
1262 }
1263
1264 // and unsigned char*:
1265 wxString(const unsigned char *psz)
1266 : m_impl(ImplStr((const char*)psz)) {}
1267 wxString(const unsigned char *psz, const wxMBConv& conv)
1268 : m_impl(ImplStr((const char*)psz, conv)) {}
1269 wxString(const unsigned char *psz, size_t nLength)
1270 { assign((const char*)psz, nLength); }
1271 wxString(const unsigned char *psz, const wxMBConv& conv, size_t nLength)
1272 {
1273 SubstrBufFromMB str(ImplStr((const char*)psz, nLength, conv));
1274 m_impl.assign(str.data, str.len);
1275 }
1276
1277 // ctors from wchar_t* strings:
1278 wxString(const wchar_t *pwz)
1279 : m_impl(ImplStr(pwz)) {}
1280 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv))
1281 : m_impl(ImplStr(pwz)) {}
1282 wxString(const wchar_t *pwz, size_t nLength)
1283 { assign(pwz, nLength); }
1284 wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv), size_t nLength)
1285 { assign(pwz, nLength); }
1286
de4983f3 1287 wxString(const wxScopedCharBuffer& buf)
38d26d60 1288 { assign(buf.data(), buf.length()); }
de4983f3 1289 wxString(const wxScopedWCharBuffer& buf)
38d26d60 1290 { assign(buf.data(), buf.length()); }
67cff9dc 1291
06e9cf13
VS
1292 // NB: this version uses m_impl.c_str() to force making a copy of the
1293 // string, so that "wxString(str.c_str())" idiom for passing strings
1294 // between threads works
67cff9dc 1295 wxString(const wxCStrData& cstr)
06e9cf13 1296 : m_impl(cstr.AsString().m_impl.c_str()) { }
67cff9dc
VZ
1297
1298 // as we provide both ctors with this signature for both char and unsigned
1299 // char string, we need to provide one for wxCStrData to resolve ambiguity
1300 wxString(const wxCStrData& cstr, size_t nLength)
1301 : m_impl(cstr.AsString().Mid(0, nLength).m_impl) {}
1302
1303 // and because wxString is convertible to wxCStrData and const wxChar *
1304 // we also need to provide this one
1305 wxString(const wxString& str, size_t nLength)
1545d942 1306 { assign(str, nLength); }
67cff9dc 1307
68482dc5
VZ
1308
1309#if wxUSE_STRING_POS_CACHE
1310 ~wxString()
1311 {
1312 // we need to invalidate our cache entry as another string could be
1313 // recreated at the same address (unlikely, but still possible, with the
1314 // heap-allocated strings but perfectly common with stack-allocated ones)
1315 InvalidateCache();
1316 }
1317#endif // wxUSE_STRING_POS_CACHE
1318
67cff9dc 1319 // even if we're not built with wxUSE_STL == 1 it is very convenient to allow
59953bf4
VS
1320 // implicit conversions from std::string to wxString and vice verse as this
1321 // allows to use the same strings in non-GUI and GUI code, however we don't
1322 // want to unconditionally add this ctor as it would make wx lib dependent on
67cff9dc
VZ
1323 // libstdc++ on some Linux versions which is bad, so instead we ask the
1324 // client code to define this wxUSE_STD_STRING symbol if they need it
59953bf4 1325#if wxUSE_STD_STRING
59953bf4
VS
1326 #if wxUSE_UNICODE_WCHAR
1327 wxString(const wxStdWideString& str) : m_impl(str) {}
1328 #else // UTF-8 or ANSI
1329 wxString(const wxStdWideString& str)
1330 { assign(str.c_str(), str.length()); }
1331 #endif
1332
1333 #if !wxUSE_UNICODE // ANSI build
1334 // FIXME-UTF8: do this in UTF8 build #if wxUSE_UTF8_LOCALE_ONLY, too
1335 wxString(const std::string& str) : m_impl(str) {}
1336 #else // Unicode
1337 wxString(const std::string& str)
1338 { assign(str.c_str(), str.length()); }
1339 #endif
7978bc72 1340#endif // wxUSE_STD_STRING
59953bf4 1341
7978bc72
VS
1342 // Unlike ctor from std::string, we provide conversion to std::string only
1343 // if wxUSE_STL and not merely wxUSE_STD_STRING (which is on by default),
1344 // because it conflicts with operator const char/wchar_t*:
1345#if wxUSE_STL
59953bf4
VS
1346 #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING
1347 // wxStringImpl is std::string in the encoding we want
1348 operator const wxStdWideString&() const { return m_impl; }
1349 #else
1350 // wxStringImpl is either not std::string or needs conversion
1351 operator wxStdWideString() const
38d26d60
VS
1352 {
1353 wxScopedWCharBuffer buf(wc_str());
1354 return wxStdWideString(buf.data(), buf.length());
1355 }
59953bf4
VS
1356 #endif
1357
111d9948 1358 #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING
59953bf4
VS
1359 // wxStringImpl is std::string in the encoding we want
1360 operator const std::string&() const { return m_impl; }
1361 #else
1362 // wxStringImpl is either not std::string or needs conversion
1363 operator std::string() const
38d26d60
VS
1364 {
1365 wxScopedCharBuffer buf(mb_str());
1366 return std::string(buf.data(), buf.length());
1367 }
59953bf4 1368 #endif
111d9948 1369#endif // wxUSE_STL
67cff9dc 1370
06e9cf13
VS
1371 wxString Clone() const
1372 {
1373 // make a deep copy of the string, i.e. the returned string will have
1374 // ref count = 1 with refcounted implementation
1375 return wxString::FromImpl(wxStringImpl(m_impl.c_str(), m_impl.length()));
1376 }
1377
8f93a29f 1378 // first valid index position
b0c4d5d7 1379 const_iterator begin() const { return const_iterator(this, m_impl.begin()); }
81727065 1380 iterator begin() { return iterator(this, m_impl.begin()); }
8f93a29f 1381 // position one after the last valid one
b0c4d5d7 1382 const_iterator end() const { return const_iterator(this, m_impl.end()); }
81727065 1383 iterator end() { return iterator(this, m_impl.end()); }
8f93a29f
VS
1384
1385 // first element of the reversed string
1386 const_reverse_iterator rbegin() const
1387 { return const_reverse_iterator(end()); }
1388 reverse_iterator rbegin()
1389 { return reverse_iterator(end()); }
1390 // one beyond the end of the reversed string
1391 const_reverse_iterator rend() const
1392 { return const_reverse_iterator(begin()); }
1393 reverse_iterator rend()
1394 { return reverse_iterator(begin()); }
1395
1396 // std::string methods:
1397#if wxUSE_UNICODE_UTF8
68482dc5
VZ
1398 size_t length() const
1399 {
1400#if wxUSE_STRING_POS_CACHE
1401 wxCACHE_PROFILE_FIELD_INC(lentot);
1402
1403 Cache::Element * const cache = GetCacheElement();
1404
1405 if ( cache->len == npos )
1406 {
1407 // it's probably not worth trying to be clever and using cache->pos
1408 // here as it's probably 0 anyhow -- you usually call length() before
1409 // starting to index the string
1410 cache->len = end() - begin();
1411 }
1412 else
1413 {
1414 wxCACHE_PROFILE_FIELD_INC(lenhits);
1415
1416 wxSTRING_CACHE_ASSERT( (int)cache->len == end() - begin() );
1417 }
1418
1419 return cache->len;
1420#else // !wxUSE_STRING_POS_CACHE
1421 return end() - begin();
1422#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE
1423 }
8f93a29f
VS
1424#else
1425 size_t length() const { return m_impl.length(); }
1426#endif
ce76f779 1427
8f93a29f
VS
1428 size_type size() const { return length(); }
1429 size_type max_size() const { return npos; }
e90c1d2a 1430
8f93a29f 1431 bool empty() const { return m_impl.empty(); }
b77afb41 1432
68482dc5
VZ
1433 // NB: these methods don't have a well-defined meaning in UTF-8 case
1434 size_type capacity() const { return m_impl.capacity(); }
1435 void reserve(size_t sz) { m_impl.reserve(sz); }
b77afb41 1436
8f93a29f
VS
1437 void resize(size_t nSize, wxUniChar ch = wxT('\0'))
1438 {
fe1b98f5
VZ
1439 const size_t len = length();
1440 if ( nSize == len)
1441 return;
1442
8f93a29f 1443#if wxUSE_UNICODE_UTF8
fe1b98f5 1444 if ( nSize < len )
8f93a29f 1445 {
68482dc5
VZ
1446 wxSTRING_INVALIDATE_CACHE();
1447
fe1b98f5
VZ
1448 // we can't use wxStringImpl::resize() for truncating the string as it
1449 // counts in bytes, not characters
1450 erase(nSize);
1451 return;
8f93a29f 1452 }
fe1b98f5
VZ
1453
1454 // we also can't use (presumably more efficient) resize() if we have to
1455 // append characters taking more than one byte
1456 if ( !ch.IsAscii() )
68482dc5 1457 {
fe1b98f5 1458 append(nSize - len, ch);
68482dc5
VZ
1459 }
1460 else // can use (presumably faster) resize() version
fe1b98f5 1461#endif // wxUSE_UNICODE_UTF8
68482dc5
VZ
1462 {
1463 wxSTRING_INVALIDATE_CACHED_LENGTH();
1464
8f93a29f 1465 m_impl.resize(nSize, (wxStringCharType)ch);
68482dc5 1466 }
8f93a29f 1467 }
e90c1d2a 1468
8f93a29f
VS
1469 wxString substr(size_t nStart = 0, size_t nLen = npos) const
1470 {
1471 size_t pos, len;
1472 PosLenToImpl(nStart, nLen, &pos, &len);
59953bf4 1473 return FromImpl(m_impl.substr(pos, len));
8f93a29f 1474 }
e90c1d2a 1475
3c67202d
VZ
1476 // generic attributes & operations
1477 // as standard strlen()
e87b7833 1478 size_t Len() const { return length(); }
3c67202d 1479 // string contains any characters?
e87b7833 1480 bool IsEmpty() const { return empty(); }
d775fa82 1481 // empty string is "false", so !str will return true
20aa779e 1482 bool operator!() const { return empty(); }
735d1db6
VZ
1483 // truncate the string to given length
1484 wxString& Truncate(size_t uiLen);
3c67202d 1485 // empty string contents
24f86c43 1486 void Empty() { clear(); }
3c67202d 1487 // empty the string and free memory
68482dc5 1488 void Clear() { clear(); }
dd1eaa89 1489
3c67202d
VZ
1490 // contents test
1491 // Is an ascii value
c801d85f 1492 bool IsAscii() const;
3c67202d 1493 // Is a number
c801d85f 1494 bool IsNumber() const;
3c67202d 1495 // Is a word
c801d85f 1496 bool IsWord() const;
c801d85f 1497
3c67202d
VZ
1498 // data access (all indexes are 0 based)
1499 // read access
8f93a29f 1500 wxUniChar at(size_t n) const
68482dc5 1501 { return wxStringOperations::DecodeChar(m_impl.begin() + PosToImpl(n)); }
c9f78968 1502 wxUniChar GetChar(size_t n) const
9d9ad673 1503 { return at(n); }
3c67202d 1504 // read/write access
8f93a29f 1505 wxUniCharRef at(size_t n)
68482dc5 1506 { return *GetIterForNthChar(n); }
c9f78968 1507 wxUniCharRef GetWritableChar(size_t n)
9d9ad673 1508 { return at(n); }
3c67202d 1509 // write access
68482dc5 1510 void SetChar(size_t n, wxUniChar ch)
dd79ca2f 1511 { at(n) = ch; }
c801d85f 1512
3c67202d 1513 // get last character
8f93a29f 1514 wxUniChar Last() const
564ab31a 1515 {
9a83f860 1516 wxASSERT_MSG( !empty(), wxT("wxString: index out of bounds") );
564ab31a
VS
1517 return *rbegin();
1518 }
09443a26 1519
3c67202d 1520 // get writable last character
c9f78968 1521 wxUniCharRef Last()
564ab31a 1522 {
9a83f860 1523 wxASSERT_MSG( !empty(), wxT("wxString: index out of bounds") );
564ab31a
VS
1524 return *rbegin();
1525 }
c801d85f 1526
d836ee96 1527 /*
9d9ad673 1528 Note that we we must define all of the overloads below to avoid
c9f78968 1529 ambiguity when using str[0].
d836ee96 1530 */
c9f78968 1531 wxUniChar operator[](int n) const
8f93a29f 1532 { return at(n); }
c1df0a3b 1533 wxUniChar operator[](long n) const
8f93a29f 1534 { return at(n); }
c9f78968 1535 wxUniChar operator[](size_t n) const
8f93a29f 1536 { return at(n); }
01398433 1537#ifndef wxSIZE_T_IS_UINT
c9f78968 1538 wxUniChar operator[](unsigned int n) const
8f93a29f 1539 { return at(n); }
01398433 1540#endif // size_t != unsigned int
01398433 1541
9d9ad673 1542 // operator versions of GetWriteableChar()
c9f78968 1543 wxUniCharRef operator[](int n)
8f93a29f 1544 { return at(n); }
c1df0a3b 1545 wxUniCharRef operator[](long n)
8f93a29f 1546 { return at(n); }
c9f78968 1547 wxUniCharRef operator[](size_t n)
8f93a29f 1548 { return at(n); }
d836ee96 1549#ifndef wxSIZE_T_IS_UINT
c9f78968 1550 wxUniCharRef operator[](unsigned int n)
8f93a29f 1551 { return at(n); }
d836ee96 1552#endif // size_t != unsigned int
c801d85f 1553
18666b42
VZ
1554
1555 /*
1556 Overview of wxString conversions, implicit and explicit:
1557
1558 - wxString has a std::[w]string-like c_str() method, however it does
1559 not return a C-style string directly but instead returns wxCStrData
1560 helper object which is convertible to either "char *" narrow string
1561 or "wchar_t *" wide string. Usually the correct conversion will be
1562 applied by the compiler automatically but if this doesn't happen you
1563 need to explicitly choose one using wxCStrData::AsChar() or AsWChar()
1564 methods or another wxString conversion function.
1565
1566 - One of the places where the conversion does *NOT* happen correctly is
1567 when c_str() is passed to a vararg function such as printf() so you
1568 must *NOT* use c_str() with them. Either use wxPrintf() (all wx
1569 functions do handle c_str() correctly, even if they appear to be
1570 vararg (but they're not, really)) or add an explicit AsChar() or, if
1571 compatibility with previous wxWidgets versions is important, add a
1572 cast to "const char *".
1573
1574 - In non-STL mode only, wxString is also implicitly convertible to
1575 wxCStrData. The same warning as above applies.
1576
1577 - c_str() is polymorphic as it can be converted to either narrow or
1578 wide string. If you explicitly need one or the other, choose to use
1579 mb_str() (for narrow) or wc_str() (for wide) instead. Notice that
1580 these functions can return either the pointer to string directly (if
1581 this is what the string uses internally) or a temporary buffer
1582 containing the string and convertible to it. Again, conversion will
1583 usually be done automatically by the compiler but beware of the
1584 vararg functions: you need an explicit cast when using them.
1585
1586 - There are also non-const versions of mb_str() and wc_str() called
1587 char_str() and wchar_str(). They are only meant to be used with
1588 non-const-correct functions and they always return buffers.
1589
1590 - Finally wx_str() returns whatever string representation is used by
1591 wxString internally. It may be either a narrow or wide string
1592 depending on wxWidgets build mode but it will always be a raw pointer
1593 (and not a buffer).
1594 */
1595
1596 // explicit conversion to wxCStrData
c9f78968 1597 wxCStrData c_str() const { return wxCStrData(this); }
8f93a29f 1598 wxCStrData data() const { return c_str(); }
c9f78968 1599
18666b42 1600 // implicit conversion to wxCStrData
c9f78968 1601 operator wxCStrData() const { return c_str(); }
7978bc72 1602
4f167b46
VZ
1603 // the first two operators conflict with operators for conversion to
1604 // std::string and they must be disabled in STL build; the next one only
1605 // makes sense if conversions to char* are also defined and not defining it
1606 // in STL build also helps us to get more clear error messages for the code
1607 // which relies on implicit conversion to char* in STL build
7978bc72 1608#if !wxUSE_STL
11aac4ba
VS
1609 operator const char*() const { return c_str(); }
1610 operator const wchar_t*() const { return c_str(); }
e015c2a3 1611
353a4edc
VZ
1612 // implicit conversion to untyped pointer for compatibility with previous
1613 // wxWidgets versions: this is the same as conversion to const char * so it
1614 // may fail!
1615 operator const void*() const { return c_str(); }
4f167b46 1616#endif // wxUSE_STL
353a4edc 1617
e015c2a3 1618 // identical to c_str(), for MFC compatibility
c9f78968
VS
1619 const wxCStrData GetData() const { return c_str(); }
1620
1621 // explicit conversion to C string in internal representation (char*,
1622 // wchar_t*, UTF-8-encoded char*, depending on the build):
81727065 1623 const wxStringCharType *wx_str() const { return m_impl.c_str(); }
e90c1d2a 1624
ef0f1387 1625 // conversion to *non-const* multibyte or widestring buffer; modifying
c67b782d
VS
1626 // returned buffer won't affect the string, these methods are only useful
1627 // for passing values to const-incorrect functions
8060b0be 1628 wxWritableCharBuffer char_str(const wxMBConv& conv = wxConvLibc) const
c67b782d
VS
1629 { return mb_str(conv); }
1630 wxWritableWCharBuffer wchar_str() const { return wc_str(); }
ef0f1387 1631
062dc5fc
VZ
1632 // conversion to the buffer of the given type T (= char or wchar_t) and
1633 // also optionally return the buffer length
1634 //
1635 // this is mostly/only useful for the template functions
1636 //
1637 // FIXME-VC6: the second argument only exists for VC6 which doesn't support
1638 // explicit template function selection, do not use it unless
1639 // you must support VC6!
1640 template <typename T>
1641 wxCharTypeBuffer<T> tchar_str(size_t *len = NULL,
1642 T * WXUNUSED(dummy) = NULL) const
1643 {
1644#if wxUSE_UNICODE
1645 // we need a helper dispatcher depending on type
1646 return wxPrivate::wxStringAsBufHelper<T>::Get(*this, len);
1647#else // ANSI
1648 // T can only be char in ANSI build
1649 if ( len )
1650 *len = length();
1651
6df09f32 1652 return wxCharTypeBuffer<T>::CreateNonOwned(wx_str(), length());
062dc5fc
VZ
1653#endif // Unicode build kind
1654 }
1655
e015c2a3
VZ
1656 // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for
1657 // converting numbers or strings which are certain not to contain special
1658 // chars (typically system functions, X atoms, environment variables etc.)
1659 //
1660 // the behaviour of these functions with the strings containing anything
1661 // else than 7 bit ASCII characters is undefined, use at your own risk.
b1ac3b56 1662#if wxUSE_UNICODE
628f9e95
VZ
1663 static wxString FromAscii(const char *ascii, size_t len);
1664 static wxString FromAscii(const char *ascii);
1665 static wxString FromAscii(char ascii);
de4983f3 1666 const wxScopedCharBuffer ToAscii() const;
e015c2a3
VZ
1667#else // ANSI
1668 static wxString FromAscii(const char *ascii) { return wxString( ascii ); }
e6310bbc
VS
1669 static wxString FromAscii(const char *ascii, size_t len)
1670 { return wxString( ascii, len ); }
628f9e95 1671 static wxString FromAscii(char ascii) { return wxString( ascii ); }
e015c2a3
VZ
1672 const char *ToAscii() const { return c_str(); }
1673#endif // Unicode/!Unicode
b1ac3b56 1674
628f9e95
VZ
1675 // also provide unsigned char overloads as signed/unsigned doesn't matter
1676 // for 7 bit ASCII characters
1677 static wxString FromAscii(const unsigned char *ascii)
1678 { return FromAscii((const char *)ascii); }
1679 static wxString FromAscii(const unsigned char *ascii, size_t len)
1680 { return FromAscii((const char *)ascii, len); }
1681
5f167b77
VS
1682 // conversion to/from UTF-8:
1683#if wxUSE_UNICODE_UTF8
cc209a51 1684 static wxString FromUTF8Unchecked(const char *utf8)
5f167b77 1685 {
f966a73d
VS
1686 if ( !utf8 )
1687 return wxEmptyString;
1688
5f167b77
VS
1689 wxASSERT( wxStringOperations::IsValidUtf8String(utf8) );
1690 return FromImpl(wxStringImpl(utf8));
1691 }
cc209a51 1692 static wxString FromUTF8Unchecked(const char *utf8, size_t len)
5f167b77 1693 {
f966a73d
VS
1694 if ( !utf8 )
1695 return wxEmptyString;
1696 if ( len == npos )
cc209a51 1697 return FromUTF8Unchecked(utf8);
f966a73d 1698
5f167b77
VS
1699 wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) );
1700 return FromImpl(wxStringImpl(utf8, len));
1701 }
cc209a51
VZ
1702
1703 static wxString FromUTF8(const char *utf8)
1704 {
1705 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8) )
1706 return "";
1707
1708 return FromImpl(wxStringImpl(utf8));
1709 }
1710 static wxString FromUTF8(const char *utf8, size_t len)
1711 {
1712 if ( len == npos )
1713 return FromUTF8(utf8);
1714
1715 if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8, len) )
1716 return "";
1717
1718 return FromImpl(wxStringImpl(utf8, len));
1719 }
1720
197380a0 1721 const wxScopedCharBuffer utf8_str() const
ed5e6560 1722 { return wxCharBuffer::CreateNonOwned(m_impl.c_str(), m_impl.length()); }
062dc5fc
VZ
1723
1724 // this function exists in UTF-8 build only and returns the length of the
1725 // internal UTF-8 representation
1726 size_t utf8_length() const { return m_impl.length(); }
5f167b77 1727#elif wxUSE_UNICODE_WCHAR
cc209a51 1728 static wxString FromUTF8(const char *utf8, size_t len = npos)
5487ff0f 1729 { return wxString(utf8, wxMBConvUTF8(), len); }
cc209a51
VZ
1730 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1731 {
1732 const wxString s(utf8, wxMBConvUTF8(), len);
1733 wxASSERT_MSG( !utf8 || !*utf8 || !s.empty(),
1734 "string must be valid UTF-8" );
1735 return s;
1736 }
de4983f3 1737 const wxScopedCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); }
5f167b77
VS
1738#else // ANSI
1739 static wxString FromUTF8(const char *utf8)
5487ff0f 1740 { return wxString(wxMBConvUTF8().cMB2WC(utf8)); }
5f167b77 1741 static wxString FromUTF8(const char *utf8, size_t len)
06c73b98 1742 {
96e93f5b 1743 size_t wlen;
de4983f3 1744 wxScopedWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen));
96e93f5b
VZ
1745 return wxString(buf.data(), wlen);
1746 }
1747 static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
1748 {
1749 size_t wlen;
de4983f3
VS
1750 wxScopedWCharBuffer buf
1751 (
1752 wxMBConvUTF8().cMB2WC
1753 (
1754 utf8,
1755 len == npos ? wxNO_LEN : len,
1756 &wlen
1757 )
1758 );
96e93f5b
VZ
1759 wxASSERT_MSG( !utf8 || !*utf8 || wlen,
1760 "string must be valid UTF-8" );
1761
1762 return wxString(buf.data(), wlen);
06c73b98 1763 }
de4983f3 1764 const wxScopedCharBuffer utf8_str() const
5487ff0f 1765 { return wxMBConvUTF8().cWC2MB(wc_str()); }
5f167b77
VS
1766#endif
1767
ed5e6560
VS
1768 const wxScopedCharBuffer ToUTF8() const { return utf8_str(); }
1769
cb2f2135
VS
1770 // functions for storing binary data in wxString:
1771#if wxUSE_UNICODE
1772 static wxString From8BitData(const char *data, size_t len)
1773 { return wxString(data, wxConvISO8859_1, len); }
1774 // version for NUL-terminated data:
1775 static wxString From8BitData(const char *data)
1776 { return wxString(data, wxConvISO8859_1); }
de4983f3
VS
1777 const wxScopedCharBuffer To8BitData() const
1778 { return mb_str(wxConvISO8859_1); }
cb2f2135
VS
1779#else // ANSI
1780 static wxString From8BitData(const char *data, size_t len)
1781 { return wxString(data, len); }
1782 // version for NUL-terminated data:
1783 static wxString From8BitData(const char *data)
1784 { return wxString(data); }
1785 const char *To8BitData() const { return c_str(); }
1786#endif // Unicode/ANSI
1787
2b5f62a0 1788 // conversions with (possible) format conversions: have to return a
e90c1d2a 1789 // buffer with temporary data
f6bcfd97
BP
1790 //
1791 // the functions defined (in either Unicode or ANSI) mode are mb_str() to
1792 // return an ANSI (multibyte) string, wc_str() to return a wide string and
1793 // fn_str() to return a string which should be used with the OS APIs
1794 // accepting the file names. The return value is always the same, but the
1795 // type differs because a function may either return pointer to the buffer
1796 // directly or have to use intermediate buffer for translation.
f54cb154 1797
2bb67b80 1798#if wxUSE_UNICODE
111d9948 1799
f54cb154
VZ
1800 // this is an optimization: even though using mb_str(wxConvLibc) does the
1801 // same thing (i.e. returns pointer to internal representation as locale is
1802 // always an UTF-8 one) in wxUSE_UTF8_LOCALE_ONLY case, we can avoid the
1803 // extra checks and the temporary buffer construction by providing a
1804 // separate mb_str() overload
111d9948
VS
1805#if wxUSE_UTF8_LOCALE_ONLY
1806 const char* mb_str() const { return wx_str(); }
f54cb154
VZ
1807 const wxScopedCharBuffer mb_str(const wxMBConv& conv) const
1808 {
1809 return AsCharBuf(conv);
1810 }
1811#else // !wxUSE_UTF8_LOCALE_ONLY
1812 const wxScopedCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const
1813 {
1814 return AsCharBuf(conv);
1815 }
1816#endif // wxUSE_UTF8_LOCALE_ONLY/!wxUSE_UTF8_LOCALE_ONLY
f6bcfd97 1817
e90c1d2a
VZ
1818 const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
1819
81727065 1820#if wxUSE_UNICODE_WCHAR
6ad68ad8 1821 const wchar_t* wc_str() const { return wx_str(); }
81727065 1822#elif wxUSE_UNICODE_UTF8
f54cb154
VZ
1823 const wxScopedWCharBuffer wc_str() const
1824 { return AsWCharBuf(wxMBConvStrictUTF8()); }
81727065 1825#endif
f6bcfd97 1826 // for compatibility with !wxUSE_UNICODE version
81727065
VS
1827 const wxWX2WCbuf wc_str(const wxMBConv& WXUNUSED(conv)) const
1828 { return wc_str(); }
e90c1d2a 1829
2bb67b80 1830#if wxMBFILES
de4983f3 1831 const wxScopedCharBuffer fn_str() const { return mb_str(wxConvFile); }
e90c1d2a 1832#else // !wxMBFILES
81727065 1833 const wxWX2WCbuf fn_str() const { return wc_str(); }
e90c1d2a 1834#endif // wxMBFILES/!wxMBFILES
81727065 1835
e90c1d2a 1836#else // ANSI
f54cb154 1837 const char* mb_str() const { return wx_str(); }
f6bcfd97
BP
1838
1839 // for compatibility with wxUSE_UNICODE version
6ad68ad8 1840 const char* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); }
f6bcfd97 1841
e90c1d2a 1842 const wxWX2MBbuf mbc_str() const { return mb_str(); }
f6bcfd97 1843
f54cb154
VZ
1844 const wxScopedWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const
1845 { return AsWCharBuf(conv); }
1846
de4983f3
VS
1847 const wxScopedCharBuffer fn_str() const
1848 { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); }
e90c1d2a 1849#endif // Unicode/ANSI
c801d85f 1850
6ad68ad8 1851#if wxUSE_UNICODE_UTF8
de4983f3 1852 const wxScopedWCharBuffer t_str() const { return wc_str(); }
6ad68ad8
RR
1853#elif wxUSE_UNICODE_WCHAR
1854 const wchar_t* t_str() const { return wx_str(); }
1855#else
1856 const char* t_str() const { return wx_str(); }
cbec0f40 1857#endif
6ad68ad8
RR
1858
1859
3c67202d
VZ
1860 // overloaded assignment
1861 // from another wxString
04abe4bc 1862 wxString& operator=(const wxString& stringSrc)
68482dc5
VZ
1863 {
1864 if ( this != &stringSrc )
1865 {
1866 wxSTRING_INVALIDATE_CACHE();
1867
1868 m_impl = stringSrc.m_impl;
1869 }
1870
1871 return *this;
1872 }
1873
8f93a29f
VS
1874 wxString& operator=(const wxCStrData& cstr)
1875 { return *this = cstr.AsString(); }
3c67202d 1876 // from a character
c9f78968 1877 wxString& operator=(wxUniChar ch)
84d2877f 1878 {
68482dc5
VZ
1879 wxSTRING_INVALIDATE_CACHE();
1880
84d2877f
VS
1881#if wxUSE_UNICODE_UTF8
1882 if ( !ch.IsAscii() )
1883 m_impl = wxStringOperations::EncodeChar(ch);
1884 else
68482dc5 1885#endif // wxUSE_UNICODE_UTF8
84d2877f
VS
1886 m_impl = (wxStringCharType)ch;
1887 return *this;
1888 }
68482dc5 1889
c9f78968 1890 wxString& operator=(wxUniCharRef ch)
8f93a29f 1891 { return operator=((wxUniChar)ch); }
c9f78968 1892 wxString& operator=(char ch)
8f93a29f 1893 { return operator=(wxUniChar(ch)); }
50e02008
VS
1894 wxString& operator=(unsigned char ch)
1895 { return operator=(wxUniChar(ch)); }
c9f78968 1896 wxString& operator=(wchar_t ch)
8f93a29f 1897 { return operator=(wxUniChar(ch)); }
c57e3bd5
RN
1898 // from a C string - STL probably will crash on NULL,
1899 // so we need to compensate in that case
992527a5 1900#if wxUSE_STL_BASED_WXSTRING
04abe4bc 1901 wxString& operator=(const char *psz)
68482dc5
VZ
1902 {
1903 wxSTRING_INVALIDATE_CACHE();
1904
1905 if ( psz )
1906 m_impl = ImplStr(psz);
1907 else
1908 clear();
1909
1910 return *this;
1911 }
1912
04abe4bc 1913 wxString& operator=(const wchar_t *pwz)
68482dc5
VZ
1914 {
1915 wxSTRING_INVALIDATE_CACHE();
1916
1917 if ( pwz )
1918 m_impl = ImplStr(pwz);
1919 else
1920 clear();
1921
1922 return *this;
1923 }
1924#else // !wxUSE_STL_BASED_WXSTRING
04abe4bc 1925 wxString& operator=(const char *psz)
68482dc5
VZ
1926 {
1927 wxSTRING_INVALIDATE_CACHE();
1928
1929 m_impl = ImplStr(psz);
1930
1931 return *this;
1932 }
1933
04abe4bc 1934 wxString& operator=(const wchar_t *pwz)
68482dc5
VZ
1935 {
1936 wxSTRING_INVALIDATE_CACHE();
1937
1938 m_impl = ImplStr(pwz);
1939
1940 return *this;
1941 }
1942#endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING
1943
04abe4bc
VS
1944 wxString& operator=(const unsigned char *psz)
1945 { return operator=((const char*)psz); }
c57e3bd5 1946
de4983f3
VS
1947 // from wxScopedWCharBuffer
1948 wxString& operator=(const wxScopedWCharBuffer& s)
38d26d60 1949 { return assign(s); }
de4983f3
VS
1950 // from wxScopedCharBuffer
1951 wxString& operator=(const wxScopedCharBuffer& s)
38d26d60 1952 { return assign(s); }
3c67202d
VZ
1953
1954 // string concatenation
1955 // in place concatenation
1956 /*
1957 Concatenate and return the result. Note that the left to right
1958 associativity of << allows to write things like "str << str1 << str2
1959 << ..." (unlike with +=)
1960 */
1961 // string += string
dd1eaa89
VZ
1962 wxString& operator<<(const wxString& s)
1963 {
8f93a29f
VS
1964#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
1965 wxASSERT_MSG( s.IsValid(),
9a83f860 1966 wxT("did you forget to call UngetWriteBuf()?") );
e87b7833 1967#endif
dd1eaa89 1968
e87b7833 1969 append(s);
dd1eaa89
VZ
1970 return *this;
1971 }
3c67202d 1972 // string += C string
8f93a29f 1973 wxString& operator<<(const char *psz)
c9f78968 1974 { append(psz); return *this; }
8f93a29f
VS
1975 wxString& operator<<(const wchar_t *pwz)
1976 { append(pwz); return *this; }
c9f78968 1977 wxString& operator<<(const wxCStrData& psz)
8f93a29f 1978 { append(psz.AsString()); return *this; }
3c67202d 1979 // string += char
c9f78968
VS
1980 wxString& operator<<(wxUniChar ch) { append(1, ch); return *this; }
1981 wxString& operator<<(wxUniCharRef ch) { append(1, ch); return *this; }
1982 wxString& operator<<(char ch) { append(1, ch); return *this; }
50e02008 1983 wxString& operator<<(unsigned char ch) { append(1, ch); return *this; }
c9f78968 1984 wxString& operator<<(wchar_t ch) { append(1, ch); return *this; }
2bb67b80
OK
1985
1986 // string += buffer (i.e. from wxGetString)
de4983f3 1987 wxString& operator<<(const wxScopedWCharBuffer& s)
38d26d60 1988 { return append(s); }
de4983f3 1989 wxString& operator<<(const wxScopedCharBuffer& s)
38d26d60 1990 { return append(s); }
d7330233 1991
3c67202d 1992 // string += C string
09443a26
VZ
1993 wxString& Append(const wxString& s)
1994 {
5a8231ef
WS
1995 // test for empty() to share the string if possible
1996 if ( empty() )
09443a26
VZ
1997 *this = s;
1998 else
e87b7833 1999 append(s);
09443a26
VZ
2000 return *this;
2001 }
8f93a29f 2002 wxString& Append(const char* psz)
e87b7833 2003 { append(psz); return *this; }
8f93a29f
VS
2004 wxString& Append(const wchar_t* pwz)
2005 { append(pwz); return *this; }
d3cefe87
VS
2006 wxString& Append(const wxCStrData& psz)
2007 { append(psz); return *this; }
de4983f3 2008 wxString& Append(const wxScopedCharBuffer& psz)
d3cefe87 2009 { append(psz); return *this; }
de4983f3 2010 wxString& Append(const wxScopedWCharBuffer& psz)
d3cefe87 2011 { append(psz); return *this; }
f416695a
VS
2012 wxString& Append(const char* psz, size_t nLen)
2013 { append(psz, nLen); return *this; }
2014 wxString& Append(const wchar_t* pwz, size_t nLen)
2015 { append(pwz, nLen); return *this; }
2016 wxString& Append(const wxCStrData& psz, size_t nLen)
2017 { append(psz, nLen); return *this; }
de4983f3 2018 wxString& Append(const wxScopedCharBuffer& psz, size_t nLen)
f416695a 2019 { append(psz, nLen); return *this; }
de4983f3 2020 wxString& Append(const wxScopedWCharBuffer& psz, size_t nLen)
f416695a 2021 { append(psz, nLen); return *this; }
3c67202d 2022 // append count copies of given character
c9f78968
VS
2023 wxString& Append(wxUniChar ch, size_t count = 1u)
2024 { append(count, ch); return *this; }
2025 wxString& Append(wxUniCharRef ch, size_t count = 1u)
2026 { append(count, ch); return *this; }
2027 wxString& Append(char ch, size_t count = 1u)
2028 { append(count, ch); return *this; }
50e02008
VS
2029 wxString& Append(unsigned char ch, size_t count = 1u)
2030 { append(count, ch); return *this; }
c9f78968 2031 wxString& Append(wchar_t ch, size_t count = 1u)
e87b7833 2032 { append(count, ch); return *this; }
3c67202d
VZ
2033
2034 // prepend a string, return the string itself
2035 wxString& Prepend(const wxString& str)
2036 { *this = str + *this; return *this; }
2037
2038 // non-destructive concatenation
1fc8cfbd
VZ
2039 // two strings
2040 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string1,
2041 const wxString& string2);
2042 // string with a single char
c9f78968 2043 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
1fc8cfbd 2044 // char with a string
c9f78968 2045 friend wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
1fc8cfbd
VZ
2046 // string with C string
2047 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
8f93a29f
VS
2048 const char *psz);
2049 friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string,
2050 const wchar_t *pwz);
1fc8cfbd 2051 // C string with string
8f93a29f
VS
2052 friend wxString WXDLLIMPEXP_BASE operator+(const char *psz,
2053 const wxString& string);
2054 friend wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz,
1fc8cfbd 2055 const wxString& string);
3c67202d
VZ
2056
2057 // stream-like functions
2058 // insert an int into string
3ce65f6c 2059 wxString& operator<<(int i)
9a83f860 2060 { return (*this) << Format(wxT("%d"), i); }
3ce65f6c
VZ
2061 // insert an unsigned int into string
2062 wxString& operator<<(unsigned int ui)
9a83f860 2063 { return (*this) << Format(wxT("%u"), ui); }
3ce65f6c
VZ
2064 // insert a long into string
2065 wxString& operator<<(long l)
9a83f860 2066 { return (*this) << Format(wxT("%ld"), l); }
3ce65f6c
VZ
2067 // insert an unsigned long into string
2068 wxString& operator<<(unsigned long ul)
9a83f860 2069 { return (*this) << Format(wxT("%lu"), ul); }
3fc1adbd
MW
2070#if defined wxLongLong_t && !defined wxLongLongIsLong
2071 // insert a long long if they exist and aren't longs
2072 wxString& operator<<(wxLongLong_t ll)
b7859132 2073 {
65154866 2074 return (*this) << Format("%" wxLongLongFmtSpec "d", ll);
b7859132 2075 }
3fc1adbd
MW
2076 // insert an unsigned long long
2077 wxString& operator<<(wxULongLong_t ull)
b7859132 2078 {
65154866 2079 return (*this) << Format("%" wxLongLongFmtSpec "u" , ull);
b7859132 2080 }
68482dc5 2081#endif // wxLongLong_t && !wxLongLongIsLong
3c67202d 2082 // insert a float into string
3ce65f6c 2083 wxString& operator<<(float f)
9a83f860 2084 { return (*this) << Format(wxT("%f"), f); }
3c67202d 2085 // insert a double into string
3ce65f6c 2086 wxString& operator<<(double d)
9a83f860 2087 { return (*this) << Format(wxT("%g"), d); }
c84c52de 2088
3c67202d 2089 // string comparison
30b21f9a 2090 // case-sensitive comparison (returns a value < 0, = 0 or > 0)
8f93a29f
VS
2091 int Cmp(const char *psz) const
2092 { return compare(psz); }
2093 int Cmp(const wchar_t *pwz) const
2094 { return compare(pwz); }
2095 int Cmp(const wxString& s) const
2096 { return compare(s); }
d3cefe87
VS
2097 int Cmp(const wxCStrData& s) const
2098 { return compare(s); }
de4983f3 2099 int Cmp(const wxScopedCharBuffer& s) const
d3cefe87 2100 { return compare(s); }
de4983f3 2101 int Cmp(const wxScopedWCharBuffer& s) const
d3cefe87 2102 { return compare(s); }
3c67202d 2103 // same as Cmp() but not case-sensitive
dcb68102 2104 int CmpNoCase(const wxString& s) const;
2712e317 2105
3c67202d
VZ
2106 // test for the string equality, either considering case or not
2107 // (if compareWithCase then the case matters)
11aac4ba 2108 bool IsSameAs(const wxString& str, bool compareWithCase = true) const
2712e317
VS
2109 {
2110#if !wxUSE_UNICODE_UTF8
2111 // in UTF-8 build, length() is O(n) and doing this would be _slower_
2112 if ( length() != str.length() )
2113 return false;
2114#endif
2115 return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0;
2116 }
11aac4ba
VS
2117 bool IsSameAs(const char *str, bool compareWithCase = true) const
2118 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2119 bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const
2120 { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
2712e317 2121
d3cefe87
VS
2122 bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const
2123 { return IsSameAs(str.AsString(), compareWithCase); }
de4983f3 2124 bool IsSameAs(const wxScopedCharBuffer& str, bool compareWithCase = true) const
d3cefe87 2125 { return IsSameAs(str.data(), compareWithCase); }
de4983f3 2126 bool IsSameAs(const wxScopedWCharBuffer& str, bool compareWithCase = true) const
d3cefe87 2127 { return IsSameAs(str.data(), compareWithCase); }
20aa779e 2128 // comparison with a single character: returns true if equal
52de37c7 2129 bool IsSameAs(wxUniChar c, bool compareWithCase = true) const;
11aac4ba
VS
2130 // FIXME-UTF8: remove these overloads
2131 bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const
2132 { return IsSameAs(wxUniChar(c), compareWithCase); }
2133 bool IsSameAs(char c, bool compareWithCase = true) const
2134 { return IsSameAs(wxUniChar(c), compareWithCase); }
2135 bool IsSameAs(unsigned char c, bool compareWithCase = true) const
2136 { return IsSameAs(wxUniChar(c), compareWithCase); }
2137 bool IsSameAs(wchar_t c, bool compareWithCase = true) const
2138 { return IsSameAs(wxUniChar(c), compareWithCase); }
2139 bool IsSameAs(int c, bool compareWithCase = true) const
2140 { return IsSameAs(wxUniChar(c), compareWithCase); }
3c67202d
VZ
2141
2142 // simple sub-string extraction
2143 // return substring starting at nFirst of length nCount (or till the end
2144 // if nCount = default value)
e87b7833 2145 wxString Mid(size_t nFirst, size_t nCount = npos) const;
3c67202d 2146
f6bcfd97 2147 // operator version of Mid()
3c67202d
VZ
2148 wxString operator()(size_t start, size_t len) const
2149 { return Mid(start, len); }
2150
3affcd07
VZ
2151 // check if the string starts with the given prefix and return the rest
2152 // of the string in the provided pointer if it is not NULL; otherwise
2153 // return false
c5e7a7d7 2154 bool StartsWith(const wxString& prefix, wxString *rest = NULL) const;
3affcd07
VZ
2155 // check if the string ends with the given suffix and return the
2156 // beginning of the string before the suffix in the provided pointer if
2157 // it is not NULL; otherwise return false
c5e7a7d7 2158 bool EndsWith(const wxString& suffix, wxString *rest = NULL) const;
f6bcfd97 2159
3c67202d 2160 // get first nCount characters
c801d85f 2161 wxString Left(size_t nCount) const;
3c67202d 2162 // get last nCount characters
c801d85f 2163 wxString Right(size_t nCount) const;
7929902d 2164 // get all characters before the first occurrence of ch
3c67202d 2165 // (returns the whole string if ch not found)
c9f78968 2166 wxString BeforeFirst(wxUniChar ch) const;
7929902d 2167 // get all characters before the last occurrence of ch
3c67202d 2168 // (returns empty string if ch not found)
c9f78968 2169 wxString BeforeLast(wxUniChar ch) const;
7929902d 2170 // get all characters after the first occurrence of ch
3c67202d 2171 // (returns empty string if ch not found)
c9f78968 2172 wxString AfterFirst(wxUniChar ch) const;
7929902d 2173 // get all characters after the last occurrence of ch
3c67202d 2174 // (returns the whole string if ch not found)
c9f78968 2175 wxString AfterLast(wxUniChar ch) const;
3c67202d
VZ
2176
2177 // for compatibility only, use more explicitly named functions above
c9f78968
VS
2178 wxString Before(wxUniChar ch) const { return BeforeLast(ch); }
2179 wxString After(wxUniChar ch) const { return AfterFirst(ch); }
3c67202d
VZ
2180
2181 // case conversion
c84c52de 2182 // convert to upper case in place, return the string itself
c801d85f 2183 wxString& MakeUpper();
c84c52de 2184 // convert to upper case, return the copy of the string
0c7db140 2185 wxString Upper() const { return wxString(*this).MakeUpper(); }
c84c52de 2186 // convert to lower case in place, return the string itself
c801d85f 2187 wxString& MakeLower();
c84c52de 2188 // convert to lower case, return the copy of the string
0c7db140
VZ
2189 wxString Lower() const { return wxString(*this).MakeLower(); }
2190 // convert the first character to the upper case and the rest to the
2191 // lower one, return the modified string itself
2192 wxString& MakeCapitalized();
2193 // convert the first character to the upper case and the rest to the
2194 // lower one, return the copy of the string
2195 wxString Capitalize() const { return wxString(*this).MakeCapitalized(); }
c801d85f 2196
3c67202d
VZ
2197 // trimming/padding whitespace (either side) and truncating
2198 // remove spaces from left or from right (default) side
d775fa82 2199 wxString& Trim(bool bFromRight = true);
3c67202d 2200 // add nCount copies chPad in the beginning or at the end (default)
c9f78968 2201 wxString& Pad(size_t nCount, wxUniChar chPad = wxT(' '), bool bFromRight = true);
dd1eaa89 2202
3c67202d
VZ
2203 // searching and replacing
2204 // searching (return starting index, or -1 if not found)
c9f78968 2205 int Find(wxUniChar ch, bool bFromEnd = false) const; // like strchr/strrchr
8a540c88
VS
2206 int Find(wxUniCharRef ch, bool bFromEnd = false) const
2207 { return Find(wxUniChar(ch), bFromEnd); }
2208 int Find(char ch, bool bFromEnd = false) const
2209 { return Find(wxUniChar(ch), bFromEnd); }
2210 int Find(unsigned char ch, bool bFromEnd = false) const
2211 { return Find(wxUniChar(ch), bFromEnd); }
2212 int Find(wchar_t ch, bool bFromEnd = false) const
2213 { return Find(wxUniChar(ch), bFromEnd); }
3c67202d 2214 // searching (return starting index, or -1 if not found)
8a540c88
VS
2215 int Find(const wxString& sub) const // like strstr
2216 {
2217 size_type idx = find(sub);
2218 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2219 }
2220 int Find(const char *sub) const // like strstr
2221 {
2222 size_type idx = find(sub);
2223 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2224 }
2225 int Find(const wchar_t *sub) const // like strstr
2226 {
2227 size_type idx = find(sub);
2228 return (idx == npos) ? wxNOT_FOUND : (int)idx;
2229 }
2230
d3cefe87
VS
2231 int Find(const wxCStrData& sub) const
2232 { return Find(sub.AsString()); }
de4983f3 2233 int Find(const wxScopedCharBuffer& sub) const
d3cefe87 2234 { return Find(sub.data()); }
de4983f3 2235 int Find(const wxScopedWCharBuffer& sub) const
d3cefe87
VS
2236 { return Find(sub.data()); }
2237
7929902d 2238 // replace first (or all of bReplaceAll) occurrences of substring with
3c67202d 2239 // another string, returns the number of replacements made
8a540c88
VS
2240 size_t Replace(const wxString& strOld,
2241 const wxString& strNew,
d775fa82 2242 bool bReplaceAll = true);
3c67202d
VZ
2243
2244 // check if the string contents matches a mask containing '*' and '?'
8a540c88 2245 bool Matches(const wxString& mask) const;
c801d85f 2246
529e491c
FM
2247 // conversion to numbers: all functions return true only if the whole
2248 // string is a number and put the value of this number into the pointer
2249 // provided, the base is the numeric base in which the conversion should be
2250 // done and must be comprised between 2 and 36 or be 0 in which case the
2251 // standard C rules apply (leading '0' => octal, "0x" => hex)
2252 // convert to a signed integer
2253 bool ToLong(long *val, int base = 10) const;
2254 // convert to an unsigned integer
2255 bool ToULong(unsigned long *val, int base = 10) const;
2256 // convert to wxLongLong
d6718dd1 2257#if defined(wxLongLong_t)
529e491c
FM
2258 bool ToLongLong(wxLongLong_t *val, int base = 10) const;
2259 // convert to wxULongLong
2260 bool ToULongLong(wxULongLong_t *val, int base = 10) const;
d6718dd1 2261#endif // wxLongLong_t
529e491c
FM
2262 // convert to a double
2263 bool ToDouble(double *val) const;
2264
2265#if wxUSE_XLOCALE
2266 // conversions to numbers using C locale
2267 // convert to a signed integer
2268 bool ToCLong(long *val, int base = 10) const;
2269 // convert to an unsigned integer
2270 bool ToCULong(unsigned long *val, int base = 10) const;
2271 // convert to a double
2272 bool ToCDouble(double *val) const;
2273#endif
03647350 2274
c9f78968 2275#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
90e572f1 2276 // formatted input/output
3c67202d 2277 // as sprintf(), returns the number of characters written or < 0 on error
7357f981 2278 // (take 'this' into account in attribute parameter count)
2523e9b7 2279 // int Printf(const wxString& format, ...);
1528e0b8 2280 WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),
d1f6e2cf 2281 DoPrintfWchar, DoPrintfUtf8)
2523e9b7 2282#ifdef __WATCOMC__
59a14f69
VS
2283 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
2284 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxString&),
2285 (wxFormatString(f1)));
2286 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxCStrData&),
2287 (wxFormatString(f1)));
2288 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const char*),
2289 (wxFormatString(f1)));
2290 WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wchar_t*),
2291 (wxFormatString(f1)));
2523e9b7 2292#endif
c9f78968 2293#endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN
3c67202d 2294 // as vprintf(), returns the number of characters written or < 0 on error
c9f78968 2295 int PrintfV(const wxString& format, va_list argptr);
dd1eaa89 2296
c9f78968 2297#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
341e7d28 2298 // returns the string containing the result of Printf() to it
862bb5c7 2299 // static wxString Format(const wxString& format, ...) WX_ATTRIBUTE_PRINTF_1;
1528e0b8 2300 WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxFormatString&),
d1f6e2cf 2301 DoFormatWchar, DoFormatUtf8)
2523e9b7
VS
2302#ifdef __WATCOMC__
2303 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
59a14f69
VS
2304 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxString&),
2305 (wxFormatString(f1)));
2306 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxCStrData&),
2307 (wxFormatString(f1)));
2308 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const char*),
2309 (wxFormatString(f1)));
2310 WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wchar_t*),
2311 (wxFormatString(f1)));
2523e9b7 2312#endif
c9f78968 2313#endif
341e7d28 2314 // the same as above, but takes a va_list
c9f78968 2315 static wxString FormatV(const wxString& format, va_list argptr);
341e7d28 2316
3c67202d
VZ
2317 // raw access to string memory
2318 // ensure that string has space for at least nLen characters
dd1eaa89 2319 // only works if the data of this string is not shared
0367b928 2320 bool Alloc(size_t nLen) { reserve(nLen); return capacity() >= nLen; }
3c67202d 2321 // minimize the string's memory
dd1eaa89 2322 // only works if the data of this string is not shared
b1801e0e 2323 bool Shrink();
8f93a29f 2324#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
d8a4b666
VS
2325 // These are deprecated, use wxStringBuffer or wxStringBufferLength instead
2326 //
3c67202d
VZ
2327 // get writable buffer of at least nLen bytes. Unget() *must* be called
2328 // a.s.a.p. to put string back in a reasonable state!
c87a0bc8 2329 wxDEPRECATED( wxStringCharType *GetWriteBuf(size_t nLen) );
3c67202d 2330 // call this immediately after GetWriteBuf() has been used
d8a4b666
VS
2331 wxDEPRECATED( void UngetWriteBuf() );
2332 wxDEPRECATED( void UngetWriteBuf(size_t nLen) );
8f93a29f 2333#endif // WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && wxUSE_UNICODE_UTF8
c801d85f 2334
77ffb593 2335 // wxWidgets version 1 compatibility functions
3c67202d
VZ
2336
2337 // use Mid()
2338 wxString SubString(size_t from, size_t to) const
2339 { return Mid(from, (to - from + 1)); }
2340 // values for second parameter of CompareTo function
c801d85f 2341 enum caseCompare {exact, ignoreCase};
3c67202d 2342 // values for first parameter of Strip function
c801d85f 2343 enum stripType {leading = 0x1, trailing = 0x2, both = 0x3};
8870c26e 2344
c9f78968 2345#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
7357f981
GD
2346 // use Printf()
2347 // (take 'this' into account in attribute parameter count)
862bb5c7 2348 // int sprintf(const wxString& format, ...) WX_ATTRIBUTE_PRINTF_2;
1528e0b8 2349 WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),
d1f6e2cf 2350 DoPrintfWchar, DoPrintfUtf8)
2523e9b7
VS
2351#ifdef __WATCOMC__
2352 // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
59a14f69
VS
2353 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxString&),
2354 (wxFormatString(f1)));
2355 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxCStrData&),
2356 (wxFormatString(f1)));
2357 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const char*),
2358 (wxFormatString(f1)));
2359 WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wchar_t*),
2360 (wxFormatString(f1)));
2523e9b7 2361#endif
c9f78968 2362#endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
c801d85f 2363
3c67202d 2364 // use Cmp()
68482dc5 2365 int CompareTo(const wxChar* psz, caseCompare cmp = exact) const
6b95b20d 2366 { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); }
c801d85f 2367
8f93a29f 2368 // use length()
e87b7833 2369 size_t Length() const { return length(); }
3c67202d 2370 // Count the number of characters
c9f78968 2371 int Freq(wxUniChar ch) const;
3c67202d 2372 // use MakeLower
c801d85f 2373 void LowerCase() { MakeLower(); }
3c67202d 2374 // use MakeUpper
c801d85f 2375 void UpperCase() { MakeUpper(); }
3c67202d 2376 // use Trim except that it doesn't change this string
c801d85f
KB
2377 wxString Strip(stripType w = trailing) const;
2378
3c67202d 2379 // use Find (more general variants not yet supported)
2bb67b80 2380 size_t Index(const wxChar* psz) const { return Find(psz); }
c9f78968 2381 size_t Index(wxUniChar ch) const { return Find(ch); }
3c67202d 2382 // use Truncate
c801d85f 2383 wxString& Remove(size_t pos) { return Truncate(pos); }
e87b7833 2384 wxString& RemoveLast(size_t n = 1) { return Truncate(length() - n); }
c801d85f 2385
e87b7833
MB
2386 wxString& Remove(size_t nStart, size_t nLen)
2387 { return (wxString&)erase( nStart, nLen ); }
dd1eaa89 2388
3c67202d 2389 // use Find()
8f93a29f 2390 int First( wxUniChar ch ) const { return Find(ch); }
8a540c88 2391 int First( wxUniCharRef ch ) const { return Find(ch); }
c9f78968 2392 int First( char ch ) const { return Find(ch); }
50e02008 2393 int First( unsigned char ch ) const { return Find(ch); }
c9f78968 2394 int First( wchar_t ch ) const { return Find(ch); }
8a540c88 2395 int First( const wxString& str ) const { return Find(str); }
8f93a29f 2396 int Last( wxUniChar ch ) const { return Find(ch, true); }
d775fa82 2397 bool Contains(const wxString& str) const { return Find(str) != wxNOT_FOUND; }
c801d85f 2398
5a8231ef
WS
2399 // use empty()
2400 bool IsNull() const { return empty(); }
c801d85f 2401
3c67202d 2402 // std::string compatibility functions
dd1eaa89 2403
3c67202d
VZ
2404 // take nLen chars starting at nPos
2405 wxString(const wxString& str, size_t nPos, size_t nLen)
1545d942 2406 { assign(str, nPos, nLen); }
f2a1b1bd 2407 // take all characters from first to last
e87b7833 2408 wxString(const_iterator first, const_iterator last)
cf9a878b 2409 : m_impl(first.impl(), last.impl()) { }
67cff9dc
VZ
2410#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2411 // the 2 overloads below are for compatibility with the existing code using
2412 // pointers instead of iterators
f2a1b1bd
VZ
2413 wxString(const char *first, const char *last)
2414 {
2415 SubstrBufFromMB str(ImplStr(first, last - first));
2416 m_impl.assign(str.data, str.len);
2417 }
2418 wxString(const wchar_t *first, const wchar_t *last)
2419 {
2420 SubstrBufFromWC str(ImplStr(first, last - first));
2421 m_impl.assign(str.data, str.len);
2422 }
67cff9dc
VZ
2423 // and this one is needed to compile code adding offsets to c_str() result
2424 wxString(const wxCStrData& first, const wxCStrData& last)
cf9a878b
VS
2425 : m_impl(CreateConstIterator(first).impl(),
2426 CreateConstIterator(last).impl())
67cff9dc
VZ
2427 {
2428 wxASSERT_MSG( first.m_str == last.m_str,
9a83f860 2429 wxT("pointers must be into the same string") );
67cff9dc
VZ
2430 }
2431#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
5460b9e3 2432
3c67202d 2433 // lib.string.modifiers
3c67202d
VZ
2434 // append elements str[pos], ..., str[pos+n]
2435 wxString& append(const wxString& str, size_t pos, size_t n)
8f93a29f 2436 {
68482dc5
VZ
2437 wxSTRING_UPDATE_CACHED_LENGTH(n);
2438
2439 size_t from, len;
2440 str.PosLenToImpl(pos, n, &from, &len);
2441 m_impl.append(str.m_impl, from, len);
2442 return *this;
8f93a29f 2443 }
e87b7833
MB
2444 // append a string
2445 wxString& append(const wxString& str)
68482dc5
VZ
2446 {
2447 wxSTRING_UPDATE_CACHED_LENGTH(str.length());
2448
2449 m_impl.append(str.m_impl);
2450 return *this;
2451 }
2452
3c67202d 2453 // append first n (or all if n == npos) characters of sz
8f93a29f 2454 wxString& append(const char *sz)
68482dc5
VZ
2455 {
2456 wxSTRING_INVALIDATE_CACHED_LENGTH();
2457
2458 m_impl.append(ImplStr(sz));
2459 return *this;
2460 }
2461
8f93a29f 2462 wxString& append(const wchar_t *sz)
68482dc5
VZ
2463 {
2464 wxSTRING_INVALIDATE_CACHED_LENGTH();
2465
2466 m_impl.append(ImplStr(sz));
2467 return *this;
2468 }
2469
8f93a29f
VS
2470 wxString& append(const char *sz, size_t n)
2471 {
68482dc5
VZ
2472 wxSTRING_INVALIDATE_CACHED_LENGTH();
2473
2474 SubstrBufFromMB str(ImplStr(sz, n));
2475 m_impl.append(str.data, str.len);
2476 return *this;
8f93a29f
VS
2477 }
2478 wxString& append(const wchar_t *sz, size_t n)
2479 {
68482dc5
VZ
2480 wxSTRING_UPDATE_CACHED_LENGTH(n);
2481
2482 SubstrBufFromWC str(ImplStr(sz, n));
2483 m_impl.append(str.data, str.len);
2484 return *this;
8f93a29f 2485 }
d3cefe87
VS
2486
2487 wxString& append(const wxCStrData& str)
2488 { return append(str.AsString()); }
de4983f3 2489 wxString& append(const wxScopedCharBuffer& str)
38d26d60 2490 { return append(str.data(), str.length()); }
de4983f3 2491 wxString& append(const wxScopedWCharBuffer& str)
38d26d60 2492 { return append(str.data(), str.length()); }
f416695a
VS
2493 wxString& append(const wxCStrData& str, size_t n)
2494 { return append(str.AsString(), 0, n); }
de4983f3 2495 wxString& append(const wxScopedCharBuffer& str, size_t n)
f416695a 2496 { return append(str.data(), n); }
de4983f3 2497 wxString& append(const wxScopedWCharBuffer& str, size_t n)
f416695a 2498 { return append(str.data(), n); }
d3cefe87 2499
3c67202d 2500 // append n copies of ch
c9f78968 2501 wxString& append(size_t n, wxUniChar ch)
8f93a29f
VS
2502 {
2503#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2504 if ( !ch.IsAscii() )
2505 {
2506 wxSTRING_INVALIDATE_CACHED_LENGTH();
2507
2508 m_impl.append(wxStringOperations::EncodeNChars(n, ch));
2509 }
2510 else // ASCII
8f93a29f 2511#endif
68482dc5
VZ
2512 {
2513 wxSTRING_UPDATE_CACHED_LENGTH(n);
2514
2515 m_impl.append(n, (wxStringCharType)ch);
2516 }
2517
2518 return *this;
8f93a29f 2519 }
68482dc5 2520
f416695a
VS
2521 wxString& append(size_t n, wxUniCharRef ch)
2522 { return append(n, wxUniChar(ch)); }
2523 wxString& append(size_t n, char ch)
2524 { return append(n, wxUniChar(ch)); }
2525 wxString& append(size_t n, unsigned char ch)
2526 { return append(n, wxUniChar(ch)); }
2527 wxString& append(size_t n, wchar_t ch)
2528 { return append(n, wxUniChar(ch)); }
2529
e87b7833
MB
2530 // append from first to last
2531 wxString& append(const_iterator first, const_iterator last)
68482dc5
VZ
2532 {
2533 wxSTRING_INVALIDATE_CACHED_LENGTH();
2534
2535 m_impl.append(first.impl(), last.impl());
2536 return *this;
2537 }
67cff9dc 2538#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
f2a1b1bd
VZ
2539 wxString& append(const char *first, const char *last)
2540 { return append(first, last - first); }
2541 wxString& append(const wchar_t *first, const wchar_t *last)
2542 { return append(first, last - first); }
67cff9dc
VZ
2543 wxString& append(const wxCStrData& first, const wxCStrData& last)
2544 { return append(CreateConstIterator(first), CreateConstIterator(last)); }
2545#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
3c67202d
VZ
2546
2547 // same as `this_string = str'
735d1db6 2548 wxString& assign(const wxString& str)
68482dc5
VZ
2549 {
2550 wxSTRING_SET_CACHED_LENGTH(str.length());
2551
2552 m_impl = str.m_impl;
2553
2554 return *this;
2555 }
2556
11aac4ba
VS
2557 wxString& assign(const wxString& str, size_t len)
2558 {
68482dc5
VZ
2559 wxSTRING_SET_CACHED_LENGTH(len);
2560
2561 m_impl.assign(str.m_impl, 0, str.LenToImpl(len));
2562
2563 return *this;
11aac4ba 2564 }
68482dc5 2565
3c67202d
VZ
2566 // same as ` = str[pos..pos + n]
2567 wxString& assign(const wxString& str, size_t pos, size_t n)
8f93a29f 2568 {
68482dc5
VZ
2569 size_t from, len;
2570 str.PosLenToImpl(pos, n, &from, &len);
2571 m_impl.assign(str.m_impl, from, len);
2572
2573 // it's important to call this after PosLenToImpl() above in case str is
2574 // the same string as this one
2575 wxSTRING_SET_CACHED_LENGTH(n);
2576
2577 return *this;
8f93a29f 2578 }
68482dc5 2579
3c67202d 2580 // same as `= first n (or all if n == npos) characters of sz'
8f93a29f 2581 wxString& assign(const char *sz)
68482dc5
VZ
2582 {
2583 wxSTRING_INVALIDATE_CACHE();
2584
2585 m_impl.assign(ImplStr(sz));
2586
2587 return *this;
2588 }
2589
8f93a29f 2590 wxString& assign(const wchar_t *sz)
68482dc5
VZ
2591 {
2592 wxSTRING_INVALIDATE_CACHE();
2593
2594 m_impl.assign(ImplStr(sz));
2595
2596 return *this;
2597 }
2598
8f93a29f
VS
2599 wxString& assign(const char *sz, size_t n)
2600 {
68482dc5
VZ
2601 wxSTRING_SET_CACHED_LENGTH(n);
2602
2603 SubstrBufFromMB str(ImplStr(sz, n));
2604 m_impl.assign(str.data, str.len);
2605
2606 return *this;
8f93a29f 2607 }
68482dc5 2608
8f93a29f
VS
2609 wxString& assign(const wchar_t *sz, size_t n)
2610 {
68482dc5
VZ
2611 wxSTRING_SET_CACHED_LENGTH(n);
2612
2613 SubstrBufFromWC str(ImplStr(sz, n));
2614 m_impl.assign(str.data, str.len);
2615
2616 return *this;
8f93a29f 2617 }
d3cefe87
VS
2618
2619 wxString& assign(const wxCStrData& str)
2620 { return assign(str.AsString()); }
de4983f3 2621 wxString& assign(const wxScopedCharBuffer& str)
38d26d60 2622 { return assign(str.data(), str.length()); }
de4983f3 2623 wxString& assign(const wxScopedWCharBuffer& str)
38d26d60 2624 { return assign(str.data(), str.length()); }
d3cefe87
VS
2625 wxString& assign(const wxCStrData& str, size_t len)
2626 { return assign(str.AsString(), len); }
de4983f3 2627 wxString& assign(const wxScopedCharBuffer& str, size_t len)
d3cefe87 2628 { return assign(str.data(), len); }
de4983f3 2629 wxString& assign(const wxScopedWCharBuffer& str, size_t len)
d3cefe87
VS
2630 { return assign(str.data(), len); }
2631
3c67202d 2632 // same as `= n copies of ch'
c9f78968 2633 wxString& assign(size_t n, wxUniChar ch)
8f93a29f 2634 {
68482dc5
VZ
2635 wxSTRING_SET_CACHED_LENGTH(n);
2636
8f93a29f 2637#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2638 if ( !ch.IsAscii() )
2639 m_impl.assign(wxStringOperations::EncodeNChars(n, ch));
2640 else
8f93a29f 2641#endif
68482dc5
VZ
2642 m_impl.assign(n, (wxStringCharType)ch);
2643
2644 return *this;
8f93a29f 2645 }
11aac4ba
VS
2646
2647 wxString& assign(size_t n, wxUniCharRef ch)
2648 { return assign(n, wxUniChar(ch)); }
2649 wxString& assign(size_t n, char ch)
2650 { return assign(n, wxUniChar(ch)); }
2651 wxString& assign(size_t n, unsigned char ch)
2652 { return assign(n, wxUniChar(ch)); }
2653 wxString& assign(size_t n, wchar_t ch)
2654 { return assign(n, wxUniChar(ch)); }
2655
e87b7833
MB
2656 // assign from first to last
2657 wxString& assign(const_iterator first, const_iterator last)
68482dc5
VZ
2658 {
2659 wxSTRING_INVALIDATE_CACHE();
2660
2661 m_impl.assign(first.impl(), last.impl());
2662
2663 return *this;
2664 }
67cff9dc 2665#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
f2a1b1bd
VZ
2666 wxString& assign(const char *first, const char *last)
2667 { return assign(first, last - first); }
2668 wxString& assign(const wchar_t *first, const wchar_t *last)
2669 { return assign(first, last - first); }
67cff9dc
VZ
2670 wxString& assign(const wxCStrData& first, const wxCStrData& last)
2671 { return assign(CreateConstIterator(first), CreateConstIterator(last)); }
2672#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
e87b7833
MB
2673
2674 // string comparison
8f93a29f 2675 int compare(const wxString& str) const;
d3cefe87
VS
2676 int compare(const char* sz) const;
2677 int compare(const wchar_t* sz) const;
2678 int compare(const wxCStrData& str) const
2679 { return compare(str.AsString()); }
de4983f3 2680 int compare(const wxScopedCharBuffer& str) const
d3cefe87 2681 { return compare(str.data()); }
de4983f3 2682 int compare(const wxScopedWCharBuffer& str) const
d3cefe87 2683 { return compare(str.data()); }
e87b7833 2684 // comparison with a substring
8f93a29f 2685 int compare(size_t nStart, size_t nLen, const wxString& str) const;
e87b7833
MB
2686 // comparison of 2 substrings
2687 int compare(size_t nStart, size_t nLen,
8f93a29f 2688 const wxString& str, size_t nStart2, size_t nLen2) const;
e87b7833
MB
2689 // substring comparison with first nCount characters of sz
2690 int compare(size_t nStart, size_t nLen,
8f93a29f
VS
2691 const char* sz, size_t nCount = npos) const;
2692 int compare(size_t nStart, size_t nLen,
2693 const wchar_t* sz, size_t nCount = npos) const;
3c67202d
VZ
2694
2695 // insert another string
e87b7833 2696 wxString& insert(size_t nPos, const wxString& str)
68482dc5 2697 { insert(GetIterForNthChar(nPos), str.begin(), str.end()); return *this; }
3c67202d
VZ
2698 // insert n chars of str starting at nStart (in str)
2699 wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n)
8f93a29f 2700 {
68482dc5
VZ
2701 wxSTRING_UPDATE_CACHED_LENGTH(n);
2702
2703 size_t from, len;
2704 str.PosLenToImpl(nStart, n, &from, &len);
2705 m_impl.insert(PosToImpl(nPos), str.m_impl, from, len);
2706
2707 return *this;
8f93a29f 2708 }
68482dc5 2709
3c67202d 2710 // insert first n (or all if n == npos) characters of sz
8f93a29f 2711 wxString& insert(size_t nPos, const char *sz)
68482dc5
VZ
2712 {
2713 wxSTRING_INVALIDATE_CACHE();
2714
2715 m_impl.insert(PosToImpl(nPos), ImplStr(sz));
2716
2717 return *this;
2718 }
2719
8f93a29f 2720 wxString& insert(size_t nPos, const wchar_t *sz)
68482dc5
VZ
2721 {
2722 wxSTRING_INVALIDATE_CACHE();
2723
2724 m_impl.insert(PosToImpl(nPos), ImplStr(sz)); return *this;
2725 }
2726
8f93a29f
VS
2727 wxString& insert(size_t nPos, const char *sz, size_t n)
2728 {
68482dc5
VZ
2729 wxSTRING_UPDATE_CACHED_LENGTH(n);
2730
2731 SubstrBufFromMB str(ImplStr(sz, n));
2732 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2733
2734 return *this;
8f93a29f 2735 }
68482dc5 2736
8f93a29f
VS
2737 wxString& insert(size_t nPos, const wchar_t *sz, size_t n)
2738 {
68482dc5
VZ
2739 wxSTRING_UPDATE_CACHED_LENGTH(n);
2740
2741 SubstrBufFromWC str(ImplStr(sz, n));
2742 m_impl.insert(PosToImpl(nPos), str.data, str.len);
2743
2744 return *this;
8f93a29f 2745 }
68482dc5 2746
3c67202d 2747 // insert n copies of ch
c9f78968 2748 wxString& insert(size_t nPos, size_t n, wxUniChar ch)
8f93a29f 2749 {
68482dc5
VZ
2750 wxSTRING_UPDATE_CACHED_LENGTH(n);
2751
8f93a29f 2752#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2753 if ( !ch.IsAscii() )
2754 m_impl.insert(PosToImpl(nPos), wxStringOperations::EncodeNChars(n, ch));
2755 else
8f93a29f 2756#endif
68482dc5
VZ
2757 m_impl.insert(PosToImpl(nPos), n, (wxStringCharType)ch);
2758 return *this;
8f93a29f 2759 }
68482dc5 2760
c9f78968 2761 iterator insert(iterator it, wxUniChar ch)
81727065 2762 {
68482dc5
VZ
2763 wxSTRING_UPDATE_CACHED_LENGTH(1);
2764
81727065 2765#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2766 if ( !ch.IsAscii() )
2767 {
2768 size_t pos = IterToImplPos(it);
2769 m_impl.insert(pos, wxStringOperations::EncodeChar(ch));
2770 return iterator(this, m_impl.begin() + pos);
2771 }
2772 else
81727065 2773#endif
68482dc5 2774 return iterator(this, m_impl.insert(it.impl(), (wxStringCharType)ch));
81727065 2775 }
68482dc5 2776
e87b7833 2777 void insert(iterator it, const_iterator first, const_iterator last)
68482dc5
VZ
2778 {
2779 wxSTRING_INVALIDATE_CACHE();
2780
2781 m_impl.insert(it.impl(), first.impl(), last.impl());
2782 }
2783
67cff9dc 2784#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
f2a1b1bd
VZ
2785 void insert(iterator it, const char *first, const char *last)
2786 { insert(it - begin(), first, last - first); }
67cff9dc 2787 void insert(iterator it, const wchar_t *first, const wchar_t *last)
f2a1b1bd 2788 { insert(it - begin(), first, last - first); }
67cff9dc
VZ
2789 void insert(iterator it, const wxCStrData& first, const wxCStrData& last)
2790 { insert(it, CreateConstIterator(first), CreateConstIterator(last)); }
2791#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER
2792
c9f78968 2793 void insert(iterator it, size_type n, wxUniChar ch)
8f93a29f 2794 {
68482dc5
VZ
2795 wxSTRING_UPDATE_CACHED_LENGTH(n);
2796
8f93a29f 2797#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2798 if ( !ch.IsAscii() )
2799 m_impl.insert(IterToImplPos(it), wxStringOperations::EncodeNChars(n, ch));
2800 else
8f93a29f 2801#endif
68482dc5 2802 m_impl.insert(it.impl(), n, (wxStringCharType)ch);
8f93a29f 2803 }
3c67202d
VZ
2804
2805 // delete characters from nStart to nStart + nLen
e87b7833 2806 wxString& erase(size_type pos = 0, size_type n = npos)
8f93a29f 2807 {
68482dc5
VZ
2808 wxSTRING_INVALIDATE_CACHE();
2809
2810 size_t from, len;
2811 PosLenToImpl(pos, n, &from, &len);
2812 m_impl.erase(from, len);
2813
2814 return *this;
8f93a29f 2815 }
68482dc5 2816
f2a1b1bd 2817 // delete characters from first up to last
e87b7833 2818 iterator erase(iterator first, iterator last)
68482dc5
VZ
2819 {
2820 wxSTRING_INVALIDATE_CACHE();
2821
2822 return iterator(this, m_impl.erase(first.impl(), last.impl()));
2823 }
2824
e87b7833 2825 iterator erase(iterator first)
68482dc5
VZ
2826 {
2827 wxSTRING_UPDATE_CACHED_LENGTH(-1);
2828
2829 return iterator(this, m_impl.erase(first.impl()));
2830 }
e87b7833
MB
2831
2832#ifdef wxSTRING_BASE_HASNT_CLEAR
2833 void clear() { erase(); }
8f93a29f 2834#else
68482dc5
VZ
2835 void clear()
2836 {
2837 wxSTRING_SET_CACHED_LENGTH(0);
2838
2839 m_impl.clear();
2840 }
e87b7833 2841#endif
3c67202d
VZ
2842
2843 // replaces the substring of length nLen starting at nStart
8f93a29f
VS
2844 wxString& replace(size_t nStart, size_t nLen, const char* sz)
2845 {
68482dc5
VZ
2846 wxSTRING_INVALIDATE_CACHE();
2847
2848 size_t from, len;
2849 PosLenToImpl(nStart, nLen, &from, &len);
2850 m_impl.replace(from, len, ImplStr(sz));
2851
2852 return *this;
8f93a29f 2853 }
68482dc5 2854
8f93a29f
VS
2855 wxString& replace(size_t nStart, size_t nLen, const wchar_t* sz)
2856 {
68482dc5
VZ
2857 wxSTRING_INVALIDATE_CACHE();
2858
2859 size_t from, len;
2860 PosLenToImpl(nStart, nLen, &from, &len);
2861 m_impl.replace(from, len, ImplStr(sz));
2862
2863 return *this;
8f93a29f 2864 }
68482dc5 2865
e87b7833
MB
2866 // replaces the substring of length nLen starting at nStart
2867 wxString& replace(size_t nStart, size_t nLen, const wxString& str)
8f93a29f 2868 {
68482dc5
VZ
2869 wxSTRING_INVALIDATE_CACHE();
2870
2871 size_t from, len;
2872 PosLenToImpl(nStart, nLen, &from, &len);
2873 m_impl.replace(from, len, str.m_impl);
2874
2875 return *this;
8f93a29f 2876 }
68482dc5 2877
3c67202d 2878 // replaces the substring with nCount copies of ch
c9f78968 2879 wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxUniChar ch)
8f93a29f 2880 {
68482dc5
VZ
2881 wxSTRING_INVALIDATE_CACHE();
2882
2883 size_t from, len;
2884 PosLenToImpl(nStart, nLen, &from, &len);
8f93a29f 2885#if wxUSE_UNICODE_UTF8
68482dc5
VZ
2886 if ( !ch.IsAscii() )
2887 m_impl.replace(from, len, wxStringOperations::EncodeNChars(nCount, ch));
2888 else
8f93a29f 2889#endif
68482dc5
VZ
2890 m_impl.replace(from, len, nCount, (wxStringCharType)ch);
2891
2892 return *this;
8f93a29f 2893 }
68482dc5 2894
3c67202d
VZ
2895 // replaces a substring with another substring
2896 wxString& replace(size_t nStart, size_t nLen,
e87b7833 2897 const wxString& str, size_t nStart2, size_t nLen2)
8f93a29f 2898 {
68482dc5 2899 wxSTRING_INVALIDATE_CACHE();
64a044d5 2900
68482dc5
VZ
2901 size_t from, len;
2902 PosLenToImpl(nStart, nLen, &from, &len);
64a044d5 2903
68482dc5
VZ
2904 size_t from2, len2;
2905 str.PosLenToImpl(nStart2, nLen2, &from2, &len2);
2906
2907 m_impl.replace(from, len, str.m_impl, from2, len2);
2908
2909 return *this;
8f93a29f 2910 }
68482dc5 2911
e87b7833 2912 // replaces the substring with first nCount chars of sz
fb20fa43 2913 wxString& replace(size_t nStart, size_t nLen,
8f93a29f
VS
2914 const char* sz, size_t nCount)
2915 {
68482dc5 2916 wxSTRING_INVALIDATE_CACHE();
64a044d5 2917
68482dc5
VZ
2918 size_t from, len;
2919 PosLenToImpl(nStart, nLen, &from, &len);
64a044d5 2920
68482dc5
VZ
2921 SubstrBufFromMB str(ImplStr(sz, nCount));
2922
2923 m_impl.replace(from, len, str.data, str.len);
2924
2925 return *this;
8f93a29f 2926 }
68482dc5 2927
8f93a29f
VS
2928 wxString& replace(size_t nStart, size_t nLen,
2929 const wchar_t* sz, size_t nCount)
2930 {
68482dc5 2931 wxSTRING_INVALIDATE_CACHE();
64a044d5 2932
68482dc5
VZ
2933 size_t from, len;
2934 PosLenToImpl(nStart, nLen, &from, &len);
64a044d5 2935
68482dc5
VZ
2936 SubstrBufFromWC str(ImplStr(sz, nCount));
2937
2938 m_impl.replace(from, len, str.data, str.len);
2939
2940 return *this;
8f93a29f 2941 }
68482dc5 2942
11aac4ba
VS
2943 wxString& replace(size_t nStart, size_t nLen,
2944 const wxString& s, size_t nCount)
2945 {
68482dc5
VZ
2946 wxSTRING_INVALIDATE_CACHE();
2947
2948 size_t from, len;
2949 PosLenToImpl(nStart, nLen, &from, &len);
2950 m_impl.replace(from, len, s.m_impl.c_str(), s.LenToImpl(nCount));
2951
2952 return *this;
11aac4ba
VS
2953 }
2954
8f93a29f 2955 wxString& replace(iterator first, iterator last, const char* s)
68482dc5
VZ
2956 {
2957 wxSTRING_INVALIDATE_CACHE();
2958
2959 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2960
2961 return *this;
2962 }
2963
8f93a29f 2964 wxString& replace(iterator first, iterator last, const wchar_t* s)
68482dc5
VZ
2965 {
2966 wxSTRING_INVALIDATE_CACHE();
2967
2968 m_impl.replace(first.impl(), last.impl(), ImplStr(s));
2969
2970 return *this;
2971 }
2972
8f93a29f
VS
2973 wxString& replace(iterator first, iterator last, const char* s, size_type n)
2974 {
68482dc5
VZ
2975 wxSTRING_INVALIDATE_CACHE();
2976
2977 SubstrBufFromMB str(ImplStr(s, n));
2978 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2979
2980 return *this;
8f93a29f 2981 }
68482dc5 2982
8f93a29f
VS
2983 wxString& replace(iterator first, iterator last, const wchar_t* s, size_type n)
2984 {
68482dc5
VZ
2985 wxSTRING_INVALIDATE_CACHE();
2986
2987 SubstrBufFromWC str(ImplStr(s, n));
2988 m_impl.replace(first.impl(), last.impl(), str.data, str.len);
2989
2990 return *this;
8f93a29f 2991 }
68482dc5 2992
e87b7833 2993 wxString& replace(iterator first, iterator last, const wxString& s)
68482dc5
VZ
2994 {
2995 wxSTRING_INVALIDATE_CACHE();
2996
2997 m_impl.replace(first.impl(), last.impl(), s.m_impl);
2998
2999 return *this;
3000 }
3001
8f93a29f
VS
3002 wxString& replace(iterator first, iterator last, size_type n, wxUniChar ch)
3003 {
68482dc5
VZ
3004 wxSTRING_INVALIDATE_CACHE();
3005
8f93a29f 3006#if wxUSE_UNICODE_UTF8
68482dc5
VZ
3007 if ( !ch.IsAscii() )
3008 m_impl.replace(first.impl(), last.impl(),
3009 wxStringOperations::EncodeNChars(n, ch));
3010 else
8f93a29f 3011#endif
68482dc5
VZ
3012 m_impl.replace(first.impl(), last.impl(), n, (wxStringCharType)ch);
3013
3014 return *this;
8f93a29f 3015 }
68482dc5 3016
e87b7833
MB
3017 wxString& replace(iterator first, iterator last,
3018 const_iterator first1, const_iterator last1)
cf9a878b 3019 {
68482dc5
VZ
3020 wxSTRING_INVALIDATE_CACHE();
3021
3022 m_impl.replace(first.impl(), last.impl(), first1.impl(), last1.impl());
3023
3024 return *this;
cf9a878b 3025 }
68482dc5 3026
f2a1b1bd
VZ
3027 wxString& replace(iterator first, iterator last,
3028 const char *first1, const char *last1)
3029 { replace(first, last, first1, last1 - first1); return *this; }
3030 wxString& replace(iterator first, iterator last,
3031 const wchar_t *first1, const wchar_t *last1)
3032 { replace(first, last, first1, last1 - first1); return *this; }
8f93a29f
VS
3033
3034 // swap two strings
3035 void swap(wxString& str)
68482dc5 3036 {
af912213
VZ
3037#if wxUSE_STRING_POS_CACHE
3038 // we modify not only this string but also the other one directly so we
3039 // need to invalidate cache for both of them (we could also try to
3040 // exchange their cache entries but it seems unlikely to be worth it)
3041 InvalidateCache();
3042 str.InvalidateCache();
3043#endif // wxUSE_STRING_POS_CACHE
68482dc5
VZ
3044
3045 m_impl.swap(str.m_impl);
3046 }
8f93a29f
VS
3047
3048 // find a substring
3049 size_t find(const wxString& str, size_t nStart = 0) const
3050 { return PosFromImpl(m_impl.find(str.m_impl, PosToImpl(nStart))); }
3051
3052 // find first n characters of sz
3053 size_t find(const char* sz, size_t nStart = 0, size_t n = npos) const
3054 {
3055 SubstrBufFromMB str(ImplStr(sz, n));
3056 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
3057 }
3058 size_t find(const wchar_t* sz, size_t nStart = 0, size_t n = npos) const
3059 {
3060 SubstrBufFromWC str(ImplStr(sz, n));
3061 return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len));
3062 }
de4983f3 3063 size_t find(const wxScopedCharBuffer& s, size_t nStart = 0, size_t n = npos) const
d3cefe87 3064 { return find(s.data(), nStart, n); }
de4983f3 3065 size_t find(const wxScopedWCharBuffer& s, size_t nStart = 0, size_t n = npos) const
d3cefe87
VS
3066 { return find(s.data(), nStart, n); }
3067 size_t find(const wxCStrData& s, size_t nStart = 0, size_t n = npos) const
3068 { return find(s.AsWChar(), nStart, n); }
8f93a29f 3069
7929902d 3070 // find the first occurrence of character ch after nStart
8f93a29f 3071 size_t find(wxUniChar ch, size_t nStart = 0) const
467175ab 3072 {
84d2877f
VS
3073#if wxUSE_UNICODE_UTF8
3074 if ( !ch.IsAscii() )
3075 return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch),
3076 PosToImpl(nStart)));
3077 else
3078#endif
3079 return PosFromImpl(m_impl.find((wxStringCharType)ch,
3080 PosToImpl(nStart)));
3081
467175ab 3082 }
8f93a29f
VS
3083 size_t find(wxUniCharRef ch, size_t nStart = 0) const
3084 { return find(wxUniChar(ch), nStart); }
3085 size_t find(char ch, size_t nStart = 0) const
3086 { return find(wxUniChar(ch), nStart); }
50e02008
VS
3087 size_t find(unsigned char ch, size_t nStart = 0) const
3088 { return find(wxUniChar(ch), nStart); }
8f93a29f
VS
3089 size_t find(wchar_t ch, size_t nStart = 0) const
3090 { return find(wxUniChar(ch), nStart); }
3091
3092 // rfind() family is exactly like find() but works right to left
3093
3094 // as find, but from the end
3095 size_t rfind(const wxString& str, size_t nStart = npos) const
3096 { return PosFromImpl(m_impl.rfind(str.m_impl, PosToImpl(nStart))); }
3097
3098 // as find, but from the end
3099 size_t rfind(const char* sz, size_t nStart = npos, size_t n = npos) const
3100 {
3101 SubstrBufFromMB str(ImplStr(sz, n));
3102 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
3103 }
3104 size_t rfind(const wchar_t* sz, size_t nStart = npos, size_t n = npos) const
3105 {
3106 SubstrBufFromWC str(ImplStr(sz, n));
3107 return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len));
3108 }
de4983f3 3109 size_t rfind(const wxScopedCharBuffer& s, size_t nStart = npos, size_t n = npos) const
d3cefe87 3110 { return rfind(s.data(), nStart, n); }
de4983f3 3111 size_t rfind(const wxScopedWCharBuffer& s, size_t nStart = npos, size_t n = npos) const
d3cefe87
VS
3112 { return rfind(s.data(), nStart, n); }
3113 size_t rfind(const wxCStrData& s, size_t nStart = npos, size_t n = npos) const
3114 { return rfind(s.AsWChar(), nStart, n); }
8f93a29f
VS
3115 // as find, but from the end
3116 size_t rfind(wxUniChar ch, size_t nStart = npos) const
467175ab 3117 {
84d2877f
VS
3118#if wxUSE_UNICODE_UTF8
3119 if ( !ch.IsAscii() )
3120 return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch),
3121 PosToImpl(nStart)));
3122 else
3123#endif
3124 return PosFromImpl(m_impl.rfind((wxStringCharType)ch,
3125 PosToImpl(nStart)));
467175ab 3126 }
8f93a29f
VS
3127 size_t rfind(wxUniCharRef ch, size_t nStart = npos) const
3128 { return rfind(wxUniChar(ch), nStart); }
3129 size_t rfind(char ch, size_t nStart = npos) const
3130 { return rfind(wxUniChar(ch), nStart); }
50e02008
VS
3131 size_t rfind(unsigned char ch, size_t nStart = npos) const
3132 { return rfind(wxUniChar(ch), nStart); }
8f93a29f
VS
3133 size_t rfind(wchar_t ch, size_t nStart = npos) const
3134 { return rfind(wxUniChar(ch), nStart); }
3135
7929902d 3136 // find first/last occurrence of any character (not) in the set:
8f93a29f
VS
3137#if wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
3138 // FIXME-UTF8: this is not entirely correct, because it doesn't work if
3139 // sizeof(wchar_t)==2 and surrogates are present in the string;
3140 // should we care? Probably not.
3141 size_t find_first_of(const wxString& str, size_t nStart = 0) const
a962cdf4 3142 { return m_impl.find_first_of(str.m_impl, nStart); }
8f93a29f
VS
3143 size_t find_first_of(const char* sz, size_t nStart = 0) const
3144 { return m_impl.find_first_of(ImplStr(sz), nStart); }
3145 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const
3146 { return m_impl.find_first_of(ImplStr(sz), nStart); }
a962cdf4 3147 size_t find_first_of(const char* sz, size_t nStart, size_t n) const
8f93a29f 3148 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
a962cdf4 3149 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const
8f93a29f
VS
3150 { return m_impl.find_first_of(ImplStr(sz), nStart, n); }
3151 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3152 { return m_impl.find_first_of((wxChar)c, nStart); }
3153
a962cdf4
VS
3154 size_t find_last_of(const wxString& str, size_t nStart = npos) const
3155 { return m_impl.find_last_of(str.m_impl, nStart); }
8f93a29f
VS
3156 size_t find_last_of(const char* sz, size_t nStart = npos) const
3157 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3158 size_t find_last_of(const wchar_t* sz, size_t nStart = npos) const
3159 { return m_impl.find_last_of(ImplStr(sz), nStart); }
3160 size_t find_last_of(const char* sz, size_t nStart, size_t n) const
3161 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3162 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const
3163 { return m_impl.find_last_of(ImplStr(sz), nStart, n); }
3164 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3165 { return m_impl.find_last_of((wxChar)c, nStart); }
3166
a962cdf4 3167 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
8f93a29f 3168 { return m_impl.find_first_not_of(str.m_impl, nStart); }
a962cdf4 3169 size_t find_first_not_of(const char* sz, size_t nStart = 0) const
8f93a29f 3170 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
a962cdf4 3171 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const
8f93a29f 3172 { return m_impl.find_first_not_of(ImplStr(sz), nStart); }
a962cdf4 3173 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const
8f93a29f 3174 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
a962cdf4 3175 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const
8f93a29f 3176 { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); }
a962cdf4 3177 size_t find_first_not_of(wxUniChar c, size_t nStart = 0) const
8f93a29f
VS
3178 { return m_impl.find_first_not_of((wxChar)c, nStart); }
3179
a962cdf4 3180 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
8f93a29f 3181 { return m_impl.find_last_not_of(str.m_impl, nStart); }
a962cdf4 3182 size_t find_last_not_of(const char* sz, size_t nStart = npos) const
8f93a29f 3183 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
a962cdf4 3184 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const
8f93a29f 3185 { return m_impl.find_last_not_of(ImplStr(sz), nStart); }
a962cdf4 3186 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const
8f93a29f 3187 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
a962cdf4 3188 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const
8f93a29f 3189 { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); }
a962cdf4 3190 size_t find_last_not_of(wxUniChar c, size_t nStart = npos) const
8f93a29f
VS
3191 { return m_impl.find_last_not_of((wxChar)c, nStart); }
3192#else
3193 // we can't use std::string implementation in UTF-8 build, because the
3194 // character sets would be interpreted wrongly:
3195
3196 // as strpbrk() but starts at nStart, returns npos if not found
3197 size_t find_first_of(const wxString& str, size_t nStart = 0) const
81727065 3198#if wxUSE_UNICODE // FIXME-UTF8: temporary
d3cefe87 3199 { return find_first_of(str.wc_str(), nStart); }
81727065 3200#else
d3cefe87 3201 { return find_first_of(str.mb_str(), nStart); }
81727065 3202#endif
8f93a29f
VS
3203 // same as above
3204 size_t find_first_of(const char* sz, size_t nStart = 0) const;
3205 size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const;
3206 size_t find_first_of(const char* sz, size_t nStart, size_t n) const;
3207 size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const;
3208 // same as find(char, size_t)
3209 size_t find_first_of(wxUniChar c, size_t nStart = 0) const
3210 { return find(c, nStart); }
3211 // find the last (starting from nStart) char from str in this string
3212 size_t find_last_of (const wxString& str, size_t nStart = npos) const
81727065 3213#if wxUSE_UNICODE // FIXME-UTF8: temporary
d3cefe87 3214 { return find_last_of(str.wc_str(), nStart); }
81727065 3215#else
d3cefe87 3216 { return find_last_of(str.mb_str(), nStart); }
81727065 3217#endif
8f93a29f
VS
3218 // same as above
3219 size_t find_last_of (const char* sz, size_t nStart = npos) const;
3220 size_t find_last_of (const wchar_t* sz, size_t nStart = npos) const;
3221 size_t find_last_of(const char* sz, size_t nStart, size_t n) const;
3222 size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const;
3223 // same as above
3224 size_t find_last_of(wxUniChar c, size_t nStart = npos) const
3225 { return rfind(c, nStart); }
3226
7929902d 3227 // find first/last occurrence of any character not in the set
8f93a29f
VS
3228
3229 // as strspn() (starting from nStart), returns npos on failure
3230 size_t find_first_not_of(const wxString& str, size_t nStart = 0) const
81727065 3231#if wxUSE_UNICODE // FIXME-UTF8: temporary
d3cefe87 3232 { return find_first_not_of(str.wc_str(), nStart); }
81727065 3233#else
d3cefe87 3234 { return find_first_not_of(str.mb_str(), nStart); }
81727065 3235#endif
8f93a29f
VS
3236 // same as above
3237 size_t find_first_not_of(const char* sz, size_t nStart = 0) const;
3238 size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const;
3239 size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const;
3240 size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3241 // same as above
3242 size_t find_first_not_of(wxUniChar ch, size_t nStart = 0) const;
3243 // as strcspn()
3244 size_t find_last_not_of(const wxString& str, size_t nStart = npos) const
81727065 3245#if wxUSE_UNICODE // FIXME-UTF8: temporary
d3cefe87 3246 { return find_last_not_of(str.wc_str(), nStart); }
81727065 3247#else
d3cefe87 3248 { return find_last_not_of(str.mb_str(), nStart); }
81727065 3249#endif
8f93a29f
VS
3250 // same as above
3251 size_t find_last_not_of(const char* sz, size_t nStart = npos) const;
3252 size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const;
3253 size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const;
3254 size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const;
3255 // same as above
3256 size_t find_last_not_of(wxUniChar ch, size_t nStart = npos) const;
3257#endif // wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 or not
3258
3259 // provide char/wchar_t/wxUniCharRef overloads for char-finding functions
3260 // above to resolve ambiguities:
3261 size_t find_first_of(wxUniCharRef ch, size_t nStart = 0) const
3262 { return find_first_of(wxUniChar(ch), nStart); }
3263 size_t find_first_of(char ch, size_t nStart = 0) const
3264 { return find_first_of(wxUniChar(ch), nStart); }
50e02008
VS
3265 size_t find_first_of(unsigned char ch, size_t nStart = 0) const
3266 { return find_first_of(wxUniChar(ch), nStart); }
8f93a29f
VS
3267 size_t find_first_of(wchar_t ch, size_t nStart = 0) const
3268 { return find_first_of(wxUniChar(ch), nStart); }
3269 size_t find_last_of(wxUniCharRef ch, size_t nStart = npos) const
3270 { return find_last_of(wxUniChar(ch), nStart); }
3271 size_t find_last_of(char ch, size_t nStart = npos) const
3272 { return find_last_of(wxUniChar(ch), nStart); }
50e02008
VS
3273 size_t find_last_of(unsigned char ch, size_t nStart = npos) const
3274 { return find_last_of(wxUniChar(ch), nStart); }
8f93a29f
VS
3275 size_t find_last_of(wchar_t ch, size_t nStart = npos) const
3276 { return find_last_of(wxUniChar(ch), nStart); }
3277 size_t find_first_not_of(wxUniCharRef ch, size_t nStart = 0) const
3278 { return find_first_not_of(wxUniChar(ch), nStart); }
3279 size_t find_first_not_of(char ch, size_t nStart = 0) const
3280 { return find_first_not_of(wxUniChar(ch), nStart); }
50e02008
VS
3281 size_t find_first_not_of(unsigned char ch, size_t nStart = 0) const
3282 { return find_first_not_of(wxUniChar(ch), nStart); }
8f93a29f
VS
3283 size_t find_first_not_of(wchar_t ch, size_t nStart = 0) const
3284 { return find_first_not_of(wxUniChar(ch), nStart); }
3285 size_t find_last_not_of(wxUniCharRef ch, size_t nStart = npos) const
3286 { return find_last_not_of(wxUniChar(ch), nStart); }
3287 size_t find_last_not_of(char ch, size_t nStart = npos) const
3288 { return find_last_not_of(wxUniChar(ch), nStart); }
50e02008
VS
3289 size_t find_last_not_of(unsigned char ch, size_t nStart = npos) const
3290 { return find_last_not_of(wxUniChar(ch), nStart); }
8f93a29f
VS
3291 size_t find_last_not_of(wchar_t ch, size_t nStart = npos) const
3292 { return find_last_not_of(wxUniChar(ch), nStart); }
3c67202d 3293
d3cefe87
VS
3294 // and additional overloads for the versions taking strings:
3295 size_t find_first_of(const wxCStrData& sz, size_t nStart = 0) const
3296 { return find_first_of(sz.AsString(), nStart); }
de4983f3 3297 size_t find_first_of(const wxScopedCharBuffer& sz, size_t nStart = 0) const
d3cefe87 3298 { return find_first_of(sz.data(), nStart); }
de4983f3 3299 size_t find_first_of(const wxScopedWCharBuffer& sz, size_t nStart = 0) const
d3cefe87
VS
3300 { return find_first_of(sz.data(), nStart); }
3301 size_t find_first_of(const wxCStrData& sz, size_t nStart, size_t n) const
3302 { return find_first_of(sz.AsWChar(), nStart, n); }
de4983f3 3303 size_t find_first_of(const wxScopedCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87 3304 { return find_first_of(sz.data(), nStart, n); }
de4983f3 3305 size_t find_first_of(const wxScopedWCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87
VS
3306 { return find_first_of(sz.data(), nStart, n); }
3307
3308 size_t find_last_of(const wxCStrData& sz, size_t nStart = 0) const
3309 { return find_last_of(sz.AsString(), nStart); }
de4983f3 3310 size_t find_last_of(const wxScopedCharBuffer& sz, size_t nStart = 0) const
d3cefe87 3311 { return find_last_of(sz.data(), nStart); }
de4983f3 3312 size_t find_last_of(const wxScopedWCharBuffer& sz, size_t nStart = 0) const
d3cefe87
VS
3313 { return find_last_of(sz.data(), nStart); }
3314 size_t find_last_of(const wxCStrData& sz, size_t nStart, size_t n) const
3315 { return find_last_of(sz.AsWChar(), nStart, n); }
de4983f3 3316 size_t find_last_of(const wxScopedCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87 3317 { return find_last_of(sz.data(), nStart, n); }
de4983f3 3318 size_t find_last_of(const wxScopedWCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87
VS
3319 { return find_last_of(sz.data(), nStart, n); }
3320
3321 size_t find_first_not_of(const wxCStrData& sz, size_t nStart = 0) const
3322 { return find_first_not_of(sz.AsString(), nStart); }
de4983f3 3323 size_t find_first_not_of(const wxScopedCharBuffer& sz, size_t nStart = 0) const
d3cefe87 3324 { return find_first_not_of(sz.data(), nStart); }
de4983f3 3325 size_t find_first_not_of(const wxScopedWCharBuffer& sz, size_t nStart = 0) const
d3cefe87
VS
3326 { return find_first_not_of(sz.data(), nStart); }
3327 size_t find_first_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3328 { return find_first_not_of(sz.AsWChar(), nStart, n); }
de4983f3 3329 size_t find_first_not_of(const wxScopedCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87 3330 { return find_first_not_of(sz.data(), nStart, n); }
de4983f3 3331 size_t find_first_not_of(const wxScopedWCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87
VS
3332 { return find_first_not_of(sz.data(), nStart, n); }
3333
3334 size_t find_last_not_of(const wxCStrData& sz, size_t nStart = 0) const
3335 { return find_last_not_of(sz.AsString(), nStart); }
de4983f3 3336 size_t find_last_not_of(const wxScopedCharBuffer& sz, size_t nStart = 0) const
d3cefe87 3337 { return find_last_not_of(sz.data(), nStart); }
de4983f3 3338 size_t find_last_not_of(const wxScopedWCharBuffer& sz, size_t nStart = 0) const
d3cefe87
VS
3339 { return find_last_not_of(sz.data(), nStart); }
3340 size_t find_last_not_of(const wxCStrData& sz, size_t nStart, size_t n) const
3341 { return find_last_not_of(sz.AsWChar(), nStart, n); }
de4983f3 3342 size_t find_last_not_of(const wxScopedCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87 3343 { return find_last_not_of(sz.data(), nStart, n); }
de4983f3 3344 size_t find_last_not_of(const wxScopedWCharBuffer& sz, size_t nStart, size_t n) const
d3cefe87
VS
3345 { return find_last_not_of(sz.data(), nStart, n); }
3346
e87b7833
MB
3347 // string += string
3348 wxString& operator+=(const wxString& s)
68482dc5
VZ
3349 {
3350 wxSTRING_INVALIDATE_CACHED_LENGTH();
3351
3352 m_impl += s.m_impl;
3353 return *this;
3354 }
e87b7833 3355 // string += C string
8f93a29f 3356 wxString& operator+=(const char *psz)
68482dc5
VZ
3357 {
3358 wxSTRING_INVALIDATE_CACHED_LENGTH();
3359
3360 m_impl += ImplStr(psz);
3361 return *this;
3362 }
8f93a29f 3363 wxString& operator+=(const wchar_t *pwz)
68482dc5
VZ
3364 {
3365 wxSTRING_INVALIDATE_CACHED_LENGTH();
3366
3367 m_impl += ImplStr(pwz);
3368 return *this;
3369 }
c9f78968 3370 wxString& operator+=(const wxCStrData& s)
68482dc5
VZ
3371 {
3372 wxSTRING_INVALIDATE_CACHED_LENGTH();
3373
3374 m_impl += s.AsString().m_impl;
3375 return *this;
3376 }
de4983f3 3377 wxString& operator+=(const wxScopedCharBuffer& s)
38d26d60 3378 { return append(s); }
de4983f3 3379 wxString& operator+=(const wxScopedWCharBuffer& s)
38d26d60 3380 { return append(s); }
e87b7833 3381 // string += char
c9f78968 3382 wxString& operator+=(wxUniChar ch)
84d2877f 3383 {
68482dc5
VZ
3384 wxSTRING_UPDATE_CACHED_LENGTH(1);
3385
84d2877f 3386#if wxUSE_UNICODE_UTF8
68482dc5
VZ
3387 if ( !ch.IsAscii() )
3388 m_impl += wxStringOperations::EncodeChar(ch);
3389 else
84d2877f 3390#endif
68482dc5
VZ
3391 m_impl += (wxStringCharType)ch;
3392 return *this;
84d2877f 3393 }
c9f78968 3394 wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); }
b7d40341 3395 wxString& operator+=(int ch) { return *this += wxUniChar(ch); }
c9f78968 3396 wxString& operator+=(char ch) { return *this += wxUniChar(ch); }
b77afb41 3397 wxString& operator+=(unsigned char ch) { return *this += wxUniChar(ch); }
c9f78968 3398 wxString& operator+=(wchar_t ch) { return *this += wxUniChar(ch); }
d8a4b666
VS
3399
3400private:
2523e9b7 3401#if !wxUSE_STL_BASED_WXSTRING
d8a4b666 3402 // helpers for wxStringBuffer and wxStringBufferLength
8f93a29f 3403 wxStringCharType *DoGetWriteBuf(size_t nLen)
68482dc5
VZ
3404 {
3405 return m_impl.DoGetWriteBuf(nLen);
3406 }
3407
8f93a29f 3408 void DoUngetWriteBuf()
68482dc5
VZ
3409 {
3410 wxSTRING_INVALIDATE_CACHE();
3411
3412 m_impl.DoUngetWriteBuf();
3413 }
3414
8f93a29f 3415 void DoUngetWriteBuf(size_t nLen)
68482dc5
VZ
3416 {
3417 wxSTRING_SET_CACHED_LENGTH(nLen);
3418
3419 m_impl.DoUngetWriteBuf(nLen);
3420 }
2523e9b7 3421#endif // !wxUSE_STL_BASED_WXSTRING
c9f78968
VS
3422
3423#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
d1f6e2cf
VS
3424 #if !wxUSE_UTF8_LOCALE_ONLY
3425 int DoPrintfWchar(const wxChar *format, ...);
3426 static wxString DoFormatWchar(const wxChar *format, ...);
3427 #endif
3428 #if wxUSE_UNICODE_UTF8
3429 int DoPrintfUtf8(const char *format, ...);
3430 static wxString DoFormatUtf8(const char *format, ...);
3431 #endif
c9f78968 3432#endif
8f93a29f
VS
3433
3434#if !wxUSE_STL_BASED_WXSTRING
3435 // check string's data validity
3436 bool IsValid() const { return m_impl.GetStringData()->IsValid(); }
3437#endif
3438
3439private:
3440 wxStringImpl m_impl;
132276cf
VS
3441
3442 // buffers for compatibility conversion from (char*)c_str() and
f54cb154
VZ
3443 // (wchar_t*)c_str(): the pointers returned by these functions should remain
3444 // valid until the string itself is modified for compatibility with the
3445 // existing code and consistency with std::string::c_str() so returning a
3446 // temporary buffer won't do and we need to cache the conversion results
3447
3448 // TODO-UTF8: benchmark various approaches to keeping compatibility buffers
132276cf
VS
3449 template<typename T>
3450 struct ConvertedBuffer
3451 {
f54cb154
VZ
3452 // notice that there is no need to initialize m_len here as it's unused
3453 // as long as m_str is NULL
3454 ConvertedBuffer() : m_str(NULL) {}
132276cf 3455 ~ConvertedBuffer()
f54cb154 3456 { free(m_str); }
132276cf 3457
f54cb154
VZ
3458 bool Extend(size_t len)
3459 {
3460 // add extra 1 for the trailing NUL
3461 void * const str = realloc(m_str, sizeof(T)*(len + 1));
3462 if ( !str )
3463 return false;
3464
3465 m_str = static_cast<T *>(str);
3466 m_len = len;
3467
3468 return true;
3469 }
132276cf 3470
f54cb154 3471 const wxScopedCharTypeBuffer<T> AsScopedBuffer() const
132276cf 3472 {
f54cb154 3473 return wxScopedCharTypeBuffer<T>::CreateNonOwned(m_str, m_len);
132276cf
VS
3474 }
3475
f54cb154
VZ
3476 T *m_str; // pointer to the string data
3477 size_t m_len; // length, not size, i.e. in chars and without last NUL
132276cf 3478 };
f54cb154
VZ
3479
3480
3481#if wxUSE_UNICODE
3482 // common mb_str() and wxCStrData::AsChar() helper: performs the conversion
3483 // and returns either m_convertedToChar.m_str (in which case its m_len is
3484 // also updated) or NULL if it failed
3485 //
3486 // there is an important exception: in wxUSE_UNICODE_UTF8 build if conv is a
3487 // UTF-8 one, we return m_impl.c_str() directly, without doing any conversion
3488 // as optimization and so the caller needs to check for this before using
3489 // m_convertedToChar
3490 //
3491 // NB: AsChar() returns char* in any build, unlike mb_str()
3492 const char *AsChar(const wxMBConv& conv) const;
3493
3494 // mb_str() implementation helper
3495 wxScopedCharBuffer AsCharBuf(const wxMBConv& conv) const
3496 {
3497#if wxUSE_UNICODE_UTF8
3498 // avoid conversion if we can
3499 if ( conv.IsUTF8() )
3500 {
3501 return wxScopedCharBuffer::CreateNonOwned(m_impl.c_str(),
3502 m_impl.length());
3503 }
3504#endif // wxUSE_UNICODE_UTF8
3505
3506 // call this solely in order to fill in m_convertedToChar as AsChar()
3507 // updates it as a side effect: this is a bit ugly but it's a completely
3508 // internal function so the users of this class shouldn't care or know
3509 // about it and doing it like this, i.e. having a separate AsChar(),
3510 // allows us to avoid the creation and destruction of a temporary buffer
3511 // when using wxCStrData without duplicating any code
582886dd
VZ
3512 if ( !AsChar(conv) )
3513 {
3514 // although it would be probably more correct to return NULL buffer
3515 // from here if the conversion fails, a lot of existing code doesn't
3516 // expect mb_str() (or wc_str()) to ever return NULL so return an
3517 // empty string otherwise to avoid crashes in it
3518 //
3519 // also, some existing code does check for the conversion success and
3520 // so asserting here would be bad too -- even if it does mean that
3521 // silently losing data is possible for badly written code
3522 return wxScopedCharBuffer::CreateNonOwned("", 0);
3523 }
f54cb154
VZ
3524
3525 return m_convertedToChar.AsScopedBuffer();
3526 }
3527
132276cf 3528 ConvertedBuffer<char> m_convertedToChar;
f54cb154
VZ
3529#endif // !wxUSE_UNICODE
3530
132276cf 3531#if !wxUSE_UNICODE_WCHAR
f54cb154
VZ
3532 // common wc_str() and wxCStrData::AsWChar() helper for both UTF-8 and ANSI
3533 // builds: converts the string contents into m_convertedToWChar and returns
3534 // NULL if the conversion failed (this can only happen in ANSI build)
3535 //
3536 // NB: AsWChar() returns wchar_t* in any build, unlike wc_str()
3537 const wchar_t *AsWChar(const wxMBConv& conv) const;
3538
3539 // wc_str() implementation helper
3540 wxScopedWCharBuffer AsWCharBuf(const wxMBConv& conv) const
3541 {
582886dd
VZ
3542 if ( !AsWChar(conv) )
3543 return wxScopedWCharBuffer::CreateNonOwned(L"", 0);
f54cb154
VZ
3544
3545 return m_convertedToWChar.AsScopedBuffer();
3546 }
3547
132276cf 3548 ConvertedBuffer<wchar_t> m_convertedToWChar;
f54cb154 3549#endif // !wxUSE_UNICODE_WCHAR
2523e9b7 3550
b0c4d5d7
VS
3551#if wxUSE_UNICODE_UTF8
3552 // FIXME-UTF8: (try to) move this elsewhere (TLS) or solve differently
f54cb154 3553 // assigning to character pointer to by wxString::iterator may
b0c4d5d7
VS
3554 // change the underlying wxStringImpl iterator, so we have to
3555 // keep track of all iterators and update them as necessary:
3556 struct wxStringIteratorNodeHead
3557 {
3558 wxStringIteratorNodeHead() : ptr(NULL) {}
3559 wxStringIteratorNode *ptr;
300b44a9
VS
3560
3561 // copying is disallowed as it would result in more than one pointer into
3562 // the same linked list
c0c133e1 3563 wxDECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead);
b0c4d5d7
VS
3564 };
3565
3566 wxStringIteratorNodeHead m_iterators;
3567
b5dbe15d
VS
3568 friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode;
3569 friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef;
b0c4d5d7
VS
3570#endif // wxUSE_UNICODE_UTF8
3571
b5dbe15d 3572 friend class WXDLLIMPEXP_FWD_BASE wxCStrData;
6798451b
VS
3573 friend class wxStringInternalBuffer;
3574 friend class wxStringInternalBufferLength;
c801d85f
KB
3575};
3576
c9f78968 3577#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
fb330c2e 3578 #pragma warning (pop)
c9f78968
VS
3579#endif
3580
a962cdf4
VS
3581// string iterator operators that satisfy STL Random Access Iterator
3582// requirements:
b5343e06 3583inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i)
a962cdf4 3584 { return i + n; }
b5343e06 3585inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i)
a962cdf4 3586 { return i + n; }
b5343e06 3587inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i)
a962cdf4 3588 { return i + n; }
b5343e06 3589inline wxString::const_reverse_iterator operator+(ptrdiff_t n, wxString::const_reverse_iterator i)
a962cdf4
VS
3590 { return i + n; }
3591
1fc8cfbd
VZ
3592// notice that even though for many compilers the friend declarations above are
3593// enough, from the point of view of C++ standard we must have the declarations
3594// here as friend ones are not injected in the enclosing namespace and without
3595// them the code fails to compile with conforming compilers such as xlC or g++4
c9f78968 3596wxString WXDLLIMPEXP_BASE operator+(const wxString& string1, const wxString& string2);
8f93a29f
VS
3597wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const char *psz);
3598wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const wchar_t *pwz);
3599wxString WXDLLIMPEXP_BASE operator+(const char *psz, const wxString& string);
3600wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz, const wxString& string);
1fc8cfbd 3601
c9f78968
VS
3602wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch);
3603wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string);
3604
3605inline wxString operator+(const wxString& string, wxUniCharRef ch)
3606 { return string + (wxUniChar)ch; }
3607inline wxString operator+(const wxString& string, char ch)
3608 { return string + wxUniChar(ch); }
3609inline wxString operator+(const wxString& string, wchar_t ch)
3610 { return string + wxUniChar(ch); }
3611inline wxString operator+(wxUniCharRef ch, const wxString& string)
3612 { return (wxUniChar)ch + string; }
3613inline wxString operator+(char ch, const wxString& string)
3614 { return wxUniChar(ch) + string; }
3615inline wxString operator+(wchar_t ch, const wxString& string)
3616 { return wxUniChar(ch) + string; }
3617
b7146cbe 3618
a0f63de9 3619#define wxGetEmptyString() wxString()
07170120 3620
062dc5fc
VZ
3621// ----------------------------------------------------------------------------
3622// helper functions which couldn't be defined inline
3623// ----------------------------------------------------------------------------
3624
3625namespace wxPrivate
3626{
3627
3628#if wxUSE_UNICODE_WCHAR
3629
3630template <>
3631struct wxStringAsBufHelper<char>
3632{
de4983f3 3633 static wxScopedCharBuffer Get(const wxString& s, size_t *len)
062dc5fc 3634 {
de4983f3 3635 wxScopedCharBuffer buf(s.mb_str());
062dc5fc
VZ
3636 if ( len )
3637 *len = buf ? strlen(buf) : 0;
3638 return buf;
3639 }
3640};
3641
3642template <>
3643struct wxStringAsBufHelper<wchar_t>
3644{
de4983f3 3645 static wxScopedWCharBuffer Get(const wxString& s, size_t *len)
062dc5fc 3646 {
6df09f32 3647 const size_t length = s.length();
062dc5fc 3648 if ( len )
6df09f32
VS
3649 *len = length;
3650 return wxScopedWCharBuffer::CreateNonOwned(s.wx_str(), length);
062dc5fc
VZ
3651 }
3652};
3653
3654#elif wxUSE_UNICODE_UTF8
3655
3656template <>
3657struct wxStringAsBufHelper<char>
3658{
de4983f3 3659 static wxScopedCharBuffer Get(const wxString& s, size_t *len)
062dc5fc 3660 {
6df09f32 3661 const size_t length = s.utf8_length();
062dc5fc 3662 if ( len )
6df09f32
VS
3663 *len = length;
3664 return wxScopedCharBuffer::CreateNonOwned(s.wx_str(), length);
062dc5fc
VZ
3665 }
3666};
3667
3668template <>
3669struct wxStringAsBufHelper<wchar_t>
3670{
de4983f3 3671 static wxScopedWCharBuffer Get(const wxString& s, size_t *len)
062dc5fc 3672 {
de4983f3 3673 wxScopedWCharBuffer wbuf(s.wc_str());
062dc5fc
VZ
3674 if ( len )
3675 *len = wxWcslen(wbuf);
3676 return wbuf;
3677 }
3678};
3679
3680#endif // Unicode build kind
3681
3682} // namespace wxPrivate
3683
1d218550
VZ
3684// ----------------------------------------------------------------------------
3685// wxStringBuffer: a tiny class allowing to get a writable pointer into string
3686// ----------------------------------------------------------------------------
3687
2523e9b7
VS
3688#if !wxUSE_STL_BASED_WXSTRING
3689// string buffer for direct access to string data in their native
3690// representation:
6798451b 3691class wxStringInternalBuffer
2523e9b7
VS
3692{
3693public:
3694 typedef wxStringCharType CharType;
3695
6798451b 3696 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
2523e9b7
VS
3697 : m_str(str), m_buf(NULL)
3698 { m_buf = m_str.DoGetWriteBuf(lenWanted); }
3699
6798451b 3700 ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); }
2523e9b7
VS
3701
3702 operator wxStringCharType*() const { return m_buf; }
941b78cc 3703
2523e9b7
VS
3704private:
3705 wxString& m_str;
3706 wxStringCharType *m_buf;
3707
c0c133e1 3708 wxDECLARE_NO_COPY_CLASS(wxStringInternalBuffer);
2523e9b7
VS
3709};
3710
6798451b 3711class wxStringInternalBufferLength
941b78cc
MB
3712{
3713public:
2523e9b7
VS
3714 typedef wxStringCharType CharType;
3715
6798451b 3716 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
2523e9b7
VS
3717 : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false)
3718 {
3719 m_buf = m_str.DoGetWriteBuf(lenWanted);
3720 wxASSERT(m_buf != NULL);
3721 }
3722
6798451b 3723 ~wxStringInternalBufferLength()
2523e9b7
VS
3724 {
3725 wxASSERT(m_lenSet);
3726 m_str.DoUngetWriteBuf(m_len);
3727 }
3728
3729 operator wxStringCharType*() const { return m_buf; }
3730 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3731
3732private:
3733 wxString& m_str;
3734 wxStringCharType *m_buf;
3735 size_t m_len;
3736 bool m_lenSet;
3737
c0c133e1 3738 wxDECLARE_NO_COPY_CLASS(wxStringInternalBufferLength);
2523e9b7
VS
3739};
3740
3741#endif // !wxUSE_STL_BASED_WXSTRING
3742
3743template<typename T>
6b583d40 3744class wxStringTypeBufferBase
2523e9b7
VS
3745{
3746public:
3747 typedef T CharType;
3748
3749 wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024)
e87b7833 3750 : m_str(str), m_buf(lenWanted)
062dc5fc
VZ
3751 {
3752 // for compatibility with old wxStringBuffer which provided direct
3753 // access to wxString internal buffer, initialize ourselves with the
3754 // string initial contents
3755
3756 // FIXME-VC6: remove the ugly (CharType *)NULL and use normal
3757 // tchar_str<CharType>
3758 size_t len;
3759 const wxCharTypeBuffer<CharType> buf(str.tchar_str(&len, (CharType *)NULL));
3760 if ( buf )
3761 {
3762 if ( len > lenWanted )
3763 {
3764 // in this case there is not enough space for terminating NUL,
3765 // ensure that we still put it there
3766 m_buf.data()[lenWanted] = 0;
3767 len = lenWanted - 1;
3768 }
3769
39329d2e 3770 memcpy(m_buf.data(), buf, (len + 1)*sizeof(CharType));
062dc5fc
VZ
3771 }
3772 //else: conversion failed, this can happen when trying to get Unicode
3773 // string contents into a char string
3774 }
941b78cc 3775
2523e9b7 3776 operator CharType*() { return m_buf.data(); }
941b78cc 3777
2523e9b7 3778protected:
941b78cc 3779 wxString& m_str;
2523e9b7 3780 wxCharTypeBuffer<CharType> m_buf;
941b78cc
MB
3781};
3782
2523e9b7 3783template<typename T>
6b583d40 3784class wxStringTypeBufferLengthBase : public wxStringTypeBufferBase<T>
941b78cc
MB
3785{
3786public:
2523e9b7 3787 wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024)
062dc5fc
VZ
3788 : wxStringTypeBufferBase<T>(str, lenWanted),
3789 m_len(0),
3790 m_lenSet(false)
941b78cc
MB
3791 { }
3792
062dc5fc
VZ
3793 ~wxStringTypeBufferLengthBase()
3794 {
3795 wxASSERT_MSG( this->m_lenSet, "forgot to call SetLength()" );
3796 }
3797
941b78cc
MB
3798 void SetLength(size_t length) { m_len = length; m_lenSet = true; }
3799
2523e9b7 3800protected:
062dc5fc
VZ
3801 size_t m_len;
3802 bool m_lenSet;
941b78cc
MB
3803};
3804
2523e9b7
VS
3805template<typename T>
3806class wxStringTypeBuffer : public wxStringTypeBufferBase<T>
3807{
3808public:
3809 wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024)
062dc5fc
VZ
3810 : wxStringTypeBufferBase<T>(str, lenWanted)
3811 { }
3812
2523e9b7
VS
3813 ~wxStringTypeBuffer()
3814 {
3815 this->m_str.assign(this->m_buf.data());
3816 }
941b78cc 3817
c0c133e1 3818 wxDECLARE_NO_COPY_CLASS(wxStringTypeBuffer);
2523e9b7
VS
3819};
3820
3821template<typename T>
3822class wxStringTypeBufferLength : public wxStringTypeBufferLengthBase<T>
1d218550
VZ
3823{
3824public:
2523e9b7 3825 wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024)
062dc5fc
VZ
3826 : wxStringTypeBufferLengthBase<T>(str, lenWanted)
3827 { }
e015c2a3 3828
2523e9b7
VS
3829 ~wxStringTypeBufferLength()
3830 {
2523e9b7
VS
3831 this->m_str.assign(this->m_buf.data(), this->m_len);
3832 }
e015c2a3 3833
c0c133e1 3834 wxDECLARE_NO_COPY_CLASS(wxStringTypeBufferLength);
2523e9b7 3835};
e015c2a3 3836
2523e9b7 3837#if wxUSE_STL_BASED_WXSTRING
7c77f334
VZ
3838
3839WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<wxStringCharType> )
3840
6798451b 3841class wxStringInternalBuffer : public wxStringTypeBufferBase<wxStringCharType>
2523e9b7
VS
3842{
3843public:
6798451b 3844 wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)
2523e9b7 3845 : wxStringTypeBufferBase<wxStringCharType>(str, lenWanted) {}
6798451b 3846 ~wxStringInternalBuffer()
2523e9b7 3847 { m_str.m_impl.assign(m_buf.data()); }
e015c2a3 3848
c0c133e1 3849 wxDECLARE_NO_COPY_CLASS(wxStringInternalBuffer);
1d218550
VZ
3850};
3851
7c77f334
VZ
3852WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE(
3853 wxStringTypeBufferLengthBase<wxStringCharType> )
3854
3855class wxStringInternalBufferLength
3856 : public wxStringTypeBufferLengthBase<wxStringCharType>
941b78cc
MB
3857{
3858public:
6798451b 3859 wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)
2523e9b7 3860 : wxStringTypeBufferLengthBase<wxStringCharType>(str, lenWanted) {}
941b78cc 3861
6798451b 3862 ~wxStringInternalBufferLength()
941b78cc 3863 {
2523e9b7 3864 m_str.m_impl.assign(m_buf.data(), m_len);
941b78cc
MB
3865 }
3866
c0c133e1 3867 wxDECLARE_NO_COPY_CLASS(wxStringInternalBufferLength);
941b78cc 3868};
062dc5fc 3869
2523e9b7
VS
3870#endif // wxUSE_STL_BASED_WXSTRING
3871
941b78cc 3872
2523e9b7
VS
3873#if wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8
3874typedef wxStringTypeBuffer<wxChar> wxStringBuffer;
3875typedef wxStringTypeBufferLength<wxChar> wxStringBufferLength;
3876#else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
6798451b
VS
3877typedef wxStringInternalBuffer wxStringBuffer;
3878typedef wxStringInternalBufferLength wxStringBufferLength;
8f93a29f 3879#endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
941b78cc 3880
7d46f92b 3881#if wxUSE_UNICODE_UTF8
628f87da
VS
3882typedef wxStringInternalBuffer wxUTF8StringBuffer;
3883typedef wxStringInternalBufferLength wxUTF8StringBufferLength;
7d46f92b 3884#elif wxUSE_UNICODE_WCHAR
7c77f334
VZ
3885
3886WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<char> )
3887
6b583d40
VZ
3888// Note about inlined dtors in the classes below: this is done not for
3889// performance reasons but just to avoid linking errors in the MSVC DLL build
3890// under Windows: if a class has non-inline methods it must be declared as
3891// being DLL-exported but, due to an extremely interesting feature of MSVC 7
3892// and later, any template class which is used as a base of a DLL-exported
3893// class is implicitly made DLL-exported too, as explained at the bottom of
3894// http://msdn.microsoft.com/en-us/library/twa2aw10.aspx (just to confirm: yes,
3895// _inheriting_ from a class can change whether it is being exported from DLL)
3896//
3897// But this results in link errors because the base template class is not DLL-
3898// exported, whether it is declared with WXDLLIMPEXP_BASE or not, because it
3899// does have only inline functions. So the simplest fix is to just make all the
3900// functions of these classes inline too.
3901
3902class wxUTF8StringBuffer : public wxStringTypeBufferBase<char>
628f87da
VS
3903{
3904public:
3905 wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024)
3906 : wxStringTypeBufferBase<char>(str, lenWanted) {}
6b583d40
VZ
3907 ~wxUTF8StringBuffer()
3908 {
3909 wxMBConvStrictUTF8 conv;
3910 size_t wlen = conv.ToWChar(NULL, 0, m_buf);
3911 wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" );
3912
3913 wxStringInternalBuffer wbuf(m_str, wlen);
3914 conv.ToWChar(wbuf, wlen, m_buf);
3915 }
628f87da 3916
c0c133e1 3917 wxDECLARE_NO_COPY_CLASS(wxUTF8StringBuffer);
628f87da
VS
3918};
3919
7c77f334
VZ
3920WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase<char> )
3921
6b583d40 3922class wxUTF8StringBufferLength : public wxStringTypeBufferLengthBase<char>
628f87da
VS
3923{
3924public:
3925 wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024)
3926 : wxStringTypeBufferLengthBase<char>(str, lenWanted) {}
6b583d40
VZ
3927 ~wxUTF8StringBufferLength()
3928 {
3929 wxCHECK_RET(m_lenSet, "length not set");
3930
3931 wxMBConvStrictUTF8 conv;
3932 size_t wlen = conv.ToWChar(NULL, 0, m_buf, m_len);
3933 wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" );
3934
3935 wxStringInternalBufferLength wbuf(m_str, wlen);
3936 conv.ToWChar(wbuf, wlen, m_buf, m_len);
3937 wbuf.SetLength(wlen);
3938 }
628f87da 3939
c0c133e1 3940 wxDECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength);
628f87da 3941};
7d46f92b 3942#endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR
628f87da
VS
3943
3944
c801d85f 3945// ---------------------------------------------------------------------------
3c67202d 3946// wxString comparison functions: operator versions are always case sensitive
c801d85f 3947// ---------------------------------------------------------------------------
f33fee2a 3948
831faf97 3949#define wxCMP_WXCHAR_STRING(p, s, op) 0 op s.Cmp(p)
fb52c2b6
VZ
3950
3951wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING)
3952
3953#undef wxCMP_WXCHAR_STRING
3954
0cbe5ee3 3955inline bool operator==(const wxString& s1, const wxString& s2)
2712e317 3956 { return s1.IsSameAs(s2); }
0cbe5ee3 3957inline bool operator!=(const wxString& s1, const wxString& s2)
2712e317 3958 { return !s1.IsSameAs(s2); }
0cbe5ee3
VZ
3959inline bool operator< (const wxString& s1, const wxString& s2)
3960 { return s1.Cmp(s2) < 0; }
0cbe5ee3
VZ
3961inline bool operator> (const wxString& s1, const wxString& s2)
3962 { return s1.Cmp(s2) > 0; }
0cbe5ee3
VZ
3963inline bool operator<=(const wxString& s1, const wxString& s2)
3964 { return s1.Cmp(s2) <= 0; }
0cbe5ee3
VZ
3965inline bool operator>=(const wxString& s1, const wxString& s2)
3966 { return s1.Cmp(s2) >= 0; }
3c67202d 3967
5d0fac27
VS
3968inline bool operator==(const wxString& s1, const wxCStrData& s2)
3969 { return s1 == s2.AsString(); }
3970inline bool operator==(const wxCStrData& s1, const wxString& s2)
3971 { return s1.AsString() == s2; }
3972inline bool operator!=(const wxString& s1, const wxCStrData& s2)
3973 { return s1 != s2.AsString(); }
3974inline bool operator!=(const wxCStrData& s1, const wxString& s2)
3975 { return s1.AsString() != s2; }
3976
de4983f3 3977inline bool operator==(const wxString& s1, const wxScopedWCharBuffer& s2)
f33fee2a 3978 { return (s1.Cmp((const wchar_t *)s2) == 0); }
de4983f3 3979inline bool operator==(const wxScopedWCharBuffer& s1, const wxString& s2)
f33fee2a 3980 { return (s2.Cmp((const wchar_t *)s1) == 0); }
de4983f3 3981inline bool operator!=(const wxString& s1, const wxScopedWCharBuffer& s2)
f6bcfd97 3982 { return (s1.Cmp((const wchar_t *)s2) != 0); }
de4983f3 3983inline bool operator!=(const wxScopedWCharBuffer& s1, const wxString& s2)
f6bcfd97 3984 { return (s2.Cmp((const wchar_t *)s1) != 0); }
be39d6f5 3985
de4983f3 3986inline bool operator==(const wxString& s1, const wxScopedCharBuffer& s2)
f33fee2a 3987 { return (s1.Cmp((const char *)s2) == 0); }
de4983f3 3988inline bool operator==(const wxScopedCharBuffer& s1, const wxString& s2)
f33fee2a 3989 { return (s2.Cmp((const char *)s1) == 0); }
de4983f3 3990inline bool operator!=(const wxString& s1, const wxScopedCharBuffer& s2)
f6bcfd97 3991 { return (s1.Cmp((const char *)s2) != 0); }
de4983f3 3992inline bool operator!=(const wxScopedCharBuffer& s1, const wxString& s2)
f6bcfd97 3993 { return (s2.Cmp((const char *)s1) != 0); }
5ca07d0f 3994
de4983f3 3995inline wxString operator+(const wxString& string, const wxScopedWCharBuffer& buf)
e90c1d2a 3996 { return string + (const wchar_t *)buf; }
de4983f3 3997inline wxString operator+(const wxScopedWCharBuffer& buf, const wxString& string)
e90c1d2a 3998 { return (const wchar_t *)buf + string; }
be39d6f5 3999
de4983f3 4000inline wxString operator+(const wxString& string, const wxScopedCharBuffer& buf)
e90c1d2a 4001 { return string + (const char *)buf; }
de4983f3 4002inline wxString operator+(const wxScopedCharBuffer& buf, const wxString& string)
e90c1d2a 4003 { return (const char *)buf + string; }
f04f3991 4004
a962cdf4 4005// comparison with char
c9f78968
VS
4006inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); }
4007inline bool operator==(const wxUniCharRef& c, const wxString& s) { return s.IsSameAs(c); }
4008inline bool operator==(char c, const wxString& s) { return s.IsSameAs(c); }
4009inline bool operator==(wchar_t c, const wxString& s) { return s.IsSameAs(c); }
4010inline bool operator==(int c, const wxString& s) { return s.IsSameAs(c); }
4011inline bool operator==(const wxString& s, const wxUniChar& c) { return s.IsSameAs(c); }
4012inline bool operator==(const wxString& s, const wxUniCharRef& c) { return s.IsSameAs(c); }
4013inline bool operator==(const wxString& s, char c) { return s.IsSameAs(c); }
4014inline bool operator==(const wxString& s, wchar_t c) { return s.IsSameAs(c); }
4015inline bool operator!=(const wxUniChar& c, const wxString& s) { return !s.IsSameAs(c); }
4016inline bool operator!=(const wxUniCharRef& c, const wxString& s) { return !s.IsSameAs(c); }
4017inline bool operator!=(char c, const wxString& s) { return !s.IsSameAs(c); }
4018inline bool operator!=(wchar_t c, const wxString& s) { return !s.IsSameAs(c); }
4019inline bool operator!=(int c, const wxString& s) { return !s.IsSameAs(c); }
4020inline bool operator!=(const wxString& s, const wxUniChar& c) { return !s.IsSameAs(c); }
4021inline bool operator!=(const wxString& s, const wxUniCharRef& c) { return !s.IsSameAs(c); }
4022inline bool operator!=(const wxString& s, char c) { return !s.IsSameAs(c); }
4023inline bool operator!=(const wxString& s, wchar_t c) { return !s.IsSameAs(c); }
e51b17a1 4024
d7330233
VS
4025// comparison with C string in Unicode build
4026#if wxUSE_UNICODE
fb52c2b6
VZ
4027
4028#define wxCMP_CHAR_STRING(p, s, op) wxString(p) op s
4029
4030wxDEFINE_ALL_COMPARISONS(const char *, const wxString&, wxCMP_CHAR_STRING)
4031
4032#undef wxCMP_CHAR_STRING
4033
d7330233
VS
4034#endif // wxUSE_UNICODE
4035
fb52c2b6
VZ
4036// we also need to provide the operators for comparison with wxCStrData to
4037// resolve ambiguity between operator(const wxChar *,const wxString &) and
4038// operator(const wxChar *, const wxChar *) for "p == s.c_str()"
4039//
4040// notice that these are (shallow) pointer comparisons, not (deep) string ones
4041#define wxCMP_CHAR_CSTRDATA(p, s, op) p op s.AsChar()
4042#define wxCMP_WCHAR_CSTRDATA(p, s, op) p op s.AsWChar()
4043
11aac4ba
VS
4044wxDEFINE_ALL_COMPARISONS(const wchar_t *, const wxCStrData&, wxCMP_WCHAR_CSTRDATA)
4045wxDEFINE_ALL_COMPARISONS(const char *, const wxCStrData&, wxCMP_CHAR_CSTRDATA)
fb52c2b6
VZ
4046
4047#undef wxCMP_CHAR_CSTRDATA
4048#undef wxCMP_WCHAR_CSTRDATA
4049
c801d85f 4050// ---------------------------------------------------------------------------
3c67202d 4051// Implementation only from here until the end of file
c801d85f
KB
4052// ---------------------------------------------------------------------------
4053
e87b7833 4054#if wxUSE_STD_IOSTREAM
c801d85f 4055
65f19af1 4056#include "wx/iosfwrap.h"
c801d85f 4057
bddd7a8d 4058WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxString&);
c9f78968 4059WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCStrData&);
de4983f3 4060WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxScopedCharBuffer&);
04abe4bc 4061#ifndef __BORLANDC__
de4983f3 4062WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxScopedWCharBuffer&);
04abe4bc 4063#endif
c801d85f 4064
6a6ea041 4065#if wxUSE_UNICODE && defined(HAVE_WOSTREAM)
6b61b594
VZ
4066
4067WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxString&);
4068WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxCStrData&);
de4983f3 4069WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxScopedWCharBuffer&);
6b61b594 4070
6a6ea041 4071#endif // wxUSE_UNICODE && defined(HAVE_WOSTREAM)
6b61b594 4072
7a906e1a 4073#endif // wxUSE_STD_IOSTREAM
c801d85f 4074
c9f78968
VS
4075// ---------------------------------------------------------------------------
4076// wxCStrData implementation
4077// ---------------------------------------------------------------------------
4078
4079inline wxCStrData::wxCStrData(char *buf)
4080 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
4081inline wxCStrData::wxCStrData(wchar_t *buf)
4082 : m_str(new wxString(buf)), m_offset(0), m_owned(true) {}
4083
92a17abc
VS
4084inline wxCStrData::wxCStrData(const wxCStrData& data)
4085 : m_str(data.m_owned ? new wxString(*data.m_str) : data.m_str),
4086 m_offset(data.m_offset),
4087 m_owned(data.m_owned)
4088{
4089}
4090
c9f78968
VS
4091inline wxCStrData::~wxCStrData()
4092{
4093 if ( m_owned )
5c33522f 4094 delete const_cast<wxString*>(m_str); // cast to silence warnings
c9f78968
VS
4095}
4096
f54cb154
VZ
4097// AsChar() and AsWChar() implementations simply forward to wxString methods
4098
c9f78968 4099inline const wchar_t* wxCStrData::AsWChar() const
11aac4ba 4100{
f54cb154
VZ
4101 const wchar_t * const p =
4102#if wxUSE_UNICODE_WCHAR
4103 m_str->wc_str();
4104#elif wxUSE_UNICODE_UTF8
4105 m_str->AsWChar(wxMBConvStrictUTF8());
4106#else
4107 m_str->AsWChar(wxConvLibc);
4108#endif
11aac4ba 4109
f54cb154
VZ
4110 // in Unicode build the string always has a valid Unicode representation
4111 // and even if a conversion is needed (as in UTF8 case) it can't fail
4112 //
4113 // but in ANSI build the string contents might be not convertible to
4114 // Unicode using the current locale encoding so we do need to check for
4115 // errors
bfb6847d 4116#if !wxUSE_UNICODE
f54cb154
VZ
4117 if ( !p )
4118 {
4119 // if conversion fails, return empty string and not NULL to avoid
4120 // crashes in code written with either wxWidgets 2 wxString or
4121 // std::string behaviour in mind: neither of them ever returns NULL
4122 // from its c_str() and so we shouldn't neither
4123 //
4124 // notice that the same is done in AsChar() below and
4125 // wxString::wc_str() and mb_str() for the same reasons
4126 return L"";
4127 }
11aac4ba 4128#endif // !wxUSE_UNICODE
c9f78968 4129
f54cb154 4130 return p + m_offset;
bfb6847d 4131}
bfb6847d 4132
f54cb154 4133inline const char* wxCStrData::AsChar() const
681e4412 4134{
f54cb154
VZ
4135#if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY
4136 const char * const p = m_str->AsChar(wxConvLibc);
4137 if ( !p )
4138 return "";
4139#else // !wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY
4140 const char * const p = m_str->mb_str();
4141#endif // wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY
4142
4143 return p + m_offset;
681e4412
VS
4144}
4145
c9f78968
VS
4146inline wxString wxCStrData::AsString() const
4147{
4148 if ( m_offset == 0 )
4149 return *m_str;
4150 else
4151 return m_str->Mid(m_offset);
4152}
4153
2523e9b7
VS
4154inline const wxStringCharType *wxCStrData::AsInternal() const
4155{
bfb6847d 4156#if wxUSE_UNICODE_UTF8
2523e9b7 4157 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
bfb6847d
VS
4158#else
4159 return m_str->wx_str() + m_offset;
4160#endif
2523e9b7
VS
4161}
4162
c9f78968
VS
4163inline wxUniChar wxCStrData::operator*() const
4164{
4165 if ( m_str->empty() )
9a83f860 4166 return wxUniChar(wxT('\0'));
c9f78968
VS
4167 else
4168 return (*m_str)[m_offset];
4169}
4170
4171inline wxUniChar wxCStrData::operator[](size_t n) const
4172{
378964d4
VS
4173 // NB: we intentionally use operator[] and not at() here because the former
4174 // works for the terminating NUL while the latter does not
4175 return (*m_str)[m_offset + n];
c9f78968
VS
4176}
4177
cc5bd48e
VZ
4178// ----------------------------------------------------------------------------
4179// more wxCStrData operators
4180// ----------------------------------------------------------------------------
4181
4182// we need to define those to allow "size_t pos = p - s.c_str()" where p is
4183// some pointer into the string
4184inline size_t operator-(const char *p, const wxCStrData& cs)
4185{
4186 return p - cs.AsChar();
4187}
4188
4189inline size_t operator-(const wchar_t *p, const wxCStrData& cs)
4190{
4191 return p - cs.AsWChar();
4192}
4193
414b7b40
VZ
4194// ----------------------------------------------------------------------------
4195// implementation of wx[W]CharBuffer inline methods using wxCStrData
4196// ----------------------------------------------------------------------------
4197
11aac4ba
VS
4198// FIXME-UTF8: move this to buffer.h
4199inline wxCharBuffer::wxCharBuffer(const wxCStrData& cstr)
681e4412 4200 : wxCharTypeBufferBase(cstr.AsCharBuf())
11aac4ba
VS
4201{
4202}
4203
4204inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr)
681e4412 4205 : wxCharTypeBufferBase(cstr.AsWCharBuf())
414b7b40
VZ
4206{
4207}
4208
b0c4d5d7
VS
4209#if wxUSE_UNICODE_UTF8
4210// ----------------------------------------------------------------------------
4211// implementation of wxStringIteratorNode inline methods
4212// ----------------------------------------------------------------------------
4213
39c20230
VS
4214void wxStringIteratorNode::DoSet(const wxString *str,
4215 wxStringImpl::const_iterator *citer,
4216 wxStringImpl::iterator *iter)
b0c4d5d7 4217{
42fc0309 4218 m_prev = NULL;
39c20230
VS
4219 m_iter = iter;
4220 m_citer = citer;
4221 m_str = str;
4222 if ( str )
4223 {
4224 m_next = str->m_iterators.ptr;
5c33522f 4225 const_cast<wxString*>(m_str)->m_iterators.ptr = this;
39c20230
VS
4226 if ( m_next )
4227 m_next->m_prev = this;
4228 }
42fc0309
VS
4229 else
4230 {
4231 m_next = NULL;
4232 }
b0c4d5d7
VS
4233}
4234
39c20230 4235void wxStringIteratorNode::clear()
b0c4d5d7
VS
4236{
4237 if ( m_next )
4238 m_next->m_prev = m_prev;
4239 if ( m_prev )
4240 m_prev->m_next = m_next;
39c20230 4241 else if ( m_str ) // first in the list
5c33522f 4242 const_cast<wxString*>(m_str)->m_iterators.ptr = m_next;
39c20230
VS
4243
4244 m_next = m_prev = NULL;
4245 m_citer = NULL;
4246 m_iter = NULL;
4247 m_str = NULL;
b0c4d5d7
VS
4248}
4249#endif // wxUSE_UNICODE_UTF8
4250
e3e8e3c0 4251#if WXWIN_COMPATIBILITY_2_8
3a3dde0d 4252 // lot of code out there doesn't explicitly include wx/crt.h, but uses
e3e8e3c0
VS
4253 // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h,
4254 // so let's include this header now that wxString is defined and it's safe
4255 // to do it:
3a3dde0d 4256 #include "wx/crt.h"
e3e8e3c0
VS
4257#endif
4258
cbec0f40
FM
4259// ----------------------------------------------------------------------------
4260// Checks on wxString characters
4261// ----------------------------------------------------------------------------
4262
4263template<bool (T)(const wxUniChar& c)>
4264 inline bool wxStringCheck(const wxString& val)
4265 {
4266 for ( wxString::const_iterator i = val.begin();
4267 i != val.end();
4268 ++i )
4269 if (T(*i) == 0)
4270 return false;
4271 return true;
4272 }
4273
11aac4ba 4274#endif // _WX_WXSTRING_H_