add XRC handler for wxSearchCtrl (extended patch 1751516)

[wxWidgets.git] / src / common / string.cpp

diff --git a/src/common/string.cpp b/src/common/string.cpp
index 9fc84b087850b87d8793504c6db697c59a6dd2ff..d36f85bd5a2cb4259e40df65fdd87a9cc3bfb7f8 100644 (file)
--- a/src/common/string.cpp
+++ b/src/common/string.cpp
@@ -23,6 +23,7 @@
 
 #ifndef WX_PRECOMP
     #include "wx/string.h"
 
 #ifndef WX_PRECOMP
     #include "wx/string.h"

+    #include "wx/wxcrtvararg.h"

 #endif
 
 #include <ctype.h>
 #endif
 
 #include <ctype.h>


@@ -219,9 +220,16 @@ wxString::~wxString()
 }
 #endif
 
 }
 #endif
 

-#if wxUSE_UNICODE
+#if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY

 const char* wxCStrData::AsChar() const
 {
 const char* wxCStrData::AsChar() const
 {

+#if wxUSE_UNICODE_UTF8
+    if ( wxLocaleIsUtf8 )
+        return AsInternal();
+#endif
+    // under non-UTF8 locales, we have to convert the internal UTF-8
+    // representation using wxConvLibc and cache the result
+

     wxString *str = wxConstCast(m_str, wxString);
 
     // convert the string:
     wxString *str = wxConstCast(m_str, wxString);
 
     // convert the string:


@@ -243,7 +251,7 @@ const char* wxCStrData::AsChar() const
     // and keep it:
     return str->m_convertedToChar + m_offset;
 }
     // and keep it:
     return str->m_convertedToChar + m_offset;
 }

-#endif // wxUSE_UNICODE
+#endif // wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY

 
 #if !wxUSE_UNICODE_WCHAR
 const wchar_t* wxCStrData::AsWChar() const
 
 #if !wxUSE_UNICODE_WCHAR
 const wchar_t* wxCStrData::AsWChar() const


@@ -305,14 +313,23 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
 wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
                                                const wxMBConv& conv)
 {
 wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
                                                const wxMBConv& conv)
 {

-    // FIXME-UTF8: return as-is without copying under UTF8 locale, return
-    //             converted string under other locales - needs wxCharBuffer
-    //             changes
-

     // anything to do?
     if ( !psz || nLength == 0 )
         return SubstrBufFromMB("", 0);
 
     // anything to do?
     if ( !psz || nLength == 0 )
         return SubstrBufFromMB("", 0);
 

+    // if psz is already in UTF-8, we don't have to do the roundtrip to
+    // wchar_t* and back:
+    if ( conv.IsUTF8() )
+    {
+        // we need to validate the input because UTF8 iterators assume valid
+        // UTF-8 sequence and psz may be invalid:
+        if ( wxStringOperations::IsValidUtf8String(psz, nLength) )
+        {
+            return SubstrBufFromMB(wxCharBuffer::CreateNonOwned(psz), nLength);
+        }
+        // else: do the roundtrip through wchar_t*
+    }
+

     if ( nLength == npos )
         nLength = wxNO_LEN;
 
     if ( nLength == npos )
         nLength = wxNO_LEN;
 


@@ -323,7 +340,7 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
         return SubstrBufFromMB("", 0);
 
     // and then to UTF-8:
         return SubstrBufFromMB("", 0);
 
     // and then to UTF-8:

-    SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxConvUTF8));
+    SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxMBConvStrictUTF8()));

     // widechar -> UTF-8 conversion isn't supposed to ever fail:
     wxASSERT_MSG( buf.data, _T("conversion to UTF-8 failed") );
 
     // widechar -> UTF-8 conversion isn't supposed to ever fail:
     wxASSERT_MSG( buf.data, _T("conversion to UTF-8 failed") );
 


@@ -365,22 +382,28 @@ const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const
 
 const wxWCharBuffer wxString::wc_str() const
 {
 
 const wxWCharBuffer wxString::wc_str() const
 {

-    return wxConvUTF8.cMB2WC(m_impl.c_str(),
-                             m_impl.length() + 1 /* size, not length */,
-                             NULL);
+    return wxMBConvStrictUTF8().cMB2WC
+                                (
+                                    m_impl.c_str(),
+                                    m_impl.length() + 1, // size, not length
+                                    NULL
+                                );

 }
 
 const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const
 {
 }
 
 const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const
 {

-    // FIXME-UTF8: optimize the case when conv==wxConvUTF8 or wxConvLibc
-    //             under UTF8 locale
+    if ( conv.IsUTF8() )
+        return wxCharBuffer::CreateNonOwned(m_impl.c_str());
+

     // FIXME-UTF8: use wc_str() here once we have buffers with length
 
     size_t wcLen;
     // FIXME-UTF8: use wc_str() here once we have buffers with length
 
     size_t wcLen;

-    wxWCharBuffer wcBuf(
-            wxConvUTF8.cMB2WC(m_impl.c_str(),
-                              m_impl.length() + 1 /* size, not length */,
-                              &wcLen));
+    wxWCharBuffer wcBuf(wxMBConvStrictUTF8().cMB2WC
+                                             (
+                                                m_impl.c_str(),
+                                                m_impl.length() + 1, // size
+                                                &wcLen
+                                             ));

     if ( !wcLen )
         return wxCharBuffer("");
 
     if ( !wcLen )
         return wxCharBuffer("");
 


@@ -408,7 +431,7 @@ bool wxString::Shrink()
 
 // deprecated compatibility code:
 #if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
 
 // deprecated compatibility code:
 #if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8

-wxChar *wxString::GetWriteBuf(size_t nLen)
+wxStringCharType *wxString::GetWriteBuf(size_t nLen)

 {
     return DoGetWriteBuf(nLen);
 }
 {
     return DoGetWriteBuf(nLen);
 }


@@ -547,6 +570,12 @@ wxString operator+(const wchar_t *pwz, const wxString& str)
 // string comparison
 // ---------------------------------------------------------------------------
 
 // string comparison
 // ---------------------------------------------------------------------------
 

+bool wxString::IsSameAs(wxUniChar c, bool compareWithCase) const
+{
+    return (length() == 1) && (compareWithCase ? GetChar(0u) == c
+                               : wxToupper(GetChar(0u)) == wxToupper(c));
+}
+

 #ifdef HAVE_STD_STRING_COMPARE
 
 // NB: Comparison code (both if HAVE_STD_STRING_COMPARE and if not) works with
 #ifdef HAVE_STD_STRING_COMPARE
 
 // NB: Comparison code (both if HAVE_STD_STRING_COMPARE and if not) works with


@@ -949,63 +978,69 @@ int wxString::CmpNoCase(const wxString& s) const
 #endif
 #endif
 
 #endif
 #endif
 

-wxString wxString::FromAscii(const char *ascii)
+wxString wxString::FromAscii(const char *ascii, size_t len)

 {
 {

-    if (!ascii)
+    if (!ascii || len == 0)

        return wxEmptyString;
 
        return wxEmptyString;
 

-    size_t len = strlen( ascii );

     wxString res;
 
     wxString res;
 

-    if ( len )

     {
     {

-        wxStringBuffer buf(res, len);
+        wxImplStringBuffer buf(res, len);
+        wxStringCharType *dest = buf;

 
 

-        wchar_t *dest = buf;
-
-        for ( ;; )
+        for ( ; len > 0; --len )

         {
         {

-           if ( (*dest++ = (wchar_t)(unsigned char)*ascii++) == L'\0' )
-               break;
+            unsigned char c = (unsigned char)*ascii++;
+            wxASSERT_MSG( c < 0x80,
+                          _T("Non-ASCII value passed to FromAscii().") );
+
+            *dest++ = (wchar_t)c;

         }
     }
 
     return res;
 }
 
         }
     }
 
     return res;
 }
 

+wxString wxString::FromAscii(const char *ascii)
+{
+    return FromAscii(ascii, wxStrlen(ascii));
+}
+

 wxString wxString::FromAscii(const char ascii)
 {
     // What do we do with '\0' ?
 
 wxString wxString::FromAscii(const char ascii)
 {
     // What do we do with '\0' ?
 

-    wxString res;
-    res += (wchar_t)(unsigned char) ascii;
+    unsigned char c = (unsigned char)ascii;

 
 

-    return res;
+    wxASSERT_MSG( c < 0x80, _T("Non-ASCII value passed to FromAscii().") );
+
+    // NB: the cast to wchar_t causes interpretation of 'ascii' as Latin1 value
+    return wxString(wxUniChar((wchar_t)c));

 }
 
 const wxCharBuffer wxString::ToAscii() const
 {
     // this will allocate enough space for the terminating NUL too
     wxCharBuffer buffer(length());
 }
 
 const wxCharBuffer wxString::ToAscii() const
 {
     // this will allocate enough space for the terminating NUL too
     wxCharBuffer buffer(length());

-
-

     char *dest = buffer.data();
 
     char *dest = buffer.data();
 

-    const wchar_t *pwc = c_str();
-    for ( ;; )
+    for ( const_iterator i = begin(); i != end(); ++i )

     {
     {

-        *dest++ = (char)(*pwc > SCHAR_MAX ? wxT('_') : *pwc);
+        wxUniChar c(*i);
+        // FIXME-UTF8: unify substituted char ('_') with wxUniChar ('?')
+        *dest++ = c.IsAscii() ? (char)c : '_';

 
         // the output string can't have embedded NULs anyhow, so we can safely
         // stop at first of them even if we do have any
 
         // the output string can't have embedded NULs anyhow, so we can safely
         // stop at first of them even if we do have any

-        if ( !*pwc++ )
+        if ( !c )

             break;
     }
 
     return buffer;
 }
 
             break;
     }
 
     return buffer;
 }
 

-#endif // Unicode
+#endif // wxUSE_UNICODE

 
 // extract string of length nCount starting at nFirst
 wxString wxString::Mid(size_t nFirst, size_t nCount) const
 
 // extract string of length nCount starting at nFirst
 wxString wxString::Mid(size_t nFirst, size_t nCount) const


@@ -1041,28 +1076,15 @@ wxString wxString::Mid(size_t nFirst, size_t nCount) const
 
 // check that the string starts with prefix and return the rest of the string
 // in the provided pointer if it is not NULL, otherwise return false
 
 // check that the string starts with prefix and return the rest of the string
 // in the provided pointer if it is not NULL, otherwise return false

-bool wxString::StartsWith(const wxChar *prefix, wxString *rest) const
+bool wxString::StartsWith(const wxString& prefix, wxString *rest) const

 {
 {

-    wxASSERT_MSG( prefix, _T("invalid parameter in wxString::StartsWith") );
-
-    // first check if the beginning of the string matches the prefix: note
-    // that we don't have to check that we don't run out of this string as
-    // when we reach the terminating NUL, either prefix string ends too (and
-    // then it's ok) or we break out of the loop because there is no match
-    const wxChar *p = c_str();
-    while ( *prefix )
-    {
-        if ( *prefix++ != *p++ )
-        {
-            // no match
-            return false;
-        }
-    }
+    if ( compare(0, prefix.length(), prefix) != 0 )
+        return false;

 
     if ( rest )
     {
         // put the rest of the string into provided pointer
 
     if ( rest )
     {
         // put the rest of the string into provided pointer

-        *rest = p;
+        rest->assign(*this, prefix.length(), npos);

     }
 
     return true;
     }
 
     return true;


@@ -1071,11 +1093,9 @@ bool wxString::StartsWith(const wxChar *prefix, wxString *rest) const
 
 // check that the string ends with suffix and return the rest of it in the
 // provided pointer if it is not NULL, otherwise return false
 
 // check that the string ends with suffix and return the rest of it in the
 // provided pointer if it is not NULL, otherwise return false

-bool wxString::EndsWith(const wxChar *suffix, wxString *rest) const
+bool wxString::EndsWith(const wxString& suffix, wxString *rest) const

 {
 {

-    wxASSERT_MSG( suffix, _T("invalid parameter in wxString::EndssWith") );
-
-    int start = length() - wxStrlen(suffix);
+    int start = length() - suffix.length();

 
     if ( start < 0 || compare(start, npos, suffix) != 0 )
         return false;
 
     if ( start < 0 || compare(start, npos, suffix) != 0 )
         return false;


@@ -1367,14 +1387,15 @@ int wxString::Find(wxUniChar ch, bool bFromEnd) const
 // conversion to numbers
 // ----------------------------------------------------------------------------
 
 // conversion to numbers
 // ----------------------------------------------------------------------------
 

-// the implementation of all the functions below is exactly the same so factor
-// it out
+// The implementation of all the functions below is exactly the same so factor
+// it out. Note that number extraction works correctly on UTF-8 strings, so
+// we can use wxStringCharType and wx_str() for maximum efficiency.

 
 

-template <typename T, typename F>
-bool wxStringToIntType(const wxChar *start,
+template <typename T>
+bool wxStringToIntType(const wxStringCharType *start,

                        T *val,
                        int base,
                        T *val,
                        int base,

-                       F func)
+                       T (*func)(const wxStringCharType*, wxStringCharType**, int))

 {
     wxCHECK_MSG( val, false, _T("NULL output pointer") );
     wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );
 {
     wxCHECK_MSG( val, false, _T("NULL output pointer") );
     wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );


@@ -1383,7 +1404,7 @@ bool wxStringToIntType(const wxChar *start,
     errno = 0;
 #endif
 
     errno = 0;
 #endif
 

-    wxChar *end;
+    wxStringCharType *end;

     *val = (*func)(start, &end, base);
 
     // return true only if scan was stopped by the terminating NUL and if the
     *val = (*func)(start, &end, base);
 
     // return true only if scan was stopped by the terminating NUL and if the


@@ -1397,22 +1418,22 @@ bool wxStringToIntType(const wxChar *start,
 
 bool wxString::ToLong(long *val, int base) const
 {
 
 bool wxString::ToLong(long *val, int base) const
 {

-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtol);
+    return wxStringToIntType(wx_str(), val, base, wxStrtol);

 }
 
 bool wxString::ToULong(unsigned long *val, int base) const
 {
 }
 
 bool wxString::ToULong(unsigned long *val, int base) const
 {

-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoul);
+    return wxStringToIntType(wx_str(), val, base, wxStrtoul);

 }
 
 bool wxString::ToLongLong(wxLongLong_t *val, int base) const
 {
 }
 
 bool wxString::ToLongLong(wxLongLong_t *val, int base) const
 {

-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoll);
+    return wxStringToIntType(wx_str(), val, base, wxStrtoll);

 }
 
 bool wxString::ToULongLong(wxULongLong_t *val, int base) const
 {
 }
 
 bool wxString::ToULongLong(wxULongLong_t *val, int base) const
 {

-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoull);
+    return wxStringToIntType(wx_str(), val, base, wxStrtoull);

 }
 
 bool wxString::ToDouble(double *val) const
 }
 
 bool wxString::ToDouble(double *val) const


@@ -1440,11 +1461,12 @@ bool wxString::ToDouble(double *val) const
 // formatted output
 // ---------------------------------------------------------------------------
 
 // formatted output
 // ---------------------------------------------------------------------------
 

+#if !wxUSE_UTF8_LOCALE_ONLY

 /* static */
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
 /* static */
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN

-wxString wxStringPrintfMixinBase::DoFormat(const wxChar *format, ...)
+wxString wxStringPrintfMixinBase::DoFormatWchar(const wxChar *format, ...)

 #else
 #else

-wxString wxString::DoFormat(const wxChar *format, ...)
+wxString wxString::DoFormatWchar(const wxChar *format, ...)

 #endif
 {
     va_list argptr;
 #endif
 {
     va_list argptr;


@@ -1457,6 +1479,23 @@ wxString wxString::DoFormat(const wxChar *format, ...)
 
     return s;
 }
 
     return s;
 }

+#endif // !wxUSE_UTF8_LOCALE_ONLY
+
+#if wxUSE_UNICODE_UTF8
+/* static */
+wxString wxString::DoFormatUtf8(const char *format, ...)
+{
+    va_list argptr;
+    va_start(argptr, format);
+
+    wxString s;
+    s.PrintfV(format, argptr);
+
+    va_end(argptr);
+
+    return s;
+}
+#endif // wxUSE_UNICODE_UTF8

 
 /* static */
 wxString wxString::FormatV(const wxString& format, va_list argptr)
 
 /* static */
 wxString wxString::FormatV(const wxString& format, va_list argptr)


@@ -1466,10 +1505,11 @@ wxString wxString::FormatV(const wxString& format, va_list argptr)
     return s;
 }
 
     return s;
 }
 

+#if !wxUSE_UTF8_LOCALE_ONLY

 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN

-int wxStringPrintfMixinBase::DoPrintf(const wxChar *format, ...)
+int wxStringPrintfMixinBase::DoPrintfWchar(const wxChar *format, ...)

 #else
 #else

-int wxString::DoPrintf(const wxChar *format, ...)
+int wxString::DoPrintfWchar(const wxChar *format, ...)

 #endif
 {
     va_list argptr;
 #endif
 {
     va_list argptr;


@@ -1490,19 +1530,52 @@ int wxString::DoPrintf(const wxChar *format, ...)
 
     return iLen;
 }
 
     return iLen;
 }

+#endif // !wxUSE_UTF8_LOCALE_ONLY

 
 

-int wxString::PrintfV(const wxString& format, va_list argptr)
+#if wxUSE_UNICODE_UTF8
+int wxString::DoPrintfUtf8(const char *format, ...)
+{
+    va_list argptr;
+    va_start(argptr, format);
+
+    int iLen = PrintfV(format, argptr);
+
+    va_end(argptr);
+
+    return iLen;
+}
+#endif // wxUSE_UNICODE_UTF8
+
+#if wxUSE_UNICODE_UTF8
+template<typename BufferType>
+#else
+// we only need one version in non-UTF8 builds and at least two Windows
+// compilers have problems with this function template, so use just one
+// normal function here
+#endif
+static int DoStringPrintfV(wxString& str,
+                           const wxString& format, va_list argptr)

 {
     int size = 1024;
 
     for ( ;; )
     {
 {
     int size = 1024;
 
     for ( ;; )
     {

-        wxStringBuffer tmp(*this, size + 1);
+#if wxUSE_UNICODE_UTF8
+        BufferType tmp(str, size + 1);
+        typename BufferType::CharType *buf = tmp;
+#else
+        wxStringBuffer tmp(str, size + 1);

         wxChar *buf = tmp;
         wxChar *buf = tmp;

+#endif

 
         if ( !buf )
         {
             // out of memory
 
         if ( !buf )
         {
             // out of memory

+
+            // in UTF-8 build, leaving uninitialized junk in the buffer
+            // could result in invalid non-empty UTF-8 string, so just
+            // reset the string to empty on failure:
+            buf[0] = '\0';

             return -1;
         }
 
             return -1;
         }
 


@@ -1510,7 +1583,7 @@ int wxString::PrintfV(const wxString& format, va_list argptr)
         // only a copy
         va_list argptrcopy;
         wxVaCopy(argptrcopy, argptr);
         // only a copy
         va_list argptrcopy;
         wxVaCopy(argptrcopy, argptr);

-        int len = wxVsnprintf(buf, size, (const wxChar*)/*FIXME-UTF8*/format, argptrcopy);
+        int len = wxVsnprintf(buf, size, format, argptrcopy);

         va_end(argptrcopy);
 
         // some implementations of vsnprintf() don't NUL terminate
         va_end(argptrcopy);
 
         // some implementations of vsnprintf() don't NUL terminate


@@ -1523,14 +1596,20 @@ int wxString::PrintfV(const wxString& format, va_list argptr)
         // buffer were large enough (newer standards such as Unix98)
         if ( len < 0 )
         {
         // buffer were large enough (newer standards such as Unix98)
         if ( len < 0 )
         {

+            // NB: wxVsnprintf() may call either wxCRT_VsnprintfW or
+            //     wxCRT_VsnprintfA in UTF-8 build; wxUSE_WXVSNPRINTF
+            //     is true if *both* of them use our own implementation,
+            //     otherwise we can't be sure

 #if wxUSE_WXVSNPRINTF
             // we know that our own implementation of wxVsnprintf() returns -1
             // only for a format error - thus there's something wrong with
             // the user's format string
 #if wxUSE_WXVSNPRINTF
             // we know that our own implementation of wxVsnprintf() returns -1
             // only for a format error - thus there's something wrong with
             // the user's format string

+            buf[0] = '\0';

             return -1;
             return -1;

-#else // assume that system version only returns error if not enough space
-            // still not enough, as we don't know how much we need, double the
-            // current size of the buffer
+#else // possibly using system version
+            // assume it only returns error if there is not enough space, but
+            // as we don't know how much we need, double the current size of
+            // the buffer

             size *= 2;
 #endif // wxUSE_WXVSNPRINTF/!wxUSE_WXVSNPRINTF
         }
             size *= 2;
 #endif // wxUSE_WXVSNPRINTF/!wxUSE_WXVSNPRINTF
         }


@@ -1554,9 +1633,34 @@ int wxString::PrintfV(const wxString& format, va_list argptr)
     }
 
     // we could have overshot
     }
 
     // we could have overshot

-    Shrink();
+    str.Shrink();

 
 

-    return length();
+    return str.length();
+}
+
+int wxString::PrintfV(const wxString& format, va_list argptr)
+{
+#if wxUSE_UNICODE_UTF8
+    #if wxUSE_STL_BASED_WXSTRING
+        typedef wxStringTypeBuffer<char> Utf8Buffer;
+    #else
+        typedef wxImplStringBuffer Utf8Buffer;
+    #endif
+#endif
+
+#if wxUSE_UTF8_LOCALE_ONLY
+    return DoStringPrintfV<Utf8Buffer>(*this, format, argptr);
+#else
+    #if wxUSE_UNICODE_UTF8
+    if ( wxLocaleIsUtf8 )
+        return DoStringPrintfV<Utf8Buffer>(*this, format, argptr);
+    else
+        // wxChar* version
+        return DoStringPrintfV<wxStringBuffer>(*this, format, argptr);
+    #else
+        return DoStringPrintfV(*this, format, argptr);
+    #endif // UTF8/WCHAR
+#endif

 }
 
 // ----------------------------------------------------------------------------
 }
 
 // ----------------------------------------------------------------------------