]> git.saurik.com Git - wxWidgets.git/blobdiff - src/common/string.cpp
added support for reading resolution information from TIFF, JPEG and BMP formats...
[wxWidgets.git] / src / common / string.cpp
index 34453a87c3d8df365c1c90b476c75375e5e7be3b..d0a6a445691b1b242f75676a89aaeb1d821a2a10 100644 (file)
@@ -340,7 +340,7 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
         return SubstrBufFromMB("", 0);
 
     // and then to UTF-8:
-    SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxMBConvUTF8()));
+    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") );
 
@@ -382,9 +382,12 @@ const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const
 
 const wxWCharBuffer wxString::wc_str() const
 {
-    return wxMBConvUTF8().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
@@ -395,14 +398,16 @@ const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const
     // FIXME-UTF8: use wc_str() here once we have buffers with length
 
     size_t wcLen;
-    wxWCharBuffer wcBuf(
-            wxMBConvUTF8().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("");
 
-    return conv.cWC2MB(wcBuf, wcLen, NULL);
+    return conv.cWC2MB(wcBuf, wcLen+1, NULL);
 }
 
 #else // ANSI
@@ -565,6 +570,12 @@ wxString operator+(const wchar_t *pwz, const wxString& str)
 // 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
@@ -967,36 +978,36 @@ int wxString::CmpNoCase(const wxString& s) const
 #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;
 
-    size_t len = strlen(ascii);
     wxString res;
 
-    if ( len )
     {
-        wxImplStringBuffer buf(res, len);
+        wxStringInternalBuffer buf(res, len);
         wxStringCharType *dest = buf;
 
-        for ( ;; )
+        for ( ; len > 0; --len )
         {
             unsigned char c = (unsigned char)*ascii++;
             wxASSERT_MSG( c < 0x80,
                           _T("Non-ASCII value passed to FromAscii().") );
 
             *dest++ = (wchar_t)c;
-
-            if ( c == '\0' )
-                break;
         }
     }
 
     return res;
 }
 
-wxString wxString::FromAscii(const char ascii)
+wxString wxString::FromAscii(const char *ascii)
+{
+    return FromAscii(ascii, wxStrlen(ascii));
+}
+
+wxString wxString::FromAscii(char ascii)
 {
     // What do we do with '\0' ?
 
@@ -1376,52 +1387,50 @@ int wxString::Find(wxUniChar ch, bool bFromEnd) const
 // conversion to numbers
 // ----------------------------------------------------------------------------
 
-// the implementation of all the functions below is exactly the same so factor
-// it out
-
-template <typename T, typename F>
-bool wxStringToIntType(const wxChar *start,
-                       T *val,
-                       int base,
-                       F func)
-{
-    wxCHECK_MSG( val, false, _T("NULL output pointer") );
-    wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );
+// 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.
 
 #ifndef __WXWINCE__
-    errno = 0;
+    #define DO_IF_NOT_WINCE(x) x
+#else
+    #define DO_IF_NOT_WINCE(x)
 #endif
 
-    wxChar *end;
-    *val = (*func)(start, &end, base);
-
-    // return true only if scan was stopped by the terminating NUL and if the
-    // string was not empty to start with and no under/overflow occurred
-    return !*end && (end != start)
-#ifndef __WXWINCE__
-        && (errno != ERANGE)
-#endif
-    ;
-}
+#define WX_STRING_TO_INT_TYPE(val, base, func)                              \
+    wxCHECK_MSG( val, false, _T("NULL output pointer") );                   \
+    wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") );  \
+                                                                            \
+    DO_IF_NOT_WINCE( errno = 0; )                                           \
+                                                                            \
+    const wxStringCharType *start = wx_str();                               \
+    wxStringCharType *end;                                                  \
+    *val = func(start, &end, base);                                         \
+                                                                            \
+    /* return true only if scan was stopped by the terminating NUL and */   \
+    /* if the string was not empty to start with and no under/overflow */   \
+    /* occurred: */                                                         \
+    return !*end && (end != start)                                          \
+        DO_IF_NOT_WINCE( && (errno != ERANGE) )
 
 bool wxString::ToLong(long *val, int base) const
 {
-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtol);
+    WX_STRING_TO_INT_TYPE(val, base, wxStrtol);
 }
 
 bool wxString::ToULong(unsigned long *val, int base) const
 {
-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoul);
+    WX_STRING_TO_INT_TYPE(val, base, wxStrtoul);
 }
 
 bool wxString::ToLongLong(wxLongLong_t *val, int base) const
 {
-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoll);
+    WX_STRING_TO_INT_TYPE(val, base, wxStrtoll);
 }
 
 bool wxString::ToULongLong(wxULongLong_t *val, int base) const
 {
-    return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoull);
+    WX_STRING_TO_INT_TYPE(val, base, wxStrtoull);
 }
 
 bool wxString::ToDouble(double *val) const
@@ -1559,6 +1568,11 @@ static int DoStringPrintfV(wxString& str,
         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;
         }
 
@@ -1579,14 +1593,20 @@ static int DoStringPrintfV(wxString& str,
         // 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
+            buf[0] = '\0';
             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
         }
@@ -1621,7 +1641,7 @@ int wxString::PrintfV(const wxString& format, va_list argptr)
     #if wxUSE_STL_BASED_WXSTRING
         typedef wxStringTypeBuffer<char> Utf8Buffer;
     #else
-        typedef wxImplStringBuffer Utf8Buffer;
+        typedef wxStringInternalBuffer Utf8Buffer;
     #endif
 #endif
 
@@ -1809,3 +1829,32 @@ wxString wxString::Upper() const
 
 // convert to lower case, return the copy of the string
 wxString wxString::Lower() const { wxString s(*this); return s.MakeLower(); }
+
+// ----------------------------------------------------------------------------
+// wxUTF8StringBuffer
+// ----------------------------------------------------------------------------
+
+#if wxUSE_UNICODE_WCHAR
+wxUTF8StringBuffer::~wxUTF8StringBuffer()
+{
+    wxMBConvStrictUTF8 conv;
+    size_t wlen = conv.ToWChar(NULL, 0, m_buf);
+    wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" );
+
+    wxStringInternalBuffer wbuf(m_str, wlen);
+    conv.ToWChar(wbuf, wlen, m_buf);
+}
+
+wxUTF8StringBufferLength::~wxUTF8StringBufferLength()
+{
+    wxCHECK_RET(m_lenSet, "length not set");
+
+    wxMBConvStrictUTF8 conv;
+    size_t wlen = conv.ToWChar(NULL, 0, m_buf, m_len);
+    wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" );
+
+    wxStringInternalBufferLength wbuf(m_str, wlen);
+    conv.ToWChar(wbuf, wlen, m_buf, m_len);
+    wbuf.SetLength(wlen);
+}
+#endif // wxUSE_UNICODE_WCHAR