fixes or embedded NULs handling in wxString

git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@59940 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

diff --git a/include/wx/string.h b/include/wx/string.h
index 1dc72edeb7da204cc9d95140ef7d6b8b5547aed8..2ebce253632e9d2ad5f7c1ee3e891800472bd5af 100644 (file)
--- a/include/wx/string.h
+++ b/include/wx/string.h
@@ -1277,9 +1277,9 @@ public:
     { assign(pwz, nLength); }
 
   wxString(const wxScopedCharBuffer& buf)
-    { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
+    { assign(buf.data(), buf.length()); }
   wxString(const wxScopedWCharBuffer& buf)
-    { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
+    { assign(buf.data(), buf.length()); }
 
     // NB: this version uses m_impl.c_str() to force making a copy of the
     //     string, so that "wxString(str.c_str())" idiom for passing strings
@@ -1341,8 +1341,10 @@ public:
   #else
     // wxStringImpl is either not std::string or needs conversion
     operator wxStdWideString() const
-        // FIXME-UTF8: broken for embedded NULs
-        { return wxStdWideString(wc_str()); }
+    {
+        wxScopedWCharBuffer buf(wc_str());
+        return wxStdWideString(buf.data(), buf.length());
+    }
   #endif
 
   #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING
@@ -1351,8 +1353,10 @@ public:
   #else
     // wxStringImpl is either not std::string or needs conversion
     operator std::string() const
-        // FIXME-UTF8: broken for embedded NULs
-        { return std::string(mb_str()); }
+    {
+        wxScopedCharBuffer buf(mb_str());
+        return std::string(buf.data(), buf.length());
+    }
   #endif
 #endif // wxUSE_STL
 
@@ -1921,10 +1925,10 @@ public:
 
     // from wxScopedWCharBuffer
   wxString& operator=(const wxScopedWCharBuffer& s)
-    { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
+    { return assign(s); }
     // from wxScopedCharBuffer
   wxString& operator=(const wxScopedCharBuffer& s)
-    { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs
+    { return assign(s); }
 
   // string concatenation
     // in place concatenation
@@ -1960,9 +1964,9 @@ public:
 
       // string += buffer (i.e. from wxGetString)
   wxString& operator<<(const wxScopedWCharBuffer& s)
-    { return operator<<((const wchar_t *)s); }
+    { return append(s); }
   wxString& operator<<(const wxScopedCharBuffer& s)
-    { return operator<<((const char *)s); }
+    { return append(s); }
 
     // string += C string
   wxString& Append(const wxString& s)
@@ -2464,9 +2468,9 @@ public:
   wxString& append(const wxCStrData& str)
     { return append(str.AsString()); }
   wxString& append(const wxScopedCharBuffer& str)
-    { return append(str.data()); }
+    { return append(str.data(), str.length()); }
   wxString& append(const wxScopedWCharBuffer& str)
-    { return append(str.data()); }
+    { return append(str.data(), str.length()); }
   wxString& append(const wxCStrData& str, size_t n)
     { return append(str.AsString(), 0, n); }
   wxString& append(const wxScopedCharBuffer& str, size_t n)
@@ -2596,9 +2600,9 @@ public:
   wxString& assign(const wxCStrData& str)
     { return assign(str.AsString()); }
   wxString& assign(const wxScopedCharBuffer& str)
-    { return assign(str.data()); }
+    { return assign(str.data(), str.length()); }
   wxString& assign(const wxScopedWCharBuffer& str)
-    { return assign(str.data()); }
+    { return assign(str.data(), str.length()); }
   wxString& assign(const wxCStrData& str, size_t len)
     { return assign(str.AsString(), len); }
   wxString& assign(const wxScopedCharBuffer& str, size_t len)
@@ -3352,9 +3356,9 @@ public:
       return *this;
   }
   wxString& operator+=(const wxScopedCharBuffer& s)
-    { return operator+=(s.data()); }
+    { return append(s); }
   wxString& operator+=(const wxScopedWCharBuffer& s)
-    { return operator+=(s.data()); }
+    { return append(s); }
       // string += char
   wxString& operator+=(wxUniChar ch)
   {

diff --git a/src/common/string.cpp b/src/common/string.cpp
index b66d5bbeb0a29537fe554cb35a635c8113430685..220df16994c19f153915e6b74bf7d27f52c16e79 100644 (file)
--- a/src/common/string.cpp
+++ b/src/common/string.cpp
@@ -499,7 +499,7 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
             // we must pass the real string length to SubstrBufFromMB ctor
             if ( nLength == npos )
                 nLength = psz ? strlen(psz) : 0;
-            return SubstrBufFromMB(wxCharBuffer::CreateNonOwned(psz, nLength),
+            return SubstrBufFromMB(wxScopedCharBuffer::CreateNonOwned(psz, nLength),
                                    nLength);
         }
         // else: do the roundtrip through wchar_t*
@@ -550,17 +550,25 @@ wxString::SubstrBufFromWC wxString::ConvertStr(const wchar_t *pwz, size_t nLengt
 //Convert wxString in Unicode mode to a multi-byte string
 const wxScopedCharBuffer wxString::mb_str(const wxMBConv& conv) const
 {
-    return conv.cWC2MB(wx_str(), length() + 1 /* size, not length */, NULL);
+    // NB: Length passed to cWC2MB() doesn't include terminating NUL, it's
+    //     added by it automatically. If we passed length()+1 here, it would
+    //     create a buffer with 2 trailing NULs of length one greater than
+    //     expected.
+    return conv.cWC2MB(wx_str(), length(), NULL);
 }
 
 #elif wxUSE_UNICODE_UTF8
 
 const wxScopedWCharBuffer wxString::wc_str() const
 {
+    // NB: Length passed to cMB2WC() doesn't include terminating NUL, it's
+    //     added by it automatically. If we passed length()+1 here, it would
+    //     create a buffer with 2 trailing NULs of length one greater than
+    //     expected.
     return wxMBConvStrictUTF8().cMB2WC
                                 (
                                     m_impl.c_str(),
-                                    m_impl.length() + 1, // size, not length
+                                    m_impl.length(),
                                     NULL
                                 );
 }
@@ -570,22 +578,11 @@ const wxScopedCharBuffer wxString::mb_str(const wxMBConv& conv) const
     if ( conv.IsUTF8() )
         return wxScopedCharBuffer::CreateNonOwned(m_impl.c_str(), m_impl.length());
 
-    // FIXME-UTF8: use wc_str() here once we have buffers with length
-
-    size_t wcLen;
-    wxScopedWCharBuffer wcBuf
-                        (
-                          wxMBConvStrictUTF8().cMB2WC
-                                               (
-                                                 m_impl.c_str(),
-                                                 m_impl.length() + 1, // size
-                                                 &wcLen
-                                               )
-                        );
-    if ( !wcLen )
+    wxScopedWCharBuffer wcBuf(wc_str());
+    if ( !wcBuf.length() )
         return wxCharBuffer("");
 
-    return conv.cWC2MB(wcBuf, wcLen+1, NULL);
+    return conv.cWC2MB(wcBuf.data(), wcBuf.length(), NULL);
 }
 
 #else // ANSI
@@ -594,7 +591,11 @@ const wxScopedCharBuffer wxString::mb_str(const wxMBConv& conv) const
 //mode is not enabled and wxUSE_WCHAR_T is enabled
 const wxScopedWCharBuffer wxString::wc_str(const wxMBConv& conv) const
 {
-    return conv.cMB2WC(wx_str(), length() + 1 /* size, not length */, NULL);
+    // NB: Length passed to cMB2WC() doesn't include terminating NUL, it's
+    //     added by it automatically. If we passed length()+1 here, it would
+    //     create a buffer with 2 trailing NULs of length one greater than
+    //     expected.
+    return conv.cMB2WC(wx_str(), length(), NULL);
 }
 
 #endif // Unicode/ANSI