Use gdk_window_invalidate instead of gtk_widget_queue_draw as the latter doesn't...

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

diff --git a/src/common/string.cpp b/src/common/string.cpp
index a287759822baabf199d1bc79f2845cba128072b3..56af51ae22c28dbe3e8306171f575729270236ac 100644 (file)
--- a/src/common/string.cpp
+++ b/src/common/string.cpp
@@ -50,6 +50,18 @@
     #define wxStringStrlen   wxStrlen
 #endif
 
     #define wxStringStrlen   wxStrlen
 #endif
 

+// ----------------------------------------------------------------------------
+// global variables
+// ----------------------------------------------------------------------------
+
+namespace wxPrivate
+{
+
+static UntypedBufferData s_untypedNullData(NULL);
+
+UntypedBufferData * const untypedNullDataPtr = &s_untypedNullData;
+
+} // namespace wxPrivate

 
 // ---------------------------------------------------------------------------
 // static class variables definition
 
 // ---------------------------------------------------------------------------
 // static class variables definition


@@ -58,6 +70,105 @@
 //According to STL _must_ be a -1 size_t
 const size_t wxString::npos = (size_t) -1;
 
 //According to STL _must_ be a -1 size_t
 const size_t wxString::npos = (size_t) -1;
 

+#if wxUSE_STRING_POS_CACHE
+
+#ifdef wxHAS_COMPILER_TLS
+
+wxTLS_TYPE(wxString::Cache) wxString::ms_cache;
+
+#else // !wxHAS_COMPILER_TLS
+
+struct wxStrCacheInitializer
+{
+    wxStrCacheInitializer()
+    {
+        // calling this function triggers s_cache initialization in it, and
+        // from now on it becomes safe to call from multiple threads
+        wxString::GetCache();
+    }
+};
+
+/*
+wxString::Cache& wxString::GetCache()
+{
+    static wxTLS_TYPE(Cache) s_cache;
+
+    return wxTLS_VALUE(s_cache);
+}
+*/
+
+static wxStrCacheInitializer gs_stringCacheInit;
+
+#endif // wxHAS_COMPILER_TLS/!wxHAS_COMPILER_TLS
+
+// gdb seems to be unable to display thread-local variables correctly, at least
+// not my 6.4.98 version under amd64, so provide this debugging helper to do it
+#ifdef __WXDEBUG__
+
+struct wxStrCacheDumper
+{
+    static void ShowAll()
+    {
+        puts("*** wxString cache dump:");
+        for ( unsigned n = 0; n < wxString::Cache::SIZE; n++ )
+        {
+            const wxString::Cache::Element&
+                c = wxString::GetCacheBegin()[n];
+
+            printf("\t%u%s\t%p: pos=(%lu, %lu), len=%ld\n",
+                   n,
+                   n == wxString::LastUsedCacheElement() ? " [*]" : "",
+                   c.str,
+                   (unsigned long)c.pos,
+                   (unsigned long)c.impl,
+                   (long)c.len);
+        }
+    }
+};
+
+void wxDumpStrCache() { wxStrCacheDumper::ShowAll(); }
+
+#endif // __WXDEBUG__
+
+#ifdef wxPROFILE_STRING_CACHE
+
+wxString::CacheStats wxString::ms_cacheStats;
+
+struct wxStrCacheStatsDumper
+{
+    ~wxStrCacheStatsDumper()
+    {
+        const wxString::CacheStats& stats = wxString::ms_cacheStats;
+
+        if ( stats.postot )
+        {
+            puts("*** wxString cache statistics:");
+            printf("\tTotal non-trivial calls to PosToImpl(): %u\n",
+                   stats.postot);
+            printf("\tHits %u (of which %u not used) or %.2f%%\n",
+                   stats.poshits,
+                   stats.mishits,
+                   100.*float(stats.poshits - stats.mishits)/stats.postot);
+            printf("\tAverage position requested: %.2f\n",
+                   float(stats.sumpos) / stats.postot);
+            printf("\tAverage offset after cached hint: %.2f\n",
+                   float(stats.sumofs) / stats.postot);
+        }
+
+        if ( stats.lentot )
+        {
+            printf("\tNumber of calls to length(): %u, hits=%.2f%%\n",
+                   stats.lentot, 100.*float(stats.lenhits)/stats.lentot);
+        }
+    }
+};
+
+static wxStrCacheStatsDumper s_showCacheStats;
+
+#endif // wxPROFILE_STRING_CACHE
+
+#endif // wxUSE_STRING_POS_CACHE
+

 // ----------------------------------------------------------------------------
 // global functions
 // ----------------------------------------------------------------------------
 // ----------------------------------------------------------------------------
 // global functions
 // ----------------------------------------------------------------------------


@@ -69,7 +180,13 @@ const size_t wxString::npos = (size_t) -1;
 wxSTD ostream& operator<<(wxSTD ostream& os, const wxCStrData& str)
 {
 #if wxUSE_UNICODE && !wxUSE_UNICODE_UTF8
 wxSTD ostream& operator<<(wxSTD ostream& os, const wxCStrData& str)
 {
 #if wxUSE_UNICODE && !wxUSE_UNICODE_UTF8

-    return os << (const char *)str.AsCharBuf();
+    const wxCharBuffer buf(str.AsCharBuf());
+    if ( !buf )
+        os.clear(wxSTD ios_base::failbit);
+    else
+        os << buf.data();
+
+    return os;

 #else
     return os << str.AsInternal();
 #endif
 #else
     return os << str.AsInternal();
 #endif


@@ -123,22 +240,30 @@ void wxString::PosLenToImpl(size_t pos, size_t len,
                             size_t *implPos, size_t *implLen) const
 {
     if ( pos == npos )
                             size_t *implPos, size_t *implLen) const
 {
     if ( pos == npos )

+    {

         *implPos = npos;
         *implPos = npos;

-    else
+    }
+    else // have valid start position

     {
     {

-        const_iterator i = begin() + pos;
-        *implPos = wxStringImpl::const_iterator(i.impl()) - m_impl.begin();
+        const const_iterator b = GetIterForNthChar(pos);
+        *implPos = wxStringImpl::const_iterator(b.impl()) - m_impl.begin();

         if ( len == npos )
         if ( len == npos )

+        {

             *implLen = npos;
             *implLen = npos;

-        else
+        }
+        else // have valid length too

         {
         {

-            // too large length is interpreted as "to the end of the string"
-            // FIXME-UTF8: verify this is the case in std::string, assert
-            // otherwise
-            if ( pos + len > length() )
-                len = length() - pos;
-
-            *implLen = (i + len).impl() - i.impl();
+            // we need to handle the case of length specifying a substring
+            // going beyond the end of the string, just as std::string does
+            const const_iterator e(end());
+            const_iterator i(b);
+            while ( len && i <= e )
+            {
+                ++i;
+                --len;
+            }
+
+            *implLen = i.impl() - b.impl();

         }
     }
 }
         }
     }
 }


@@ -369,6 +494,9 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength,
         // UTF-8 sequence and psz may be invalid:
         if ( wxStringOperations::IsValidUtf8String(psz, nLength) )
         {
         // UTF-8 sequence and psz may be invalid:
         if ( wxStringOperations::IsValidUtf8String(psz, 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);
         }
         // else: do the roundtrip through wchar_t*
             return SubstrBufFromMB(wxCharBuffer::CreateNonOwned(psz), nLength);
         }
         // else: do the roundtrip through wchar_t*


@@ -987,6 +1115,7 @@ size_t wxString::find_last_not_of(const wxOtherCharType* sz, size_t nStart,
 
 int wxString::CmpNoCase(const wxString& s) const
 {
 
 int wxString::CmpNoCase(const wxString& s) const
 {

+#if wxUSE_UNICODE_UTF8

     // FIXME-UTF8: use wxUniChar::ToLower/ToUpper once added
 
     const_iterator i1 = begin();
     // FIXME-UTF8: use wxUniChar::ToLower/ToUpper once added
 
     const_iterator i1 = begin();


@@ -1010,6 +1139,9 @@ int wxString::CmpNoCase(const wxString& s) const
     else if ( len1 > len2 )
         return 1;
     return 0;
     else if ( len1 > len2 )
         return 1;
     return 0;

+#else // wxUSE_UNICODE_WCHAR or ANSI
+    return wxStricmp(m_impl.c_str(), s.m_impl.c_str());
+#endif

 }
 
 
 }
 
 


@@ -1166,7 +1298,7 @@ wxString wxString::Right(size_t nCount) const
   return dest;
 }
 
   return dest;
 }
 

-// get all characters after the last occurence of ch
+// get all characters after the last occurrence of ch

 // (returns the whole string if ch not found)
 wxString wxString::AfterLast(wxUniChar ch) const
 {
 // (returns the whole string if ch not found)
 wxString wxString::AfterLast(wxUniChar ch) const
 {


@@ -1175,7 +1307,7 @@ wxString wxString::AfterLast(wxUniChar ch) const
   if ( iPos == wxNOT_FOUND )
     str = *this;
   else
   if ( iPos == wxNOT_FOUND )
     str = *this;
   else

-    str = wx_str() + iPos + 1;
+    str.assign(*this, iPos + 1, npos);

 
   return str;
 }
 
   return str;
 }


@@ -1193,16 +1325,17 @@ wxString wxString::Left(size_t nCount) const
   return dest;
 }
 
   return dest;
 }
 

-// get all characters before the first occurence of ch
+// get all characters before the first occurrence of ch

 // (returns the whole string if ch not found)
 wxString wxString::BeforeFirst(wxUniChar ch) const
 {
   int iPos = Find(ch);
 // (returns the whole string if ch not found)
 wxString wxString::BeforeFirst(wxUniChar ch) const
 {
   int iPos = Find(ch);

-  if ( iPos == wxNOT_FOUND ) iPos = length();
+  if ( iPos == wxNOT_FOUND )
+      iPos = length();

   return wxString(*this, 0, iPos);
 }
 
   return wxString(*this, 0, iPos);
 }
 

-/// get all characters before the last occurence of ch
+/// get all characters before the last occurrence of ch

 /// (returns empty string if ch not found)
 wxString wxString::BeforeLast(wxUniChar ch) const
 {
 /// (returns empty string if ch not found)
 wxString wxString::BeforeLast(wxUniChar ch) const
 {


@@ -1214,19 +1347,19 @@ wxString wxString::BeforeLast(wxUniChar ch) const
   return str;
 }
 
   return str;
 }
 

-/// get all characters after the first occurence of ch
+/// get all characters after the first occurrence of ch

 /// (returns empty string if ch not found)
 wxString wxString::AfterFirst(wxUniChar ch) const
 {
   wxString str;
   int iPos = Find(ch);
   if ( iPos != wxNOT_FOUND )
 /// (returns empty string if ch not found)
 wxString wxString::AfterFirst(wxUniChar ch) const
 {
   wxString str;
   int iPos = Find(ch);
   if ( iPos != wxNOT_FOUND )

-    str = wx_str() + iPos + 1;
+      str.assign(*this, iPos + 1, npos);

 
   return str;
 }
 
 
   return str;
 }
 

-// replace first (or all) occurences of some substring with another one
+// replace first (or all) occurrences of some substring with another one

 size_t wxString::Replace(const wxString& strOld,
                          const wxString& strNew, bool bReplaceAll)
 {
 size_t wxString::Replace(const wxString& strOld,
                          const wxString& strNew, bool bReplaceAll)
 {


@@ -1234,29 +1367,60 @@ size_t wxString::Replace(const wxString& strOld,
     wxCHECK_MSG( !strOld.empty(), 0,
                  _T("wxString::Replace(): invalid parameter") );
 
     wxCHECK_MSG( !strOld.empty(), 0,
                  _T("wxString::Replace(): invalid parameter") );
 

+    wxSTRING_INVALIDATE_CACHE();
+

     size_t uiCount = 0;   // count of replacements made
 
     size_t uiCount = 0;   // count of replacements made
 

-    size_t uiOldLen = strOld.length();
-    size_t uiNewLen = strNew.length();
+    // optimize the special common case: replacement of one character by
+    // another one (in UTF-8 case we can only do this for ASCII characters)
+    //
+    // benchmarks show that this special version is around 3 times faster
+    // (depending on the proportion of matching characters and UTF-8/wchar_t
+    // build)
+    if ( strOld.m_impl.length() == 1 && strNew.m_impl.length() == 1 )
+    {
+        const wxStringCharType chOld = strOld.m_impl[0],
+                               chNew = strNew.m_impl[0];

 
 

-    for ( size_t dwPos = 0; dwPos < length(); )
+        // this loop is the simplified version of the one below
+        for ( size_t pos = 0; ; )
+        {
+            pos = m_impl.find(chOld, pos);
+            if ( pos == npos )
+                break;
+
+            m_impl[pos++] = chNew;
+
+            uiCount++;
+
+            if ( !bReplaceAll )
+                break;
+        }
+    }
+    else // general case

     {
     {

-        dwPos = find(strOld, dwPos);
-        if ( dwPos == npos )
-            break;
+        const size_t uiOldLen = strOld.m_impl.length();
+        const size_t uiNewLen = strNew.m_impl.length();

 
 

-        // replace this occurance of the old string with the new one
-        replace(dwPos, uiOldLen, strNew, uiNewLen);
+        for ( size_t pos = 0; ; )
+        {
+            pos = m_impl.find(strOld.m_impl, pos);
+            if ( pos == npos )
+                break;

 
 

-        // move up pos past the string that was replaced
-        dwPos += uiNewLen;
+            // replace this occurrence of the old string with the new one
+            m_impl.replace(pos, uiOldLen, strNew.m_impl);

 
 

-        // increase replace count
-        ++uiCount;
+            // move up pos past the string that was replaced
+            pos += uiNewLen;

 
 

-        // stop after the first one?
-        if ( !bReplaceAll )
-            break;
+            // increase replace count
+            uiCount++;
+
+            // stop after the first one?
+            if ( !bReplaceAll )
+                break;
+        }

     }
 
     return uiCount;
     }
 
     return uiCount;


@@ -1331,6 +1495,20 @@ wxString& wxString::MakeLower()
   return *this;
 }
 
   return *this;
 }
 

+wxString& wxString::MakeCapitalized()
+{
+    const iterator en = end();
+    iterator it = begin();
+    if ( it != en )
+    {
+        *it = (wxChar)wxToupper(*it);
+        for ( ++it; it != en; ++it )
+            *it = (wxChar)wxTolower(*it);
+    }
+
+    return *this;
+}
+

 // ---------------------------------------------------------------------------
 // trimming and padding
 // ---------------------------------------------------------------------------
 // ---------------------------------------------------------------------------
 // trimming and padding
 // ---------------------------------------------------------------------------


@@ -1935,38 +2113,3 @@ int wxString::Freq(wxUniChar ch) const
     return count;
 }
 
     return count;
 }
 

-// convert to upper case, return the copy of the string
-wxString wxString::Upper() const
-{ wxString s(*this); return s.MakeUpper(); }
-
-// 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