fix unicode decomposition (bugs 3313 and 3498):

* src/support/docstring.{cpp,h}:
	- rename normalize_kc to normalize_c and use normalization form C (instead of KC)

* src/callback.cpp:
* src/Trans.cpp:
* src/frontends/qt4/GuiClipboard.cpp:
* src/frontends/qt4/GuiSelection.cpp:
* lib/lyx2lyx/lyx_1_5.py:
	- use normalization forms C and D instead of KC and KD.

git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@18991 a592a061-630c-0410-9148-cb99ea01b6c8

lib/lyx2lyx/lyx_1_5.py

@@ -921,7 +921,7 @@ def _convert_accent(accent, accented_char):
         return ''
     a = accent_map.get(type)
     if a:
-        return unicodedata.normalize("NFKC", "%s%s" % (char, a))
+        return unicodedata.normalize("NFC", "%s%s" % (char, a))
     return ''
 
 
@@ -1027,9 +1027,9 @@ def revert_accent(document):
         # because we never use u'xxx' for string literals, but 'xxx'.
         # Therefore we may have to try two times to normalize the data.
         try:
-            document.body[i] = unicodedata.normalize("NFKD", document.body[i])
+            document.body[i] = unicodedata.normalize("NFD", document.body[i])
         except TypeError:
-            document.body[i] = unicodedata.normalize("NFKD", unicode(document.body[i], 'utf-8'))
+            document.body[i] = unicodedata.normalize("NFD", unicode(document.body[i], 'utf-8'))
 
     # Replace accented characters with InsetLaTeXAccent
     # Do not convert characters that can be represented in the chosen
@@ -1086,7 +1086,7 @@ def revert_accent(document):
                     accented_char = inverse_accented_map[accented_char]
                 accent = document.body[i][j]
                 try:
-                    dummy = unicodedata.normalize("NFKC", accented_char + accent).encode(encoding_stack[-1])
+                    dummy = unicodedata.normalize("NFC", accented_char + accent).encode(encoding_stack[-1])
                 except UnicodeEncodeError:
                     # Insert the rest of the line as new line
                     if j < len(document.body[i]) - 1:
@@ -1101,7 +1101,7 @@ def revert_accent(document):
                     break
     # Normalize to "Normal form C" (NFC, pre-composed characters) again
     for i in range(numberoflines):
-        document.body[i] = unicodedata.normalize("NFKC", document.body[i])
+        document.body[i] = unicodedata.normalize("NFC", document.body[i])
 
 
 def normalize_font_whitespace_259(document):

src/Trans.cpp

@@ -124,7 +124,7 @@ static docstring const doAccent(docstring const & s, tex_accent accent)
 			       << lyx_accent_table[accent].name << '.' << std::endl;
 		os << s.substr(1);
 	}
-	return normalize_kc(os.str());
+	return normalize_c(os.str());
 }
 
 

src/callback.cpp

@@ -409,7 +409,7 @@ docstring const getContentsOfPlaintextFile(BufferView * bv, string const & f,
 		file_content = from_local8bit(tmpstr);
 	}
 
-	return normalize_kc(file_content);
+	return normalize_c(file_content);
 }
 
 

src/frontends/qt4/GuiClipboard.cpp

@@ -62,7 +62,7 @@ docstring const GuiClipboard::getAsText() const
 {
 	// text data from other applications
 	QString const str = qApp->clipboard()->text(QClipboard::Clipboard)
-				.normalized(QString::NormalizationForm_KC);
+				.normalized(QString::NormalizationForm_C);
 	LYXERR(Debug::ACTION) << "GuiClipboard::getAsText(): `"
 			      << fromqstr(str) << "'" << endl;
 	if (str.isNull())

src/frontends/qt4/GuiSelection.cpp

@@ -59,7 +59,7 @@ void GuiSelection::haveSelection(bool own)
 docstring const GuiSelection::get() const
 {
 	QString const str = qApp->clipboard()->text(QClipboard::Selection)
-				.normalized(QString::NormalizationForm_KC);
+				.normalized(QString::NormalizationForm_C);
 	LYXERR(Debug::ACTION) << "GuiSelection::get: " << fromqstr(str)
 			      << endl;
 	if (str.isNull())

src/support/docstring.cpp

@@ -140,9 +140,9 @@ std::string const to_filesystem8bit(docstring const & s)
 }
 
 
-docstring const normalize_kc(docstring const & s)
+docstring const normalize_c(docstring const & s)
 {
-	return qstring_to_ucs4(toqstr(s).normalized(QString::NormalizationForm_KC));
+	return qstring_to_ucs4(toqstr(s).normalized(QString::NormalizationForm_C));
 }
 
 

src/support/docstring.h

@@ -62,8 +62,8 @@ docstring const from_filesystem8bit(std::string const & s);
 /// convert \p s from ucs4 to the encoding of the file system.
 std::string const to_filesystem8bit(docstring const & s);
 
-/// normalize \p s to precomposed form kc
-docstring const normalize_kc(docstring const & s);
+/// normalize \p s to precomposed form c
+docstring const normalize_c(docstring const & s);
 
 /// Compare a docstring with a C string of ASCII characters
 bool operator==(lyx::docstring const &, char const *);