Why use sal_Unicode for this when OUStringLiteral1 takes a char anyway?

The compiler generates an error anyway if we would pass a sal_Unicode that is
outside the range of 'char' as the template parameter to OUStringLiteral1, so
if for some reason we would want cSequenceMarkSeparator to be outside of
ASCII, the code would need changes.

Change-Id: I1f6be560d3347cb957e1283dea2f7f7920d3aba7

sw/inc/swtypes.hxx

@@ -169,7 +169,7 @@ SW_DLLPUBLIC Size GetGraphicSizeTwip( const Graphic&, OutputDevice* pOutDev );
 // Separator for jumps to different content types in document.
 const sal_Unicode cMarkSeparator = '|';
 // Sequences names for jumps are <name of sequence>!<no>
-const sal_Unicode cSequenceMarkSeparator = '!';
+const char cSequenceMarkSeparator = '!';
 
 #define DB_DELIM ((sal_Unicode)0xff)        // Database <-> table separator.