C++: Avoid looking ahead when lexing u8"literal"

This makes things slightly more efficient. But it will be more
significant when we introduce R"rawliterals" since we would avoid
an even further lookahead for cases like u8R"string".

Change-Id: Id4bad8b917752d23daf2f4989330434979cf602f
Reviewed-by: Roberto Raggi <roberto.raggi@nokia.com>
Reviewed-by: hjk <qthjk@ovi.com>

1 files changed, 6 insertions, 10 deletions

diff --git a/src/libs/3rdparty/cplusplus/Lexer.cpp b/src/libs/3rdparty/cplusplus/Lexer.cpp
index 5179d83b54..5214a345f7 100644
--- a/src/libs/3rdparty/cplusplus/Lexer.cpp
+++ b/src/libs/3rdparty/cplusplus/Lexer.cpp
@@ -580,19 +580,15 @@ void Lexer::scan_helper(Token *tok)
                 yyinp();
                 scanCharLiteral(tok, ch);
             } else if (ch == 'u' && _yychar == '8') {
-                unsigned char la = 0;
-                if (_currentChar + 1 != _lastChar)
-                    la = *(_currentChar + 1);
-                if (la == '"') {
-                    yyinp();
+                yyinp();
+                if (_yychar == '"') {
                     yyinp();
                     scanStringLiteral(tok, '8');
-                } else if (la == '\'') {
-                    yyinp();
+                } else if (_yychar == '\'') {
                     yyinp();
                     scanCharLiteral(tok, '8');
                 } else {
-                    scanIdentifier(tok);
+                    scanIdentifier(tok, 1);
                 }
             } else {
                 scanIdentifier(tok);
@@ -691,9 +687,9 @@ void Lexer::scanNumericLiteral(Token *tok)
         tok->number = control()->numericLiteral(yytext, yylen);
 }
 
-void Lexer::scanIdentifier(Token *tok)
+void Lexer::scanIdentifier(Token *tok, unsigned extraProcessedChars)
 {
-    const char *yytext = _currentChar - 1;
+    const char *yytext = _currentChar - 1 - extraProcessedChars;
     while (std::isalnum(_yychar) || _yychar == '_' || _yychar == '$')
         yyinp();
     int yylen = _currentChar - yytext;