[cfe-dev] c++0x char/string literals

[Kona Blend]

Hi,

I was reading the cfe-dev archives and there was some talk of c++0x char/string literals about 8 months ago. Is this work still happening or has it maybe stalled?

I wouldn't mind taking a stab at modify clang's lexer and have some questions which might be moot if there's already coding happening:

There is code in lex which enforces some string literal concatenation rules. With c++0x there may be combinations which are ill-formed and I'm wondering if the parser would be a better place for that logic; I'm exploring what can be done to reduce lex code which needs to differentiate between string literal token kinds and potentially drop tok::wide_string_literal -- there don't seem to be too many places where it's needed.

As a simple test, I have expanded the number of string_literal token kinds to a total of 10 for c++0x and added lex code for the various incantations. It's workable but does seem heavy handed. I'd rather see if these bits can go somewhere else.

Any guidance or insight would be appreciated.

--kb