[cfe-commits] r68911 - in /cfe/trunk: include/clang/Basic/DiagnosticParseKinds.td lib/Parse/ParseDecl.cpp test/Lexer/block_cmt_end.c test/Parser/declarators.c test/Parser/objc-foreach-syntax.m test/SemaObjC/exception-go-boom.m

[Sebastian Redl]

Chris Lattner wrote:
> Author: lattner
> Date: Sun Apr 12 15:42:31 2009
> New Revision: 68911
>
> URL: http://llvm.org/viewvc/llvm-project?rev=68911&view=rev
> Log:
> Implement the first set of changes for PR3963 and rdar://6759604,
> which tries to do better error recovery when it is "obvious" that an
> identifier is a mis-typed typename.

Very, very cool. I think you have to work with MSVC's awful error
recovery to *really* appreciate how nice this is.

> I'd really appreciate any scrutiny possible on this, it 
> is a tricky area.
>
>   

Looks pretty solid to me. One comment below.

> +/// isValidAfterIdentifierInDeclaratorAfterDeclSpec - Return true if the
> +/// specified token is valid after the identifier in a declarator which
> +/// immediately follows the declspec.  For example, these things are valid:
> +///
> +///      int x   [             4];         // direct-declarator
> +///      int x   (             int y);     // direct-declarator
> +///  int(int x   )                         // direct-declarator
> +///      int x   ;                         // simple-declaration
> +///      int x   =             17;         // init-declarator-list
> +///      int x   ,             y;          // init-declarator-list
> +///      int x   __asm__       ("foo");    // init-declarator-list
> +///
>   

We don't parse it yet, but in C++0x there's the unified initialization
syntax, which makes this valid:

int x {5};

which is equivalent to

int x = {5};

and

int x(5);

(Yay for redundancy.)

Sebastian