Re: [PATCH] libcpp deferred pragma hack (PR c++/25294)

[Mark Mitchell <mark at codesourcery dot com>]

Richard Henderson wrote:
> On Thu, Dec 29, 2005 at 01:41:04PM -0800, Mark Mitchell wrote:
> 
>>(Perhaps you could merge those pieces to mainline, since
>>they're presumably independent of OpenMP?)
> 
> 
> There is a rather largish patch pending review from c++ front end
> maintainers...  ;-)
> 
>   http://gcc.gnu.org/ml/gcc-patches/2005-11/msg01888.html

Patch OK; nits below.

+      /* We smuggled the cpp_token->u.pragma value in an INTEGER_CST.  */
+      token->pragma_kind = TREE_INT_CST_LOW (token->value);

I guess we have no convenient choice, but making value a union would be
nicer; I hate to add places where we wrap things in trees....

+cp_parser_skip_to_pragma_eol (cp_parser* parser, cp_token *pragma_tok)

Need documentation for new function paramater pragma_tok.

+		     bool in_compound)

And here for in_compound.  (There are several functions which now have
an in_compound parameter.)

-      cp_parser_statement_seq_opt (parser, false);
+      cp_parser_statement_seq_opt (parser, NULL_TREE);

Wow, thanks.  Using a boolean where we wanted a tree; ugh.

Thanks,

-- 
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
(650) 331-3385 x713