Re: [gomp] Parser for OpenMP directives

["Joseph S. Myers" <joseph at codesourcery dot com>]

On Mon, 28 Mar 2005, Daniel Jacobowitz wrote:

> > This is going to have to be done differently.  Note that section 2.1
> > specifies that preprocessing tokens following "#pragma omp" are 
> > subject to macro replacement.  Thus
> > 
> > #define foo barrier
> > #pragma omp foo
> >   
> > is legal.  AFAIK, the interface currently provided by libcpp contains
> > the raw text line.
> 
> You're right; but see c_register_pragma_with_expansion.

In this regard note bug 18638: #pragma macro expansion is disabled when 
deferred pragma handling is in effect (for C++) to avoid expanding macros 
in the wrong context of macro definitions.  We will want to do lexing up 
from for C as well at some point, and this issue with #pragma macro 
expansion will need to be fixed, as it will if an OpenMP parser for C++ is 
to be implemented.

I concur with Zack's view that cpplib should end up expanding all 
non-ignored pragmas as a token sequence (with or without macro expansion) 
which is then parsed by the parser however it likes.  This would start 
with a special token for #pragma, end with a special token for the end of 
a #pragma (I prefer such a special token to ending with an ordinary ';' 
token) and inbetween have the tokens of the pragma.  The pragma namespace 
("GCC", "STDC", "omp") might or might not be a property of the #pragma 
special token rather than a separate token, as might the following 
identifier naming the pragma in question.

-- 
Joseph S. Myers               http://www.srcf.ucam.org/~jsm28/gcc/
    jsm@polyomino.org.uk (personal mail)
    joseph@codesourcery.com (CodeSourcery mail)
    jsm28@gcc.gnu.org (Bugzilla assignments and CCs)