Does anyone care enough to look at this?  I'm not opposed to a code
review, but I don't want to sit around waiting.  Everything compiles
and all regressions pass.  The main issue is that I've changed the
slicc grammar slightly to accept :: in types.  Seems like a good thing
to me, but I just wanted to know if anyone saw a problem with this.

If you need time, that's fine, at least let me know.

  Nate

On Wed, Mar 31, 2010 at 5:11 PM, nathan binkert <[email protected]> wrote:
> I got rid of "using namespace std;" in ruby headers, but to do so, I
> need to change string to std::string in slicc.  This requires that I
> change the slicc parser to accept :: as part of a type.  The diff to
> slicc is below.  (The whole diff is like 3000 lines, and is attached
>
> Anyone have a problem with this?
>
>  Nate
>
> diff --git a/src/mem/slicc/parser.py b/src/mem/slicc/parser.py
> --- a/src/mem/slicc/parser.py
> +++ b/src/mem/slicc/parser.py
> @@ -394,8 +394,16 @@
>         "types : empty"
>         p[0] = []
>
> -    def p_type(self, p):
> -        "type : ident"
> +    def p_typestr__multi(self, p):
> +        "typestr : typestr DOUBLE_COLON ident"
> +        p[0] = '%s::%s' % (p[1], p[3])
> +
> +    def p_typestr__single(self, p):
> +        "typestr : ident"
> +        p[0] = p[1]
> +
> +    def p_type__one(self, p):
> +        "type : typestr"
>         p[0] = ast.TypeAST(self, p[1])
>
>     def p_void(self, p):
> @@ -670,7 +678,7 @@
>
>     def p_literal__string(self, p):
>         "literal : STRING"
> -        p[0] = ast.LiteralExprAST(self, p[1], "string")
> +        p[0] = ast.LiteralExprAST(self, p[1], "std::string")
>
>     def p_literal__number(self, p):
>         "literal : NUMBER"
>
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to