Re[1]: [cedet-semantic] Calling functions for default values in C++.

[Eric M. L. <er...@si...>]

>>> Thomas Maier <T....@ep...> seems to think that:
>Hello again :)
>
>and yet another thing semantic does not parse.  Not too bad as it
>doesn't keep me from using ecb, but anyway.  In C++ I can write
>
>int g();
>void f(int i=g());
>
>However, semantic doesn't parse the "()" from "g()".  I wrote a very
>very simple fix for that because I do not have time to really create
>something better.  And, even worse, I do not know if this breaks
>something else.  Anyway, it might be the right place where a better fix
>should be applied.
>
>I added function-call and changed expression:
>
>
>function-call: symbol semantic-list
>             ;
>
># Use expression for parsing only.  Don't actually return anything
># for now.  Hopefully we can fix this later.
>expression : number
>	     ( (identity start) (identity end) )
>           | function-call
>	     ( (identity start) (identity end) )
>	   | symbol
>	     ( (identity start) (identity end) )
>	   | string
>	     ( (identity start) (identity end))
>           | type-cast expression  # A cast to some other type
>	     ( (identity start) (identity end) )
>	   | semantic-list
>	     ( (identity start) (identity end) )
>	   | punctuation "[-+*/%^|&]" expression
>	     ( (identity start) (identity end) )
>	   ;
>
>
>But unfortunately I'm still not too comfortable with semantic and c.bnf
>so beware and maybe you better ignore the fix and write a real one.
  [ ... ]

That seems like a reasonable fix for this situation.  The expression
parser has always been a bit overly simple for the reason you cite
above.

Thanks
Eric

-- 
          Eric Ludlam:                 za...@gn..., er...@si...
   Home: www.ultranet.com/~zappo            Siege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: www.gnu.org