Re: [aspectj-users] Pointcut general form

[James Howe <jwh@xxxxxxxxxxxxxx>]

I see the definition for MethodPat and TypePat, but I don't see a 
definition for either ModifiersPat or IdPat.  I'll check out the 
programming guide.

On Tue, 13 May 2003 13:35:46 -0700, Wes Isberg <wes@xxxxxxxxxxxxxx> wrote:

> Hi James -
>
> Good eye!
>
> The definitions for MethodPat, et al. are at the end of that section on 
> the second page, where, e.g., MethodPat is:
>
> [ModifiersPat] TypePat [TypePat . ] IdPat ( TypePat , … ) [ throws 
> ThrowsPat ]
>
> As for execution(* *.*(..))
>
> Thank you for pointing out an omission in the quick reference.  Here:
>
> MethodPat :     *         *    .   *   (    ..    )
> MethodPat : {TypePat} {TypePat}.{IdPat}({TypePat?})
>
> We forgot to specify that (..) in method or constructor patterns means 
> the same as it does in the args(..) pattern: any number of arguments.
>
> The best source for grammar and usage is the programming guide. There's a 
> quick reference appendix and a semantics guide
> which discusses things like args(..).  However, we don't publish a 
> grammar in BNF form; the parser is a bit more careful, e.g., to avoid 
> keywords.
>
> Wes
>
>
> James Howe wrote:
> > I'm trying to get a better understanding of the various ways that 
> > pointcuts
> > can be defined, particularly when using wildcards.  I was looking at the
> > Quick Reference PDF and noticed that the definition of the general form 
> > for
> > pointcuts seems to be incomplete (or I'm overlooking something).  In the
> > section called "Primitive Pointcuts" in the general form section, the
> > following grammar is given (I'm substituting <> for italicized 
> > informatino)
> > :
> >
> > execution(<MethodPat>)
> >
> > where
> >
> > <MethodPat>: [<ModifiersPat>] <TypePat> [<TypePat> .] <IdPat>(<TypePat> 
> > ,
> > ... ) [ throws <ThrowsPat> ]
> > <TypePat> :  ... type pattern definition
> >
> > However, I can't find a definition for <MethodPat>, <IdPat> or 
> > <ThrowsPat>
> >
> > I'm also not convinced that the grammar given is accurate.  For example,
> > I've seen definitions which look like this:
> >
> > execution(* *.*(..))
> >
> > but I can't seem to construct that definition from the given grammar
> > although perhaps what is missing is the definition of IdPat.
> >
> > Is there a better document which describes the grammar used to parse
> > pointcut definitions?
> >
> > Thanks.
> >
> > --
> > James Howe
> > _______________________________________________
> > aspectj-users mailing list
> > aspectj-users@xxxxxxxxxxx
> > http://dev.eclipse.org/mailman/listinfo/aspectj-users
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/aspectj-users



--
James Howe