[Versa] equality

[Jeremy Kloth]

On Monday 26 September 2005 12:46 pm, Michael Olson wrote:
> On Sep 26, 2005, at 12:39 PM, Jeremy Kloth wrote:
> > On Monday 26 September 2005 12:29 pm, Michael Olson wrote:
> >>>> 1)  now we could use this binding language for other queries such as
> >>>> XPath
> >>>> 2)  by separating the grammars we are free to reuse symbols.  In
> >>>> this
> >>>> case the '@' is fair game again, and we could go back to "=" for
> >>>> equality (I think)
> >>
> >> Just an FYI, not so lucky here with the "=",  as the reduce/reduce
> >> conflict comes from keyword arguments not the declarations
> >
> > You do need precedence, you bastard :)
>
> I don't think it will help in this case as precedence is not an issue
>
> the production
>
> qname '=' expression
>
> is really ambiguous in our grammar.  In one case, we say (roughly)
>
> expression :== qname '=' expression
>
> and in another we say
>
> keyword-argument :== qname '=' expression
>
> so the query
>
> function_call ( foo = bar)
>
> can always be interpreted as either
>
> "call the function "function_call" with the keyword argument "foo" set
> to "bar"
>
> or
>
> "call the function "function_call" with a single boolean parameter"
>

Maybe look to the Python grammar for some help.  Although it really depends on 
how the grammar is interpreted.  If there are two levels, a "line" and a 
series of "lines" then it would be easy enough.  In Python speak, a statement 
maps to a "line" and a "suite" maps to a sequence of "lines".


-- 
Jeremy Kloth
Fourthought, Inc.
http://fourthought.com/
http://4suite.org/