Re: SPARQL grammar 2004-10-12 lex/yacc conflicts feedback (was Re: SPARQL 2004-10-12 syntax and grammar issues)

On Mon, 8 Nov 2004 14:26:04 -0000, "Seaborne, Andy" <andy.seaborne@hp.com> wrote:

> Ah! A somewhat different timescale.  Oct 12th this year.
> 
> The first one is resolved by having {braces} as the grouping symbols.
> It is in rq23/ as noted at:
> 
> Revision 1.126  2004/11/04 15:20:29  aseaborne
> ...
> + triple pattern grouping is now {braces} (except for [] optionals)
> ...

Thanks, that should fix it.  Closes this issue for me.

> Can't do much about the second - it's because of unary and binary '+'
> and '-'.

Yes, I said as much in the
  http://lists.w3.org/Archives/Public/public-rdf-dawg/2004OctDec/0170.html
message

> -------------------
> 
> The important thing for the SPARQL grammar is to communicate the
> language - it is not there to be exactly as an implementation would want
> for yacc.  The implementer is going tohave to do some work to turn the
> grammar in the document into yacc/javacc/antlr/hand coded parser/....

Yes, that's the work I am reporting on.  I provided a bunch of
suggestions only based on lex/yacc work, some changes are due to that
implementation, some are more general such as non-terminals with no
purpose I can see.  All of those I called as suggestions and not
necessarily bugs.

This shift/reduce problem I had found I wanted confirmation it was a
grammar ambiguity problem, which you've done, and not an
implementation one.

<snip/>

Dave

Received on Monday, 8 November 2004 14:42:42 UTC