Re: [PATCH nft] parser_bison: simplify syntax to list all sets in table

[Pablo Neira Ayuso <pablo@netfilter.org>]

On Sat, Jan 11, 2025 at 05:24:45PM +0100, Florian Westphal wrote:
> Pablo Neira Ayuso <pablo@netfilter.org> wrote:
> > Revisit f99ccda252fa ("parser: allow listing sets in one table") to add
> > an alias to list all sets in a given table, eg.
> 
> [..]
> 
> > diff --git a/src/parser_bison.y b/src/parser_bison.y
> > index c8714812532d..ac8de398f8a7 100644
> > --- a/src/parser_bison.y
> > +++ b/src/parser_bison.y
> > @@ -1590,8 +1590,13 @@ list_cmd		:	TABLE		table_spec
> >  			{
> >  				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_SETS, &$2, &@$, NULL);
> >  			}
> > +			|	SETS		table_spec
> > +			{
> > +				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_SETS, &$2, &@$, NULL);
> > +			}
> >  			|	SETS		TABLE	table_spec
> >  			{
> > +				/* alias of previous rule. */
> >  				$$ = cmd_alloc(CMD_LIST, CMD_OBJ_SETS, &$3, &@$, NULL);
> >  			}
> 
> I have concerns wrt. to ever expanding list of aliases, it causes divergence as
> to what is allowed:
> 
> <keyword>
> <keyword> inet
> <keyword> inet foo
> <keyword> table inet foo
> 
> In some cases, all of these work  (4 being aliased to 3).
> In others, e.g. "counters" or "quotas" "inet foo" won't work.
> 
> It would be good to at least be consistent.  What about this?
> 
> It would be good to compact it further, this is all because of
> copypastry +slow divergence on later addendums to command subsets.

Consolidating aliases is indeed better approach, please go ahead
submit this patch.

Thanks.