Re: [nft PATCH 3/4] xt: Rewrite unsupported compat expression dumping

[Florian Westphal <fw@strlen.de>]

Phil Sutter <phil@nwl.cc> wrote:
> On Thu, Nov 24, 2022 at 06:04:54PM +0100, Florian Westphal wrote:
> > Phil Sutter <phil@nwl.cc> wrote:
> > > diff --git a/src/scanner.l b/src/scanner.l
> > > index 1371cd044b65a..4edd729c80dab 100644
> > > --- a/src/scanner.l
> > > +++ b/src/scanner.l
> > > @@ -214,6 +214,7 @@ addrstring	({macaddr}|{ip4addr}|{ip6addr})
> > >  %s SCANSTATE_TCP
> > >  %s SCANSTATE_TYPE
> > >  %s SCANSTATE_VLAN
> > > +%s SCANSTATE_XT
> > >  %s SCANSTATE_CMD_EXPORT
> > >  %s SCANSTATE_CMD_IMPORT
> > >  %s SCANSTATE_CMD_LIST
> > > @@ -799,6 +800,8 @@ addrstring	({macaddr}|{ip4addr}|{ip6addr})
> > >  
> > >  "secmark"		{ scanner_push_start_cond(yyscanner, SCANSTATE_SECMARK); return SECMARK; }
> > >  
> > > +"xt"			{ scanner_push_start_cond(yyscanner, SCANSTATE_XT); return XT; }
> > 
> > Why is there a new scanner state?  It has no tokens, so it doesn't do
> > anything.
> > 
> > Perhaps a leftover?  Or did you plan to make match/target/watcher scoped
> > tokens?
> 
> We want to accept two arbitrary strings after "xt" and just complain
> about the whole thing. Without the scope, something like "xt target
> exthdr" will make the parser complain about the unexpected keyword.

I see.  FWIW I'm fine with the series.