On Thu, Jan 31, 2019 at 05:23:05PM +0100, Florian Westphal wrote: > Phil Sutter <phil@xxxxxx> wrote: > > Aligning arptables-nft output with legacy one is rather important IMO > > since there is no way to check existence of a rule (like with 'iptables > > -C'), so one should expect existence of scripts parsing list/save > > output. Therefore I think it is acceptable to carry quite a bit of extra > > code in MARK and CLASSIFY targets. > > Sucks but I agree. I will review (and likely apply) this series later today. Thanks! And yes, it is rather ugly. Luckily arptables doesn't have many extensions, that's also the reason why I went with the copy'n'paste programming approach in patches 2 and 3 instead of extending the guided option parser in libxtables. Cheers, Phil
