Re: [PATCH 2/3] xtables: deactivate intrapositional support

[Patrick McHardy <kaber@trash.net>]

Jozsef Kadlecsik wrote:
> On Mon, 2 Nov 2009, Patrick McHardy wrote:
> 
>> Jan Engelhardt wrote:
>>> In going to fix NF bug #611, "argv" would be needed in
>>> xtables_check_inverse to set "optarg" to the right spot in case of an
>>> intrapositional negation. Adding argv to the parameter list would
>>> cause an API change, which I would like to avoid because it is
>>> planned to be thrown out anyway.
>>>
>>> So deactive intrapositional support now already, but leave the error
>>> message (which has already been there for two releases). Slightly
>>> adjust it, because some users have been wondering whether iptables
>>> or they are at "fault".
>> I'm worried that this is too early, two releases is not particulary
>> long and I'd expect a lot of people haven't noticed the warning yet,
>> especially on headless systems. Looking at bug #611, we've never
>> claimed "-ptcp" would be supported, so this hardly justifies the
>> risk IMO.
> 
> I agree with Patrick - it's no good to break unknown number of firewall 
> setups out there. We did support intrapositional negation and a lot of 
> scripts may use that syntax. It's a too high price for "fixing" what's 
> actually a feature request.

OK, lets leave the warning in for now and reconsider this in a year
or something like that.