On Wed, Jul 10, 2019 at 09:40:31AM -0700, John Hardin wrote: > On Wed, 10 Jul 2019, Henrik K wrote: > > >On Wed, Jul 10, 2019 at 06:04:39PM +0200, Giovanni Bechis wrote: > >>On 7/10/19 5:54 PM, Mark London wrote: > >>>I'm sorry for not using bugzilla, but the new rule for PDS_NO_HELO_DNS is > >>>mostly hittng real emails at my site 1168 real emails versus 219 spam mls. > >>> Luckily, the score is not high, to be making any difference. FWIW. - > >>>Mark > >>> > >>ruleqa has the same opinion: > >>https://ruleqa.spamassassin.org/?daterev=20190709-r1862790-n&rule=PDS_NO_HELO_DNS&srcpath=&g=Change > >> Giovanni > > > >PDS_NO_HELO_DNS should not even run for that nightly masscheck, since it's > >net rule. Not sure why SA doesn't currently ignore net flagged rules > >completely when net is disabled. Probably needs fixing.. > > What if you have a meta that depends on a NET rule and a non-NET rule?
I think we have larger issues to think about, since even this will always hit: meta FOO !NONEXISTINGRULE Reasonable solution would be ignoring the meta completely, if all rule dependencies are not met. Of course this could break lots of metas between sandboxes etc, it should be implemented so everyone is aware of it. Perhaps a new tflag require_meta_deps or similar could be implemented to maintain legacy compatibility..
