On Sun, Aug 9, 2009 at 15:06, Mark Martinec<[email protected]> wrote: > On Sunday 09 August 2009 15:33:00 Justin Mason wrote: >> > Updated t/data/01_test_rules.cf in trunk, r802274. >> does it fix the bug? > > I believe it does, according to my tests. > >> if so, why? > > It adds the missing rules: > > header USER_IN_DEF_WHITELIST eval:check_from_in_default_whitelist() > > header SPF_PASS eval:check_for_spf_pass() > header SPF_NEUTRAL eval:check_for_spf_neutral() > header SPF_FAIL eval:check_for_spf_fail() > header SPF_SOFTFAIL eval:check_for_spf_softfail() > header SPF_HELO_PASS eval:check_for_spf_helo_pass() > header SPF_HELO_NEUTRAL eval:check_for_spf_helo_neutral() > header SPF_HELO_FAIL eval:check_for_spf_helo_fail() > header SPF_HELO_SOFTFAIL eval:check_for_spf_helo_softfail() > > header USER_IN_SPF_WHITELIST eval:check_for_spf_whitelist_from() > header USER_IN_DEF_SPF_WL eval:check_for_def_spf_whitelist_from() > >> it appears most of the changes >> regarding those rules in that change were to move them around, remove >> descriptions, and add "ifplugin" scope around them... > > Yes, the rest is cosmetics. > >> I'm still mystified why this didn't cause probs for me. > > It didn't cause problems for me either, for two reasons: > - I only tried it initially under my svn directory, which did contain > the full set of rules > - when testing the fetched alpha2, I didn't turn on the > run_long_tests (controlling the whitelist_from.t), > and the run_net_tests (spf.t)
ah. long tests! I'd forgotten all about those when testing the tarball :( time to add a new step to the release protocol. -- --j.
