spamassassin-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject [Bug 5119] New: 3.1.6 breaks sa-update, if admins set rule scores in site config
Date Fri, 06 Oct 2006 12:51:45 GMT

           Summary: 3.1.6 breaks sa-update, if admins set rule scores in
                    site config
           Product: Spamassassin
           Version: SVN Trunk (Latest Devel Version)
          Platform: Other
        OS/Version: other
            Status: NEW
          Severity: critical
          Priority: P5
         Component: sa-update

(as discussed on the dev list --) it appears that if the site
config changes a setting that refers to a system rule -- for example,

   score ALL_TRUSTED -2.0

in /etc/mail/spamassassin/, then sa-update will fail to perform an
update successfully.

discussion so far:

> Theo Van Dinter wrote:
> > On Thu, Oct 05, 2006 at 03:31:33PM -0400, Daryl C. W. O'Shea wrote:
> >> Sigh.  None of the machines I tested the change on redefined any scores.
> >>  I forgot that this actually causes an error and not info, so I didn't
> >> even think of trying it.
> >>
> >> For now you'll have to disable the score lines while you do sa-update.
> >> Temporarily renaming to is probably the
> >> easiest way to go about it.
> >
> > For such a major issue, I would suggest we fix it ASAP and release 3.1.7.
> > People will likely just stop doing sa-update if they have to go through this.
> I agree (and I feel like an ass -- I should have caught this).
> So far these are the options I've thought of:
> 1) only load pre files:
> I originally didn't do this since I wanted a successful sa-update to
> signify that the rest of the config was good too and you could restart
> spamd with no problems.
> 2) pass in a flag to lint to ignore warnings about score/tflags/describe
> for non-existent rules:
> There could be other things that would cause lint errors, when the
> default (or sa-update'd) rulesets aren't loaded, that I'm not thinking of.
> 3) load the current ruleset in the local state dir (or the default rule
> set if sa-update hasn't been run before):
> We can't do it... we run into the same problem we have now if someone
> has zeroed out a score in their local config and the same rule gets
> removed in an sa-update ruleset.
> 4) revert the change for now
> So, option 1 seems the safest.  Option 2 would be nice, but I'm not
> certain I'm not missing something.  Option 3 isn't an option.  I'd like
> to avoid option 4, but it's better than nothing if 1 or 2 or something
> else doesn't work out shortly.
> Unfortunately, as like when anything goes wrong, I'm out the door in
> less than an hour to go pickup SWMBO.  I'll be away for 36+ hours. :(

#2 seems unreliable to me, since new plugins can define new things that depend
on existing rule definitions.  #3, ugh.   #1 I think should work, though.

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

View raw message