Mailing List Archive

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From jm@jmason.org 2004-03-14 20:12 -------
two things:

1. I think check_tick() will need to be called more often -- best thing to do is
check this in and we can worry about that afterwards.

2. I notice that the compiled perl fns use priority in their name. Does this
work if the pris are < 0?




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From jm@jmason.org 2004-03-14 20:13 -------
oh yeah, otherwise +1, looks good!



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From parkerm@pobox.com 2004-03-14 21:51 -------
Subject: Re: [review] RFE: allow test order to be controlling using a test 'priority' setting

On Sun, Mar 14, 2004 at 08:12:49PM -0800, bugzilla-daemon@bugzilla.spamassassin.org wrote:
>
> 2. I notice that the compiled perl fns use priority in their name. Does this
> work if the pris are < 0?
>

Good question. It works for sure, but could be wierd. You certainly
can't create and call a subroutine named _head_tests_-100. I'll fix
it to it changes the - to neg or something similar.

Michael





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From matt@nightrealms.com 2004-03-15 00:49 -------
Asside from the negative-priority thing, as noted by Justin, +1 from me.



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From duncf@debian.org 2004-03-15 07:43 -------
How difficult would it be to implement short-circuiting too?

It seems if we are changing the way we order tests, it wouldn't be too hard to
run all negative tests first and then positive tests and break out when score >
threshold (if the option is enabled...).

We'd have to skip auto-learning and possibly AWL, but I think that it'd be worth
it. (In my case, for example, auto-learning is useless -- I train on all my
incoming mail.)

More info: bug 62, bug 304



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From parkerm@pobox.com 2004-03-15 10:15 -------
Subject: Re: [review] RFE: allow test order to be controlling using a test 'priority' setting

Well, I found a problem with the plugin interaction. Specifically
this section in URIDNSBL::parsed_metadata:

# only hit DNSBLs for which score != 0
foreach my $rulename (keys %{$scanner->{conf}->{uridnsbls}}) {
# trim_rules() will remove the head_evals entry for this
next unless ($scanner->{conf}->{head_evals}->{$rulename});
# set score to 0 should also block the rule
next unless ($scanner->{conf}->{scores}->{$rulename});
$scanstate->{active_rules}->{$rulename} = 1;
}

The check on the head_evals will fail since it needs a priority.

I think it would be better to possibly have a method in Conf.pm that
can be used for this sort of thing, pass in a test type, a rulename
and an optional priority. If no priority is given it will loop over
all of the available priorities trying to find the given rulename.

Michael





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912





------- Additional Comments From jm@jmason.org 2004-03-15 10:49 -------
Subject: Re: [review] RFE: allow test order to be controlling using a test 'priority' setting

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


>The check on the head_evals will fail since it needs a priority.
>
>I think it would be better to possibly have a method in Conf.pm that
>can be used for this sort of thing, pass in a test type, a rulename
>and an optional priority. If no priority is given it will loop over
>all of the available priorities trying to find the given rulename.

+1 -- some kind of API is better than external code crawling through
what should be a private data structure.

Even an "is rule active" API in Conf would do fine.

- --j.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Exmh CVS

iD8DBQFAVfrDQTcbUG5Y7woRAlolAJ97nnrBG1x0U8N2fwQ9s1UPdZnDegCeLgkY
EKEjg7rwgB3kLwEzOWvWBcU=
=oeH5
-----END PGP SIGNATURE-----





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912

parkerm@pobox.com changed:

What |Removed |Added
----------------------------------------------------------------------------
Attachment #1841 is|0 |1
obsolete| |



------- Additional Comments From parkerm@pobox.com 2004-03-16 11:22 -------
Created an attachment (id=1852)
--> (http://bugzilla.spamassassin.org/attachment.cgi?id=1852&action=view)
Patch File

Fix - in subroutine names. Also add some constants for the level in which DNS
RBL queries will be harvested and meta tests will run.



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912

parkerm@pobox.com changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED



------- Additional Comments From parkerm@pobox.com 2004-03-17 09:33 -------
Committed revision 9571.




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

http://bugzilla.spamassassin.org/show_bug.cgi?id=2912

jm@jmason.org changed:

What |Removed |Added
----------------------------------------------------------------------------
OtherBugsDependingO| |3208
nThis| |





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.