https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7735
--- Comment #75 from Wolfgang Breyha <wbre...@gmx.net> ---
I checked out rev 1905124 today, built a local package and tried to test it
with some of my test E-Mails I used before.
I compared the results with my currently running 4.0-rc3 with the patches
Henrik provided before and one patch I sent him privately (4 in sum).
The results from trunk are worse then the results of my 4.0-rc3.
I can't tell what's wrong exactly yet, but many rules are marked as "not hit"
while the patched rc3 correctly hits them. On first sight it looks like some
rules are not rerun if others (with dependencies) got hit after they were run.
After reapplying this:
----------
--- /usr/local/spamassassin/share/perl5/Mail/SpamAssassin/Plugin/Check.pm.orig
2022-11-07 16:27:36.063764282 +0100
+++ /usr/local/spamassassin/share/perl5/Mail/SpamAssassin/Plugin/Check.pm
2022-11-07 16:29:54.476177667 +0100
@@ -297,6 +297,7 @@
my $mr = $pms->{meta_check_ready};
my $mp = $pms->{meta_pending};
my $md = $pms->{conf}->{meta_dependencies};
+ my $mdr = $pms->{conf}->{meta_deprules};
my $mt = $pms->{conf}->{meta_tests};
my $h = $pms->{tests_already_hit};
my $retry;
@@ -318,6 +319,9 @@
dbg("rules-all: ran meta rule $rulename, no hit") if
$would_log_rules_all;
$h->{$rulename} = 0; # mark meta done
}
+ foreach (keys %{$mdr->{$rulename}}) {
+ $mr->{$_} = 1;
+ }
delete $mr->{$rulename};
delete $mp->{$rulename};
# Reiterate all metas again, in case some meta depended on us
----------
... I get better results.
--
You are receiving this mail because:
You are the assignee for the bug.
