svn commit: r532038 - in /spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen: generate-new-scores lock-scores masses-Makefile.patch

[do...@apache.org]

Author: dos
Date: Tue Apr 24 11:40:59 2007
New Revision: 532038

URL: http://svn.apache.org/viewvc?view=rev&rev=532038
Log:
lock scores to zero for no longer active rules that appear in mass-check logs that were based on older active.list logs (like what happens mid-week when rules that were active in the net check are no longer active)

Modified:
    spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/generate-new-scores
    spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/lock-scores
    spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/masses-Makefile.patch

Modified: spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/generate-new-scores
URL: http://svn.apache.org/viewvc/spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/generate-new-scores?view=diff&rev=532038&r1=532037&r2=532038
==============================================================================
--- spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/generate-new-scores (original)
+++ spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/generate-new-scores Tue Apr 24 11:40:59 2007
@@ -94,6 +94,7 @@
 
 svn co -r $REVISION https://svn.apache.org/repos/asf/spamassassin/trunk trunk-new-rules-set$SCORESET
 svn co https://svn.apache.org/repos/asf/spamassassin/tags/spamassassin_release_3_2_0_rc_2/rules trunk-new-rules-set$SCORESET/rules-base
+svn co https://svn.apache.org/repos/asf/spamassassin/trunk/rules trunk-new-rules-set$SCORESET/rules-current
 
 cd trunk-new-rules-set$SCORESET
 svn up -r $REVISION rulesrc/

Modified: spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/lock-scores
URL: http://svn.apache.org/viewvc/spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/lock-scores?view=diff&rev=532038&r1=532037&r2=532038
==============================================================================
--- spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/lock-scores (original)
+++ spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/lock-scores Tue Apr 24 11:40:59 2007
@@ -1,11 +1,23 @@
 #!/usr/bin/perl
 
+# locks the score ranges for the base release rules to their original scores
+# from 50_scores.cf
+
+# if called with a 1 parameter new rules that aren't in the most current copy
+# of the active.list file will be locked to zero so that the GA can ignore
+# rules that aren't in the most current update (this is used for zeroing rules
+# found in the weekly net checks that are no longer in the nightly non-net
+# checks which 6 of 7 updates a week are based on)
+
 use strict;
 use warnings;
 
 my $scoreset = 0; # default
 
 my %rulescores;
+my %currently_active;
+
+my $only_currently_active_rules = (defined $ARGV[0] && $ARGV[0] == 1 ? 1 : 0);
 
 open(CONFIG, "config") or die "Cannot open config file: $!";
 while (<CONFIG>) {
@@ -32,6 +44,13 @@
 }
 close ORIG;
 
+if ($only_currently_active_rules) {
+  open(ACTIVE, "../rules-current/active.list") or die "Cannot open rules-current/active.list: $!";
+  while(<ACTIVE>) {
+    $currently_active{$1} = undef if (/^(?!#)(\S+)$/);
+  }
+  close ACTIVE;
+}
 
 open(ORIG, "tmp/ranges.data") or die "Cannot open original range.data file: $!";
 open(NEW, ">tmp/ranges.data-new") or die "Cannot open range.data-new file: $!";
@@ -40,7 +59,15 @@
     if (defined $rulescores{$1}) {
       print NEW "$rulescores{$1} $rulescores{$1} 0 $1\n";
     } else {
-      print NEW $_;
+      if ($only_currently_active_rules) {
+        if (exists $currently_active{$1}) {
+          print NEW $_;
+        } else {
+          print NEW "0 0 0 $1\n";
+        }
+      } else {
+        print NEW $_;
+      }
     }
   } else {
     print NEW $_;

Modified: spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/masses-Makefile.patch
URL: http://svn.apache.org/viewvc/spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/masses-Makefile.patch?view=diff&rev=532038&r1=532037&r2=532038
==============================================================================
--- spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/masses-Makefile.patch (original)
+++ spamassassin/rules/trunk/sandbox/dos/new-rule-score-gen/masses-Makefile.patch Tue Apr 24 11:40:59 2007
@@ -8,7 +8,7 @@
  tmp/ranges.data: tmp/.created freqs score-ranges-from-freqs
 +	perl add-hitless-active-to-freqs
  	perl score-ranges-from-freqs $(RULES) $(SCORESET) < freqs
-+	perl lock-scores
++	perl lock-scores 1
 +	mv tmp/ranges.data-new tmp/ranges.data
  
  freqs: spam.log ham.log