You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@spamassassin.apache.org by do...@apache.org on 2006/09/30 08:52:00 UTC

svn commit: r451554 - in /spamassassin/branches/3.1: lib/Mail/SpamAssassin/HTML.pm rules/20_html_tests.cf rules/30_text_de.cf rules/30_text_fr.cf rules/30_text_nl.cf rules/30_text_pl.cf rules/50_scores.cf

Author: dos
Date: Fri Sep 29 23:52:00 2006
New Revision: 451554

URL: http://svn.apache.org/viewvc?view=rev&rev=451554
Log:
bug 4418: remove no longer useful HTML_WEB_BUGS and HTML_LINK_IMAGE_BUG rules

Modified:
    spamassassin/branches/3.1/lib/Mail/SpamAssassin/HTML.pm
    spamassassin/branches/3.1/rules/20_html_tests.cf
    spamassassin/branches/3.1/rules/30_text_de.cf
    spamassassin/branches/3.1/rules/30_text_fr.cf
    spamassassin/branches/3.1/rules/30_text_nl.cf
    spamassassin/branches/3.1/rules/30_text_pl.cf
    spamassassin/branches/3.1/rules/50_scores.cf

Modified: spamassassin/branches/3.1/lib/Mail/SpamAssassin/HTML.pm
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/lib/Mail/SpamAssassin/HTML.pm?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/lib/Mail/SpamAssassin/HTML.pm (original)
+++ spamassassin/branches/3.1/lib/Mail/SpamAssassin/HTML.pm Fri Sep 29 23:52:00 2006
@@ -657,9 +657,6 @@
   if ($tag eq "img" && exists $self->{inside}{a} && $self->{inside}{a} > 0) {
     $self->{uri}->{$self->{anchor_last}}->{anchor_text}->[-1] .= "<img>\n";
     $self->{anchor}->[-1] .= "<img>\n";
-    if ($self->{anchor_last} =~ /\.(?:pl|cgi|php|asp|jsp|cfm)\b/i) {
-      $self->put_results(anchor_image_bug => 1);
-    }
   }
 
   if ($tag eq "img" && exists $attr->{width} && exists $attr->{height}) {
@@ -682,12 +679,6 @@
     if ($width > 0 && $height > 0) {
       $area = $width * $height;
       $self->{image_area} += $area;
-    }
-    # this is intended to match any width and height if they're specified
-    if (exists $attr->{src} &&
-	$attr->{src} =~ /\.(?:pl|cgi|php|asp|jsp|cfm)\b/i)
-    {
-      $self->put_results(web_bugs => 1);
     }
   }
   if ($tag eq "form" && exists $attr->{action}) {

Modified: spamassassin/branches/3.1/rules/20_html_tests.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/20_html_tests.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/20_html_tests.cf (original)
+++ spamassassin/branches/3.1/rules/20_html_tests.cf Fri Sep 29 23:52:00 2006
@@ -152,9 +152,6 @@
 describe HTML_IMAGE_RATIO_06	HTML has a low ratio of text to image area
 describe HTML_IMAGE_RATIO_08	HTML has a low ratio of text to image area
 
-body HTML_LINK_IMAGE_BUG	eval:html_test('anchor_image_bug')
-describe HTML_LINK_IMAGE_BUG	HTML link plus image plus web bug
-
 body HTML_LINK_PUSH_HERE	eval:html_text_match('anchor', '(?i)(?:push|go|cl[1l]ck)\s*(?:here|this)')
 describe HTML_LINK_PUSH_HERE	HTML link text says "push here" or similar
 
@@ -196,9 +193,6 @@
 describe HTML_ATTR_BAD		HTML has many bad attributes in tags
 body HTML_ATTR_UNIQUE		eval:html_range('attr_unique_bad','0.5','1.0')
 describe HTML_ATTR_UNIQUE	HTML appears to have random attributes in tags
-
-body HTML_WEB_BUGS		eval:html_test('web_bugs')
-describe HTML_WEB_BUGS		Image tag intended to identify you
 
 body HTML_TAG_BALANCE_BODY	eval:html_tag_balance('body', '!= 0')
 describe HTML_TAG_BALANCE_BODY	HTML has unbalanced "body" tags

Modified: spamassassin/branches/3.1/rules/30_text_de.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/30_text_de.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/30_text_de.cf (original)
+++ spamassassin/branches/3.1/rules/30_text_de.cf Fri Sep 29 23:52:00 2006
@@ -363,7 +363,6 @@
 lang de describe HTML_BACKHAIR_8 HTML-Markierungen sollen Wörter tarnen
 lang de describe HTML_ATTR_BAD HTML-Markierungen mit vielen ungültigen Attributen
 lang de describe HTML_ATTR_UNIQUE HTML-Markierungen mit zufälligen Attributen
-lang de describe HTML_WEB_BUGS Bildhyperlink mit Identifizierungsnummer
 lang de describe HTML_TAG_BALANCE_BODY Anzahl "body"-Tags nicht ausgeglichen
 lang de describe HTML_TAG_BALANCE_HEAD Anzahl "head"-Tags nicht ausgeglichen
 lang de describe HTML_TAG_EXIST_MARQUEE HTML-Element "marquee" gefunden

Modified: spamassassin/branches/3.1/rules/30_text_fr.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/30_text_fr.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/30_text_fr.cf (original)
+++ spamassassin/branches/3.1/rules/30_text_fr.cf Fri Sep 29 23:52:00 2006
@@ -237,7 +237,6 @@
 lang fr describe HTML_TAG_BALANCE_HEAD	Le tag de fermeture de "head" HTML est manquant
 lang fr describe HTML_TITLE_EMPTY	Le titre HTML ne contient aucun texte
 lang fr describe HTML_TITLE_UNTITLED	Le titre HTML contient "untitled" (sans-titre)
-lang fr describe HTML_WEB_BUGS		Tag image HTML avec un code ID permettant de vous identifier
 lang fr describe HTML_00_10		Le message est composé de 0 à 10% de HTML
 lang fr describe HTML_10_20		Le message est composé de 10 à 20% de HTML
 lang fr describe HTML_20_30		Le message est composé de 20 à 30% de HTML

Modified: spamassassin/branches/3.1/rules/30_text_nl.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/30_text_nl.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/30_text_nl.cf (original)
+++ spamassassin/branches/3.1/rules/30_text_nl.cf Fri Sep 29 23:52:00 2006
@@ -233,7 +233,6 @@
 lang nl describe HTML_OBFUSCATE_70_80            Bericht bevat 70% tot 80% HTML verdoezeling
 lang nl describe HTML_OBFUSCATE_80_90            Bericht bevat 80% tot 90% HTML verdoezeling
 lang nl describe HTML_OBFUSCATE_90_100           Bericht bevat 90% tot 100% HTML verdoezeling
-lang nl describe HTML_WEB_BUGS                   Plaatje om je te kunnen identificeren
 lang nl describe HTML_TAG_BALANCE_BODY           HTML bevat ongebalanceerde "body" tags
 lang nl describe HTML_TAG_BALANCE_HEAD           HTML bevat ongebalanceerde "head" tags
 lang nl describe HTML_TAG_EXIST_MARQUEE          HTML bevat een "marquee" tag

Modified: spamassassin/branches/3.1/rules/30_text_pl.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/30_text_pl.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/30_text_pl.cf (original)
+++ spamassassin/branches/3.1/rules/30_text_pl.cf Fri Sep 29 23:52:00 2006
@@ -230,7 +230,6 @@
 lang pl describe HTML_TAG_BALANCE_HEAD	HTML posiada niepozamykane znaczniki w "head"
 lang pl describe HTML_TITLE_EMPTY	Tytu³ HTML nie zawiera tekstu
 lang pl describe HTML_TITLE_UNTITLED	Tytu³ HTML zawiera "bez tytu³u" (ang.)
-lang pl describe HTML_WEB_BUGS		Znacznik grafiki maj±cy na celu identyfikacjê osobow±
 lang pl describe HTTP_CTRL_CHARS_HOST	Sekwencje steruj±ce wewn±trz nazwy hosta w URL
 lang pl describe HTTP_ESCAPED_HOST	U¿ywa % wewn±trz nazwy hosta w URL
 lang pl describe HTTP_EXCESSIVE_ESCAPES	Ca³kowicie zbêdne % wewn±trz URL

Modified: spamassassin/branches/3.1/rules/50_scores.cf
URL: http://svn.apache.org/viewvc/spamassassin/branches/3.1/rules/50_scores.cf?view=diff&rev=451554&r1=451553&r2=451554
==============================================================================
--- spamassassin/branches/3.1/rules/50_scores.cf (original)
+++ spamassassin/branches/3.1/rules/50_scores.cf Fri Sep 29 23:52:00 2006
@@ -294,7 +294,6 @@
 score HTML_IMAGE_RATIO_04 0.877 0 1.057 0
 score HTML_IMAGE_RATIO_06 0 0 0.139 0
 score HTML_IMAGE_RATIO_08 0
-score HTML_LINK_IMAGE_BUG 0 0 0.110 0
 score HTML_LINK_OPT_OUT 1.151 0 0.823 0
 score HTML_LINK_PUSH_HERE 1.762 0.402 1.920 0.397
 score HTML_MIME_NO_HTML_TAG 0 0.512 0 1.082
@@ -342,7 +341,6 @@
 score HTML_TITLE_LONG 0.482 1.329 0.458 1.655
 score HTML_TITLE_SUBJ_DIFF 1.547 0.551 1.224 0.266
 score HTML_TITLE_UNTITLED 0 0 0.168 0.514
-score HTML_WEB_BUGS 0 0 0.329 0
 score HTTPS_IP_MISMATCH 1.920 1.920 2.220 2.400
 score HTTP_77 2.242 2.658 2.644 2.346
 score HTTP_CTRL_CHARS_HOST 1.280 1.259 1.480 1.600