SA Bugzilla – Bug 6675
HTML_TITLE_SUBJ_DIFF hits subject "<3"
Last modified: 2019-06-19 15:01:55 UTC
I just got a false positive on an email largely because its entire subject was "<3" (ascii heart), which hits HTML_TITLE_SUBJ_DIFF, which has a score of 2.2. Doesn't seem like it should have hit that rule. In 20_html_tests.cf: meta HTML_TITLE_SUBJ_DIFF __HTML_TITLE_SUBJ_DIFF && !__MIME_ATTACHMENT body __HTML_TITLE_SUBJ_DIFF eval:html_title_subject_ratio('3.5') No description. html_title_subject_ratio seems to come from the HTMLEval plugin, lacking a man page ( Mail::SpamAssassin::Plugin::HTMLEval ). Looks like it's converting "<3" to html, coming up with a result of an empty string by ignoring the fact that there is no ">", and deciding that because the length ratio of an empty string to the original 2 character string is bad. http://ruleqa.spamassassin.org/?daterev=20111008-r1180336-n&rule=HTML_TITLE_SUBJ_DIFF&srcpath=&g=Change The ham to spam ratio of this rule is terrible. Why is it in the default rule set, and why does it have such a high score?
This rule currently has a score of 1.15 so it's possible that this has been corrected in the past 6 years.
Closing old bug. Works fine here.