At 10:33 AM 6/1/2004, Charles Gregory wrote:
Yesterday I set up a test for a URI (in SA 2.60, procmail, RH9): uri LOC_DIETSITE /[0-9]{3,6}diet\.biz/i describe LOC_DIETSITE diet website score LOC_DIETSITE 2
Today I had a piece of spam with http://www.3434diet.biz in it as plain text. To put it another way, I expect my rule to trigger on *this* message when I get it back.
Correct.
If it does, then once again, I have to ask what 'games' are being played with character sets?
It does, I just tested your rule against a text/plain test email using spamassassin -t <test and it matched just fine. I'm using version 2.63.
If it doesn't trigger, then I have to ask if the URI tests are designed to only trigger
on URI's within an HTML part (ie. 'href=http://....')? This seems like
something far too easily avoided in just the way it happened today.
SA matches uri rules against text/plain sections provided they begin with http:// or www.
The rule matches the following plain texts: http://www.3434diet.biz http://3434diet.biz www.3434diet.biz
but does not match plain text: 3434diet.biz
So if this is a shortcoming, is it addressed in SA 3.0?
It's not a problem, so no need to speculate about 3.0.