At 8/9/03 04:46 PM , Robert Menschel wrote:
Proposed rule:
body      RM_bl_ConsWord8  /\b[bcdfghjklmnpqrstvwxz]{8}\b/
describe  RM_bl_ConsWord8  Body contains "word" of 8 consec consonants
score     RM_bl_ConsWord8  0.01

Problem: SA appears to be matching this rule against emails where I think
it shouldn't.  One example is attached. I have tried to find the matching
location for this regex in the attached email, and cannot. But SA thinks
it matched.

I also tried to find where it was matching, using the following method:


[EMAIL PROTECTED] kmactane$ perl -e '$ret = open FILE, "./newtest.lst"; print "ret: $ret\n"; $count=0; while (<FILE>) { if ($_ =~ /[bcdfghjklmnpqrstvwxyz]{5,} /) { print "line $count: $&\n"; } $count++; }'

For clarity, this is how that code looks when pretty-printed:

$ret = open FILE, "./newtest.lst";
print "ret: $ret\n";
$count=0;
while (<FILE>) {
        if ($_ =~ /[bcdfghjklmnpqrstvwxyz]{5,} /) {
                print "line $count: $&\n";
        }
        $count++;
}

The line printing the value of $ret was for debugging purposes; my initial tests without quoting the filename failed the open() call.

The regex used here is, obviously, much more permissive than yours; it omits the word boundaries and matches 5-or-more characters, instead of 8-and-only-8. Here's the output it gave me:

ret: 1
line 4: bsmtp
line 36: ffffff
line 39: ffffff
line 41: ffffff
line 43: ffffff
line 44: ffffff
line 96: ffffff

The "bsmtp" is from your first Received: header ("with local-bsmtp"); the others are obviously HTML color specifications. There's nothing past 6 characters. I have no idea why SA is triggering on this rule with this input.

                                                --Kai MacTane
----------------------------------------------------------------------
"I looked Death in the face last night,/I saw him in a mirror,
 And he simply smiled,/He told me not to worry:
 He told me just to take my time."
                                                --Oingo Boingo,
                                                 "We Close Our Eyes"



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to