Filtering problem

Glenn Steen glenn.steen at gmail.com
Thu Nov 16 11:51:34 GMT 2006


On 16/11/06, Martin Hepworth <martinh at solidstatelogic.com> wrote:
> Howard Robinson wrote:
> > Dear list
> > This (coincidentally) is another filter query. I have looked at EXAMPLES & README and it appears to confirm what I have set.
> >
> > I am having a problem trying to stop MailScanner filtering out HTML & script
> > I am sending myself an email with html script in it for testing.
> > In my MailScanner.conf I have included
> >
> > Allow Script Tags = %rules-dir%/script.rules
> > Non Spam Actions = %rules-dir%/striphtml.rules
> > note %rules-dir% = /etc/MailScanner/rules
> >
> > in /etc/MailScanner/rules/script.rules I have
> >
> > To:     hrobinson at harper-adams.ac.uk    yes
> > To:     howard at harper-adams.ac.uk       yes
Note that the above address....

> > # I tried FromOrTo: on both these with no different result
> > FromOrTo:       default no
> >
> > in /etc/MailScanner/rules/striphtml.rules I have
> >
> > To:     hrobinson at harper-adams.ac.uk deliver
> > To:     howard at harper-adams.ac.uk deliver
... as well as the one above here...

> > # I tried FromOrTo: on both these with no different result
> > FromOrTo:       default striphtml deliver
> > (spaces are tabs by the way)
> >
> > In Maillog the following lines relate to the email, Validipdaddress & Validrelay changed from real one.
> >
> > Nov 15 16:25:17 blackhole2 MailScanner[29444]: Message kAFGOisP029475 from VALIDIPADDRESS (hrobinson at harper-adams.ac.uk) to harper-adams.ac.uk is not spam, SpamAssassin (score=-2.53, required 5, autolearn=not spam, ALL_TRUSTED -3.30, BAYES_00 -0.05, HTML_80_90 0.15, HTML_MESSAGE 0.00, HTML_NONELEMENT_00_10 0.00, NO_REAL_NAME 0.01, SUBJ_ALL_CAPS 0.67)
> > A few lines later
> > Nov 15 16:27:22 blackhole2 MailScanner[29444]: Content Checks: Detected and will convert HTML message to plain text in kAFGOisP029475
> > A few lines later
> > Nov 15 16:28:30 blackhole2 sendmail[30284]: kAFGOisP029475: to=howard at gw.harper-adams.ac.uk, delay=00:03:46, xdelay=00:00:00, mailer=esmtp, pri=125149, relay=VALIDRELAY. [VALIDIPADDRESS], dsn=2.0.0, stat=Sent (Ok)

... don't match with the recipient (envelope) address here. So I'm
guessing this is why you are seeing this behavior (just another case
of it doing exactly what you tell it to do...:-)


> > On Mailwatch the details of the message has the line  kAFGOisP029475
> > Non Spam Actions = %rules-dir%/striphtml.rules
> >
> > It must have html in as it's being picked up in Maillog.
> > Mailscanner is using the striphtml.rules as mailwatch says it is and that's confirmed in maillog but so it should follow the rule!
> > I can't help feeling I am missing something that will be blatently obvious but so far it remains a mistery.
> > Am I right in thinking that the filters will be carried out before it looks at the aliases list. Logically I can't see it working otherwise.
> > Any ideas on what else could be amiss.
> > Thanks
> >
> >
>
> Howard
>
> I think the default action action on striphtml.rules is invalid. See the
> comments above the "Non Spam Actions" for valid actions
>
Haven't looked, so this might be true too:-).
But first see to it that  howard at harper-adams.ac.uk and
howard at gw.harper-adams.ac.uk match up better... Then look at Martins
suggestion;-).

-- 
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se


More information about the MailScanner mailing list