On Jul 2, 2016 03:18, Rob Szarka <[email protected]> wrote:
>
> In one test with a different address, I did get this message: 
>
> Cutthrough cancelled by presence of DKIM signing 

Yes; dkim fell into the too-hard basket.
I don't recall why. It might need thinking around the possible cases again.

>
> (
> not so great, since I assume that means spammers that DKIM sign 
> their spam will not get the cutthrough treatment?) 

Exactly.

> But in testing with real sender/recipient/IP data pulled from my logs, I 
> see no mention at all of cutthrough after processing the warn: 
>
> using ACL "check_recipient" 
> processing "warn" 
> check control = cutthrough_delivery 
> warn: condition test succeeded in ACL "check_recipient" 
>
> then we eventually hit the relevant acl block and trigger callout 
> verification: 

> Verifying CUSTOMER 

> callout cache: found domain record for CUSTOMER 
> dbfn_read: key=CUSTOMER/<[email protected]> 
> callout cache: found address record for CUSTOMER/<[email protected]> 
> callout cache: address record is positive 
> ----------- end verify ------------ 
> accept: condition test succeeded in ACL "check_recipient" 
> end of ACL "check_recipient": ACCEPT 
> SMTP>> 250 Accepted 
> 250 Accepted

We had a cache record for the
recipient so do not need to do an actual callout.

> and then we proceed through DATA


Except that on finishing the RCPT acl we should have noted the lack of open 
connection left by a verify callout, and made one for the cutthrough.

This case, of the cached positive callout record, probably is not in the 
testsuite and is buggy. I'll check when I get back from holiday.

In the meantime, exim_dumpdb the callout database and grep for this recipient, 
then editdb to delete the record.
-- 
Cheers,
  Jeremy
-- 
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Reply via email to