Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Thomas Eckardt
, if it is the result of a tempfail or not that a recipient is not valid - and the decision that a recipient is not valid, is done as last resort. Thomas Von:Charles Marcus An: ASSP development mailing list Datum: 15.04.2010 23:12 Betreff:Re: [Assp-test] fixes and changes in

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Charles Marcus
On 2010-04-15 4:36 PM, Thomas Eckardt wrote: >> 'NoValidRecipient' is used in every case. > ?? Why? You should *never* reject with a 5XX Invalid Recipient in a TEMP > fail situation. > > This is a really bad decision. > How ever - I never send a 5xx on a unknown recipient. Maybe this is just a

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Thomas Eckardt
How ever - I never send a 5xx on a unknown recipient. Thomas Von:Charles Marcus An: assp-test@lists.sourceforge.net Datum: 15.04.2010 21:49 Betreff:Re: [Assp-test] fixes and changes in 2.0.2_1.0.10 On 2010-04-15 2:23 PM, Thomas Eckardt wrote: > On 2010-04-15 4:19

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Scott MacLean
Looking into this some more, I see a lot of valid messages rejected with SpoofedSender errors, when the sender domain was not a local domain. I suspect this is related to the ForgedHELO problem - both are treating remote domains as local, and incorrectly rejecting valid email as a result. At 0

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Charles Marcus
On 2010-04-15 2:23 PM, Thomas Eckardt wrote: > On 2010-04-15 4:19 AM, Thomas Eckardt wrote: >>> - if VRFY or RCPT TO check is used and the MTA is not available - >>> assp now uses 'LDAPFail' to decide if the adrress is valid or >>> not. >>> >>> 'LDAPFail','LDAP/VRFY failures return >>> false',20,

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Scott MacLean
Oh, I should mention that .10 was the most stable version of ASSP I have seen on my server in a long time. It stayed up for over 6 hours before I shut it down to revert to .08. Of course, it was rejecting most of the inbound mail during that time, but it didn't crash (regex out of memory) once

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Scott MacLean
I installed the two LIB files and installed .10 this morning. My clients immediately started screaming, because almost all of the mail people were sending them was bouncing. I had a look at the logs, and to my horror, almost every connection was being reported as a forged HELO, even though the

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Thomas Eckardt
>Does this result in a TEMPfail (ie, 4XX), or permanent fail (ie 5XX)? 'NoValidRecipient' is used in every case. Thomas Von:Charles Marcus An: assp-test@lists.sourceforge.net Datum: 15.04.2010 17:06 Betreff: Re: [Assp-test] fixes and changes in 2.0.2_1.0.10 On

Re: [Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Charles Marcus
On 2010-04-15 4:19 AM, Thomas Eckardt wrote: > - if VRFY or RCPT TO check is used and the MTA is not available - assp now > uses 'LDAPFail' to decide if the adrress is valid or not. > > 'LDAPFail','LDAP/VRFY failures return > false',20,\&checkbox,'','(.*)',undef,'If checked, when an error occurs

[Assp-test] fixes and changes in 2.0.2_1.0.10

2010-04-15 Thread Thomas Eckardt
Hi all, fixed in 2.0.2_1.0.10: - if noSPFRe is used an error 'Unmatched ( in regex; marked by <-- HERE in m/( <-- HERE (?-xism:(?is:.))/ at assp.pl line 14653. ' cause the thread (Worker) to die. changed: - if VRFY or RCPT TO check is used and the MTA is not available - assp now uses '