Bug Tracker item #2981066, was opened at 2010-04-02 15:05
Message generated for change (Comment added) made by ls999
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1126467&aid=2981066&group_id=250683

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: daemon
Group: v3.9.0
Status: Closed
Resolution: None
Priority: 5
Private: No
Submitted By:  LS999 (ls999)
Assigned to: Stevan Bajic (sbajic)
Summary: email-addr upper|lower case 

Initial Comment:
I'm use now hash_drv for testing dspam & found next problem: 'To' & 'From' mail 
fields can use uppper & lower case symbols (it not impotant for smtp). As 
result DB can contain some info for one email (a...@mydomain.no, 
a...@mydomain.no and more more, more...). 

Result: 
- DB contain all versions of a...@mydomain.no in all variantes in five days in 
my case  (Oo-disk space is not rubber!)
- if mail send to a...@myemail.no i can't retrain DB when send message later 
from a...@myemail.no to s...@mydomain.no. will see:

"Apr  2 13:37:37 main dspam[26224]: Unable to open file for reading: 
/var/local/dspam/data/mydomain.no/aa/aa.sig/4bb554fb268827715621127.sig: No   
such file or directory
Apr  2 13:37:37 main dspam[26224]: Signature retrieval for 
'4bb554fb268827715621127' failed
Apr  2 13:37:37 main dspam[26224]: Unable to find a valid signature. Aborting.
Apr  2 13:37:37 main dspam[26224]: process_message returned error -5.  dropping 
message."

but real signature is present in the 
/var/local/dspam/data/MyEmail.NO/Aa/Aa.sig/ directory

IMHO it some smtp-server depended & realy possibility limits when use another 
processing what described in README (i realy can rewrite this data using 
smtp-server abilites, but it will global for all another processes after dspam. 
and it not bes way for me as example. it MUST work internally)

one-side: all email-addr must processed as is
another-side: subsequent analysis must be case-insensitive

IMHO daemon CAN (&mb must) process incoming email-addr as is, for collect data, 
but database processing (exactly - signatures find) must be 
case-insensitive|config options regulated

PS as add: require to future 'opt-in' 'opt-out' domain based checks when 
compuled with --enable-domain-scale options  (global l level is, user level is, 
where is domain level?)

PSS Oooo and more: custom queries to DB  in conf. I think not only me already 
have & use self user+servises databases. Another DB from U & permanent synk on 
every UID  change not show best (therefore i try hash_drv engine for testing 
now).

TY & Best regards.

----------------------------------------------------------------------

Comment By:  LS999 (ls999)
Date: 2010-05-31 01:35

Message:
I think it about item 2. I'm try using another words.

If i have thousands users (m...@domain.com, m...@domain.com, ...,
my1...@domain.com, ..., m...@domain.com) then i must perform (as you say)
for everyone:
     # dspam_admin change preference m...@domain.no localStore m...@domain.no
     # dspam_admin change preference m...@domain.no localStore m...@domain.no
It manually or script's work (don't forget that I can't use a web-if). 

This way is impossible in my case:
1) manually processing is simple delirium  for thousands users;
2) script processing is possible as fact, but realtime reaction to changes
in mail system (add new e-mail client for example) is impossible (cron) or
very mail system depended. Mail system here not ONLY smtp-server, but ALL
mail system components & services & it configuration. => many
CURRENT-mail-system depended scripts.
2a) this realization type is time critical when any crash of mail system
is occured (because it very system depended & stupid support can't real
make anything => call me... call me...call me... m...@domain.com
...myn+?anym...@domain.com -  i want sleep  sometimes :).

Therefore i say again - this is dspam-server problem:
1) if dspam CAN act as simple filter for add marks then it CAN'T & MUST
don't change ANY another info in incoming mail on self output (it deforms
the information received by following filters) in this mode. The final
decision on mail processing MUST be dspam independent and data
falsification inadmissible here.
1a) this mode must be configurable by server options (not all need such
functioning and i understand it)

PS I don't track this project long time (considering speed of
development). Mb it all already resolved. While I am compelled to use other
decision, but I don't leave dspam ). 
TY 4 all & Personal 4 U sbajic.




----------------------------------------------------------------------

Comment By: Stevan Bajic (sbajic)
Date: 2010-05-29 12:42

Message:
> 2. what about MY1,My2,My...,MyN? manual processing? 
>
I don't understand that. Can you rephrase that question?

----------------------------------------------------------------------

Comment By:  LS999 (ls999)
Date: 2010-05-29 05:44

Message:
Your are right 100%. Not my native language.

1. i'm don't use web (& can't)
2. what about MY1,My2,My...,MyN? manual processing? 

----------------------------------------------------------------------

Comment By: Stevan Bajic (sbajic)
Date: 2010-05-13 22:49

Message:
I have real hard time to understand your writing. Probably English is not
your native language (it is not my native language too)?

Anyway... if I understand you right, then all you want is that DSPAM does
NOT change any original 'To' headers but use one single store for signature
writing/reading. Right?

Well... DSPAM has already that functionality. Assume you have the email
m...@domain.no and want it to be threated the same way as m...@domain.no then
just do this:
- DO NOT enable "Broken case"
- use dspam_admin to change the store for both emails:
  # dspam_admin change preference m...@domain.no localStore m...@domain.no
  # dspam_admin change preference m...@domain.no localStore m...@domain.no

That's it. All mails now arriving to "m...@domain.no" or "m...@domain.no" will
use "m...@domain.no" as their store.

I am now going to close this bug report since DSPAM already offers out of
the box the functionality you asked for.

----------------------------------------------------------------------

Comment By:  LS999 (ls999)
Date: 2010-04-06 02:39

Message:
I don't think what You are right.
It's very situation depended. DSPAM acts as the filter in most cases for
add X-headers only. Therefore in this situation:

data->[DSPAM]->[filter_1]->[filter_2]->...

already "filter_1" (and all following it) don't have original message data
(addr modified by DSPAM). 

Therefore m...@example.com and m...@example.com IS different addresses in
this case (as DSPAM outgoing data). And it is very important for adequate
data processing with following filters.

IMHO must be one more config option for forward original address as is.
Thus if this option OFF, the "broken case" acts as now. When new option is
ON, the "broken case" is used for internal data process only, but we have
the original "To" (and similar) field(s) on DSPAM output.

And i don't see RFC contradiction here, but feature improvement only .

Adequate data processing

Best regards.

----------------------------------------------------------------------

Comment By: Ion-Mihai "IOnut" Tetcu (itetcu)
Date: 2010-04-04 12:40

Message:
Per RFC:
- the right-hand part (domain) is case insensitive, so we could lowercase
it.
- the left-hand part case-sensitiveness is local site policy. Personally I
don't think we should alter anything here, as it's valid (although not
recommended) to have m...@example.com and m...@example.com different
addresses.


----------------------------------------------------------------------

Comment By:  LS999 (ls999)
Date: 2010-04-04 05:01

Message:
Add.

must be
m...@domain.no->[DSPAM m...@domain.no (can be config option regulated)
DSPAM]->m...@domain.no

but not
m...@domain.no->[DSPAM]->m...@domain.no
or
m...@domain.no->[DSPAM]->m...@domain.no


----------------------------------------------------------------------

Comment By:  LS999 (ls999)
Date: 2010-04-03 04:54

Message:
TY 4 "broken case" option - i missed it - &sorry). But "broken case" option
in this case is partial decision of a problem. For example:

- DSAPM off. Here i see all letter cases was perfect keeped:

Return-Path: <m...@rambler.ru>
X-Original-To: m...@domain.no
Delivered-To: m...@domain.no
Received: from mcgi64.rambler.ru (mcgi64.rambler.ru [81.19.67.198])
     by mx.domain.no (MTA) with ESMTP id C5F6818006E
     for <m...@domain.no>; Sat, 3 Apr 2010 02:32:15 +0400 (MSD)
Received: from [78.31.73.87] by mcgi64.rambler.ru with HTTP (mailimap);
Sat, 03 Apr 2010 02:32:14 +0400
From: <m...@rambler.ru>
To: <m...@domain.no>
Subject: dspam off

- DSPAM on, "broken case" option off. Now before DSPAM we have
"m...@domain.no" recipient adress, but after reinjecting mail from DSPAM back
to the mail server (smpt used) we have already "for <m...@domain.no>":

Return-Path: <m...@rambler.ru>
X-Original-To: m...@domain.no
Delivered-To: m...@domain.no
Received: from localhost (localhost.localdomain [127.0.0.1])
     by mx.domain.no (MTA) with SMTP id CFB1A180076
     for <m...@domain.no>; Sat, 3 Apr 2010 02:25:43 +0400 (MSD)
Received: from mxa.rambler.ru (mxa.rambler.ru [81.19.66.231])
     by [mx.domain.no] (MTA) with ESMTP id 8F0CA18006E
     for <m...@domain.no>; Sat, 3 Apr 2010 02:25:41 +0400 (MSD)
...
From: <m...@rambler.ru>
To: <m...@domain.no>
Subject: broken case off

- DSPAM on, "broken case" option on. And now we have "m...@domain.no" in &
"m...@domain.no" out. Predictably:

Return-Path: <m...@rambler.ru>
X-Original-To: m...@domain.no
Delivered-To: m...@domain.no
Received: from localhost (localhost.localdomain [127.0.0.1])
     by m...@domain.no (MTA) with SMTP id 0E095180076
     for <m...@domain.no>; Sat, 3 Apr 2010 02:29:30 +0400 (MSD)
Received: from mcgi20.rambler.ru (mcgi20.rambler.ru [81.19.67.53])
     by main.eletek.ru (MTA) with ESMTP id BB71318006E
     for <m...@domain.no>; Sat, 3 Apr 2010 02:29:27 +0400 (MSD)
Received: from [78.31.73.87] by mcgi20.rambler.ru with HTTP (mailimap);
Sat, 03 Apr 2010 02:29:27 +0400
From: <m...@rambler.ru>
To: <m...@domain.no>
Subject: broken case on

But in ALL cases with DSPAM i don't have ORIGINAL message headers on
output & this restriction for the filters following it.

IMHO DSPAM don't need change any original message headers. 

Best regards.

----------------------------------------------------------------------

Comment By: Stevan Bajic (sbajic)
Date: 2010-04-02 16:32

Message:
Have you tried to use "Broken case"?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1126467&aid=2981066&group_id=250683

------------------------------------------------------------------------------

_______________________________________________
Dspam-devel mailing list
Dspam-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dspam-devel

Reply via email to