postfix nodup daemon configuration

[Gianni Sandigliano]

I am trying to implement nodup postfix daemon on my linux (Fedora) box 
with no success up2now...



Here shortly described what I've done:

0) I've found some documentation on this page (or on similar pages...): 
http://osdir.com/ml/mail.postfix.devel/2007-05/msg00010.html


1) On another internet page I've read nodup is available starting from 
postfix 2.5. My postfix package version is 2.5.5. It should be OK...


2) After reading the synopsis "nodup [generic Postfix daemon options]" 
I've added to my master.cf the following line:


  nodup   unix   -   -   n   -   -   nodup -v

In my understanding it should be OK...

3) In main.cf file I've added the following lines:

   duplicate_filter_limit = 1000

   duplicate_filter_style = pragmatic

4) Documentation reports all mailing list should be "owned". All my 
local mailing lists are owned by root adding a proper line into my 
/etc/aliases file:


mylist1   myuser1,myuser2
owner-mylist1root
In my understanding it should be OK...
5 - Question) I've tried to check availability of a nodup executable on 
my linux box but with no success!!! The only reference to a nodup file 
in my system is /var/spool/postfix/private/nodup after trying to start 
it; it was not available before. Is this right??? I was expecting a 
nodup executable where the other postfix executables are located into 
/usr/libexec/postfix. I've checked also in the latest postfix package 
included into latest version but a nodup executable in not available 
either. Is nodup daemon really available on my system? What/where should 
I check for it?
6 - Question) I was expecting some output either in /var/log/maillog or 
in /var/log/messages marked as postfix/nodup but nothing up2now. Is this 
a confirmation that nodup is not available ???


Where am I wrong ?
Is a more detailed documentation available somewhere ?
Could someone point me in the right direction ?

Tnks! Regards

Gianni

Re: use smtp_bind_address or defer

[Sven Strickroth]

Hi,

Am 26.09.2009 04:37 schrieb Victor Duchovni:
> Not without source code changes:

I thought that...

> Why does the address get de-configured? You should be able to bind
> it provided the interface is administratively up, even if it is not
> working.

I've two DSL lines and use RP-PPPoE. The IP is static but I receive it
using PPP (like DHCP) and if the DSL line is down, the pppX-device
"vanishes" -> that's why no process can bind to the ip.

-- 
Best regards,
 Sven  mailto:s...@clamav.net
 ClamAV, a GPL anti-virus toolkit   http://www.clamav.net

Re: use smtp_bind_address or defer

[Sven Strickroth]

Hi,

Am 26.09.2009 15:04 schrieb Wietse Venema:
>> Is it possible to tell postfix not to send mails if the
>> smtp_bind_address is not available?

> All reasonable TCP/IP stacks will automatically use the right source
> IP address when sending out mail over the dialup link or over the
> local network.

I've two dial-up lines and no other internet-connection. But if the
static-ip connection is down, the TCP/IP-STACK uses the other connection
with a dynamic ip. With the smtp_bind_address I need to force postfix to
use the static ip, instead of using the dynamic ip for sending.

Using the dynamic IP has lots of disadvantages: Higher spam cores,
bounces and so on.

-- 
Best regards,
 Sven  mailto:s...@clamav.net
 ClamAV, a GPL anti-virus toolkit   http://www.clamav.net

Re: use smtp_bind_address or defer

[Wietse Venema]

Sven Strickroth:
> Hi,
> 
> Am 26.09.2009 15:04 schrieb Wietse Venema:
> >> Is it possible to tell postfix not to send mails if the
> >> smtp_bind_address is not available?
> 
> > All reasonable TCP/IP stacks will automatically use the right source
> > IP address when sending out mail over the dialup link or over the
> > local network.
> 
> I've two dial-up lines and no other internet-connection. But if the
> static-ip connection is down, the TCP/IP-STACK uses the other connection
> with a dynamic ip. With the smtp_bind_address I need to force postfix to
> use the static ip, instead of using the dynamic ip for sending.

OK, so you send out IP packets with your static IP source address
out over the dynamic interface.

How do the return packets from remote sites (with destination your
source IP address) know that they have to travel via your dynamic
interface? 

What routing arrangements have you made with your dynamic provider
to make that work?

Wietse

> Using the dynamic IP has lots of disadvantages: Higher spam cores,
> bounces and so on.
> 
> -- 
> Best regards,
>  Sven  mailto:s...@clamav.net
>  ClamAV, a GPL anti-virus toolkit   http://www.clamav.net
> 
> 
> 

Re: Larger Messages Failing to Send

[mouss]

Trekker101 a écrit :
> Hi 
> 
> I am new to postfix and I am experiencing a problem sending larger e- 
> mails through google mail. 
> 
> 
> Small messages go through without any problem. However, when sending 
> mails of more than two or three lines in length I get the following 
> message in the log. 
> 
> 
> Sep 21 20:22:06 server postfix/smtp[29948]: 788861A241E6: 
> to=, relay=smtp.gmail.com[74.125.79.109]:587, 
> delay=271795, delays=270915/0.84/279/600, dsn=4.4.2, status=deferred 
> (host smtp.gmail.com[74.125.79.109] said: 451 4.4.2 Timeout - closing 
> connection. 10sm5431eyz.34 (in reply to end of DATA command)) 
> 
> 
> The server is running Ubuntu 8.04 server LTS x64. It is fully updated 
> and running kernel 2.6.24-24-server. I have used the same setup on Ubuntu
> 9.04 x64 with no problem.
> 
> 

Maybe something in the path is playing bad TCP games. some devices out
there have bugs regarding tcp window scaling, delayed ack, ... and some
firewalls are misconfigured to drop all ICMP traffic (which breaks  PMTU
discovery).

You'll need to sniff traffic... (can you upload large files via ftp/scp
from the postfix machine to an _external_ site?).

Re: postfix nodup daemon configuration

[Noel Jones]

On 9/26/2009 7:24 AM, Gianni Sandigliano wrote:


I am trying to implement nodup postfix daemon on my linux (Fedora) box
with no success up2now...


Not surprising it doesn't work since there is no such software.

The old message you found is proposed documentation for a 
non-existent feature on the developer's list.  It was never 
implemented as described.



Could someone point me in the right direction ?


All postfix features are documented here:
http://www.postfix.org/documentation.html

  -- Noel Jones

Re: use smtp_bind_address or defer

[Wietse Venema]

Sven Strickroth:
> Hi,
> 
> I've a server with multiple ip addresses (all dial-in). But only one
> static ip-address. So if the static ip line is down postfix prints out
> "warning: smtp_connect_addr: bind x.x.x.x: Cannot assign requested
> address" and uses the other (dynamic) ip for sending.
> 
> Is it possible to tell postfix not to send mails if the
> smtp_bind_address is not available?

I never need to use smtp_bind_address when my own mailhost needs
to fall back to dialup (which is fortunately rarely).

All reasonable TCP/IP stacks will automatically use the right source
IP address when sending out mail over the dialup link or over the
local network.

Wietse

Re: use smtp_bind_address or defer

[Sven Strickroth]

Am 26.09.2009 16:18 schrieb Wietse Venema:
> OK, so you send out IP packets with your static IP source address
> out over the dynamic interface.

no, and there's the problem. if the static-ip connection is down,
postfix reports "warning: smtp_connect_addr: bind x.x.x.x: Cannot assign
requested address" and then it binds to the dynamic ip and uses that.

Why not add an option so that if the binding to the smtp_bind_address
fails that postfix refuses to send mail.

-- 
Best regards,
 Sven  mailto:s...@clamav.net
 ClamAV, a GPL anti-virus toolkit   http://www.clamav.net

Re: Semi-OT: recipient delimiter spec/std?

[LuKreme]

On Sep 26, 2009, at 0:08, Barney Desmond   
wrote:



LuKreme: sure, it's easy to describe the generally-expected behaviour,
but I suspect Wietse's point is that you're welcome to write the patch
and make sure nothing breaks. *grin*


Aye, there's the rub.

Larger Messages Failing to Send

[Trekker101]

Hi 

I am new to postfix and I am experiencing a problem sending larger e- 
mails through google mail. 


Small messages go through without any problem. However, when sending 
mails of more than two or three lines in length I get the following 
message in the log. 


Sep 21 20:22:06 server postfix/smtp[29948]: 788861A241E6: 
to=, relay=smtp.gmail.com[74.125.79.109]:587, 
delay=271795, delays=270915/0.84/279/600, dsn=4.4.2, status=deferred 
(host smtp.gmail.com[74.125.79.109] said: 451 4.4.2 Timeout - closing 
connection. 10sm5431eyz.34 (in reply to end of DATA command)) 


The server is running Ubuntu 8.04 server LTS x64. It is fully updated 
and running kernel 2.6.24-24-server. I have used the same setup on Ubuntu
9.04 x64 with no problem.


Thanks very much for any help that you can offer. 

-- 
View this message in context: 
http://www.nabble.com/Larger-Messages-Failing-to-Send-tp25581804p25581804.html
Sent from the Postfix mailing list archive at Nabble.com.

Any way to avoid message duplication after mailing list expansion?

[Gianni Sandigliano]

Is there any way to avoid message duplication after mailing list expansion?


A quick example:


A message is sent:to: al...@mycompany.com

cc: theb...@mycompany.com

Because theboss is included in Alist (and the sender does not know 
it...) two copies of the message are delivered to theboss.


The desired feature would be:
1) avoid delivering to an address in bcc: list when the same address is 
already listed in cc: or to:
2) avoid delivering to an address in cc: list when the same address is 
already listed in to:


Is this possible in postfix? Could someone point me in the right direction?

Tnks,
Gianni

Re: use smtp_bind_address or defer

[Victor Duchovni]

On Sat, Sep 26, 2009 at 05:36:30PM +0200, Sven Strickroth wrote:

> Am 26.09.2009 16:18 schrieb Wietse Venema:
> > OK, so you send out IP packets with your static IP source address
> > out over the dynamic interface.
> 
> no, and there's the problem. if the static-ip connection is down,
> postfix reports "warning: smtp_connect_addr: bind x.x.x.x: Cannot assign
> requested address" and then it binds to the dynamic ip and uses that.
> 
> Why not add an option so that if the binding to the smtp_bind_address
> fails that postfix refuses to send mail.

Add filters to block outbound port 25 on the dynamic interface. Pretend
your ISP has already done that for you (as many do).

-- 
Viktor.

Disclaimer: off-list followups get on-list replies or get ignored.
Please do not ignore the "Reply-To" header.

To unsubscribe from the postfix-users list, visit
http://www.postfix.org/lists.html or click the link below:


If my response solves your problem, the best way to thank me is to not
send an "it worked, thanks" follow-up. If you must respond, please put
"It worked, thanks" in the "Subject" so I can delete these quickly.

Re: Semi-OT: recipient delimiter spec/std?

[mouss]

Erick Calder wrote:
> On Sep 25, 2009, at 3:07 PM, Wietse Venema wrote:
> 
>> Erick Calder:
>>> On Sep 25, 2009, at 2:30 PM, Wietse Venema wrote:
> 
> 
> 
 You can't replace the delimiter. That would break other people's
 transit mail, among many things.
>>>
>>> I'm not sure I understand... perhaps we're speaking of 2 different
>>> things.  what I mean is that when an e-mail first arrives at the
>>> server, before it gets processed, it could be rewritten to use the
>>> known delimiter i.e. a mail arriving for e/j...@arix.com could be
>>> rewritten as e+j...@arix.com (since + is what postfix uses to delimit)
>>
>> On an infrastructure server, can't replace the delimiter. That
>> would break other people's mail.
> 
> oh, I think i get it.  if server A is just relaying to server B, it will
> get e/j...@arix.com and hand e+j...@arix.com to B.  I'm not sure I
> understand how that would break the mail (since e+j...@arix.com) is
> valid and will still be received.  of course, if B is configured to use
> delimiter | then it will break since it will receive e+j...@arix.com
> when it expects e|j...@arix.com - but that is easily fixed since server
> A knows whether it's relaying or delivering to a local account, no?  so
> the rewrite could happen for local deliveries only.
> 
>> On an end-node server, you can use a regexp map in one of the Postfix
>> address rewriting features that already exist.
> 
> I figured there was already some such capability.  I'll need to research
> (for my own purposes)



== virtual_alias_maps:

/^(joe|jim|jane)-(.*)@(example\.net|example\.com)$/$1...@$3

this converts joe-...@example.com to joe+...@example.com

If you don't want to generate the file (and update it when you add
users), you can use mysql or friends.

PS. in your examples, you use '/' and '|'. but those sites that refuse
'+' won't accept these either.

Re: use smtp_bind_address or defer

[Stan Hoeppner]

Victor Duchovni put forth on 9/26/2009 1:36 PM:
> On Sat, Sep 26, 2009 at 05:36:30PM +0200, Sven Strickroth wrote:
> 
>> Am 26.09.2009 16:18 schrieb Wietse Venema:
>>> OK, so you send out IP packets with your static IP source address
>>> out over the dynamic interface.
>> no, and there's the problem. if the static-ip connection is down,
>> postfix reports "warning: smtp_connect_addr: bind x.x.x.x: Cannot assign
>> requested address" and then it binds to the dynamic ip and uses that.
>>
>> Why not add an option so that if the binding to the smtp_bind_address
>> fails that postfix refuses to send mail.
> 
> Add filters to block outbound port 25 on the dynamic interface. Pretend
> your ISP has already done that for you (as many do).

Or...

Might it be possible, via transport_maps or something, to relay via SASL
auth to your dynamic IP DSL provider's submission/outbound SMTP relay
server, either on TCP 25 or 587 (whichever they support), whenever your
static IP interface is down?  This will definitely help to mitigate some
delivery issues you mentioned WRT receivers' anti spam measures.  I ran
this way for almost 4 years on an SBC/AT&T PPPoE dynamic IP DSL
connection with great success, although it was a single line setup.

--
Stan

Re: use smtp_bind_address or defer

[Wietse Venema]

Stan Hoeppner:
> Victor Duchovni put forth on 9/26/2009 1:36 PM:
> > On Sat, Sep 26, 2009 at 05:36:30PM +0200, Sven Strickroth wrote:
> > 
> >> Am 26.09.2009 16:18 schrieb Wietse Venema:
> >>> OK, so you send out IP packets with your static IP source address
> >>> out over the dynamic interface.
> >> no, and there's the problem. if the static-ip connection is down,
> >> postfix reports "warning: smtp_connect_addr: bind x.x.x.x: Cannot assign
> >> requested address" and then it binds to the dynamic ip and uses that.
> >>
> >> Why not add an option so that if the binding to the smtp_bind_address
> >> fails that postfix refuses to send mail.
> > 
> > Add filters to block outbound port 25 on the dynamic interface. Pretend
> > your ISP has already done that for you (as many do).
> 
> Or...
> 
> Might it be possible, via transport_maps or something, to relay via SASL
> auth to your dynamic IP DSL provider's submission/outbound SMTP relay
> server, either on TCP 25 or 587 (whichever they support), whenever your
> static IP interface is down?  This will definitely help to mitigate some
> delivery issues you mentioned WRT receivers' anti spam measures.  I ran
> this way for almost 4 years on an SBC/AT&T PPPoE dynamic IP DSL
> connection with great success, although it was a single line setup.

Postfix supports smtp_fallback_relay (set it to the dynamic ISP);
and Postfix SASL passwords are always tied to specific SMTP servers.

Wietse

Re: Virtual Aliases Usage and Syntax Question

[Erick Calder]

so I tried this:

/^([^._\/-]*)[._\/-](.*)@arix\.com$/ ${1}+$...@arix.com

which seemed to work fine, except for the underscores.  thinking that  
they might be meaningful to the set declaration I tried:


/^e_j...@arix.com$/ e...@arix.com

which also failed... so I'm mystified but I'm happy that now e/j...@arix.com 
, e.j...@arix.com and e-j...@arix.com all land in my box... curiously  
the address isn't actually rewritten (so I actually see the character  
used).  also e...@arix.com and e+j...@arix.com still work.


any light on the underscore would be cool.

On Sep 25, 2009, at 10:10 PM, LuKreme wrote:


On 25-Sep-2009, at 19:49, Erick Calder wrote:
do I understand it correctly that when a message arrives with an  
address like ekkis_j...@arix.com you convert it to ekkis+j...@arix.com 
?


Exactly.

if so, would you share your pcre?  this I would really like to do  
because a lot of sites don't allow + in the address


Which is exactly why I did it.

The virtual_sql.pcre file is basically this:

/^sales_(.*)@example\.com$/ sales+$...@example.com
/^info_(.*)@example\.com$/ info+$...@example.com
/^bob_(.*)@example\.com$/ bob+$...@example.com

I have one line for every email address

The SQL one is the same, but is for all the sql stored users. I  
thought about making this a lookup, but I don't have enough users to  
make it worthwhile, and building the list is pretty fast.


ls -1 /usr/local/virtual/ | grep "@" | \
sed 's/^\([...@]*\)@\([^.]*\)\.\(.*\)$/\/^\1_\(.*\)@\2\\.\3$\/ \1+${1}@ 
\2.\3/' > \

/etc/postfix/virtual_address.pcre

I'd suggest reading the thread "Re: address rewriting with pcre?"  
from March of this year since Wietse referred to my method as  
'unsafe'.


What I have works for me and I have few enough addresses that I can  
easily spot any problems. (for example, no email addresses on my  
system have odd characters that need escaping in the username  
portion, and no domains have multiple .'s)



--
The Salvation Army Band played and the children drunk
lemonade and the morning lasted all day, all day.
And through an open window came like Sinatra in a
younger day pushing the town away

Re: Semi-OT: recipient delimiter spec/std?

[Erick Calder]

On Sep 26, 2009, at 12:30 PM, mouss wrote:




== virtual_alias_maps:

/^(joe|jim|jane)-(.*)@(example\.net|example\.com)$/$1...@$3

this converts joe-...@example.com to joe+...@example.com

If you don't want to generate the file (and update it when you add
users), you can use mysql or friends.

PS. in your examples, you use '/' and '|'. but those sites that refuse
'+' won't accept these either.


yeah, I tried : too but it confuses the headers :) but a dot, dash or  
underscore are all pretty safe

Re: Virtual Aliases Usage and Syntax Question

[Wietse Venema]

Erick Calder:
> so I tried this:
> 
> /^([^._\/-]*)[._\/-](.*)@arix\.com$/ ${1}+$...@arix.com

Unfortunately, this causes Postfix to accept mail for non-existent
recipients (the virtual alias matches a bogus of username portion,
and Postfix discovers only upon delivery that the address is no
good).

To work around that, you'd need to specify explicit usernames in
the pattern, or run a policy daemon that queries a table with
known-good usernames after lopping off the extensions, and that
rejects the bogus ones.

> which seemed to work fine, except for the underscores.  thinking that  
> they might be meaningful to the set declaration I tried:
> 
> /^e_j...@arix.com$/ e...@arix.com
> 
> which also failed... so I'm mystified but I'm happy that now e/j...@arix.com 
> , e.j...@arix.com and e-j...@arix.com all land in my box... curiously  
> the address isn't actually rewritten (so I actually see the character  
> used).  also e...@arix.com and e+j...@arix.com still work.
> 
> any light on the underscore would be cool.

Try using a hexdump tool, and see if you entered a non-ASCII code.

Wietse

Re: Virtual Aliases Usage and Syntax Question

[Erick Calder]

On Sep 26, 2009, at 6:01 PM, Wietse Venema wrote:


Erick Calder:

so I tried this:

/^([^._\/-]*)[._\/-](.*)@arix\.com$/ ${1}+$...@arix.com


Unfortunately, this causes Postfix to accept mail for non-existent
recipients (the virtual alias matches a bogus of username portion,
and Postfix discovers only upon delivery that the address is no
good).


by the "virtual alias" you mean the incoming address? if a message  
were to arrive for noex...@arix.com it would fail to match the above  
regex (since it must have a dot, underscore, slash or dash and it  
doesn't), in which case regular handling occurs (message bounces).  if  
the message were for no-ex...@arix.com then it gets rerouted to n...@arix.com 
 which also doesn't exist and therefor bounces.


I'm clearly not understanding how this works.



To work around that, you'd need to specify explicit usernames in
the pattern, or run a policy daemon that queries a table with
known-good usernames after lopping off the extensions, and that
rejects the bogus ones.


which seemed to work fine, except for the underscores.  thinking that
they might be meaningful to the set declaration I tried:

/^e_j...@arix.com$/ e...@arix.com

which also failed... so I'm mystified but I'm happy that now e/j...@arix.com
, e.j...@arix.com and e-j...@arix.com all land in my box... curiously
the address isn't actually rewritten (so I actually see the character
used).  also e...@arix.com and e+j...@arix.com still work.

any light on the underscore would be cool.


Try using a hexdump tool, and see if you entered a non-ASCII code.

Wietse


well, the actual contents of the file are as follows:

/^e_j...@arix.com$/ e...@arix.com
# the following expression converts e.j...@arix.com, e-j...@arix.com
# e/j...@arix.com to e+j...@arix.com and leaves e...@arix.com alone, as
# well as respecting e+j...@arix.com
/^([^._\/-]*)[._\/-](.*)@(.*)$/ ${1}+$...@${3}

and the hexdump shows the output below.  you can see that the _ is 0x5f

  2f 5e 65 5f 6a 75 6e 6b  40 61 72 69 78 2e 63 6f  |/^e_j...@arix.co 
|
0010  6d 24 2f 20 65 40 61 72  69 78 2e 63 6f 6d 0a 23  |m$/ e...@arix.com 
.#|
0020  20 74 68 65 20 66 6f 6c  6c 6f 77 69 6e 67 20 65  | the  
following e|
0030  78 70 72 65 73 73 69 6f  6e 20 63 6f 6e 76 65 72  |xpression  
conver|
0040  74 73 20 65 2e 6a 75 6e  6b 40 61 72 69 78 2e 63  |ts e.j...@arix.c 
|
0050  6f 6d 2c 20 65 2d 6a 75  6e 6b 40 61 72 69 78 2e  |om, e- 
j...@arix.|
0060  63 6f 6d 0a 23 20 65 2f  6a 75 6e 6b 40 61 72 69  |com.# e/ 
j...@ari|
0070  78 2e 63 6f 6d 20 74 6f  20 65 2b 6a 75 6e 6b 40  |x.com to e 
+junk@|
0080  61 72 69 78 2e 63 6f 6d  20 61 6e 64 20 6c 65 61  |arix.com  
and lea|
0090  76 65 73 20 65 40 61 72  69 78 2e 63 6f 6d 20 61  |ves e...@arix.com 
 a|
00a0  6c 6f 6e 65 2c 20 61 73  0a 23 20 77 65 6c 6c 20  |lone,  
as.# well |
00b0  61 73 20 72 65 73 70 65  63 74 69 6e 67 20 65 2b  |as  
respecting e+|
00c0  6a 75 6e 6b 40 61 72 69  78 2e 63 6f 6d 0a 2f 5e  |j...@arix.com 
./^|
00d0  28 5b 5e 2e 5f 5c 2f 2d  5d 2a 29 5b 2e 5f 5c 2f  |([^._ 
\/-]*)[._\/|
00e0  2d 5d 28 2e 2a 29 40 28  2e 2a 29 24 2f 20 24 7b  |-] 
(.*)@(.*)$/ ${|
00f0  31 7d 2b 24 7b 32 7d 40  24 7b 33 7d 0a   |1}+${2}@ 
${3}.|

00fd

Re: Virtual Aliases Usage and Syntax Question

[Erick Calder]

haha, clearly my regex is borked.  I tried replying and got the mail  
back with:


This Message was undeliverable due to the following reason:

Each of the following recipients was rejected by a remote mail server.
The reasons given by the server are included to help you determine why
each recipient was rejected.

   Recipient: 
   Reason:5.1.1 : Recipient address  
rejected: User unknown in virtual alias table



The following websites may contain more information to assist you:

http://help.rr.com/HMSLogic/rrmail.aspx

http://security.rr.com/help.htm

http://security.rr.com/contact.htm

Please do not reply to this message, as it will go to an unread
mailbox
Reporting-MTA: dns; cdptpa-omta04.mail.rr.com
Arrival-Date: Sun, 27 Sep 2009 06:42:34 +
Received-From-MTA: dns; mail.arix.com (98.149.115.77)

On Sep 26, 2009, at 6:01 PM, Wietse Venema wrote:


Erick Calder:

so I tried this:

/^([^._\/-]*)[._\/-](.*)@arix\.com$/ ${1}+$...@arix.com


Unfortunately, this causes Postfix to accept mail for non-existent
recipients (the virtual alias matches a bogus of username portion,
and Postfix discovers only upon delivery that the address is no
good).

To work around that, you'd need to specify explicit usernames in
the pattern, or run a policy daemon that queries a table with
known-good usernames after lopping off the extensions, and that
rejects the bogus ones.


which seemed to work fine, except for the underscores.  thinking that
they might be meaningful to the set declaration I tried:

/^e_j...@arix.com$/ e...@arix.com

which also failed... so I'm mystified but I'm happy that now e/j...@arix.com
, e.j...@arix.com and e-j...@arix.com all land in my box... curiously
the address isn't actually rewritten (so I actually see the character
used).  also e...@arix.com and e+j...@arix.com still work.

any light on the underscore would be cool.


Try using a hexdump tool, and see if you entered a non-ASCII code.

Wietse