Re: BUG: sieve does not set seen-Flag

[Jakobus Schürz]

Am 07.12.18 um 08:10 schrieb Noel Butler:
>
> On 07/12/2018 16:44, Aki Tuomi wrote:
>
>>
>> On 6.12.2018 6.54, Noel Butler wrote:
>>>
>>> On 06/12/2018 07:29, Jakobus Schürz wrote:
>>>
>>>  that all and every Flag is set, except \Seen... I tried to
>>> figure out, whats happening here...
>>>
>>> Paste what your sieve file contains now (no, I'm not going back over
>>> this thread - its becoming as long as war and peace, and you may
>>> have changed it since then)
>>>
>>>
>>> Please understand me right... It is nice for you, if dovecot
>>> does, what you expect... It is nice. But here it does not work
>>> correctly. dovecot makes a big mistake. And i try to give as
>>> much information, as i'm possible to give...
>>>
>>> I doubt its dovecot, since no one else has reported this problem
>>> that I can see - without going back to find the start of the thread.
>>>
>>> my dovecot does not copy the Seen-flag. It ignores it. But WHY.
>>> Fucking WHY???
>>>
>>> Mind your tongue if you want help here, despite frustrations (man I
>>> must be getting old and mellowing), no one here has to do shit for
>>> you, the fact it works for everyone else, indicates there is a
>>> problem with your configuration and yours alone - somewhere, and
>>> because you're the only one experiencing this, it may be harder to
>>> trace the origin of.
>>>
>>>
>>> -- 
>>>
>>> Kind Regards,
>>>
>>> Noel Butler
>>>
>>> This Email, including any attachments, may contain legally
>>> privileged information, therefore remains confidential and subject
>>> to copyright protected under international law. You may not
>>> disseminate, discuss, or reveal, any part, to anyone, without the
>>> authors express written authority to do so. If you are not the
>>> intended recipient, please notify the sender then delete all copies
>>> of this message including attachments, immediately. Confidentiality,
>>> copyright, and legal privilege are not waived or lost by reason of
>>> the mistaken delivery of this message. Only PDF
>>>  and ODF
>>>  documents accepted,
>>> please do not send proprietary formatted documents
>>>
>>
>> I finally had some time to try this out, and wasn't able to reproduce
>> the problem with 2.3.4 and 0.5.4.
>>
>> I tried both Sieve and IMAPSieve, but I wonder if this is something
>> particular in your environment or settings, so I have to ask you to
>> post your `doveconf -n` once more.
>>
>> Aki
>>
>
> Did he ever tell us what this is on, I saw stretch somewhere in thread
> so I gather its debian, but is it on real hardware, or rpi, has he
> tried using the source, who knows what happens when distros butcher
> things up into 70 different sub packages :)
>
Hi. sorry for my anger a few days ago...
And Aki... i reviewd the thread... you never asked me before for dovecot
-n. It is the first time.
I also wrote, that i use the packages from the dovecot-repo for debian.
So i was thinking, it is clear, which version i use. It's not
debian-repo, it is dovecot-repo, which i got from the dovecot-release-notes.

My anger was, i wrote details, logmessages, behaviour... again and
again... and i got the every similar message "for me it works"... and "i
dont want to read the whole thread"... so i was angry, how often again i
should post the same again...

Great sorry for my tongue.

My hardware is a rented virtual server from a cloudprovider in germany,
where i have full permissions on it. The filesystem is ext4.
I atteched my dovecot -n

The mails all are stored in maildir in /var/mail.
There is an extra dir /var/lib/dovecot/db... where index and control are
in separate directories.

The owner and group from all of this directories are all vmail:vmail
The permissions are 0700 (only vmail is allowed to read/write/execute in
this directories)

And again... it's independed from MUA: When i move a message to another
folder, the message in the new folder is shown as recent and unseen. I
posted - i think - 3 times the logs from the copy/expunge-task, where
the "flags()" is empty on copy, but expunge from the original folder
shows the correct flags. If you want... i can do it a 4th time ;-)

I also asked for a possibility (which i do not know) to turn up the
debug-level more than i have now, to see, what happens, that i can post
it. maybe it is a permission-problem. I don't know.
Maybe there is a sieve-script working, which i don't know, which sets a
message to unseen and recent, if it arrives to a folder (i deactivated
all the sieve-scripts, but the behaviour was the same wrong).
There are two scripts for rspamd and spamassassin, which learn spam or
ham, depending a message is moved to or from Junk. I also commented the
lines out in the sievescript... no change. Every message which is new in
a Folder is set to recent and unseen.

Best regards


Jakob

# 2.3.4 (0ecbaf23d): /etc/dovecot/dovecot.conf
# Pigeonhole version 0.5.4 (60b0f48d)
# OS: Linux 4.9.0-8-amd64 x8

Re: BUG: sieve does not set seen-Flag

[Aki Tuomi]

On 7.12.2018 9.10, Noel Butler wrote:
>
> On 07/12/2018 16:44, Aki Tuomi wrote:
>
>>
>> On 6.12.2018 6.54, Noel Butler wrote:
>>>
>>> On 06/12/2018 07:29, Jakobus Schürz wrote:
>>>
>>>  that all and every Flag is set, except \Seen... I tried to
>>> figure out, whats happening here...
>>>
>>> Paste what your sieve file contains now (no, I'm not going back over
>>> this thread - its becoming as long as war and peace, and you may
>>> have changed it since then)
>>>
>>>
>>> Please understand me right... It is nice for you, if dovecot
>>> does, what you expect... It is nice. But here it does not work
>>> correctly. dovecot makes a big mistake. And i try to give as
>>> much information, as i'm possible to give...
>>>
>>> I doubt its dovecot, since no one else has reported this problem
>>> that I can see - without going back to find the start of the thread.
>>>
>>> my dovecot does not copy the Seen-flag. It ignores it. But WHY.
>>> Fucking WHY???
>>>
>>> Mind your tongue if you want help here, despite frustrations (man I
>>> must be getting old and mellowing), no one here has to do shit for
>>> you, the fact it works for everyone else, indicates there is a
>>> problem with your configuration and yours alone - somewhere, and
>>> because you're the only one experiencing this, it may be harder to
>>> trace the origin of.
>>>
>>>
>>> -- 
>>>
>>> Kind Regards,
>>>
>>> Noel Butler
>>>
>>> This Email, including any attachments, may contain legally
>>> privileged information, therefore remains confidential and subject
>>> to copyright protected under international law. You may not
>>> disseminate, discuss, or reveal, any part, to anyone, without the
>>> authors express written authority to do so. If you are not the
>>> intended recipient, please notify the sender then delete all copies
>>> of this message including attachments, immediately. Confidentiality,
>>> copyright, and legal privilege are not waived or lost by reason of
>>> the mistaken delivery of this message. Only PDF
>>>  and ODF
>>>  documents accepted,
>>> please do not send proprietary formatted documents
>>>
>>
>> I finally had some time to try this out, and wasn't able to reproduce
>> the problem with 2.3.4 and 0.5.4.
>>
>> I tried both Sieve and IMAPSieve, but I wonder if this is something
>> particular in your environment or settings, so I have to ask you to
>> post your `doveconf -n` once more.
>>
>> Aki
>>
>
> Did he ever tell us what this is on, I saw stretch somewhere in thread
> so I gather its debian, but is it on real hardware, or rpi, has he
> tried using the source, who knows what happens when distros butcher
> things up into 70 different sub packages :)
>
>
> -- 
>
> Kind Regards,
>
> Noel Butler
>
>

I am rather confident that hardware does not cause this problem.

Aki

Re: BUG: sieve does not set seen-Flag

[Noel Butler]

On 07/12/2018 16:44, Aki Tuomi wrote:

> On 6.12.2018 6.54, Noel Butler wrote: 
> 
> On 06/12/2018 07:29, Jakobus Schürz wrote: 
> 
> that all and every Flag is set, except \Seen... I tried to figure out, whats 
> happening here... 
> 
> Paste what your sieve file contains now (no, I'm not going back over this 
> thread - its becoming as long as war and peace, and you may have changed it 
> since then) 
> 
> Please understand me right... It is nice for you, if dovecot does, what you 
> expect... It is nice. But here it does not work correctly. dovecot makes a 
> big mistake. And i try to give as much information, as i'm possible to 
> give... 
> 
> I doubt its dovecot, since no one else has reported this problem that I can 
> see - without going back to find the start of the thread. 
> 
> my dovecot does not copy the Seen-flag. It ignores it. But WHY. Fucking 
> WHY??? 
> 
> Mind your tongue if you want help here, despite frustrations (man I must be 
> getting old and mellowing), no one here has to do shit for you, the fact it 
> works for everyone else, indicates there is a problem with your configuration 
> and yours alone - somewhere, and because you're the only one experiencing 
> this, it may be harder to trace the origin of. 
> 
> -- 
> Kind Regards, 
> 
> Noel Butler 
> 
> This Email, including any attachments, may contain legally privileged 
> information, therefore remains confidential and subject to copyright 
> protected under international law. You may not disseminate, discuss, or 
> reveal, any part, to anyone, without the authors express written authority to 
> do so. If you are not the intended recipient, please notify the sender then 
> delete all copies of this message including attachments, immediately. 
> Confidentiality, copyright, and legal privilege are not waived or lost by 
> reason of the mistaken delivery of this message. Only PDF [1] and ODF [2] 
> documents accepted, please do not send proprietary formatted documents

I finally had some time to try this out, and wasn't able to reproduce
the problem with 2.3.4 and 0.5.4. 

I tried both Sieve and IMAPSieve, but I wonder if this is something
particular in your environment or settings, so I have to ask you to post
your `doveconf -n` once more. 

Aki 

Did he ever tell us what this is on, I saw stretch somewhere in thread
so I gather its debian, but is it on real hardware, or rpi, has he tried
using the source, who knows what happens when distros butcher things up
into 70 different sub packages :) 

-- 
Kind Regards, 

Noel Butler 

This Email, including any attachments, may contain legally 
privileged
information, therefore remains confidential and subject to copyright
protected under international law. You may not disseminate, discuss, or
reveal, any part, to anyone, without the authors express written
authority to do so. If you are not the intended recipient, please notify
the sender then delete all copies of this message including attachments,
immediately. Confidentiality, copyright, and legal privilege are not
waived or lost by reason of the mistaken delivery of this message. Only
PDF [1] and ODF [2] documents accepted, please do not send proprietary
formatted documents 

 

Links:
--
[1] http://www.adobe.com/
[2] http://en.wikipedia.org/wiki/OpenDocument

Re: BUG: sieve does not set seen-Flag

[Aki Tuomi]

On 6.12.2018 6.54, Noel Butler wrote:
>
> On 06/12/2018 07:29, Jakobus Schürz wrote:
>
>>  that all and every Flag is set, except \Seen... I tried to figure
>> out, whats happening here...
>>
> Paste what your sieve file contains now (no, I'm not going back over
> this thread - its becoming as long as war and peace, and you may have
> changed it since then)
>
>
>> Please understand me right... It is nice for you, if dovecot does,
>> what you expect... It is nice. But here it does not work correctly.
>> dovecot makes a big mistake. And i try to give as much information,
>> as i'm possible to give...
>>
> I doubt its dovecot, since no one else has reported this problem that
> I can see - without going back to find the start of the thread.
>
>> my dovecot does not copy the Seen-flag. It ignores it. But WHY.
>> Fucking WHY???
>>
> Mind your tongue if you want help here, despite frustrations (man I
> must be getting old and mellowing), no one here has to do shit for
> you, the fact it works for everyone else, indicates there is a problem
> with your configuration and yours alone - somewhere, and because
> you're the only one experiencing this, it may be harder to trace the
> origin of.
>
>
> -- 
>
> Kind Regards,
>
> Noel Butler
>
> This Email, including any attachments, may contain legally privileged
> information, therefore remains confidential and subject to copyright
> protected under international law. You may not disseminate, discuss,
> or reveal, any part, to anyone, without the authors express written
> authority to do so. If you are not the intended recipient, please
> notify the sender then delete all copies of this message including
> attachments, immediately. Confidentiality, copyright, and legal
> privilege are not waived or lost by reason of the mistaken delivery of
> this message. Only PDF  and ODF
>  documents accepted, please
> do not send proprietary formatted documents
>

I finally had some time to try this out, and wasn't able to reproduce
the problem with 2.3.4 and 0.5.4.

I tried both Sieve and IMAPSieve, but I wonder if this is something
particular in your environment or settings, so I have to ask you to post
your `doveconf -n` once more.

Aki

Re: argonid and dovecote

[Aki Tuomi]

 
 
  
   
  
  
   
On 06 December 2018 at 22:15 Jerry <
je...@seibercom.net> wrote:
   
   

   
   

   
   
On Thu, 6 Dec 2018 19:05:59 +, Larry Rosenman stated:
   
   

   
   
>On 12/6/18, 1:03 PM, "dovecot on behalf of Jerry"
   
   
><
dovecot-boun...@dovecot.org on behalf of 
je...@seibercom.net> wrote:
   
   
>
   
   

 On Thu, 6 Dec 2018 18:38:32 +, Larry Rosenman stated:


 


 >Because it is a PASSWORD hash, not an authentication mechanism.


 


 Okay, that make sense. Perhaps, a note about that somewhere might prove


 useful.


 


 --


 Jerry

   
   
>Where on the wiki would you suggest that be put?
   
   

   
   
I think that what was confusing, at least for me, is that the "doveadm-pw"
   
   
man page says:
   
   

   
   
-l List all supported password schemes and exit successfully.
   
   
There are up to three optional password schemes: BLF-CRYPT
   
   
Blowfish crypt), SHA256-CRYPT and SHA512-CRYPT. Their
   
   
availability depends on the system's currently used libc.
   
   

   
   
I had assumed, obviously incorrectly, that I had to enter the scheme into the
   
   
dovecot conf file as indicated earlier.
   
   

   
   
Perhaps, for people like me that are not the sharpest knife in the drawer, a
   
   
notation to that affect might prove useful.
   
   

   
   
Just my 2 ￠.
   
   

   
   
--
   
   
Jerry
   
  
  
   
  
  
   I guess the manpage could use some updating. Thank you for pointing this out.
  
  
   ---
   Aki Tuomi
   
 

Re: argonid and dovecote

[Jerry]

On Thu, 6 Dec 2018 19:05:59 +, Larry Rosenman stated:

>On 12/6/18, 1:03 PM, "dovecot on behalf of Jerry"
> wrote:
>
>On Thu, 6 Dec 2018 18:38:32 +, Larry Rosenman stated:
>
>>Because it is a PASSWORD hash, not an authentication mechanism.  
>
>Okay, that make sense. Perhaps, a note about that somewhere might prove
>useful.
>
>-- 
>Jerry

>Where on the wiki would you suggest that be put?

I think that what was confusing, at least for me, is that the "doveadm-pw"
man page says:

-l List all supported password schemes and exit successfully.
   There are up to three optional password schemes: BLF-CRYPT
   Blowfish crypt), SHA256-CRYPT and SHA512-CRYPT.  Their
   availability depends on the system's currently used libc.

I had assumed, obviously incorrectly, that I had to enter the scheme into the
dovecot conf file as indicated earlier.

Perhaps, for people like me that are not the sharpest knife in the drawer, a
notation to that affect might prove useful.

Just my 2 ￠.

-- 
Jerry

Re: argonid and dovecote

[Larry Rosenman]

On 12/6/18, 1:03 PM, "dovecot on behalf of Jerry"  wrote:

On Thu, 6 Dec 2018 18:38:32 +, Larry Rosenman stated:

>Because it is a PASSWORD hash, not an authentication mechanism.

Okay, that make sense. Perhaps, a note about that somewhere might prove
useful.

-- 
Jerry


Where on the wiki would you suggest that be put? 

Re: argonid and dovecote

[Jerry]

On Thu, 6 Dec 2018 18:38:32 +, Larry Rosenman stated:

>Because it is a PASSWORD hash, not an authentication mechanism.

Okay, that make sense. Perhaps, a note about that somewhere might prove
useful.

-- 
Jerry

Re: argonid and dovecote

[Larry Rosenman]

Because it is a PASSWORD hash, not an authentication mechanism.



On 12/6/18, 12:37 PM, "dovecot on behalf of Jerry" 
 wrote:


I have the latest port. I discovered that I do not have to list "argon2id" 
in
"auth_mechanisms", although that does seem strange.

-- 
Jerry

Re: argonid and dovecote

[Jerry]

On Thu, 6 Dec 2018 18:16:05 +, Larry Rosenman stated:

>Update to the latest port in the HEAD ports tree
>
>On 12/6/18, 11:44 AM, "dovecot on behalf of Jerry"
> wrote:
>
>on a FreeBSD 11.2 amd64 machine, I am trying to get Dovecot 2.3.4 to play
>nice with "argonid" encryption.
>
>In the "10-auth.conf" file, I tried:
>
>auth_mechanisms = plain argon2id
>
>Upon restarting dovecot, I received an error message when attempting to
>actually it:
>
>auth: FATAL: Unknown authentication mechanism "ARGON2ID"
>
>Output from doveadm pw -l
>
>doveadm pw -l
>SHA1 SSHA512 BLF-CRYPT PLAIN HMAC-MD5 OTP SHA512 SHA RPA DES-CRYPT CRYPT
> SSHA MD5-CRYPT SKEY PLAIN-MD4 PLAIN-MD5 SCRAM-SHA-1 LANMAN SHA512-CRYPT
> CLEAR CLEARTEXT ARGON2I ARGON2ID SSHA256 NTLM MD5 PBKDF2 SHA256 CRAM-MD5
>PLAIN-TRUNC SHA256-CRYPT SMD5 DIGEST-MD5 LDAP-MD5
>
>I assume I am making a stupid mistake, but I do not know what it is.
>
>-- 
>Jerry

I have the latest port. I discovered that I do not have to list "argon2id" in
"auth_mechanisms", although that does seem strange.

-- 
Jerry

Dovecot 2.3.4 crashing on Solaris 10 - "Panic: output stream is missing error handling"

[Fabian Fritz]

Hi,

I just compiled Dovecot for Solaris 10 and try to run it now. Other
than setting a custom port, I did not add custom settings. When I
tried running doveconf for the first time, it crashed immediately:

doveconf -n
# 2.3.4 (): /export/dovecot/etc/dovecot/dovecot.conf
# OS: SunOS 5.10 sun4u~~
# Hostname: merlin
service pop3-login {
  inet_listener pop3 {
port = 1110
  }
}
submission_max_mail_size = 0
doveconf: Panic: output stream  is missing error handling
doveconf: Error: Raw backtrace: 0xff01b3a8 -> 0xff063b04 -> 0x1673c ->
0x1797c -> 0x14478
Abort (core dumped)

The same error appears when I run dovecot -F. Unfortunately I
currently don't have access to the core dumps (I might do later), but
I could run it with the Debugger with user process debugging enabled:

(This is the part when it crashes. I can provide the full output if requested)

truss -u "libdovecot" /export/dovecot/bin/doveconf -n

service pop3-login {
  inet_listener pop3 {
port = 1110
  }
}
submission_max_mail_size = 0
/1: write(1, " s e r v i c e   p o p 3".., 95)  = 95
/1@1:   <- libdovecot:o_stream_uncork() = 0xa1918
/1@1:   -> libdovecot:buffer_free(0xa1a10, 0x0, 0x0, 0x0)
/1@1:   -> libdovecot:pool_system_free(0xff0fb050, 0x4af10,
0xff0fb030, 0xff059470)
/1@1:   <- libdovecot:buffer_free() = 0
/1@1:   -> libdovecot:buffer_free(0xa1a18, 0x1c00, 0xfec47940, 0x0)
/1@1:   -> libdovecot:pool_system_free(0xff0fb050, 0x4af60,
0xff0fb030, 0xff059470)
/1@1:   <- libdovecot:buffer_free() = 0
/1@1:   -> libdovecot:pool_datastack_create(0xc3ed8, 0xc4200, 0x328, 0x0)
/1@1:   <- libdovecot:pool_datastack_create() = 0x42a58
/1@1:   -> libdovecot:buffer_create_dynamic(0x42a58, 0x20, 0x8000, 0xff0f44f4)
/1@1:   <- libdovecot:buffer_create_dynamic() = 0x42a68
/1@1:   -> libdovecot:buffer_append(0x42a68, 0x525c8, 0x4, 0x42a68)
/1@1:   -> libdovecot:buffer_write(0x42a68, 0x0, 0x525c8, 0x4)
/1@1:   <- libdovecot:buffer_append() = 0x42a88
/1@1:   -> libdovecot:array_sort_i(0xffbfec38, 0x1cd70, 0x0, 0x52830)
/1@1:   <- libdovecot:array_sort_i() = 0x42a88
/1@1:   -> libdovecot:buffer_append_space_unsafe(0x42a68, 0x4, 0x4, 0x1cd70)
/1@1:   -> libdovecot:buffer_get_space_unsafe(0x42a68, 0x4, 0x4, 0x1cd70)
/1@1:   <- libdovecot:buffer_append_space_unsafe() = 0x42a8c
/1@1:   -> libdovecot:o_stream_uncork(0xa1968, 0xffbfedb0, 0x0, 0x1)
/1@1:   <- libdovecot:o_stream_uncork() = 0xa1918
/1@1:   -> libdovecot:o_stream_destroy(0xffbfed3c, 0x0, 0x0, 0x1)
/1@1:   -> libdovecot:o_stream_unref(0xffbfed3c, 0x0, 0x0, 0x1)
doveconf: Panic: output stream  is missing error handling
/1: write(2, " d o v e c o n f :   P a".., 58)  = 58
/1: getcontext(0xFFBFE988)
/1: open("/proc/self/as", O_RDONLY) = 8
/1: fstat64(8, 0xFFBFE7B8)  = 0
/1: pread(8, "FFBFEB H", 4, 0xFFBFE958) = 4
/1: pread(8, "FF\0CBD0", 4, 0xFFBFE95C) = 4
/1: pread(8, "FFBFEBA8", 4, 0xFFBFEB80) = 4
/1: pread(8, "FF01B11C", 4, 0xFFBFEB84) = 4
/1: pread(8, "FFBFEC\b", 4, 0xFFBFEBE0) = 4
/1: pread(8, "FF01B3A8", 4, 0xFFBFEBE4) = 4
/1: pread(8, "FFBFEC80", 4, 0xFFBFEC40) = 4
/1: pread(8, "FF06 ;04", 4, 0xFFBFEC44) = 4
/1: pread(8, "FFBFECE0", 4, 0xFFBFECB8) = 4
/1: pread(8, "\001 g <", 4, 0xFFBFECBC) = 4
/1: pread(8, "FFBFED @", 4, 0xFFBFED18) = 4
/1: pread(8, "\001 y |", 4, 0xFFBFED1C) = 4
/1: pread(8, "FFBFEDF0", 4, 0xFFBFED78) = 4
/1: pread(8, "\001 D x", 4, 0xFFBFED7C) = 4
/1: pread(8, "\0\0\0\0", 4, 0xFFBFEE28) = 4
/1: pread(8, "\0\0\0\0", 4, 0xFFBFEE2C) = 4
/1: close(8)= 0
doveconf: Error: Raw backtrace: 0xff01b3a8 -> 0xff063b04 -> 0x1673c ->
0x1797c -> 0x14478
/1: write(2, " d o v e c o n f :   E r".., 90)  = 90
/1: sigaction(SIGABRT, 0x, 0xFFBFEB88)  = 0
/1: sigaction(SIGABRT, 0xFFBFEA38, 0xFFBFEAD8)  = 0
/1: lwp_sigmask(SIG_SETMASK, 0x, 0x) = 0xFFBFFEFF
[0x]
/1: lwp_kill(1, SIGABRT)= 0
/1: Received signal #6, SIGABRT [default]
/1:   siginfo: SIGABRT pid=25172 uid=0 code=-1

I have no idea about what that error means.

Thanks,
Fabian

Re: argonid and dovecote

[Larry Rosenman]

And enable the "libsodium" option. __ 
(I'm the maintainer for Dovecot, AKA: l...@freebsd.org).

On 12/6/18, 12:16 PM, "Larry Rosenman"  wrote:

Update to the latest port in the HEAD ports tree

On 12/6/18, 11:44 AM, "dovecot on behalf of Jerry" 
 wrote:

on a FreeBSD 11.2 amd64 machine, I am trying to get Dovecot 2.3.4 to 
play
nice with "argonid" encryption.

In the "10-auth.conf" file, I tried:

auth_mechanisms = plain argon2id

Upon restarting dovecot, I received an error message when attempting to
actually it:

auth: FATAL: Unknown authentication mechanism "ARGON2ID"

Output from doveadm pw -l

doveadm pw -l
SHA1 SSHA512 BLF-CRYPT PLAIN HMAC-MD5 OTP SHA512 SHA RPA DES-CRYPT 
CRYPT SSHA
MD5-CRYPT SKEY PLAIN-MD4 PLAIN-MD5 SCRAM-SHA-1 LANMAN SHA512-CRYPT CLEAR
CLEARTEXT ARGON2I ARGON2ID SSHA256 NTLM MD5 PBKDF2 SHA256 CRAM-MD5
PLAIN-TRUNC SHA256-CRYPT SMD5 DIGEST-MD5 LDAP-MD5

I assume I am making a stupid mistake, but I do not know what it is.

-- 
Jerry

Re: argonid and dovecote

[Larry Rosenman]

Update to the latest port in the HEAD ports tree

On 12/6/18, 11:44 AM, "dovecot on behalf of Jerry" 
 wrote:

on a FreeBSD 11.2 amd64 machine, I am trying to get Dovecot 2.3.4 to play
nice with "argonid" encryption.

In the "10-auth.conf" file, I tried:

auth_mechanisms = plain argon2id

Upon restarting dovecot, I received an error message when attempting to
actually it:

auth: FATAL: Unknown authentication mechanism "ARGON2ID"

Output from doveadm pw -l

doveadm pw -l
SHA1 SSHA512 BLF-CRYPT PLAIN HMAC-MD5 OTP SHA512 SHA RPA DES-CRYPT CRYPT 
SSHA
MD5-CRYPT SKEY PLAIN-MD4 PLAIN-MD5 SCRAM-SHA-1 LANMAN SHA512-CRYPT CLEAR
CLEARTEXT ARGON2I ARGON2ID SSHA256 NTLM MD5 PBKDF2 SHA256 CRAM-MD5
PLAIN-TRUNC SHA256-CRYPT SMD5 DIGEST-MD5 LDAP-MD5

I assume I am making a stupid mistake, but I do not know what it is.

-- 
Jerry

argonid and dovecote

[Jerry]

on a FreeBSD 11.2 amd64 machine, I am trying to get Dovecot 2.3.4 to play
nice with "argonid" encryption.

In the "10-auth.conf" file, I tried:

auth_mechanisms = plain argon2id

Upon restarting dovecot, I received an error message when attempting to
actually it:

auth: FATAL: Unknown authentication mechanism "ARGON2ID"

Output from doveadm pw -l

doveadm pw -l
SHA1 SSHA512 BLF-CRYPT PLAIN HMAC-MD5 OTP SHA512 SHA RPA DES-CRYPT CRYPT SSHA
MD5-CRYPT SKEY PLAIN-MD4 PLAIN-MD5 SCRAM-SHA-1 LANMAN SHA512-CRYPT CLEAR
CLEARTEXT ARGON2I ARGON2ID SSHA256 NTLM MD5 PBKDF2 SHA256 CRAM-MD5
PLAIN-TRUNC SHA256-CRYPT SMD5 DIGEST-MD5 LDAP-MD5

I assume I am making a stupid mistake, but I do not know what it is.

-- 
Jerry

Re: Solr

[Tanstaafl]

On Wed Dec 05 2018 07:35:39 GMT-0500 (Eastern Standard Time), Joan
Moreau via dovecot  wrote:
> Why making squat obsolete ? It is simple and straightforward

Because no one has stepped up to maintain it?

Re: BUG: sieve does not set seen-Flag

[Steffen Kaiser]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Wed, 5 Dec 2018, Jakobus Schürz wrote:


Am 05.12.18 um 22:19 schrieb Ralph Seichter:

* Jakobus Schürz:


Does it matter, if the keyword-files do not match?


I think it does not matter.


6 Seen

This does not look right to me. AFAIK, "Seen" should not appear in this
list, because it is one of the predefined flags, but that is also a



Ok. I use dovecot as a service... i did nothing, which is not said in
the official dovecot doku... so... Seen is in this file... Dovecot must
write it down there... because i did it not... i work with thunderbird,
gmail and horde... so dovecot must wrote it itself...


Yes, Dovecot write it there, but a client instructed Dovecot to do so as 
keyword (tag or whatever it is named), maybe the Sieve script was it.


Seen in this file is not the "seen" status of IMAP.

To get deeper into the problem, I would:

1) edit the dovecot-keywords files to spell "Seen" as "XSeen", in order to 
make sure this setting (local keyword) does not interfere with \Seen 
(global IMAP flag)


2) change (I know it should not matter)

addflag "Junk \\seen";

into

addflag "Junk";
addflag "\\Seen";

3) use a file dump tool to verify that "what you see is what you've 
typed", meaning that "\\Seen" is really ASCII \\Seen, e.g. in Linux:


hd sieve-script



- -- 
Steffen Kaiser

-BEGIN PGP SIGNATURE-
Version: GnuPG v1

iQEVAwUBXAjx48QnQQNheMxiAQKaFgf+NLsYdX83pDDj0JDRjXzot/jGDWSEm3Ec
IhCmFtaKf/nE2WQsgZct4pHPicGgjRX6BrLWbGlQepiQY3yMp5qlZYu6rk1wkk8q
qtmkL5mgGL53KtaHPdHGBp20+ATGVF7eAQbnaNFBK3UEBddKbq/QNbeChrHtxXuF
kbVLYN+6nyX9YGBayo1CT05jwhu77Taflx6jIxRXncY1kEYkdGz8E7K/AqSNkYBn
KTHhYfuRUL69vJrJMUpzlnJWtV7H3BCv1yLxYMAOqJFFa4pAOcezGIh2mUIOFl2L
wdzwOQ0LQKekdV5m36K2UrcxTJ94j22/g2N5H6OO3zLh0i/uXT3FNA==
=gt8A
-END PGP SIGNATURE-

Re: Restricting sending mail to domain or group

[admin]

Am Mittwoch, den 05.12.2018, 13:46 -0800 schrieb Joseph Tam:
> On Wed, 5 Dec 2018, Alexander Dalloz wrote:
> 
> [...]
> >
> > Both to be setup on MTA side.
> 
> Agreed.  You can manage your mailing lists with software (e.g.
> mailman),
Actually I tried to avoid introducing another system...

As for now round about 5 people would gain permission to send to this
all alias. We could use mailchimp, as this is already part of marketing
dpt. but I really would like to keep this very simple.

Following possibilities so far:

(1) Either I take care of having a network shared text file with all 
emplpyees up to date 
(2) Using/introducing a mailing list (mailman // mailchimp)
(3) via postfix smtpd_sender_restrictions smtpd_receiver_restrictions?
I am not sure if this is really possible
(4) I could setup an internal virtual domain, e.g. vcomapny.com and 
restrict this domain to only be able to receive mails from @company.com 
and setup a...@vcompany.com group alias.
(5) setup domain wide sieve to only accept mail from a...@company.com if
sender i eieth a, b, or c...

So far (4) and (5) seem to be easy to implement. For me more or less 
the same overhead with (1) but for the user much simpler.

In this case simplicity has priority over giving birth to a feature
that has already an app for to be used with.


> or hack in a filter between the (public) external and (secret)
> internal
> mailing list e.g.
> 
> # Forwards to grp1-secret@ if sender is approved
> grp1: |/path/to/check-sender-filter
> 
> grp1-secret: :include:/path/to/grp1.list
Thanks!

> 
> If spoofing is a concern, you'll have to endow your filter with more
> intelligence.
Its not.

> 
> Joseph Tam 

-M

Re: Restricting sending mail to domain or group

[admin]

Am Mittwoch, den 05.12.2018, 22:31 + schrieb Jochen Bern:
> On 12/05/2018 06:57 PM, admin (@awib.it) wrote:
> > I have a group alias (a...@company.com).
> > (1) Only company.com accounts should be able to send an email to
> > everybody in that company via a...@company.com.
> 
> Do you have a means to identify "some suitable account was used" - as
> opposed to a trivially forged sender address - *other* than by
> watching
> the actual MUA-to-MSA login happen?

Either way would suffice - the simpler the better.

Restricting the sender to be on the same domain if mail is sent to the
company group alias would be fine.

All I try to achieve is to protect those in the group from unwanted
"SPAM", so no one from outside the domain should be able to send to 
that alias (a...@company.com is quite generic...).
Optionally only chosen senders (whitelist) should be able to send to 
that alias. We do not want our mail system to become a craigslist or 
everbody being able to reach thousdands of people with only one email
being sent.

We are rather generous thinking of mail attachments (up to 100MB).
I do not want to see an email like this copied over to thousands of
accounts...


>  [...]
> 
> Regards,

Thanks four your thoughts!

-M