Re: SMTP extensions, was Re: [PATCH] Implement redirect

Hi !!

>> Just as a whish, it would be easy to add some minor changes so
>> exim could be used to experiment with smtp extensions:
> 
> I have added these suggestions to the Wish List. This does not mean that 
> I necessarily agree or disagree with them. The List is just a record of 
> things people would like that at first glance do not seem totally 
> unreasonable or impossible. 

does it means that if i submit a patch for some of those features it
will be incorporated on a future exim version ?

-- 
Best regards ...

----------------------------------------------------------------
    David Saez Padros                http://www.ols.es
    On-Line Services 2000 S.L.       e-mail  david <at> ols.es
    Pintor Vayreda 1                 telf    +34 902 50 29 75
    08184 Palau-Solita i Plegamans   movil   +34 670 35 27 53
----------------------------------------------------------------

--

-- 

[PATCH] smtp_code command to override SMTP response code

As promised :)

  http://www.exim.org/bugzilla/show_bug.cgi?id=115

This patch implements the smtp_code command for ACLs to override the SMTP
response codes used in defer, deny and accept.  Useful for special functionality
codes, like 551 or 251.

See also: http://exim.org/mail-archives/exim-dev/2006-June/msg00017.html

This is item 311 in Exim wishlist.

-- 
Robert Millan

--

-- 

Re: [PATCH] Implement redirect

On Thu, Jun 29, 2006 at 04:55:18PM +0100, Philip Hazel wrote:
> 
> redirect_special:
>   driver = redirect
>   check_local_user
>   require_file $home/.redirect
>   condition = ${a hairy condition that reads the file, picks the domain
>     out of the address, and checks it against +local_domains for not 
>     being local}  
>   data = :fail: Not local, try ${readfile{$home/.redirect}}   

IIUC, this will generate a bounce, not a delivery-time status code.  It is not
possible to do this outside the acl section is it?

Unfortunately, when working in the ACL section, I miss the check_local_user
facility.  Looks like a no-go.  Maybe I implemented smtp_code overrides in the
wrong place.. :/

Or am I missing something?

-- 
Robert Millan

--

-- 

Re: [PATCH] Implement redirect

On Saturday 01 July 2006 23:21, Robert Millan took the opportunity to write:
> On Thu, Jun 29, 2006 at 04:55:18PM +0100, Philip Hazel wrote:
> > redirect_special:
> >   driver = redirect
> >   check_local_user
> >   require_file $home/.redirect
> >   condition = ${a hairy condition that reads the file, picks the domain
> >     out of the address, and checks it against +local_domains for not
> >     being local}
> >   data = :fail: Not local, try ${readfile{$home/.redirect}}
>
> IIUC, this will generate a bounce, not a delivery-time status code. 

No, as long as no_verify isn't set, the router data will be processed as part 
of recipient verification, so the ":fail:" will cause a 550 status code. 
However, for it to work, the exim user must be able to read the .redirect 
file.

-- 
Magnus Holmgren        holmgren <at> lysator.liu.se
                       (No Cc of list mail needed, thanks)

--

-- 

Re: [PATCH] Implement redirect

On Sat, 1 Jul 2006, Magnus Holmgren wrote:

> > > redirect_special:
> > >   driver = redirect
> > >   check_local_user
> > >   require_file $home/.redirect
> > >   condition = ${a hairy condition that reads the file, picks the domain
> > >     out of the address, and checks it against +local_domains for not
> > >     being local}
> > >   data = :fail: Not local, try ${readfile{$home/.redirect}}
> >
> > IIUC, this will generate a bounce, not a delivery-time status code. 
> 
> No, as long as no_verify isn't set, the router data will be processed as part 
> of recipient verification, so the ":fail:" will cause a 550 status code. 
> However, for it to work, the exim user must be able to read the .redirect 
> file.

Yes, that's right. I think this is the point that you may have 
overlooked, Robert. The routers are used at SMTP ACL time when an 
address is verified - so they can influence the status code at that 
point. In fact, you might like to add verify_only to the above router, 
so that it is *not* used when actually deliverying a message.

-- 
Philip Hazel            University of Cambridge Computing Service
Get the Exim 4 book:    http://www.uit.co.uk/exim-book

--

-- 

Re: SMTP extensions, was Re: [PATCH] Implement redirect

On Sat, 1 Jul 2006, David Saez Padros wrote:

> > I have added these suggestions to the Wish List. This does not mean that 
> > I necessarily agree or disagree with them. The List is just a record of 
> > things people would like that at first glance do not seem totally 
> > unreasonable or impossible. 
> 
> does it means that if i submit a patch for some of those features it
> will be incorporated on a future exim version ?

Maybe. 

I can't promise anything other than that I'll look at the patch and make 
a judgement. On some timescale...

-- 
Philip Hazel            University of Cambridge Computing Service
Get the Exim 4 book:    http://www.uit.co.uk/exim-book

--

-- 

Re: [PATCH] Implement redirect

Robert Millan wrote:
> The question is that you could have an MTA with lots of delivery-time
> checks in recipient side, but you can't take advantage of them if the
> spammer doesn't attempt to delivery their stuff directly.

The question now is, what is this mailbox good for? I can't imagine a
real world scenario where this makes sense.

lg,
daniel

--

-- 

Re: Fix for broken WITH_CONTENT_SCAN spam scanning on OS X

On Wed, 7 Jun 2006, Erik wrote:

> The CURE:
>    To avoid re-introducing blocking, avoid a CPU burning loop, and  
> avoid the spotty implementations of poll() on OS X, one can just use  
> the natural BSDish select() instead!  

[snip]

> On a separate note in checking the result of the poll() or select(),  
> and dealing with potential "signal" interruption, it probably should  
> "goto again;" rather than "continue;"  

Since Tom doesn't seem to have picked up these patches, I have applied 
them and committed the spam.c source module. They will therefore be in
tonight's snapshot.

Philip

-- 
Philip Hazel, University of Cambridge Computing Service.

--

-- 

libsrs_alt, exim

This question may be slightly off topic -- does anybody have libsrs_alt + exim 
working in a mixed i386/x86_64 environment? I migrated a mail server to x86_64 
this morning, only to discover that there is a problem with srs on the 64 bit 
machine, as it was not able to reverse the srs address:

--------> srs_inbound router <--------
local_part=srs0=uxfnr7=au=stanford.edu=miles domain=cs.stanford.edu
SRS Rewriting as
calling srs_inbound router
SRS initialized
srs_reverse failed (srs0=uxfnr7=au=stanford.edu=miles <at> cs.stanford.edu): Hash 
verification failed (forged)
srs_inbound router forced address failure
----------- end verify ------------

Looking into it, running an srs forward on the 32-bit machine gives different 
results from the 64-bit one:

32-bit:
srs --secret=foo --alias=cs.stanford.edu --forward miles <at> stanford.edu
SRS0=vsbszV=AU=stanford.edu=miles <at> cs.stanford.edu

64-bit:
srs --secret=foo --alias=cs.stanford.edu --forward miles <at> stanford.edu
SRS0=RU4dln=AU=stanford.edu=miles <at> cs.stanford.edu

Shouldn't those be the *same*? Is there some non-64-bit safe code that I need 
to fix?

--

-- 

Re: Exim 4.62 errors on Solaris

On Jun 30, 2006, at 12:07 PM, Markus Mayer wrote:
>
> I am wondering if anybody else has tried 64 bit Exim binaries on other
> platforms/OSes and what their results were.

I've been running it on FreeBSD/SPARC64 for a while now, plain-jane  
out of ports.  All has been well.

bender% file /usr/local/sbin/exim-4.60-0
/usr/local/sbin/exim-4.60-0: setuid ELF 64-bit MSB executable, SPARC  
V9, version 1 (FreeBSD), dynamically linked (uses shared libs), stripped
bender% uname -mnr
bender.cl.msu.edu 6.1-RELEASE sparc64

./matt

--

--