logs archiveIRC Archive / Freenode / #exim / 2015 / August / 24 / 1
bjornar
henk: I sent helo
henk
bjornar: ok, I recommend you use swaks for testing such things and pastebin outputs verbatim for us to see what is happening &
bjornar
http://pastie.org/10372138
I use swaks normally, but this is not too complicated..
HELP advertises AUTH, which is not supported
jgh_
AUTH is only supported under ESMTP
bjornar
So why does HELP say "AUTH" ?
jgh_
because the system supports ESMTP AUTH
bjornar
but it is disabled in server_advertise_condition
jgh_
that controls the response to an EHLO, which advertises whether AUTH is being offered, and which mechanisms
bjornar
Ok.. anyway.. is it possible to disable AUTH from HELP?
         

jgh_
no
bjornar
How can I debug the content of a domainlist?
jgh_
start by checking it's being used when you think, using cmdline debup options
bjornar
sure it is.. but unsure of its content
since I do mysql lookup to fill it .. and seems mysql lookup dont support >:
jgh_
are you doing a string-expansion mysql lookup that returns a list, or a query-style mysql lookup that is only checked for success/fail ?
bjornar_
Are routers run on every delivery-attempt of a message?
hs12
Yes. Except when not.
bjornar_
Sounds great
hs12
grep for '^first_delivery' in the spec.txt.
There is a condition, that can be evaluated as a router pre-condition: condition = ${if first_delivery}
bjornar_
ok..
so for example for expanding aliases?
hs12
It's up to you. But I'd expand aliases for every attempt, because sometimes I change my mind about the aliases between two attempts.
bjornar_
but are the expanded aliases saved to -H the first time?
henk
bjornar_: why do you ask all that? Sounds like an xy problem &
hs12
no
bjornar: no, because on the next attempt they may be changed
bjornar_
henk, ok, so changes in data from a redirect router is not saved.. then ofcorse I need to reexpand
hs12, that was for you .. henk only speaks about "xy problems"
its like programmers speaking about "race conditions" ;)
hs12
bjornar: otoh, I'm somewhat unsure...
bjornar_
ok...
thats what I am as well, and the reason I asked..
hs12
I'll try to solve it. Moment
The expansion result is saved in the -H file. And a flag, if already delivered
bjornar_
When a message cannot be delivered to all of its recipients immediately, leading to two or more delivery attempts, redirection expansion is carried out afresh each time for those addresses whose children were not all previously delivered. If redirection is being used as a mailing list, this can lead to new members of the list receiving copies of old messages. The one_time option can be used to avoid this.
hs12
There is a description about the format of the -H file in the spect.txt
         

bjornar_
So, normally one would want the one_time flag for aliases, I would say
atleast if it generates a database lookup..
hs12
Yeah, 'one_time' is probably better than '${if first_delivery}', or maybe it leads to the same result.
bjornar_
probably
I will stick with one_time
Trying to optimize my new setup a little here
henk, this was not a xy problem
hs12
Depending on your requirements, it sounds better than the first_delivery approach. Because I'm not sure what Exim will to with the list of addresses, found in the -H file. Maybe just for duplicate delivery suppression
bjornar_
My requirement is basically, that if the rcpt is expanded, it is expanded for this message..
hs12
sounds plausible
If I check such an -H file, there seem to be just the list with the the recipients that already got that message. Thus not the complete expansion result. I think, this will change with the one_time option
bjornar_
hs12, hope so! .. this is not very well documented...
hs12
Yes, does, after using one_time, I get the expansion results in the -H and see it even with 'exipick'
bjornar_
cool
another question, is it possible to change return_path in routers or acl, or only transport?
henk
bjornar_: yeah, not quite. its still a good idea to give context to such questions &
hs12
The -H file is well documented in the spec. And the notice about re-expansion you found your self around the one_time option :)
bjornar_: If I'm not wrong, you can use the router option 'errors_to' or the transport option 'return_path'. I'm not aware of a corresponding acl control = ... But you're free to use acl variables for that.
bjornar_
hs12, Just wondered if I could combine batv signing with a "standard" smtp transport.. without needing multiple transports..
hs12
Probably you can, in your transport. Untested. return_path = ${if <some condition> {$sender_address}{ ${prvs{$sender_address}{BATV_KEY}}}
or ... return_path = ${if <want prvs>{${prvs{$return_path}{BATV_KEY}}}fail}
grep the spec for ^.return_path
and see 'force failure' there
bjornar_
yeah, I'll check if out
noz
henk, notkoos: My sincere apologies for asking a question and disappearing - I hate it when other people do that (real life intervened). Yes, I was doing 'exim -bV -C config.tmp' and was sloppy in transcribing it. I have solved it; turns out it was a corrupt libgcrypt11 library on the system. Once I replaced that, the -bV returned 0 and all was well.
« prev next »