Developers of Sympa

[Olivier Salaün <address@concealed>]

Hi Chris,

It seems that there is a hard-coded value for the TTL of scenario custom 
conditions ; this hard-coded value is 3600 seconds.
I think it's too bad this is not customizeable, therefore I can think of 
two ways to change this :

  1. provide a new global/list "custom_conditions_ttl" parameter
  2. extend the API of the custom conditions : currently the verify()
     subroutine should return a boolean value, but it could return a
     boolean + a ttl

(1) is faster to implement, whereas (2) is more flexible.
Please comment on this, Chris.
You're also welcome to add an official feature request : 
https://sourcesup.cru.fr/tracker/?group_id=23

The related Sympa documentation : 
http://www.sympa.org/manual/authorization-scenarios#custom_perl_package_conditions

Anyway I move this thread to the sympa-dev mailing list.

Chris Andrews a écrit :
> So what is the amount of time that Sympa caches the results of Custom
> Conditions?
>
> ---
>  Chris Andrews
>  Boston College
>  Collaboration Team
> ________________________________________
> From: Chris Andrews [mailto:address@concealed] 
> Sent: Monday, September 22, 2008 12:35 PM
> To: Olivier Salaün
> Cc: address@concealed
> Subject: RE: [sympa-users] How long does sympa cache the results of
>
> The 2 custom scenarios take care of the SMTP check, because every user here
> has 3 to 7 different email aliases I had to write my own custom conditions
> to make sure that we took into account all of those other email aliases.
>
> CustomCondition::checkallbcemails([sender],[listname],"subscriber") smtp  
> -> do_it
> CustomCondition::checkallbcemails([sender],[listname],"editor")     smtp  
> -> do_it
>
> The problem is that Sympa is caching the results of the custom conditions
> and reusing those results instead of re-evaluating the custom scenario.
>
> ---
>  Chris Andrews
>  Boston College
>  Collaboration Team
>
> -----Original Message-----
> From: Olivier Salaün [mailto:address@concealed] 
> Sent: Monday, September 22, 2008 7:38 AM
> To: Chris Andrews
> Cc: address@concealed
> Subject: Re: [sympa-users] How long does sympa cache the results of
> conditions?
>
> Hi Chris,
>
> I guess that your scenario failed to accept the subscriber's message
> because it is missing the 'smtp' authentication method, along with 'md5'
> and 'smime'.
> It is required because when a message is sent via the mail interface the
> trust only relies on the From: SMTP field.
> Your scenario's first line should look like this :
>
>     is_subscriber([listname],[sender])          *smtp,*md5,smime   -> do_it
>
>
> Chris Andrews a écrit :
>   
> > We had a user send a message to a list and it was rejected because the
> > email address was not a subscriber.  They then added the email address
> > as a subscriber and even ten minutes later the message was still rejected.
> >
> >
> >
> > Our scenario:
> >
> >
> >
> > title.gettext Sending and posting privileges are limited to
> > subscribers all others are forwarded to the owner for moderation.
> >
> >
> >
> > is_subscriber([listname],[sender])          md5,smime   -> do_it
> >
> > is_editor([listname],[sender])              md5,smime   -> do_it
> >
> > CustomCondition::checkallbcemails([sender],[listname],"subscriber")
> > smtp   -> do_it
> >
> > CustomCondition::checkallbcemails([sender],[listname],"editor") smtp
> > -> do_it
> >
> > is_listmaster([sender])                     md5,smime   -> do_it
> >
> > true()                                 smtp,md5,smime   ->
> > reject(reason='send_subscribers_only')
> >
> >
> >
> > Logs:
> >
> >
> >
> > [...]
> >
> > *Sep 17 15:43:28 listserv sympa[27302]: main::DoMessage() Message for
> > <LIST NAME> from <Sending Address> rejected() because sender not allowed*
> >
> > Sep 17 15:43:28 listserv sympa[27302]: Moving bad file <LIST
> > NAME>@listserv.bc.edu.1221680603.18072 to bad/
> >
> > *Sep 17 15:44:00 listserv wwsympa[6663]: [robot listserv.bc.edu]
> > [client 136.167.41.130] [address@concealed] [list <LIST NAME>]
> > main::do_add() do_add(<Sending Address>)*
> >
> > [...]
> >