The mailing list for listmasters using Sympa

["Roger B.A. Klorese" <address@concealed>]

It always surprises me these days when I hear of software being built specifically for compatibility with Sendmail, or sites setting up new Sendmail servers. Not because of any distrust about Sendmail's security -- I don't think it's significantly worse than any other common MTA -- and not because of any lack of confidence in the developers -- I'm a former employee of Sendmail, Inc.

No, it surprises me because of the slow decline of Sendmail's market share, and the rise of Postfix and Exim. Sendmail is still likely to be the leader, but with common Linux distros and other Unix versions either deprecating it or dropping it altogether, I'm surprised to see people starting with it.

That, and the fact that even coding m4, let alone cf files, makes me run screaming.

My other point, though, is this: once integration working code is contributed and observed to work in the wild, many other open source projects accept that code into the mainstream, knowing that it will be tested in the normal release process.  I'd think you'd do the same with MTA integrations.



On 1/26/10 1:01 AM, David Verdin wrote:

address@concealed"> Hi Roger,

Well Nothing changed regarding the way aliases must be set for Sympa between 5.3 and 6.0. It is a reasonnably stable part of the code.
So I guess you can use these guidances without troubles.

"Good" reasons? Well there are reasons, only you can know whether they are good or not.
First we develop and run Sympa using sendmail. So our area of expertise covers this MTA and we rely on contributions to help people using Sympa with other softs, such as Postfix or Exim.
Second, we try to make Sympa as indepedendant of its environment as possible. We already need to maintain specific code for several RDBMS, so we can't develop specific code for MTA, or Web servers. Consequently Sympa just scan its spools. You need to find the way to correctly set the aliases to redirect messages to the right spool. We can help you for sendmail, and some of our users contributed helps and howto about Postfix and Exim.

Hope this clarify things.

Regards,

David

Roger B.A. Klorese a écrit :

address@concealed"> Thanks, - though "guidance" it is, as the patches are for 5.3, and I need to find someone capable of implementing it on 6.

Is there any good reason for this, or some other approach that properly guarantees rejection and handles virtual addresses, not to be upstreamed?

On 1/22/10 8:46 AM, Christian Dahlhausen wrote: I think this http://www.folly.org.uk/sympa/sympa_config_03.html will give you some guidance. Haven't implemented it yet myself though.
Christian

On Thu, Jan 21, 2010 at 7:20 PM, Roger B.A. Klorese <address@concealed> wrote:
On 1/19/10 4:50 PM, Roger B.A. Klorese wrote:
Is there some combination of build options and patches that will enable Sympa to build per-robot and per-list-within-robot aliases and to fully support rejection (NOT bouncing) of non-existent addresses within the robot subdomains?

(And if it requires patching, this seems to be a major use case - shouldn't it be in the base distribution?)

Anyone?  Is NOBODY running the obvious combination of Sympa + Postfix + robots?

--
-----------------------------------------------------------
Christian Dahlhausen, Network Systems Engineer
University of Virginia - ITC Network Systems
PO Box 400324, 2015 Ivy Road, Charlottesville, VA 22904