I agree with you on the implicit dependencies on state hash map.
While you work on A, I propose to implement some useful message and
command handlers that we can ship with JAMES.
The experience in developing these handlers can provide inputs to long
term B approach.
Do you have any suggestions?
I would move the solution to this problem for a second refactoring, btw,
as we are here I try to plot a solution:
1) events (extension points) can be published by every component
2) the default events connect and command are handled using the
ConnectHandler and CommandHandler, while the message
I understand that the Commandhandlers and Message handlers are
stateless. What I am not clear the configuration part.
With you proposed design the DataCommandHandler needs to be configured
message handlers two ways
OR
The DataCommandHandler queries the SMTPSession to give l
Venkataraman, Narendra wrote:
How would we configure a handler class that implements both
CommandHandler and Message Handler?
Will the DataCommandHandler call all the commandhandlers to see if they
want add any header to final message?
You can specify the same class as commandHandler and as mes
How would we configure a handler class that implements both
CommandHandler and Message Handler?
Will the DataCommandHandler call all the commandhandlers to see if they
want add any header to final message?
Naren
-Original Message-
From: Stefano Bagnara [mailto:[EMAIL PROTECTED]
Sent: F
[ http://issues.apache.org/jira/browse/JAMES-442?page=all ]
Bernd Fondermann updated JAMES-442:
---
Attachment: postage.zip
+ added ant build file
+ matched & unmatched mail results are written into one common file
+ much wider error catching + recording
[ http://issues.apache.org/jira/browse/JAMES-442?page=all ]
Bernd Fondermann reassigned JAMES-442:
--
Assign To: Bernd Fondermann
> [TOOL] Stress Test Tool
> ---
>
> Key: JAMES-442
> URL: http://issues.apache.org/
I agree with you.
Here are my thots -
For instance to implement an SPF handler - that verifies sender in MAIL
Command and later point in time wants to add a header to the message
that SPF check is passed. Currently a handler can implements both
CommandHandler and MessageHandler interfaces and pro
Am Freitag, den 21.04.2006, 16:20 +0200 schrieb Stefano Bagnara:
> Norman Maurer wrote:
> >> 2) Split currently provided command handlers into multiple command
> >> handlers: the current handler already support multiple handlers per
> >> command, it simply stop after the first handler writing so
Norman Maurer wrote:
2) Split currently provided command handlers into multiple command
handlers: the current handler already support multiple handlers per
command, it simply stop after the first handler writing something as
response.
Yes that seems a good point. So we can write "diffrent Han
[ http://issues.apache.org/jira/browse/JAMES-427?page=all ]
Bernd Fondermann updated JAMES-427:
---
Attachment: unittest6.0.patch
MailetLoader unit tests
> [PATCH] Introduce Unit Testing
> --
>
> Key: JAMES-427
>
Venkataraman, Narendra wrote:
Regarding moving message handler code to DATACmdHandler code would make
the architecture inflexible, if somebody wants
to develop their own Custom Datahandler and still reuse some of the
message handlers.
Well, now you can overwrite the DataCmdHandler with a DataCm
Here are my comments:
My understanding is that current architecture allows multiple command
handlers per command. it stops processing
next set of command handlers registered for the command to ensure that
there are not more than one responses
from the rest of the command handlers.
What would be
Stefano Bagnara wrote:
Why don't we simply put the mailet in the Jar and import the package in
config.xml ?
A mailet already is a class that handle a mail and could be able to send
the message to remote service.
one has to write a new mailet for each remote service type.
if this mailet is in
Am Freitag, den 21.04.2006, 13:27 +0200 schrieb Stefano Bagnara:
> Hi all,
>
> I would like to change something in the current handlerchain code
> (formerly known as fastfail code).
>
> 1) Move some code from the SMTPHandler to the DataCmdHandler. I think
> that the MessageHandler stuff (onMess
Hello all,
Does the JDBC client of JAMES-2.1.3 support OCI?
If it supports , please teach me setting way.
and
Does the JDBC client of JAMES-2.1.3 support ORCLE10g SERVER?
I tried it , but it did not succeed.
[THE PRESENT SETTING]
as following
---
Why don't we simply put the mailet in the Jar and import the package in
config.xml ?
A mailet already is a class that handle a mail and could be able to send
the message to remote service.
Maybe we should support fully qualified class names in the class
attribute for our mailets in config.xm
Sounds ok for me.
bye
Am Freitag, den 21.04.2006, 10:02 +0200 schrieb Stefano Bagnara:
> Norman Maurer wrote:
> > Should we maybe throw a exception if someone has set
> > true but has false
> > set too ?
> > Otherwise this check cann simple passed by not sending a helo/ehlo at
> > all.
> >
Norman Maurer wrote:
Should we maybe throw a exception if someone has set
true but has false
set too ?
Otherwise this check cann simple passed by not sending a helo/ehlo at
all.
What you guys think ?
Maybe an info log saying that the checkValidHelo is true but
heloEhloEnforcement is fa
[ http://issues.apache.org/jira/browse/JAMES-465?page=all ]
Norman Maurer resolved JAMES-465:
-
Resolution: Fixed
This is allready done by findMXRecords.
Sorry
> Check for valid sender domain in mail from
> --
Author: bago
Date: Fri Apr 21 00:53:45 2006
New Revision: 395808
URL: http://svn.apache.org/viewcvs?rev=395808&view=rev
Log:
Updated config.xml to reflect findMXRecords behaviour.
Modified:
james/server/trunk/src/conf/james-config.xml
Modified: james/server/trunk/src/conf/james-config.xml
UR
[
http://issues.apache.org/jira/browse/JAMES-465?page=comments#action_12375505 ]
Norman Maurer commented on JAMES-465:
-
Forget to post the RFC-Part:
The lookup first attempts to locate an MX
record associated with the name. If a CNAME record is fou
Noel J. Bergman wrote:
Stefano Bagnara wrote:
IMHO OSGi specifications themselves are too low-level for James needs.
We probably need an abstraction layer (avalon-osgi bridge or
avalon-somelayer bridge) to simplify this move.
Well ... :-) XBeans will apparently provide a b
23 matches
Mail list logo