You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@turbine.apache.org by Martin Poeschl <mp...@marmot.at> on 2002/12/08 15:21:54 UTC

[vote] plans for turbine 2.3

i posted my plans for the next version of turbine some weeks ago .. now
it's voting time ;-)

o remove all deprecated stuff (+ code cleanup)
   incl. the freemarker and webmacro support (nobody is using it and the current versions are not 
supported)

o switch to commons-logging


+2

martin



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
"Colin Chalmers" <co...@maxware.nl> writes:

>Martin,

>Sounds simple enough.

>What should be done with enhancements, should it be done in Turbine or
>Fulcrum, both? And is there anything that should be taken into account to
>ensure we move closer to Avalon?

>I also think you mentioned a preference for releasing more often, what would
>be the time frame for 2.3? 3 months?

Hi Colin,

please be aware that "moving closer to Avalon" to me means:

"Make it possible to use our Services from Avalon" in the first place

and the in the second place

"Check out whether it is possible to replace the BaseServiceBroker with
 something more slick from Avalon"

The first thing means, that we treat our Services as "Avalon
Components" and adjust them to the Avalon Lifecycle as described on
http://jakarta.apache.org/avalon/framework/reference-the-lifecycle.html

This is mainly a matter of renaming some methods and implementing some
interfaces from the org.apache.avalon.framework package 

(Come on people, this is something that _cries_ for a commons
package. commons-lifecycle anyone?).

There are just a bunch of interfaces with well defined functions that
a service must implement. jmcnally started this for Fulcrum which is
good.

The 2nd part is more tricky as we don't want to overengineer here. The
Turbine Service Broker is an overengineered hash table with some init
function and you can replace it with something as simple as about 75
lines of code. If we find a small container (which is Avalon Speak for
"something that can start up Components") that's cool. If not, I will
strip down this glorified bugger into a single class and a single
interface which will be able to start up Avalon components in a
Turbine manner.

(I personally find the following lines on the Avalon web site
offensive as this is neither discussed on the Turbine team nor a goal
that we all agreed on. And the last sentence basically says "get lost,
I do as I want and then I push it on you without discussion. You have
no vote on that because it is developed in Plexus-land, not
Turbine-land"):

--- cut --- http://jakarta.apache.org/avalon/framework/reference-containers.html --- cut ---
nPlexus is a container under development that will become the basis of the services framework for a sister project to avalon, Turbine.

See Plexus.

Plexus is being developed outside of avalon itself as an external project.
--- cut --- http://jakarta.apache.org/avalon/framework/reference-containers.html --- cut ---

Jason, please remove this from the Avalon website. Or at least the
Turbine reference.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
"Colin Chalmers" <co...@maxware.nl> writes:

>I also think you mentioned a preference for releasing more often, what would
>be the time frame for 2.3? 3 months?

I discussed with Martin off-mailing list (as we have the same mother
tongue, it is sometimes simpler to discuss things in german. :-) ) to
maybe put out a bugfix-release to our current releases (which is 2.2)
all four to eight weeks but without a fixed time frame. We haven't
agreed on a real time window (and a naming scheme :-) ) yet, because
we might want to discuss it here.

I'd personally go for 2.2p1, 2.2p2, 2.2p3 do make clear that these
releases are really bug fixed Turbine 2.2 versions but Martin did tell
me that he prefers 2.2.1, 2.2.2, 2.2.3 and so on. As he does the
releases, it is his call, I think.

These releases should aim for maximum compatibility for 2.2 as we want
to try to (re-)build an active user base starting with this release.

2.3 might introduce some user space visible changes (see my other
mail) so at least I don't intend it as a direct replacement for
Turbine 2.2 (just like 2.2 wasn't a direct replacement for 2.1).

I'd personally consider the following short time release map:

earliest        latest          release
mid Jan'03      end Jan'03      turbine-2.2.1 - first bugfix release of T2.2
Mar'03          mid Mar'03      turbine-2.2.2 - second bugfix release of T2.2

Feb'03          mid Feb'03      turbine-2.3-alpha1 first alpha release of
                                next turbine version
Mar'03          mid Mar'03      turbine-2.3-beta1 first beta release of
                                next turbine version
Apr'03          mid Apr'03      turbine-2.3 next turbine release

and at least one, possible two more releases of turbine-2.2.x after that.

Yes, I do want to speed up development here.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jason van Zyl <ja...@zenplex.com> writes:

>On Sun, 2002-12-08 at 19:04, Age Mooy wrote:
>> > Martin Poeschl <mp...@marmot.at> writes:
>> > 
>> > >summit/plexus for future releases i prefere to keep the services in
>> > 
>> > I might be ignorant here, but I simple marked about four 
>> > months of this mailing list as "read" and deleted it. I found 
>> > out what Plexus is (it seems to become an Avalon compatible 
>> > container for components (the eighth according to
>> > http://jakarta.apache.org/avalon/framework/reference-containers.html)
>> > but what the heck is "Summit"?

>A summit is the top of a mountain, or other precipice, which on a clear
>day provides a good view.

Yep, I found this from the dictionary. But what relation has it to an
"avalon container compatible reimplementation of Turbine-3"?

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: [vote] plans for turbine 2.3

Posted by Jason van Zyl <ja...@zenplex.com>.
On Sun, 2002-12-08 at 19:04, Age Mooy wrote:
> > Martin Poeschl <mp...@marmot.at> writes:
> > 
> > >summit/plexus for future releases i prefere to keep the services in
> > 
> > I might be ignorant here, but I simple marked about four 
> > months of this mailing list as "read" and deleted it. I found 
> > out what Plexus is (it seems to become an Avalon compatible 
> > container for components (the eighth according to
> > http://jakarta.apache.org/avalon/framework/reference-containers.html)
> > but what the heck is "Summit"?

A summit is the top of a mountain, or other precipice, which on a clear
day provides a good view.

> Summit is the turbine-3 core redesigned as a Plexus (and therefor
> Avalon) component by Jason van Zyl.
> 
> It currently resides in the Plexus CVS repo at zenplex.org
> 
> Age
> 
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: [vote] plans for turbine 2.3

Posted by Age Mooy <am...@home.nl>.

> Martin Poeschl <mp...@marmot.at> writes:
> 
> >summit/plexus for future releases i prefere to keep the services in
> 
> I might be ignorant here, but I simple marked about four 
> months of this mailing list as "read" and deleted it. I found 
> out what Plexus is (it seems to become an Avalon compatible 
> container for components (the eighth according to
> http://jakarta.apache.org/avalon/framework/reference-containers.html)
> but what the heck is "Summit"?

Summit is the turbine-3 core redesigned as a Plexus (and therefor
Avalon) component by Jason van Zyl.

It currently resides in the Plexus CVS repo at zenplex.org

Age



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Martin Poeschl <mp...@marmot.at> writes:

>summit/plexus for future releases i prefere to keep the services in 

I might be ignorant here, but I simple marked about four months of
this mailing list as "read" and deleted it. I found out what Plexus is
(it seems to become an Avalon compatible container for components (the
eighth according to
http://jakarta.apache.org/avalon/framework/reference-containers.html)
but what the heck is "Summit"?

People, what's would have been so bad in calling the "fulcrum"
repository "jakarta-turbine-services" when there was the chance for
it? Do we really have to invent new (and most of the time
non-descriptive) names for projects?

BTW: According to dict.leo.org, a "fulcrum" is "a part of an animal
that serves as a hinge or support" or "the support about which a lever
turns" or "one that supplies capability for action" (the third
definition might go by a stretch to meet the function of the fulcrum
repository)

A "summit" is "the highest point". Now that's a telling description. :-)

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Daniel Rall <dl...@finemaltcoding.com>.
Martin Poeschl <mp...@marmot.at> writes:

> Colin Chalmers wrote:
> 
> >Martin,
> >
> >Sounds simple enough.
> >
> >What should be done with enhancements, should it be done in Turbine or
> >Fulcrum, both? And is there anything that should be taken into account to
> >ensure we move closer to Avalon?
> >
> i'm not sure about the future of turbine-3 and fulcrum as
> summit/plexus is discussed as an alternative ...

I'd definitely like to see the changes go into both 2.x and Fulcrum.
Please ping me when ever making changes which should be forward-ported
and I will do so.  Thanks to John McNally, Fulcrum is already mostly
Avalon-ized.

> switching form turbine 2.1 to 2.2 causes tons of code changes .. so i
> would prefere the 2.3 release to cause less pain

A quick 2.3 with minimal addition of functionality and some bug fixes
is Goodness.

> as long as it is not deiced if we will use turbine-3/fulcrum or
> summit/plexus for future releases i prefere to keep the services in
> turbine-2 and improve them there

It doesn't make sense to use either of the separate services
frameworks in the 2.x series -- such a big change is deserving of a
major point release bump.

> >I also think you mentioned a preference for releasing more often, what would
> >be the time frame for 2.3? 3 months?
> >
> maybe ;-)

If it's just a minimal set of changes, there's nothing stopping a fast
release cycle.  Martin, would you mind updating the todo.xml with your
suggested changes (or just do it as a text doc if it's eaiser)?
-- 

Daniel Rall <dl...@finemaltcoding.com>

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Martin Poeschl <mp...@marmot.at>.
Colin Chalmers wrote:

>Martin,
>
>Sounds simple enough.
>
>What should be done with enhancements, should it be done in Turbine or
>Fulcrum, both? And is there anything that should be taken into account to
>ensure we move closer to Avalon?
>
i'm not sure about the future of turbine-3 and fulcrum as summit/plexus 
is discussed as an alternative ...

switching form turbine 2.1 to 2.2 causes tons of code changes .. so i 
would prefere the 2.3 release to cause less pain

as long as it is not deiced if we will use turbine-3/fulcrum or 
summit/plexus for future releases i prefere to keep the services in 
turbine-2 and improve them there

>I also think you mentioned a preference for releasing more often, what would
>be the time frame for 2.3? 3 months?
>
maybe ;-)


martin

>
>
>/c
>
>----- Original Message -----
>From: "Martin Poeschl" <mp...@marmot.at>
>To: "Turbine Developers List" <tu...@jakarta.apache.org>
>Sent: Sunday, December 08, 2002 3:21 PM
>Subject: [vote] plans for turbine 2.3
>
>
>  
>
>>i posted my plans for the next version of turbine some weeks ago .. now
>>it's voting time ;-)
>>
>>o remove all deprecated stuff (+ code cleanup)
>>   incl. the freemarker and webmacro support (nobody is using it and the
>>    
>>
>current versions are not
>  
>
>>supported)
>>
>>o switch to commons-logging
>>
>>
>>+2
>>
>>martin
>>
>>
>>
>>--
>>To unsubscribe, e-mail:
>>    
>>
><ma...@jakarta.apache.org>
>  
>
>>For additional commands, e-mail:
>>    
>>
><ma...@jakarta.apache.org>
>  
>
>>    
>>
>
>
>--
>To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>
>
>  
>



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Colin Chalmers <co...@maxware.nl>.
Martin,

Sounds simple enough.

What should be done with enhancements, should it be done in Turbine or
Fulcrum, both? And is there anything that should be taken into account to
ensure we move closer to Avalon?

I also think you mentioned a preference for releasing more often, what would
be the time frame for 2.3? 3 months?


/c

----- Original Message -----
From: "Martin Poeschl" <mp...@marmot.at>
To: "Turbine Developers List" <tu...@jakarta.apache.org>
Sent: Sunday, December 08, 2002 3:21 PM
Subject: [vote] plans for turbine 2.3


> i posted my plans for the next version of turbine some weeks ago .. now
> it's voting time ;-)
>
> o remove all deprecated stuff (+ code cleanup)
>    incl. the freemarker and webmacro support (nobody is using it and the
current versions are not
> supported)
>
> o switch to commons-logging
>
>
> +2
>
> martin
>
>
>
> --
> To unsubscribe, e-mail:
<ma...@jakarta.apache.org>
> For additional commands, e-mail:
<ma...@jakarta.apache.org>
>
>


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Daniel Rall <dl...@finemaltcoding.com>.
"Henning P. Schmiedehausen" <hp...@intermeta.de> writes:

> Jon Scott Stevens <jo...@latchkey.com> writes:
> 
> >on 2002/12/8 2:37 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> >> What about Avalon logging? It's better suited for server-side logging,
> >> and has its own abstractions that make it possible to use as
> >> implementations log4j, logkit or whatever.
> 
> >I prefer that we use commons-logging.
> 
> >I'm -1 on Avalon logging as the dependency simply because I like to have the
> >choice as to which logging package I want to use.
> 
> I agree on this. commons-logging is a "commons" package. I'd prefer to
> use as little from other projects (where we have no influence and
> vote) as possible.

Indeed, I'm more likely to use Log4J than LogKit or the JDK 1.4
Logging API, but I would rather leave it up to the Turbine users to
decide for themselves -- thus, Commons Logging is the perfect fit.
-- 

Daniel Rall <dl...@finemaltcoding.com>

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jon Scott Stevens <jo...@latchkey.com> writes:

>on 2002/12/8 2:37 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

>> What about Avalon logging? It's better suited for server-side logging,
>> and has its own abstractions that make it possible to use as
>> implementations log4j, logkit or whatever.

>I prefer that we use commons-logging.

>I'm -1 on Avalon logging as the dependency simply because I like to have the
>choice as to which logging package I want to use.

I agree on this. commons-logging is a "commons" package. I'd prefer to
use as little from other projects (where we have no influence and
vote) as possible.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Ceki Gülcü <ce...@qos.ch>.
At 15:49 10.12.2002 -0800, Jon Scott Stevens wrote:

>Everyone who doesn't want to standardize on a single logging implementation
>should at least standardize on a single logging interface and that should be
>commons-logging. If you have a problem with commons-logging, then it should
>be addressed there, not re-implemented somewhere else.

Jon, you are absolutely right. The constructive attitude is to expose to 
c-l developers their errors (as I perceive them). I must admit that I am 
still in shock with respect to their choice of class loader based mechanism 
for selecting the logging implementation.

>-jon

--
Ceki



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jason van Zyl <ja...@zenplex.com> writes:

>I did however, for better or worse, decouple fulcrum and torque, and I
>gutted 2.x into something a lot more comprehensible which is t3. But
>always expect to throw the first one away and what is present in Summit
>is what I expect to be a long term solution.

Aha. What BTW is a component container like plexus now doing in Maven? 

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jason van Zyl <ja...@zenplex.com>.
On Thu, 2002-12-12 at 10:00, Henning P. Schmiedehausen wrote:
> Jason van Zyl <ja...@zenplex.com> writes:
> 
> >Nothing in Turbine 2.x is clear. Period.
> 
> Hearing this from a main contributor of Code for T2.x, x < 2 is
> basically telling "I did screw up and then I left to let others pick
> up the pieces".
> 
> So, please do so.

I was not a main contributor to 2.x. I had nothing to do with the way
the request life cycle was handled, the security service or the high
level of coupling that occurred naturally. I cleaned it up for release,
added some services and the code generation for Torque but the
architecture is not mine

I did however, for better or worse, decouple fulcrum and torque, and I
gutted 2.x into something a lot more comprehensible which is t3. But
always expect to throw the first one away and what is present in Summit
is what I expect to be a long term solution.

> 	Regards
> 		Henning
-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jason van Zyl <ja...@zenplex.com> writes:

>Nothing in Turbine 2.x is clear. Period.

Hearing this from a main contributor of Code for T2.x, x < 2 is
basically telling "I did screw up and then I left to let others pick
up the pieces".

So, please do so.

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jason van Zyl <ja...@zenplex.com>.
On Tue, 2002-12-10 at 18:50, Henning P. Schmiedehausen wrote:

> 
> Folks,
> 
> please. We do not try to build Avalon-light or an Avalon clone.
> Staring at foreign code and rebuilding every interface from another
> project will not help us.

I think John has gone the correct path in converting Fulcrum over to
just plain use Avalon.

> Compared to Avalon, Turbine is really lightweight and simple. 

Are you kidding? In what respect do you find Turbine more lightweight
and simple? I cannot think of a single example for which Turbine is
lighter or simpler. Can you give me any?

The entire lifecycle for Avalon components and containers is contained
in a ~60k JAR. All components are well defined and live out their
existance in a well defined manner. Nothing in Turbine 2.x is well
defined and nothing lives out its existance in a well defined manner.
The service code is one example of complete confusion and obscurity.
Torque, though I have not looked at it, began its lifecycle with a call
to Criteria. If that is not confusing and unintuitive I don't know what
is. The security service, the lifecycle of a request ... none of these
thing are clear. In a component framework like Avalon these entities are
forced, by the patterns inherent in the framework, to be clear.

Nothing in Turbine 2.x is clear. Period.

> I'll try
> to keep it that way. 

You want to keep it a completely and totally unintuitive mess?

> If we can use synergies with Avalon, fine. If
> not, well, if you need all the features of Avalon, why not use Avalon
> in the first place? We do not try to be an end-all, be-all solution
> which replaces all other web-application frameworks. We do not need to
> be "best-of-brand" in every category. We're not even trying to
> compete. :-)

Speak for yourself.

> I see logging for Turbine as a way to offer a centralized, easy to use
> and flexible facility for applications and services. All this does
> commons-logging in conjunction with log4j (or JDK 1.4) offer.
> 
> I'll keep this code simple, hardcoded and out of the lifecycle
> code. 

There is no doubt that logging is a fundamental aspect to an
application. In Avalon it's attached to each component and works fairly
well in that environment. In Turbine it definitely doesn't work well as
a service which is why it was removed as such in the turbine 3 code
base. At the very least if you're going to start making 2.x like 3.x use
the 3.x code.

> Reason: You'll end up with a mess like in the current Turbine. What if
> you want to log messages before your ServiceBroker/ComponentContainer
> has even started? 

Been discussed which is why it was removed as a service in 3.x.

> Remember the posting about the stratum message
> popping out of Stdout because of this. Logging is simply the second
> thing that we will set up right after startup (First thing is reading
> the conf because we need this for configuring the logging). That's it.
> 
> 	Regards
> 		Henning
-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Kurt Schrader <ks...@karmalab.org>.
On Tue, 10 Dec 2002, Henning P. Schmiedehausen wrote:

> Compared to Avalon, Turbine is really lightweight and simple. I'll try
> to keep it that way. If we can use synergies with Avalon, fine. If
> not, well, if you need all the features of Avalon, why not use Avalon
> in the first place?

What are you talking about?  Avalon isn't a webapp framework.

> We do not try to be an end-all, be-all solution which replaces all other
> web-application frameworks. We do not need to be "best-of-brand" in
> every category. We're not even trying to compete. :-)

What are we trying to do then?  Build a half-assed framework?

-Kurt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Kurt Schrader <ks...@karmalab.org> writes:

>On Wed, 11 Dec 2002, Nicola Ken Barozzi wrote:

>> I'm not suggesting that you use Logkit, but the Avalon logging interface
>> and facades. the Logger is part of the Avalon lifecycle, and it's very
>> important, also from a security perspective, that it's given by the
>> container, not gotten by the Component, and makes management of it easier.

>Does it tie Turbine to using Avalon exclusively or is it like
>commons-logging with some extra hooks into Avalon?

Folks,

please. We do not try to build Avalon-light or an Avalon clone.
Staring at foreign code and rebuilding every interface from another
project will not help us.

Compared to Avalon, Turbine is really lightweight and simple. I'll try
to keep it that way. If we can use synergies with Avalon, fine. If
not, well, if you need all the features of Avalon, why not use Avalon
in the first place? We do not try to be an end-all, be-all solution
which replaces all other web-application frameworks. We do not need to
be "best-of-brand" in every category. We're not even trying to
compete. :-)

I see logging for Turbine as a way to offer a centralized, easy to use
and flexible facility for applications and services. All this does
commons-logging in conjunction with log4j (or JDK 1.4) offer.

I'll keep this code simple, hardcoded and out of the lifecycle
code. 

Reason: You'll end up with a mess like in the current Turbine. What if
you want to log messages before your ServiceBroker/ComponentContainer
has even started? Remember the posting about the stratum message
popping out of Stdout because of this. Logging is simply the second
thing that we will set up right after startup (First thing is reading
the conf because we need this for configuring the logging). That's it.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jon Scott Stevens <jo...@latchkey.com> writes:

>I'm strongly -1 on using the Avalon logging interface because this is
>something that belongs in Commons. There is no reason for Avalon to have
>*yet another logging interface* when commons-logging is available.

I do agree with Jon on that (Wonder, oh Wonder. =:-) )

>StudioZ.tv /\ Bar/Nightclub/Entertainment

Must. Find. Time. Next year to visit Bay Area....

	Regards
		Henning



-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 4:55 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> Try using Commons logging in a server environment with 10 apps all using
> it, and you'll see that you will start losing control on what's
> happening to the loggings.

Works fine for Scarab (the 'apps' are the various components that it depends
on...which are quite a bit)...

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Henning P. Schmiedehausen wrote:
> Nicola Ken Barozzi <ni...@apache.org> writes:
> 
> 
>>Look in the framework:
>>http://cvs.apache.org/viewcvs/jakarta-avalon/src/java/org/apache/avalon/framework/logger/
> 
> 
>>AbstractLogEnabled.java
>>AvalonFormatter.java
>>ConsoleLogger.java
>>Jdk14Logger.java
>>Log4JLogger.java
>>LogEnabled.java
>>LogKit2AvalonLoggerAdapter.java
>>LogKitLogger.java
>>Logger.java
>>NullLogger.java
> 
> 
> Yep. Horrors. 
> That's exactly what I don't want in Turbine. The
> logging code that I plan to check in has exactly 10 lines of code to
> set up and replaces the whole LoggingService and lots of this "try to
> log before we have something like LogService but switch to it after it
> came up" horrors that spread all over Turbine and the Service Broker.

I don't understand what your problem is, we have no problem in using it 
in many projects.

Probably you were not doing it right, but I dunno actually what your 
code looked like.

Remember that static loggers are hell in a server environment. The 
Component mus tuse the logger it gets from the container, period. Easy, 
clean, safe, and easily configurable.

Try using Commons logging in a server environment with 10 apps all using 
it, and you'll see that you will start losing control on what's 
happening to the loggings.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Nicola Ken Barozzi <ni...@apache.org> writes:

>Look in the framework:
>http://cvs.apache.org/viewcvs/jakarta-avalon/src/java/org/apache/avalon/framework/logger/

>AbstractLogEnabled.java
>AvalonFormatter.java
>ConsoleLogger.java
>Jdk14Logger.java
>Log4JLogger.java
>LogEnabled.java
>LogKit2AvalonLoggerAdapter.java
>LogKitLogger.java
>Logger.java
>NullLogger.java

Yep. Horrors.  That's exactly what I don't want in Turbine. The
logging code that I plan to check in has exactly 10 lines of code to
set up and replaces the whole LoggingService and lots of this "try to
log before we have something like LogService but switch to it after it
came up" horrors that spread all over Turbine and the Service Broker.

	Regards
		Henning



-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Kurt Schrader <ks...@karmalab.org>.
On Wed, 11 Dec 2002, Nicola Ken Barozzi wrote:

> I'm not suggesting that you use Logkit, but the Avalon logging interface
> and facades. the Logger is part of the Avalon lifecycle, and it's very
> important, also from a security perspective, that it's given by the
> container, not gotten by the Component, and makes management of it easier.

Does it tie Turbine to using Avalon exclusively or is it like
commons-logging with some extra hooks into Avalon?

-Kurt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.

Henning P. Schmiedehausen wrote:
> Nicola Ken Barozzi <ni...@apache.org> writes:
> 
> 
>>Remember though that the fact that commons strongly opposes anything 
>>that has even a remote hint about avalon framework really doesn't help.
> 
> 
> I've already proposed that we create a commons-lifecycle package which
> contains the Avalon lifecycle interfaces. I got no response on that.
> 
> What do you think?

Honest?

I think it's ridiculous. What if I asked you to create a commons-turbine 
package that contains all the turbine interfaces? Heck, why Turbine at 
all? Let's dump everything in Commons, wohooo! :-P

Lifecycle interfaces only make sense if they are part of a framework. 
The framework makes them used in a consistent manner and specifies the 
contracts. Once you have done it, you have actually moved Avalon to 
Commons. Yeah, what a result.

I already replied on this point on commons-dev BTW.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Nicola Ken Barozzi <ni...@apache.org> writes:

>Remember though that the fact that commons strongly opposes anything 
>that has even a remote hint about avalon framework really doesn't help.

I've already proposed that we create a commons-lifecycle package which
contains the Avalon lifecycle interfaces. I got no response on that.

What do you think?

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Jon Scott Stevens <jo...@latchkey.com> writes:

>Avalon should drop whatever duplicate logging interfaces it has and adopt
>commons-logging. The same should go for Turbine and every other Jakarta
>application out there.

No, please. It should offer its code (which seems to be better by the
opinion of the Avalon developers) to the commons-logging project and 
_then_ use the new and improved commons-logging.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Jason van Zyl <ja...@zenplex.com>.
On Wed, 2002-12-11 at 14:31, Daniel Rall wrote:
> Martin Poeschl <mp...@marmot.at> writes:
> 
> > Nicola Ken Barozzi wrote:
> > 
> > > We didn't fork. Any commons project should work with all projects to
> > > find a common solution that all would switch to. This hasn't been
> > > done, and is happening now with configuration. We have a
> > > Configuration system, and commons configuration is still going its
> > > way. We'll see if it comes out similarly to logging.
> > 
> > 
> > i just started to look at the avalon stuff ... but i think it would be
> > good to have things like logging and configuration in commons ...
> > 
> > commons-configuration needs a XMLConfiguration implementation
> > ... avalon has one .. why don't we join forces??
> 
> What about the one in Stratum? 
> 
> jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java
> 
> Stratum's configuration stuff was just copied from Commons
> Configuration when Jason wanted to do some refactoring, so it would be
> simple to merge back into Commons Configuration.  Looking at what
> Avalon has to offer is a good idea regardless.

I'm just using the straight-up avalon configuration stuff in
Plexus/Summit in most places. It works well, though it could use a Java
Preferences API wrapper.

-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Daniel Rall <dl...@collab.net>.
On Fri, 13 Dec 2002, Martin Poeschl wrote:

> Daniel Rall wrote:
> 
> >On Thu, 12 Dec 2002, Martin Poeschl wrote:
> >
> >  
> >
> >>Daniel Rall wrote:
> >>
> >>    
> >>
> >>>Martin Poeschl <mp...@marmot.at> writes:
> >>>
> >>> 
> >>>
> >>>      
> >>>
> >>>>Nicola Ken Barozzi wrote:
> >>>>
> >>>>   
> >>>>
> >>>>        
> >>>>
> >>>>>We didn't fork. Any commons project should work with all projects to
> >>>>>find a common solution that all would switch to. This hasn't been
> >>>>>done, and is happening now with configuration. We have a
> >>>>>Configuration system, and commons configuration is still going its
> >>>>>way. We'll see if it comes out similarly to logging.
> >>>>>     
> >>>>>
> >>>>>          
> >>>>>
> >>>>i just started to look at the avalon stuff ... but i think it would be
> >>>>good to have things like logging and configuration in commons ...
> >>>>
> >>>>commons-configuration needs a XMLConfiguration implementation
> >>>>... avalon has one .. why don't we join forces??
> >>>>   
> >>>>
> >>>>        
> >>>>
> >>>What about the one in Stratum? 
> >>>
> >>>jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java
> >>>
> >>>      
> >>>
> >>the configuration package in stratum is empty as all stuff was moved to 
> >>commons-configuration ...
> >>    
> >>
> >
> >Excellent, it's already there:
> >
> >jakarta-commons-sandbox/configuration/src/java/org/apache/commons/configuration/XmlConfiguration.java
> >
> >Sounds like Commons Configuration already has a XML implementation.  Thanks 
> >Martin.
> >
> it has a placeholder ...
> someone posted a implementation using dom4j to the commons-list .... but 
> it's not checked in

Do you have a URL for this implementation?  I'll see about getting it 
commited.

- Dan



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Martin Poeschl <mp...@marmot.at>.
Daniel Rall wrote:

>On Thu, 12 Dec 2002, Martin Poeschl wrote:
>
>  
>
>>Daniel Rall wrote:
>>
>>    
>>
>>>Martin Poeschl <mp...@marmot.at> writes:
>>>
>>> 
>>>
>>>      
>>>
>>>>Nicola Ken Barozzi wrote:
>>>>
>>>>   
>>>>
>>>>        
>>>>
>>>>>We didn't fork. Any commons project should work with all projects to
>>>>>find a common solution that all would switch to. This hasn't been
>>>>>done, and is happening now with configuration. We have a
>>>>>Configuration system, and commons configuration is still going its
>>>>>way. We'll see if it comes out similarly to logging.
>>>>>     
>>>>>
>>>>>          
>>>>>
>>>>i just started to look at the avalon stuff ... but i think it would be
>>>>good to have things like logging and configuration in commons ...
>>>>
>>>>commons-configuration needs a XMLConfiguration implementation
>>>>... avalon has one .. why don't we join forces??
>>>>   
>>>>
>>>>        
>>>>
>>>What about the one in Stratum? 
>>>
>>>jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java
>>>
>>>      
>>>
>>the configuration package in stratum is empty as all stuff was moved to 
>>commons-configuration ...
>>    
>>
>
>Excellent, it's already there:
>
>jakarta-commons-sandbox/configuration/src/java/org/apache/commons/configuration/XmlConfiguration.java
>
>Sounds like Commons Configuration already has a XML implementation.  Thanks 
>Martin.
>
it has a placeholder ...
someone posted a implementation using dom4j to the commons-list .... but 
it's not checked in

martin

>
>- Dan
>
>
>
>--
>To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>
>
>  
>



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Daniel Rall <dl...@collab.net>.
On Thu, 12 Dec 2002, Martin Poeschl wrote:

> Daniel Rall wrote:
> 
> >Martin Poeschl <mp...@marmot.at> writes:
> >
> >  
> >
> >>Nicola Ken Barozzi wrote:
> >>
> >>    
> >>
> >>>We didn't fork. Any commons project should work with all projects to
> >>>find a common solution that all would switch to. This hasn't been
> >>>done, and is happening now with configuration. We have a
> >>>Configuration system, and commons configuration is still going its
> >>>way. We'll see if it comes out similarly to logging.
> >>>      
> >>>
> >>i just started to look at the avalon stuff ... but i think it would be
> >>good to have things like logging and configuration in commons ...
> >>
> >>commons-configuration needs a XMLConfiguration implementation
> >>... avalon has one .. why don't we join forces??
> >>    
> >>
> >
> >What about the one in Stratum? 
> >
> >jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java
> >
> the configuration package in stratum is empty as all stuff was moved to 
> commons-configuration ...

Excellent, it's already there:

jakarta-commons-sandbox/configuration/src/java/org/apache/commons/configuration/XmlConfiguration.java

Sounds like Commons Configuration already has a XML implementation.  Thanks 
Martin.

- Dan



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Martin Poeschl <mp...@marmot.at>.
Daniel Rall wrote:

>Martin Poeschl <mp...@marmot.at> writes:
>
>  
>
>>Nicola Ken Barozzi wrote:
>>
>>    
>>
>>>We didn't fork. Any commons project should work with all projects to
>>>find a common solution that all would switch to. This hasn't been
>>>done, and is happening now with configuration. We have a
>>>Configuration system, and commons configuration is still going its
>>>way. We'll see if it comes out similarly to logging.
>>>      
>>>
>>i just started to look at the avalon stuff ... but i think it would be
>>good to have things like logging and configuration in commons ...
>>
>>commons-configuration needs a XMLConfiguration implementation
>>... avalon has one .. why don't we join forces??
>>    
>>
>
>What about the one in Stratum? 
>
>jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java
>
the configuration package in stratum is empty as all stuff was moved to 
commons-configuration ...

>Stratum's configuration stuff was just copied from Commons
>Configuration when Jason wanted to do some refactoring, so it would be
>simple to merge back into Commons Configuration.  Looking at what
>Avalon has to offer is a good idea regardless.
>

>  
>



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Daniel Rall <dl...@finemaltcoding.com>.
Martin Poeschl <mp...@marmot.at> writes:

> Nicola Ken Barozzi wrote:
> 
> > We didn't fork. Any commons project should work with all projects to
> > find a common solution that all would switch to. This hasn't been
> > done, and is happening now with configuration. We have a
> > Configuration system, and commons configuration is still going its
> > way. We'll see if it comes out similarly to logging.
> 
> 
> i just started to look at the avalon stuff ... but i think it would be
> good to have things like logging and configuration in commons ...
> 
> commons-configuration needs a XMLConfiguration implementation
> ... avalon has one .. why don't we join forces??

What about the one in Stratum? 

jakarta-turbine-stratum/src/java/org/apache/stratum/configuration/XmlConfiguration.java

Stratum's configuration stuff was just copied from Commons
Configuration when Jason wanted to do some refactoring, so it would be
simple to merge back into Commons Configuration.  Looking at what
Avalon has to offer is a good idea regardless.
-- 

Daniel Rall <dl...@finemaltcoding.com>

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Nicola Ken Barozzi <ni...@apache.org>.

Martin Poeschl wrote:
> Nicola Ken Barozzi wrote:
> 
>> We didn't fork. Any commons project should work with all projects to 
>> find a common solution that all would switch to. This hasn't been 
>> done, and is happening now with configuration. We have a Configuration 
>> system, and commons configuration is still going its way. We'll see if 
>> it comes out similarly to logging.
> 
> 
> i just started to look at the avalon stuff ... but i think it would be 
> good to have things like logging and configuration in commons ...
> commons-configuration needs a XMLConfiguration implementation ... avalon 
> has one .. why don't we join forces??

I have posted many messages on commons-dev about doing so, and some in 
the last days, but to no avail.

I'd be more than happy to help you or anyone on this. My personal 
tentative is that of seeing that we don't divide on this and eventually 
are able to reunite also on the Logging; now that Turbine uses Avalon 
it's even more important.

Lead me, I will follow (I'm on commons.dev also).

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


configuration (was: Re: [vote] plans for turbine 2.3)

Posted by Martin Poeschl <mp...@marmot.at>.
Nicola Ken Barozzi wrote:

> We didn't fork. Any commons project should work with all projects to 
> find a common solution that all would switch to. This hasn't been 
> done, and is happening now with configuration. We have a Configuration 
> system, and commons configuration is still going its way. We'll see if 
> it comes out similarly to logging.

i just started to look at the avalon stuff ... but i think it would be 
good to have things like logging and configuration in commons ...
commons-configuration needs a XMLConfiguration implementation ... avalon 
has one .. why don't we join forces??
martin


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Daniel Rall <dl...@collab.net>.
On 12 Dec 2002, Leo Simons wrote:

> Hi Daniel,
> 
> On Wed, 2002-12-11 at 20:29, Daniel Rall wrote:
> > As both implementations are roughly functionally equivalent, we need
> > to look at their other attributes.  The Avalon project is a server
> > framework, not focussed on logging.  The entire point of Commons is to
> > develop very specific components shared as widely as possible -- its
> > Logging component is focussed _specifically_ on logging.  Unless
> > there's some technical point which I've overlooked (please speak up if
> > this is so), I'm going to choose the implementation with the narrowest
> > focus every time. 
> 
> the technical point is that avalon is quite rigorous in its application
> of Inversion of Control [1]. In an IoC application, use of static
> factories is problematic. As Commons Logging is based on a static
> factory, there's friction.
> 
> Like I said yesterday, if you go IoC/SoC/COP (as in avalon-style), you
> don't want static factories. If you go singleton factory model (as in
> current stratum/fulcrum style), you probably do.
> 
> best regards,
> 
> - Leo
> 
> [1] - http://jakarta.apache.org/avalon/framework/guide-patterns-ioc.html

Leo, will you point me to some descriptions of the problems associated with 
static factories?  I didn't see anything on the IoC page.

- Dan




--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Leo Simons <le...@apache.org>.
Hi Daniel,

On Wed, 2002-12-11 at 20:29, Daniel Rall wrote:
> As both implementations are roughly functionally equivalent, we need
> to look at their other attributes.  The Avalon project is a server
> framework, not focussed on logging.  The entire point of Commons is to
> develop very specific components shared as widely as possible -- its
> Logging component is focussed _specifically_ on logging.  Unless
> there's some technical point which I've overlooked (please speak up if
> this is so), I'm going to choose the implementation with the narrowest
> focus every time. 

the technical point is that avalon is quite rigorous in its application
of Inversion of Control [1]. In an IoC application, use of static
factories is problematic. As Commons Logging is based on a static
factory, there's friction.

Like I said yesterday, if you go IoC/SoC/COP (as in avalon-style), you
don't want static factories. If you go singleton factory model (as in
current stratum/fulcrum style), you probably do.

best regards,

- Leo

[1] - http://jakarta.apache.org/avalon/framework/guide-patterns-ioc.html

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
On Thu, 2002-12-12 at 15:00, Jon Scott Stevens wrote:
> on 2002/12/12 12:54 AM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> > That's why our implementation, logkit, is not part of the framework. The
> > framework actually just has hooks to use other implementations.
> 
> It is part of Avalon and not part of Jakarta Commons. That is the *detail*
> that you can't seem to understand for some reason.

I think it is understood Jon. I'm not sure, however, that all of us see
that as a *problem*

(referring to framework of course, logkit is of no interest to me)


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/12 12:54 AM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> That's why our implementation, logkit, is not part of the framework. The
> framework actually just has hooks to use other implementations.

It is part of Avalon and not part of Jakarta Commons. That is the *detail*
that you can't seem to understand for some reason.

-jon


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Daniel Rall wrote:
> Nicola Ken Barozzi <ni...@apache.org> writes:
> 
>>Jon Scott Stevens wrote:
>>
>>>on 2002/12/10 11:51 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
>>>
>>
>>>>Avalon is a server framework. You mean that logging is not to be a
>>>>concern of a server framework?
>>>
>>>Correct. Very much so. If Avalon was a logging framework, then
>>>logging would
>>
>>>be a concern. But it isn't.
>>
>>Logging done in server apps is a *concern* of the server framework;
>>the logging policies are definately concerns of server apps, hence of
>>the frameworks to create them.
> 
> 
> Logging is definitely a concern of server frameworks.  However, it's
> not up to a framework to supply implementations for something which
> will have very different use cases depending upon a user's
> environment.

That's why our implementation, logkit, is not part of the framework. The 
framework actually just has hooks to use other implementations.

So just choose your implementation, hook it up, and you're set. This is 
in line with Leo's proposal of using Commons Logging underneath.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/11 11:29 AM, "Daniel Rall" <dl...@finemaltcoding.com> wrote:

> Logging is definitely a concern of server frameworks.  However, it's
> not up to a framework to supply implementations for something which
> will have very different use cases depending upon a user's
> environment.
> 
> In this case, both Commons Logging and Avalong supply a set of Java
> interfaces which allow (somewhat) pluggable logging.  This confuses
> the issue, as both parties think that they're not supplying
> implementations.  In fact, both are supplying implementations of a
> conceptual, meta logging interface (imagine a checkbox on a marketing
> hand-out, but don't be sidetracked by my reference to marketing ;).
> 
> I'm always going to choose the tool which works best for the situation
> at hand.  Now, I assume that both implementations of the conceptual
> logging interface _function_ just fine -- that's not what's under
> debate.  What's under debate is why to choose an Avalon-specific
> implementation over a Commons implementation.
> 
> As both implementations are roughly functionally equivalent, we need
> to look at their other attributes.  The Avalon project is a server
> framework, not focussed on logging.  The entire point of Commons is to
> develop very specific components shared as widely as possible -- its
> Logging component is focussed _specifically_ on logging.  Unless
> there's some technical point which I've overlooked (please speak up if
> this is so), I'm going to choose the implementation with the narrowest
> focus every time.  Generally speaking, a narrow focus means I can
> count on the tool to be highly cohesive, and over time much more
> widely deployed (for instance, Commons Loggin ships with Tomcat 4).
> The wide deployment means that more of my users are likely to have
> come in contact with the tool before, and may already be using it in
> their environment.  This familiarity results in a lower learning curve
> when coming up to speed on my framework.  I could go on about the
> "softer" aspects of tool choices for days, but I'm sure we all get the
> point.  :)

+1

-jon


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Daniel Rall <dl...@finemaltcoding.com>.
Nicola Ken Barozzi <ni...@apache.org> writes:

> Jon Scott Stevens wrote:
> > on 2002/12/10 11:51 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> >
> 
> >>Avalon is a server framework. You mean that logging is not to be a
> >>concern of a server framework?
> > Correct. Very much so. If Avalon was a logging framework, then
> > logging would
> 
> > be a concern. But it isn't.
> 
> Logging done in server apps is a *concern* of the server framework;
> the logging policies are definately concerns of server apps, hence of
> the frameworks to create them.

Logging is definitely a concern of server frameworks.  However, it's
not up to a framework to supply implementations for something which
will have very different use cases depending upon a user's
environment.

In this case, both Commons Logging and Avalong supply a set of Java
interfaces which allow (somewhat) pluggable logging.  This confuses
the issue, as both parties think that they're not supplying
implementations.  In fact, both are supplying implementations of a
conceptual, meta logging interface (imagine a checkbox on a marketing
hand-out, but don't be sidetracked by my reference to marketing ;).

I'm always going to choose the tool which works best for the situation
at hand.  Now, I assume that both implementations of the conceptual
logging interface _function_ just fine -- that's not what's under
debate.  What's under debate is why to choose an Avalon-specific
implementation over a Commons implementation.

As both implementations are roughly functionally equivalent, we need
to look at their other attributes.  The Avalon project is a server
framework, not focussed on logging.  The entire point of Commons is to
develop very specific components shared as widely as possible -- its
Logging component is focussed _specifically_ on logging.  Unless
there's some technical point which I've overlooked (please speak up if
this is so), I'm going to choose the implementation with the narrowest
focus every time.  Generally speaking, a narrow focus means I can
count on the tool to be highly cohesive, and over time much more
widely deployed (for instance, Commons Loggin ships with Tomcat 4).
The wide deployment means that more of my users are likely to have
come in contact with the tool before, and may already be using it in
their environment.  This familiarity results in a lower learning curve
when coming up to speed on my framework.  I could go on about the
"softer" aspects of tool choices for days, but I'm sure we all get the
point.  :)
-- 

Daniel Rall <dl...@finemaltcoding.com>

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.

Jon Scott Stevens wrote:
> on 2002/12/10 11:51 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> 
>>Avalon is a server framework. You mean that logging is not to be a
>>concern of a server framework?
> 
> Correct. Very much so. If Avalon was a logging framework, then logging would
> be a concern. But it isn't. 

Logging done in server apps is a *concern* of the server framework; the 
logging policies are definately concerns of server apps, hence of the 
frameworks to create them.
Correct, safe, mannageable and efficient logging is very important in 
server apps, strange you don't know it.

> Why is this so difficult to understand?

Yeah, why? ;-)

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 11:51 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> Avalon is a server framework. You mean that logging is not to be a
> concern of a server framework?

Correct. Very much so. If Avalon was a logging framework, then logging would
be a concern. But it isn't. As you say, it is a server framework.
Commons-logging is a logging framework. Avalon (as a server framework)
should depend on commons-logging for its logging.

Imagine if commons-logging tried to export Avalon interfaces. that would be
just as stupid as avalon exporting commons-logging (or avalon-logging)
interfaces for people to depend on.

Why is this so difficult to understand?

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Jon Scott Stevens wrote:
> on 2002/12/10 5:22 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
>>There are technical points that Avalon logging pursues that are missing
>>from Commons Logging.
> 
> The point being that you should work in the Commons project to get them
> added instead of coming up with another logging interface and using that.
> Sure, you have the right to fork, but it shouldn't be the first thing that
> happens.

We didn't fork. Any commons project should work with all projects to 
find a common solution that all would switch to. This hasn't been done, 
and is happening now with configuration. We have a Configuration system, 
and commons configuration is still going its way. We'll see if it comes 
out similarly to logging.

>>And it's sensible that projects that use the
>>Avalon framework use the Logging that is part of the framework, I don't
>>see the "pushing" part you refer to.
> 
> It is not sensible to me. It is a terrible idea in fact. If you say you care
> about good design, then you won't base all of your decisions around using
> Avalon for things that really don't belong in Avalon.

Avalon is a server framework. You mean that logging is not to be a 
concern of a server framework?

Oh well...

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 5:22 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> It's of no help in trying to find a common solution

That is why they called the project "Jakarta Commons".

> There are technical points that Avalon logging pursues that are missing
> from Commons Logging.

The point being that you should work in the Commons project to get them
added instead of coming up with another logging interface and using that.
Sure, you have the right to fork, but it shouldn't be the first thing that
happens.

> And it's sensible that projects that use the
> Avalon framework use the Logging that is part of the framework, I don't
> see the "pushing" part you refer to.

It is not sensible to me. It is a terrible idea in fact. If you say you care
about good design, then you won't base all of your decisions around using
Avalon for things that really don't belong in Avalon.

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Jon Scott Stevens wrote:
> on 2002/12/10 4:59 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> 
>>>I'm strongly -1 on using the Avalon logging interface because this is
>>>something that belongs in Commons. There is no reason for Avalon to have
>>>*yet another logging interface* when commons-logging is available.
>>
>>Actually, the Avalon interface IIRC is older.
> 
> I don't care who is older. Like everyone else around here, I care about
> where things live.

I care about making the correct design decisions.

>>It has not been reimplemented, rather the opposite, and your attitude is
>>of no help.
> 
> My attitude is of no help towards your cause. Of course it isn't. We are not
> agreeing on this topic. I love how you imply that because I'm disagreeing
> with you that I'm no help.

It's of no help in trying to find a common solution, I have no cause to 
pursue other than solving this divide. I have already proposed to use 
the Commons Logging interface in the past, and it got voted down. In 
this perspective, by not wanting to address design concerns and pursuing 
a "I care about where things live" vision instead, you are not helping.

>>Anyway, I will propose again to our list that we use the Logger
>>interface somehow in the new Avalon 5, or try somehow to reunite the scism.
> 
> +1
> 
> Avalon should drop whatever duplicate logging interfaces it has and adopt
> commons-logging. The same should go for Turbine and every other Jakarta
> application out there.

Your opinion.

>>Remember though that the fact that commons strongly opposes anything
>>that has even a remote hint about avalon framework really doesn't help.
> 
> It does not help YOU! That isn't my problem. Don't think you can come around
> here and try to push your Avalon stuff on the people here as a way to get
> around the Commons project disagreeing with Avalon. That isn't my problem
> and I don't really care to get into those wars.

;-)   I'm not pushing anything, it's not /my/ avalon stuff, and I don't 
want to get round anything.

There are technical points that Avalon logging pursues that are missing 
from Commons Logging. And it's sensible that projects that use the 
Avalon framework use the Logging that is part of the framework, I don't 
see the "pushing" part you refer to.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
> > Anyway, I will propose again to our list that we use the Logger
> > interface somehow in the new Avalon 5, or try somehow to reunite the scism.
> 
> +1
> 
> Avalon should drop whatever duplicate logging interfaces it has and adopt
> commons-logging. The same should go for Turbine and every other Jakarta
> application out there.

Agreed! 

If I could implement LogEnabled, and have the logging facade provided by
the container be a commons-logging class, that would meet everyone's
needs I think. We could share the interfaces through commons, but choose
not to use the LogFactory class (and its associated discovery stuff) as
we wish.

-- jt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 4:59 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

>> I'm strongly -1 on using the Avalon logging interface because this is
>> something that belongs in Commons. There is no reason for Avalon to have
>> *yet another logging interface* when commons-logging is available.
> 
> Actually, the Avalon interface IIRC is older.

I don't care who is older. Like everyone else around here, I care about
where things live.

> It has not been reimplemented, rather the opposite, and your attitude is
> of no help.

My attitude is of no help towards your cause. Of course it isn't. We are not
agreeing on this topic. I love how you imply that because I'm disagreeing
with you that I'm no help.

> Anyway, I will propose again to our list that we use the Logger
> interface somehow in the new Avalon 5, or try somehow to reunite the scism.

+1

Avalon should drop whatever duplicate logging interfaces it has and adopt
commons-logging. The same should go for Turbine and every other Jakarta
application out there.

> Remember though that the fact that commons strongly opposes anything
> that has even a remote hint about avalon framework really doesn't help.

It does not help YOU! That isn't my problem. Don't think you can come around
here and try to push your Avalon stuff on the people here as a way to get
around the Commons project disagreeing with Avalon. That isn't my problem
and I don't really care to get into those wars.

I am starting to like the idea of Avalon for some stuff...mostly as a
container, but I'm certainly not going to pick an Avalon detail over a
Commons detail.

If you think of things like a tree...commons is the top and we
(Turbine/Avalon) are the branches. We should both be picking from the top.

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Jon Scott Stevens wrote:
> on 2002/12/10 3:16 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> 
>>I'm not suggesting that you use Logkit, but the Avalon logging interface
>>and facades.
> 
> 
> I'm strongly -1 on using the Avalon logging interface because this is
> something that belongs in Commons. There is no reason for Avalon to have
> *yet another logging interface* when commons-logging is available.

Actually, the Avalon interface IIRC is older.

> Everyone who doesn't want to standardize on a single logging implementation
> should at least standardize on a single logging interface and that should be
> commons-logging. If you have a problem with commons-logging, then it should
> be addressed there, not re-implemented somewhere else.

It has not been reimplemented, rather the opposite, and your attitude is 
of no help.

Anyway, I will propose again to our list that we use the Logger 
interface somehow in the new Avalon 5, or try somehow to reunite the scism.

Remember though that the fact that commons strongly opposes anything 
that has even a remote hint about avalon framework really doesn't help.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
On Thu, 2002-12-12 at 10:06, Henning P. Schmiedehausen wrote:
> James Taylor <ja...@jamestaylor.org> writes:
> 
> >The problem is that the avalon way and the commons-logging way are quite
> >different. 
> 
> >The avalon way inverts control so that the container provides logging to
> >each component. In the commons-logging / log4j way each class requests a
> >logger from a factory (ick).
> 
> I still see no reason, why a commons-logging factory should not be able
> to return the container supplied logger reference to the application.
> 
> like this:
> 
> "Container provides application with logging facility".
> 
> 
> "Turbine starts up. Notices, that it got a logging facility from the
>  container. Registers it".
> 
> User uses "Log log = LogFactory.getLog("foo")". Commons Logging
> has been configured to return a log object provided by the container-
> provided log facility.

Yes, but I want to provide a different log facility for two different
components of my app. No can do unless the component is kind enough to
request a specific category and I configure off that. The IOC way is
more elegant.

> 
> And the user never notices that he has not a "Turbine configured log4j
> logging" but a "container provided enterprise logging facility".
> 
> 	Regards
> 		Henning


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
James Taylor <ja...@jamestaylor.org> writes:

>The problem is that the avalon way and the commons-logging way are quite
>different. 

>The avalon way inverts control so that the container provides logging to
>each component. In the commons-logging / log4j way each class requests a
>logger from a factory (ick).

I still see no reason, why a commons-logging factory should not be able
to return the container supplied logger reference to the application.

like this:

"Container provides application with logging facility".


"Turbine starts up. Notices, that it got a logging facility from the
 container. Registers it".

User uses "Log log = LogFactory.getLog("foo")". Commons Logging
has been configured to return a log object provided by the container-
provided log facility.

And the user never notices that he has not a "Turbine configured log4j
logging" but a "container provided enterprise logging facility".

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.

Ceki Gülcü wrote:
> At 11:20 11.12.2002 +0100, Nicola Ken Barozzi wrote:
> 
>> This does not mean that all classes in your app must be Avalon 
>> Components, nor that every single class of your app must be 
>> LogEnabled. The Container gives the Component the Logger. Then it can 
>> supply it to its children how it wishes.
> 
> Ok, so Avalon components are LogEnabled and pass their logger (as 
> supplied by the Container) to other objects (their children) not 
> necessarily LogEnabled. Thus, the Avalon Components and their children 
> must have a mechanism of sharing their loggers. Fine.  How do you deal 
> with classes that do not follow this model? That is the question.

It's not necessarily an Avalon concern. Cocoon actually decides to pass 
the logger to the children. POI (which BTW uses Commons Logging ATM) 
instead has its own static method that its classes used.

I'm not personally aware of any Avalon recomendation on this matter, 
these are my personal thoughts.

>> I have been doing logging in Jakarta POI. Do you think that I use 
>> Logenabled in the whole memory tree objects of the POI docs? No way! ;-)
> 
> No, but the developers of POI, you in this case, are aware of the logger 
> sharing model. What do you do with code that is oblivious to the model? 
> This is a real question, not a slimy trick.

This is true with any model. If you have jarred code that logs as it 
wants, you cannot manage it. If you have an app that uses SuckyLogger 
and your app uses Log4J, you cannot manage SuckyLogger stuff with Log4J, 
the same with Commons logging.

I say that every component should be able to recieve a logger. Some say 
that every app should get it itself from commons Logging. The fact is 
that you cannot control something that has no hooks in it.

Don't know if this is the answer to the question, I'm not sure I 
understand it 100%, please rectify if it's the case.

>> But if I use POI in Cocoon, I would like to be able to set the logger 
>> on it easily, and make all the logging be delegated easily to Cocoon.
>> POI should be a Component, not necessarily its internals.
> 
> You have already made your point clear.

:-)  Good, it's not always easy.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by James Taylor <ja...@jamestaylor.org>.
On Thu, 2002-12-12 at 12:20, Henning P. Schmiedehausen wrote:
> Kurt Schrader <ks...@karmalab.org> writes:
> 
> >On Thu, 12 Dec 2002, Henning P. Schmiedehausen wrote:
> 
> >> to make this clear, once for all,
> >>
> >> I DON'T _WANT_ TO USE THEIR LIFECYCLE!
> 
> >Can you make it clear why you don't want to use their lifecycle?
> 
> Because, _I_ personally would prefer a "commons-lifecycle" package
> where all the projects that _want_ to use lifecycle interfaces can
> meet on "common ground".
> 
> I personally fear that once other projects start to use the "xxx-lifecycle
> classes", and the developers of the "xxx" project decide that they want
> to change something on these classes, they might not even bother to ask
> their external users. Mainly because it is the "xxx-lifecycle" package.

Perhaps it is only _my_ bad experience with commons, but I feel avalon
framework is much more likely to remain a stable API then most of
commons. Avalon is and has always been intended for use by other
projects, and has done an excellent job at it. Look at all the pain the
are going through to provide a migration path from ECM to more modern
containers -- backward compatibility is important to them.

Contrast this to commons where things seem to change constantly. I have
two versions of commons-lang in my classpath, as well as still having
commons-util, just to provide the same services under different names
depending on what is using them.

> Commons IMHO was always intended to be a meeting ground for all
> jakarta code developers from all jakarta projects. Here, packages are
> a lot less loaded with political or project specific issues, mainly
> because they're _intended_ to be used in lots of different projects.

> The attitude of some Avalon developers towards IoC and CoC pattern
> makes me shiver.

Please elaborate.


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Kurt Schrader <ks...@karmalab.org> writes:

>On Thu, 12 Dec 2002, Henning P. Schmiedehausen wrote:

>> to make this clear, once for all,
>>
>> I DON'T _WANT_ TO USE THEIR LIFECYCLE!

>Can you make it clear why you don't want to use their lifecycle?

Because, _I_ personally would prefer a "commons-lifecycle" package
where all the projects that _want_ to use lifecycle interfaces can
meet on "common ground".

I personally fear that once other projects start to use the "xxx-lifecycle
classes", and the developers of the "xxx" project decide that they want
to change something on these classes, they might not even bother to ask
their external users. Mainly because it is the "xxx-lifecycle" package.

Commons IMHO was always intended to be a meeting ground for all
jakarta code developers from all jakarta projects. Here, packages are
a lot less loaded with political or project specific issues, mainly
because they're _intended_ to be used in lots of different projects.

The attitude of some Avalon developers towards IoC and CoC pattern
makes me shiver.

But nevertheless, I'd think that actually going along a lifecycle with
Turbine is a good thing. And if we can't find a common(s) ground (pun
intended) then we better go with Avalon than cooking up our own
o.a.turbine.lifecycle interfaces.

As I said, everything above is IMHO.

	Regards
		Henning



-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: logging (was: Re: [vote] plans for turbine 2.3)

Posted by Ceki Gülcü <ce...@qos.ch>.
At 11:34 12.12.2002 -0600, Stephen Haberman wrote:
>I remember specifically at least one of the committers mentioning they
>have many thousands of lines of production system code based on the
>Avalon framework, as I'm going to assume they all do, and they have no
>intention of breaking that willy-nilly.
>
>Also note that they've done releases of Avalon specifically in response
>to the needs of at least Cocoon and perhaps James, depending on how well
>my memory is. So they are responsive to other communities.
>
>My impression was that they really enjoy other people using their stuff
>and are willing to sacrifice to keep from endangering that prospect
>(e.g. several of the Avalon people are lurking here on Turbine as they
>really look forward to spreading their good word to more of Jakarta).

I share the same impression. I will not be surprised to see Avalon do
well as a top-level Apache project.

Just my 2 cents from the peanut gallery.

>- Stephen

--
Ceki



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
"Stephen Haberman" <st...@chase3000.com> writes:

>Also, I lurked on Avalon-dev for awhile and will say, from what I

Ok, will do that, too. Sigh. Must. Stop. Sleeping.  :-)

>gathered, they are sticklers for backwards-compatibility. To the point
>where it kills some of the initiatives they want to do with the code
>base.

Hm. Maybe my attitude towards Avalon is unfairly biased because it
comes from developers who I found that they have "jeden Monat eine
neue Sau durchs Dorf getrieben" (... "that they came up with an new
and cunning plan every month") and never really released any stable
Turbine code (finally Martin did which I still consider a feat getting
things like your build system broken in middle-development.).

I honestly promise to take a deeper look at Avalon than I already have
and if we can really get more synergies here, I'm more than happy
to use their code.

After all, I came to this Turbine because, I wanted to write
applications for a framework, not write a framework for applications. :-)

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: logging (was: Re: [vote] plans for turbine 2.3)

Posted by Stephen Haberman <st...@chase3000.com>.
 
> If I (or any other Turbine developer) has no real chance to influence
> this, I'm afraid that at one point, some (well discussed and thought
> out change, that makes lots of sense in Avalon) leaves our user base
> with a broken or non-working Turbine. That's what I want to avoid.

What might be a good idea is to talk to some of the James developers.
Their project, as I understand it, is completely Avalon based and yet,
again, as I understand it, not many of them are active in the Avalon
community itself.

Their input on how the Avalon community has treated the issues they as a
separate project run into would provide some beneficial, be it either
comforting or discomforting, insight.

Also, I lurked on Avalon-dev for awhile and will say, from what I
gathered, they are sticklers for backwards-compatibility. To the point
where it kills some of the initiatives they want to do with the code
base.

I remember specifically at least one of the committers mentioning they
have many thousands of lines of production system code based on the
Avalon framework, as I'm going to assume they all do, and they have no
intention of breaking that willy-nilly.

Also note that they've done releases of Avalon specifically in response
to the needs of at least Cocoon and perhaps James, depending on how well
my memory is. So they are responsive to other communities.

My impression was that they really enjoy other people using their stuff
and are willing to sacrifice to keep from endangering that prospect
(e.g. several of the Avalon people are lurking here on Turbine as they
really look forward to spreading their good word to more of Jakarta).

- Stephen


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
James Taylor <ja...@jamestaylor.org> writes:

>Does it really matter that much that it is not in commons?

Look James, I got brushed down in a project which is currently under
"Turbine" as a Turbine developer with the words "You don't have a vote
in xxx-land". What might I expect in a project where I haven't
contributed a single bit yet (and my time probably does not permit to
do)?

If I (or any other Turbine developer) has no real chance to influence
this, I'm afraid that at one point, some (well discussed and thought
out change, that makes lots of sense in Avalon) leaves our user base
with a broken or non-working Turbine. That's what I want to avoid.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: logging (was: Re: [vote] plans for turbine 2.3)

Posted by Stephen Haberman <st...@chase3000.com>.
> Does it really matter that much that it is not in commons?

I agree, the Avalon lifecycles are perfectly fine where they are now. A
project does not have to be in commons for it to suddenly become
available for other projects to use.

Nicola's comment about turbine-commons was also right on the point.

Commons-logging does not have any lifecycle support. The Turbine
community wants to move towards lifecycles as there seems to be the
general consensus that it makes life much easier. The Avalon LogEnabled
interfaces give us logging via a lifecycle _regardless_ on the backend
implementation of the logging, which can very well be
commons-logging/log4j, as I understand it.

Also note that I'm operating under the assumption that I suspect not
every has realized, but that in the current T2/Torque/Fulcrum,
commons-logging/log4j is the way to go.

(So in reference to your changed logging code, Henning, I think we'd all
agree to go ahead and commit it as it'll be an improvement over the
existing approach).

Avalon/LogEnabled/etc. only comes into play when we start talking about
the potential re-write with T3/T4/Summit, which will, as proposed, be
completely Avalon-based.

And I'd think that along with the T3/T4/Summit release in it's
LogEnabled/Avalon-ized form, we'd have a Fulcrum do a corresponding
major release where even though the current version used in T2/Torque it
is not at all coupled to Avalon (e.g. it uses commons-logging/log4j),
the next mature release would be in the form of Avalon components (and
so use the lifecycles which is the whole point to this Avalon/T4/Summit
thing) that inherently use LogEnabled.

(Note that if the next version of Fulcrum gets merged into the Avalon
component repository or perhaps better yet becomes wrappers in the
Avalon component repository for commons-xxx projects that are not
Avalon-coupled, I think it'd be for all the better).

- Stephen


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by James Taylor <ja...@jamestaylor.org>.
> to make this clear, once for all,
> 
> I DON'T _WANT_ TO USE THEIR LIFECYCLE!
> 
> But it is the only really useful lifecycle interface that there is around,
> so we _WILL_ use their lifecycle.
> 
> MUCH MUCH MORE, I'd prefer that they donate their lifecycle classes to
> commons and then Avalon uses commons-lifecycle. AND WE USE COMMONS-LIFECYCLE!

Does it really matter that much that it is not in commons?

Avalon-Framework has a strong community where it is, it has been through
several releases, and has excellent policy. Hang out on avalon-dev for a
while and you get a great feel for how stable the API really is --
they'll argue for months before deprecating something =]

Moving it to commons would just upset all that. I see no reason why
EVERYTHING shared has to be in commons. Lots of jakarta (and XML)
projects use the avalon framework right where it is, why make things
more difficult than they have to be?

-- jt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by Kurt Schrader <ks...@karmalab.org>.
On Thu, 12 Dec 2002, Henning P. Schmiedehausen wrote:

> to make this clear, once for all,
>
> I DON'T _WANT_ TO USE THEIR LIFECYCLE!

Can you make it clear why you don't want to use their lifecycle?

-Kurt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
James Taylor <ja...@jamestaylor.org> writes:

>For _that_ case I would like to see the avalon idiom and interfaces
>used, especially when we are already planning on using their lifecycle. 

Look,

to make this clear, once for all,

I DON'T _WANT_ TO USE THEIR LIFECYCLE!

But it is the only really useful lifecycle interface that there is around,
so we _WILL_ use their lifecycle.

MUCH MUCH MORE, I'd prefer that they donate their lifecycle classes to
commons and then Avalon uses commons-lifecycle. AND WE USE COMMONS-LIFECYCLE!

How about that? Avalon people? Jon? Martin?

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by James Taylor <ja...@jamestaylor.org>.
> turbine is not only the turbine.jar .. we use velocity, 
> commons-packages, ...
> all this packages use different logging  packages (log4j, 
> commons-logging, ...)
> 
> so it will not be possible to have one single logging mechanism for all 
> of them ...
> i think this is a problem which should be discussed on the general list 

Yes, we are stuck with configuring many logging mechanisms now, and will
continue to be.

I thought the more immediate concern was logging WITHIN turbine and its
associated services.

For _that_ case I would like to see the avalon idiom and interfaces
used, especially when we are already planning on using their lifecycle. 

It doesn't _matter_ what the actual logging mechanism is in such a case.
I'm sure any of us with a reasonable complex application will be stuck
configuring 3 or more anyway.

Tangent: I'm all for sharing code, but what makes something in commons
better than something in avalon-framework? Commons is pretty much
unregulated, there is no reason to believe any of the commons projects
are the best solution. Avalon-Framework has been through a number of
revisions and has been fought over and refined for much longer than
anything in commons. Lets make a decision that is technically best, and
then work out the politics.





--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Martin Poeschl <mp...@marmot.at> writes:

>thanx to Ceki and Nicola for their posts!

>after reading all the post to the logging question it's still not clear 
>how we can solve the problem.

>turbine is not only the turbine.jar .. we use velocity, 
>commons-packages, ...
>all this packages use different logging  packages (log4j, 
>commons-logging, ...)

Ok, even on the risk of making myself being unpopular again...

We have three major building blocks for Turbine applications ATM. These
are "Turbine", "Velocity" and "Torque". 

"Turbine" currently uses the LogService. This is definitely going to die.

"Velocity" uses some possible loggers, one of them is log4j

"Torque" uses log4j

So I _still_ propose to integrate my 10 lines change to Turbine.java
which activates log4j as the primary logging facility for Turbine,
initialized commons-logging to actually use log4j and the convert
every Log.<xxx> reference in the Turbine code to log.<xxx>

I do have this code not only here and working bug actually deployed
and in production. Contrary to the belief of some Avalon and log4j
developers, none of my Tomcats has melted down (yet... :-) )

This might not be the ideal solutions but IMHO it is a working
solution until the religious wars about CoC and IoC models cease.

	Regards
		Henning

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: logging (was: Re: [vote] plans for turbine 2.3)

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Martin Poeschl wrote:
> thanx to Ceki and Nicola for their posts!
> 
> after reading all the post to the logging question it's still not clear 
> how we can solve the problem.
> 
> turbine is not only the turbine.jar .. we use velocity, 
> commons-packages, ...
> all this packages use different logging  packages (log4j, 
> commons-logging, ...)
> 
> so it will not be possible to have one single logging mechanism for all 
> of them ...
> i think this is a problem which should be discussed on the general list 
> .. it would be nice to have one logging mechanism used by all apache 
> java projects (it's not only jakarta)
> on the other hand i'm sure this will end in an endless discussion 
> without any solution :-((

I think that given the latest changes in Avalon and the use if Avalon by 
Turbine, we can reasonably come up with a solution.

> i really like log4j, but i understand the need to be able to use the 
> jdk1.4 stuff (i still don't understand why sun didn't add log4j to the 
> jdk .. but that's another story)
> so commons-logging seems to solve this problem ...
> as Ceki ponted out there are other problems with commons-logging ..
> 
> so maybe we should simple vote for log4j or commons-logging or avalon 
> logging ... or logging to System.out ;-)

I'd say, let's work to find a real solution that will end this 
anachronistic divide that has no technical base.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


logging (was: Re: [vote] plans for turbine 2.3)

Posted by Martin Poeschl <mp...@marmot.at>.
thanx to Ceki and Nicola for their posts!

after reading all the post to the logging question it's still not clear 
how we can solve the problem.

turbine is not only the turbine.jar .. we use velocity, 
commons-packages, ...
all this packages use different logging  packages (log4j, 
commons-logging, ...)

so it will not be possible to have one single logging mechanism for all 
of them ...
i think this is a problem which should be discussed on the general list 
.. it would be nice to have one logging mechanism used by all apache 
java projects (it's not only jakarta)
on the other hand i'm sure this will end in an endless discussion 
without any solution :-((

i really like log4j, but i understand the need to be able to use the 
jdk1.4 stuff (i still don't understand why sun didn't add log4j to the 
jdk .. but that's another story)
so commons-logging seems to solve this problem ...
as Ceki ponted out there are other problems with commons-logging ..

so maybe we should simple vote for log4j or commons-logging or avalon 
logging ... or logging to System.out ;-)

martin



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Ceki Gülcü <ce...@qos.ch>.
At 11:20 11.12.2002 +0100, Nicola Ken Barozzi wrote:
>This does not mean that all classes in your app must be Avalon Components, 
>nor that every single class of your app must be LogEnabled. The Container 
>gives the Component the Logger. Then it can supply it to its children how 
>it wishes.

Ok, so Avalon components are LogEnabled and pass their logger (as supplied 
by the Container) to other objects (their children) not necessarily 
LogEnabled. Thus, the Avalon Components and their children must have a 
mechanism of sharing their loggers. Fine.  How do you deal with classes 
that do not follow this model? That is the question.

>I have been doing logging in Jakarta POI. Do you think that I use 
>Logenabled in the whole memory tree objects of the POI docs? No way! ;-)

No, but the developers of POI, you in this case, are aware of the logger 
sharing model. What do you do with code that is oblivious to the model? 
This is a real question, not a slimy trick.

>But if I use POI in Cocoon, I would like to be able to set the logger on 
>it easily, and make all the logging be delegated easily to Cocoon.
>POI should be a Component, not necessarily its internals.

You have already made your point clear.

>--
>Nicola Ken Barozzi                   nicolaken@apache.org

--
Ceki



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Ceki Gülcü wrote:
> At 22:14 10.12.2002 -0500, James Taylor wrote:

[...]

> It is actually feasible. Please read my "Supporting the log4j 
> RepositorySelector in Servlet Containers" document at 
> http://qos.ch/logging/sc.html.
> 
>> However, with the IOC / avalon idiom, I can simply provide each of the
>> two views with a different logger -- problem solved.
> 
> The IOC approach makes a lot of sense except that it supposes a world 
> where there is a container in charge of logging and nice little obedient 
> components implementing the LogEnabled interface -- an assumption that 
> does not necessarily extend to reality. 

All Avalon Components should implement LogEnabled and be obedient in 
this sense.
Actually, the lifecycle is
all about creating order (Separation of Concerns) and correct chain
of command (Inversion of Control).

This does not mean that all classes in your app must be Avalon 
Components, nor that every single class of your app must be LogEnabled. 
The Container gives the Component the Logger. Then it can supply it to 
its children how it wishes.

I have been doing logging in Jakarta POI. Do you think that I use 
Logenabled in the whole memory tree objects of the POI docs? No way! ;-)

But if I use POI in Cocoon, I would like to be able to set the logger on 
it easily, and make all the logging be delegated easily to Cocoon.
POI should be a Component, not necessarily its internals.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Ceki Gülcü <ce...@qos.ch>.
At 22:14 10.12.2002 -0500, James Taylor wrote:
>An example which will hopefully make my argument more clear:
>
>Consider two turbine views running withing the same servlet container. I
>want all logging for view A to be logged to a database, and all the logs
>for the other view to a specific file.
>
>This is not feasible with the factory / singleton logging idiom since
>the logger is tied to the class.

It is actually feasible. Please read my "Supporting the log4j 
RepositorySelector in Servlet Containers" document at 
http://qos.ch/logging/sc.html.

>However, with the IOC / avalon idiom, I can simply provide each of the
>two views with a different logger -- problem solved.

The IOC approach makes a lot of sense except that it supposes a world where 
there is a container in charge of logging and nice little obedient 
components implementing the LogEnabled interface -- an assumption that does 
not necessarily extend to reality. I think that Jon made very similar 
points a little earlier.

>Does that make any sense?

It certainly does.

>Thanks,
>James

--
Ceki



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
James Taylor <ja...@jamestaylor.org> writes:

>Oh my god man, could this be more fugly?

I understand that your "I point the component at the logger that _I_
want" approach is for your scenario much simpler. I really do. I also
do understand the implications of the IoC concept.

I was giving you just some whipped-up example code to illustrate that
c-l does not hardcode the loggers on the class name. It was not a
pattern you should follow. I just wanted to give you the basic idea.
Of course you can also pass down a preloaded o.a.c.l.Log object. Where
is the problem?

You did seem to be keen on misunderstanding me. So do. Fine.

In 90+% of all applications ever written for Turbine, the normal
logging scenario will be "one log file containing runtime messages and
errors".  Maybe two (application and turbine).

Dynamic reconfiguration and user-based logging over arbitrary classes
IMHO don't come into play here. If you want to go Enterprise
solutions, I'd think that Jasons' Plexus/Summit will be soooo much
better than our boring bread-and-butter servlet based Turbine-2
stuff. But we will try to release on a regular base and support our
user base.

	Regards
		Henning

>> Huh?
>> 
>> consider a user specific property called "wantDbLogging".
>> 
>>   Boolean wantDbLogging = (Boolean) data.getUser().getTemp("wantDbLogging");
>>   Log log = LogFactory.getLogger(wantDbLogging.booleanValue() ? "dbLog" : "normalLog");
>> [...]
>>   log.debug("foo");
>> 
>> I can even flip that boolean switch at runtime. 

>So now all my components requiring logging have to be written to look at
>that flag? Also, I have added coupling tun Rundata, User, and thus all
>of turbine. Wahoo!

>> (actual method names might vary slightly but you get the point).
>> 
>> Commons Logging is _NOT_ "write everything onto the class name based
>> logger". This is just the most convenient approach.

>Compare this to writing my class as a black box LogEnabled. If I want
>the bugger to log to X, Y, Z, or whatever combination, I just hand it
>the appropriate Logger, which might wrap commons-logging, log4j,
>System.out, the bit bucket, or my evil twin's email account.

>The component doesn't CARE whether I want DB logging -- why should it! 

>-- jt


>--
>To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>

-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
Oh my god man, could this be more fugly?

> Huh?
> 
> consider a user specific property called "wantDbLogging".
> 
>   Boolean wantDbLogging = (Boolean) data.getUser().getTemp("wantDbLogging");
>   Log log = LogFactory.getLogger(wantDbLogging.booleanValue() ? "dbLog" : "normalLog");
> [...]
>   log.debug("foo");
> 
> I can even flip that boolean switch at runtime. 

So now all my components requiring logging have to be written to look at
that flag? Also, I have added coupling tun Rundata, User, and thus all
of turbine. Wahoo!

> (actual method names might vary slightly but you get the point).
> 
> Commons Logging is _NOT_ "write everything onto the class name based
> logger". This is just the most convenient approach.

Compare this to writing my class as a black box LogEnabled. If I want
the bugger to log to X, Y, Z, or whatever combination, I just hand it
the appropriate Logger, which might wrap commons-logging, log4j,
System.out, the bit bucket, or my evil twin's email account.

The component doesn't CARE whether I want DB logging -- why should it! 

-- jt


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
James Taylor <ja...@jamestaylor.org> writes:

>Consider two turbine views running withing the same servlet container. I
>want all logging for view A to be logged to a database, and all the logs
>for the other view to a specific file.

>This is not feasible with the factory / singleton logging idiom since
>the logger is tied to the class.

Huh?

consider a user specific property called "wantDbLogging".

  Boolean wantDbLogging = (Boolean) data.getUser().getTemp("wantDbLogging");
  Log log = LogFactory.getLogger(wantDbLogging.booleanValue() ? "dbLog" : "normalLog");
[...]
  log.debug("foo");

I can even flip that boolean switch at runtime. 


(actual method names might vary slightly but you get the point).

Commons Logging is _NOT_ "write everything onto the class name based
logger". This is just the most convenient approach.

[..]
>Does that make any sense? 

No, sorry. :-) Or I didn't understand you. 

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by John McNally <jm...@collab.net>.
The ability to pass a specific logger to a given component is a good
thing.  I have had some difficulties with the singleton approach as
well.  Consider the case of a turbine application running in a servlet
container (tomcat) which has other webapps running as well, and I do not
have complete control over the location of all the jars related to the
servlet container.

I have my logging pretty much to my liking except that I have a
component which uses beanutils which is in tomcat/commons/lib.  I have
log4j configured and the jar living within my webapps's lib dir.  But
the beanutils jar is in the parent classloader's path.  So logging in
beanutils is not using log4j because that is unavailable to
commons-logging.  I basically punted on configuring the logging for
beanutils, its going to catalina_out.log (system.out).  Maybe this is
easy to solve, but its not a high enough priority that I have tried.

Now someone comes along and adds a log4j jar to common/lib.  Now
beanutils will log according to the rules of whichever webapp first
configures a logger that will work for "org.apache.commons.beanutils"
Is there a way around this kind behavior and using the singleton
approach?  

To top all this off (and this might not be related to the topic), my
application now is logging according my configuration (it is populating
the correct log files.)  But it is also sending all logs statements to
system.out.  I now get to try to figure out what changed and I am aware
of nothing changing within my app that should affect this.  So my
inclination is to blame the new jar in common/lib and classloader
issues.

Another specific example similar to James' and also does not involve
classloaders, let's say I have two fulcrum services that are using
commons-collections.  If I have log4j configured for
"o.a.t.fulcrum.service1" and "o.a.t.fulcrum.service2" so that they are
in different files.  The logging for o.a.commons.collections cannot go
to the appropriate logfile for the service that is using it.  It could
be configured to write to one or the other, but since that would mean in
some cases the log entries might be inappropriate, I am left with the
suboptimal solution that I must have a different log file for
commons-collections.

john mcnally

On Tue, 2002-12-10 at 19:14, James Taylor wrote:
> > > The avalon way inverts control so that the container provides logging to
> > > each component. In the commons-logging / log4j way each class requests a
> > > logger from a factory (ick).
> > 
> > IOC doesn't work for logging because you may be integrating parts that are
> > not components which need their own logging. The Factory (singleton) design
> > *for logging* is the right way to go.
> 
> An example which will hopefully make my argument more clear:
> 
> Consider two turbine views running withing the same servlet container. I
> want all logging for view A to be logged to a database, and all the logs
> for the other view to a specific file.
> 
> This is not feasible with the factory / singleton logging idiom since
> the logger is tied to the class.
> 
> However, with the IOC / avalon idiom, I can simply provide each of the
> two views with a different logger -- problem solved.
> 
> Does that make any sense? 
> 
> Thanks,
> James
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
> > The avalon way inverts control so that the container provides logging to
> > each component. In the commons-logging / log4j way each class requests a
> > logger from a factory (ick).
> 
> IOC doesn't work for logging because you may be integrating parts that are
> not components which need their own logging. The Factory (singleton) design
> *for logging* is the right way to go.

An example which will hopefully make my argument more clear:

Consider two turbine views running withing the same servlet container. I
want all logging for view A to be logged to a database, and all the logs
for the other view to a specific file.

This is not feasible with the factory / singleton logging idiom since
the logger is tied to the class.

However, with the IOC / avalon idiom, I can simply provide each of the
two views with a different logger -- problem solved.

Does that make any sense? 

Thanks,
James


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
> > The avalon way inverts control so that the container provides logging to
> > each component. In the commons-logging / log4j way each class requests a
> > logger from a factory (ick).
> 
> IOC doesn't work for logging because you may be integrating parts that are
> not components which need their own logging. The Factory (singleton) design
> *for logging* is the right way to go.
> 
> If you are going to contest this statement in anyway, tell me how you can
> realistically integrate other bits of code that you may not even have the
> source code for and that are not components and still get proper logging.

If those components implement LogEnabled then I can give them a logger
of my choice.

In any other case, I have to cope with what the component does, which I
do now anyway, nothing new. 

Besides, nothing about using avalon logging in turbine and its
components prevents someone from using stuff that uses commons logging
as well, they just need to be prepared to configure it as well. 

And for the other direction (our components in XYZ app) they just hand
the component a Logger which wraps a commons.logging.Log or whatever. No
sweat.

> > So, I personally am leaning toward the avalon way because it seems more
> > flexible and reduces coupling between components and a specific logging
> > layer. 
> 
> There is no coupling between components and a specific logging layer if you
> use commons-logging. That is the whole point of commons-logging in the first
> place!

There is coupling to the LogFactory though. In the avalon model a given
LogEnabled class (need not be a component) only uses interfaces for
logging, and only what it is provided. In the commons logging model each
class must directly access another concrete class to get a logger.


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 5:09 PM, "James Taylor" <ja...@jamestaylor.org> wrote:

> The problem is that the avalon way and the commons-logging way are quite
> different. 
> 
> The avalon way inverts control so that the container provides logging to
> each component. In the commons-logging / log4j way each class requests a
> logger from a factory (ick).

IOC doesn't work for logging because you may be integrating parts that are
not components which need their own logging. The Factory (singleton) design
*for logging* is the right way to go.

If you are going to contest this statement in anyway, tell me how you can
realistically integrate other bits of code that you may not even have the
source code for and that are not components and still get proper logging.

> So, I personally am leaning toward the avalon way because it seems more
> flexible and reduces coupling between components and a specific logging
> layer. 

There is no coupling between components and a specific logging layer if you
use commons-logging. That is the whole point of commons-logging in the first
place!

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
On Tue, 2002-12-10 at 18:49, Jon Scott Stevens wrote:
> on 2002/12/10 3:16 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> > I'm not suggesting that you use Logkit, but the Avalon logging interface
> > and facades.
> 
> I'm strongly -1 on using the Avalon logging interface because this is
> something that belongs in Commons. There is no reason for Avalon to have
> *yet another logging interface* when commons-logging is available.

The problem is that the avalon way and the commons-logging way are quite
different. 

The avalon way inverts control so that the container provides logging to
each component. In the commons-logging / log4j way each class requests a
logger from a factory (ick).

Really, there are two seperate problems. What logging layer gets used,
and how do loggers get provided to components. I think the avalon
logging facade provides a better solution to the latter, while being
just as good as commons-logging at the former (perhaps even a little
more flexible).

So, I personally am leaning toward the avalon way because it seems more
flexible and reduces coupling between components and a specific logging
layer. 

> Everyone who doesn't want to standardize on a single logging implementation
> should at least standardize on a single logging interface and that should be
> commons-logging. If you have a problem with commons-logging, then it should
> be addressed there, not re-implemented somewhere else.
> 
> -jon


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/10 3:16 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> I'm not suggesting that you use Logkit, but the Avalon logging interface
> and facades.

I'm strongly -1 on using the Avalon logging interface because this is
something that belongs in Commons. There is no reason for Avalon to have
*yet another logging interface* when commons-logging is available.

Everyone who doesn't want to standardize on a single logging implementation
should at least standardize on a single logging interface and that should be
commons-logging. If you have a problem with commons-logging, then it should
be addressed there, not re-implemented somewhere else.

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
James Taylor wrote:
> On Sun, 2002-12-08 at 18:15, Jon Scott Stevens wrote:
> 
>>on 2002/12/8 2:37 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
>>
>>
>>>What about Avalon logging? It's better suited for server-side logging,
>>>and has its own abstractions that make it possible to use as
>>>implementations log4j, logkit or whatever.
>>
>>I prefer that we use commons-logging.
>>
>>I'm -1 on Avalon logging as the dependency simply because I like to have the
>>choice as to which logging package I want to use.
> 
> 
> One thing we might consider is not requiring LogKit as the logger, but
> still use an IOC style of log configuration like in Avalon. In fact, we
> could still use the LogEnabled / Logger facade without restricting log
> configuration options.

Exactly.

Avalon Logging != Logkit

I'm not suggesting that you use Logkit, but the Avalon logging interface 
and facades. the Logger is part of the Avalon lifecycle, and it's very 
important, also from a security perspective, that it's given by the 
container, not gotten by the Component, and makes management of it easier.

> This may not be appropriate for t2 but I think it will make sense for
> t3. One of the nice things about IOC in this case is that the container
> can build / obtain the logger it provides in any way. Thus, for example,
> if you have a scenario where class loading discovery is not workable,
> you can eliminate it without having to touch the components, just the
> managers.

Exactly.

The simply -1 doesn't stand because using Avalon logging it's possible 
to use any logging implementation you wish.

Look in the framework:
http://cvs.apache.org/viewcvs/jakarta-avalon/src/java/org/apache/avalon/framework/logger/

AbstractLogEnabled.java
AvalonFormatter.java
ConsoleLogger.java
Jdk14Logger.java
Log4JLogger.java
LogEnabled.java
LogKit2AvalonLoggerAdapter.java
LogKitLogger.java
Logger.java
NullLogger.java


> -- jt, who is almost completely converted to avalon now.

:-)

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


RE: logging in turbine

Posted by Stephen Haberman <st...@chase3000.com>.
> For more short-term decisions (IIUC T2.3 != plexus), you probably want
> to base that decision on the long-term goal....but OTOH turbine atm is
> very much about singletons and 'active' components, which is very
> different from IoC with passive components. It might be easier to keep
> the current pattern for 2.3, and then it is probably easiest to use
> commons-logging.
> In which case it makes even more sense to have good commons-logging
> support in avalon for when T3 hits the streets.

Well put. I like the observation that:

- With T2, lots of classes are singletons and hence static logger
creation makes a lot of sense. In such a scenario, either log4j or
commons-logging is great.

- However, in T3/T4/Summit/Plexus/any-avalon-based-system, the singleton
design pattern is not implemented via static classes, but instead is
left up to the container to handle via the lookup/lifecycle interfaces
(correct?).

In this 2nd scenario, within a server framework (what Turbine and
Fulcrum were already moving towards becoming, but we're hoping to skip
to Avalon to avoid re-creating what they've already done), the server
framework handles creation of the loggers and gives it to the component,
there-by centralizing and decoupling logger configuration in a
context-sensitive manner, which, to me, is a good thing.

Granted it would be nice if every package we ever wanted to use in
Turbine used the same logging framework, as then we could use that; but
that's not going to happen. And even if it did, the setup of the static
loggers is not context-sensitive, as John McNally pointed out.

Given that both scenarios (static creation and IoC) are best practices
in their given environment, I think it only makes sense to use the given
best practice in the environment we're subscribing to. Assuming we go
ahead with an Avalon-based system, this would be the IoC-ized
LogEnabled.

Don't get me wrong, I think log4j/commons-logging are great and would
use one them in any non-Avalon system (I won't say which as the debate,
to me, is static vs. IoC).

But the Avalon guys have spent a long time designing their interfaces
and I think it's smart to trust that _within the context of Avalon_ what
they have provided and strongly suggest using is, to me, the way to go.

(Not that we should blindly follow their lead, questioning approaches is
always good, but I don't think we should be too touchy about the whole
ordeal. If any thing, this is just logging, and while, granted, logging
is really important, both scenarios will work and picking one or the
other will not be the determining factor that makes or breaks the next
Turbine.)

- Stephen


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


logging in turbine

Posted by Leo Simons <le...@apache.org>.
Hi gang,

it's the logging debate again! Whooh-hooh :) Long boring e-mail ahead.

up front summary:
- turbine does have influence over avalon
- java logging is a mess, let's not point fingers but figure out a
migration path
- logging in avalon follows the same IoC pattern as everything else
- it should be possible to do avalon-style logging using commons-logging
- I think T2.3 should go with commons-logging, and T3/plexus with avalon
logging with commons-logging as a backend

Everyone's invited
------------------
Someone said that turbine has no influence over at avalon. Simply put:
sure you do! Avalon is an apache project and works the same as all
others :D
BTW, I'm tracking turbine-dev for the sole reason of figuring out as
early as possible what possible T3 needs wrt avalon will be and how
avalon can accommodate those.

The logging mess
----------------
The fact we have two java logging toolkits @ apache seems pretty silly.
We've probably been over that a few hundred times (and that there is
then another one in the JDK that's worse than either is truely laughable
;).

The fact that we have two near identical logging abstractions seems just
as silly. I don't know exactly why this came to be, and it doesn't
matter that much to me either (well it does a lot from the "fix the
community process" perspective but that's not the topic atm). The
important thing is fixing it.

Logging in avalon
-----------------
Someone asked whether using avalon ties you to avalon logging or avalon
logkit. Answer: nope.
If you use the avalon lifecycle (like plexus/summit does), it makes a
lot of sense (say the avalon peeps like me) to use the LogEnabled and
Logger interfaces in avalon-framework.jar to handle your logging, but
this is not required in any way. If you do, your code looks like:

MyComponent implements LogEnabled {
	m_logger logger;
	enableLogging( Logger logger ) { m_logger = logger; }
	// ...
	m_logger.debug("blaat");
	// ...
}

where enableLogging() is always the first method called right after the
constructor. There's also AbstractLogEnabled which basically saves you
some typing.

Different implementations of the avalon Logger interface are provided
using Log4J, logkit, JDK 1,4 logging and stdout. Logkit is really only a
part of the rest of the avalon projects as in that is shares the same
url and community. It's completely decoupled otherwise.

Avalon-Phoenix (an example of a container, ie the thing that calls
enableLogging()) allows easy configuration of the logging support using
either the Log4J or Logkit implementation. Not sure about JDK logging.

Whether it makes sense to have the Logger interface in avalon as opposed
to having it in commons-logging....probably not. It's there and will
need to stay there for a long time to come because of compatibility. I
personally think what would have made the most sense was not having
logging wars a few years ago and now having a merger logkit+log4j part
of the JDK, removing the need for an abstraction. Didn't work out.

Avalon logging and commons-logging
----------------------------------
Best way to go forward? Not sure. The main technical difference I see
atm is that the avalon logging abstraction supports Inversion of Control
(vital to working with avalon, this is) whereas commons-logging doesn't.
If it is possible to have an IoC layer on top of commons-logging (as
already suggested), that's probably the best route, as I don't think
that commons-logging themselves want to buy into the IOC way of doing
things.

Avalon won't be able to drop the logging interfaces in avalon for a long
time, even if we decide that makes sense, because of compatibility
issues.

all this leads to what is good for turbine...

Turbine logging
---------------
If turbine goes the 'avalon route' (ie plexus/summit or something
similar), you will likely find it makes a lot of sense to use the
LogEnabled interface.

If you really want a commons-logging interface, you could have an
org.apache.turbine.logging.LogEnabled which takes a commons-logging
logger, but that would drastically reduce portability of turbine
components with other avalon users (like cocoon) so I'd strongly
recommend against that.

The more I think of it...the best solution is probably to have an avalon
Logger interface implementation using commons-logging. What backend
implementation to use...don't have a strong opinion. I mostly use log4j.

If turbine does not go the avalon route (are you guys sure yet?), I
think the choice for commons-logging is easily made as you probably
don't want to import the whole avalon-framework.jar (60k) just to access
4 classes inside.

For more short-term decisions (IIUC T2.3 != plexus), you probably want
to base that decision on the long-term goal....but OTOH turbine atm is
very much about singletons and 'active' components, which is very
different from IoC with passive components. It might be easier to keep
the current pattern for 2.3, and then it is probably easiest to use
commons-logging.
In which case it makes even more sense to have good commons-logging
support in avalon for when T3 hits the streets.


just my .02 euro,

cheers,

- Leo


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by James Taylor <ja...@jamestaylor.org>.
On Sun, 2002-12-08 at 18:15, Jon Scott Stevens wrote:
> on 2002/12/8 2:37 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:
> 
> > What about Avalon logging? It's better suited for server-side logging,
> > and has its own abstractions that make it possible to use as
> > implementations log4j, logkit or whatever.
> 
> I prefer that we use commons-logging.
> 
> I'm -1 on Avalon logging as the dependency simply because I like to have the
> choice as to which logging package I want to use.

One thing we might consider is not requiring LogKit as the logger, but
still use an IOC style of log configuration like in Avalon. In fact, we
could still use the LogEnabled / Logger facade without restricting log
configuration options.

This may not be appropriate for t2 but I think it will make sense for
t3. One of the nice things about IOC in this case is that the container
can build / obtain the logger it provides in any way. Thus, for example,
if you have a scenario where class loading discovery is not workable,
you can eliminate it without having to touch the components, just the
managers.

-- jt, who is almost completely converted to avalon now.


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Jon Scott Stevens <jo...@latchkey.com>.
on 2002/12/8 2:37 PM, "Nicola Ken Barozzi" <ni...@apache.org> wrote:

> What about Avalon logging? It's better suited for server-side logging,
> and has its own abstractions that make it possible to use as
> implementations log4j, logkit or whatever.

I prefer that we use commons-logging.

I'm -1 on Avalon logging as the dependency simply because I like to have the
choice as to which logging package I want to use.

-jon

-- 
StudioZ.tv /\ Bar/Nightclub/Entertainment
314 11th Street @ Folsom /\ San Francisco
        http://studioz.tv/


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Nicola Ken Barozzi <ni...@apache.org>.
Henning P. Schmiedehausen wrote:
> Martin Poeschl <mp...@marmot.at> writes:
> 

[...]

>>o switch to commons-logging

What about Avalon logging? It's better suited for server-side logging, 
and has its own abstractions that make it possible to use as 
implementations log4j, logkit or whatever.

> - Commons Configuration integration. 

What about Avalon Configuration? It's part of the lifecycle too.

[...]

Just 2 cents from the peanut gallery.

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Laurie Harper <zo...@holoweb.net>.
Henning P. Schmiedehausen wrote:

> Good question. I understood from the mail on the users list that all
> direct class references are a problem from JDK 1.3 to JDK 1.4 I don't
> see no difference between directly using foo.class and
> foo.class.getName().


My reading of that message was that direct class references *in nested 
(inner) classes* which *need to be serialised* were a problem. I'd have 
to go dig out the message to be sure, but that's how I remeber it...

L.


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Martin Poeschl <mp...@marmot.at> writes:

>Henning P. Schmiedehausen wrote:

>>- Commons Logging using log4j. This is ready, we use it all the time
>>  here (couldn't live without) and is basically adding some startup
>>  code to Turbine.java, some small glue code to the Velocity Service and
>>  then a big replacement Log -> log and adding lines like this
>>
>>  private static Log log = LogFactory.getLog(foo.class);
>>
>>  to all the classes that need logging. That's why I asked about the
>>  JDK 1.3 / JDK 1.4 incompatibilities regarding to direct class 
>>  references. If we really do have problems here, we must either offer
>>  1.3 and 1.4 compiled binary packages or use 
>>
>>  private Log log = LogFactory.getLog(this.getClass()) 
>>
>>  which needs more memory as we have more objects and is less performant.
>>  Opinions? I'd go for the "private static" option, mainly because building
>>  with maven is really not critical and providing two binary distributions
>>  is no problem either. 
>>  
>>
>will

>private static Log log = LogFactory.getLog(foo.class.getName());

>solve the problem??

Good question. I understood from the mail on the users list that all
direct class references are a problem from JDK 1.3 to JDK 1.4 I don't
see no difference between directly using foo.class and
foo.class.getName().

Anyone with deeper knowledge of the JDK here?

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by Martin Poeschl <mp...@marmot.at>.
Henning P. Schmiedehausen wrote:

>- Commons Logging using log4j. This is ready, we use it all the time
>  here (couldn't live without) and is basically adding some startup
>  code to Turbine.java, some small glue code to the Velocity Service and
>  then a big replacement Log -> log and adding lines like this
>
>  private static Log log = LogFactory.getLog(foo.class);
>
>  to all the classes that need logging. That's why I asked about the
>  JDK 1.3 / JDK 1.4 incompatibilities regarding to direct class 
>  references. If we really do have problems here, we must either offer
>  1.3 and 1.4 compiled binary packages or use 
>
>  private Log log = LogFactory.getLog(this.getClass()) 
>
>  which needs more memory as we have more objects and is less performant.
>  Opinions? I'd go for the "private static" option, mainly because building
>  with maven is really not critical and providing two binary distributions
>  is no problem either. 
>  
>
will

private static Log log = LogFactory.getLog(foo.class.getName());

solve the problem??


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: [vote] plans for turbine 2.3

Posted by "Henning P. Schmiedehausen" <hp...@intermeta.de>.
Martin Poeschl <mp...@marmot.at> writes:

>i posted my plans for the next version of turbine some weeks ago .. now
>it's voting time ;-)

>o remove all deprecated stuff (+ code cleanup)
>   incl. the freemarker and webmacro support (nobody is using it and the current versions are not 
>supported)

>o switch to commons-logging

I'm obviously +1 on this. :-)

Here is some more that I _have_ (no need to do coding, just pulling
out of our internal CVS, pushing to jakarta and integrating with the
current code).

- Commons Logging using log4j. This is ready, we use it all the time
  here (couldn't live without) and is basically adding some startup
  code to Turbine.java, some small glue code to the Velocity Service and
  then a big replacement Log -> log and adding lines like this

  private static Log log = LogFactory.getLog(foo.class);

  to all the classes that need logging. That's why I asked about the
  JDK 1.3 / JDK 1.4 incompatibilities regarding to direct class 
  references. If we really do have problems here, we must either offer
  1.3 and 1.4 compiled binary packages or use 

  private Log log = LogFactory.getLog(this.getClass()) 

  which needs more memory as we have more objects and is less performant.
  Opinions? I'd go for the "private static" option, mainly because building
  with maven is really not critical and providing two binary distributions
  is no problem either. 

- Commons Configuration integration. 

  I removed the TurbineResources Service completely but obviously one
  could write a very thin compatbility layer with simply queries the 
  central configuration object and offers the same methods. 

  I'm just 0 on keeping TurbineResources as rewriting is pretty simple 
  (With my code  you just change TurbineResources.getString() 
  to Turbine.getConfiguration().getString()) and you can run in some
  "gotchas" because commons-configuration behaves different in some corner
  cases from TurbineResources (I did).

  I do consider this a breaking change and people should do a "Heads up" here. 
 
- Decoupling the Services and use Fulcrum. Yeah, I know, there are discussions
  about Avalon, Services repositories in common yadda, yadda, yadda. My main
  point here is: I have code that works. And I have about a bazillion patches
  to Fulcrum which I will not move into the Turbine code. This is too much
  work and I plan on working only on one code base. To me this is not a 
  2.3 change but I plan on making a branch tag  out from the post-2.2 code
  (Once I have everything in place that I need from the main branch, mainly
  logging and configuration) and putting the code onto the jakarta repository
  to let people decide. I plan on supporting the tree for quite a while and
  even if we dissolve Fulcrum at a later date, this will be a nice working
  base.

- Revamping the URI handling. The current state of the URI handling is a 
  mess. I rewrote the whole ContentURI/DynamicURI/RelativeDynamicURI code,
  bundled it under turbine/util/uri and made the whole code base use this
  in a consistent manner. I removed lots and lots of gotchas from the whole
  handling and the nice by-effect of this is, that user visible is only a
  small change in the tools to load in the context. 
  _AND_ my code is fully documented. :-)
  If you every tried to implement cookie-less session handling with Turbine,
  you will really appreciate this. 

- Pull Service rewrite, scope cleanup and rework. Most of the pull service
  is simply poor to very poor code quality. I rewrote this, cleaned up the
  refresh handling, added a fifth scope and a second tool class. 
  This BTW made it possible to use pull tools as objects in the JSP pages
  (which is untested but should work)

- RunData. Don't get me started on RunData. :-) I factored out the
  ServerData, moved lots of code to actually use ServerData objects which
  makes the whole Turbine startup pretty slick and allowed the URI code
  to take advantage of this. 

- Use commons whereever possible. This removes some helper classes.

- Remove the whole Stratum code. Most of it can be replaced with Lifecycle or
  with other commons code. 

... and there are just the plans for Turbine. Fulcrum is some more... :-)

Martin, that's why I proposed to do Turbine 2.2.x changes along the 
branch tag that you did in the repository. If users start writing code
on Turbine-2.2, they _should_ check out from the TURBINE_2_2_BRANCH tag
and can be sure that the checked out code (and the further 2.2 releases
(either 2.2.1, 2.2.2 or 2.2p1, 2.2p2) will work without changes on their
applications. These releases will only be bug-fix releases without any new
features.  

If you check out from the HEAD, you must read the changes and be
prepared to adjust your code. That's the price for bleeding edge. The
deprecation scheme simply does not work for bigger changes (like
ripping out LogService and ResourcesService). We must do this, endure
the pain and live on. :-) We won't get up to speed otherwise. 

So, people, I propose if you want to live on the T-2.2 bleeding edge,
go for the TURBINE_2_2_BRANCH tag, if you want to live dangerously, go
for HEAD.

Ah, and please add a "<release>" tag to project.xml (or did this
disappear somewhere past maven 1.0b4?) and maybe some documentation
about the branch tags.

	Regards
		Henning


-- 
Dipl.-Inf. (Univ.) Henning P. Schmiedehausen       -- Geschaeftsfuehrer
INTERMETA - Gesellschaft fuer Mehrwertdienste mbH     hps@intermeta.de

Am Schwabachgrund 22  Fon.: 09131 / 50654-0   info@intermeta.de
D-91054 Buckenhof     Fax.: 09131 / 50654-20   

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>