[naming] Inquiring about status

[Alex Karasulu <ao...@bellsouth.net>]

Hi Henri and Phil,

I was just chatting to Noel about the naming code and where it is in
terms of usability.  Is there anything left to get the code to a point
so others can readily use it.  Perhaps use it over at Geronimo, on
Tomcat or elsewhere like in commons configuration?

Alex

RE: [naming] Inquiring about status

[Alex Karasulu <ao...@bellsouth.net>]

On Tue, 2004-08-31 at 01:21, Brett Porter wrote:
> > The Jakarta Commons Configuration folks wanted to know if it was ready for a
> > Release.  Considering the origin of the code, and the fact that we've
> > already released it as part of Tomcat before relocating it, I'd suggest that
> > the Incubator PMC would be willing to allow a Release of that code, if it is
> > ready.
> 
> I agree. I've been using it in production for a year now (3 years if you count
> it as part of Tomcat 4 :).

Very cool I had no idea the code was that mature or I would have used it
myself in a few places.  This is good news.  I think we should see if
the Incubator PMC would allow a release if Phil and Henri feel ok with
that. 

Alex

Re: [naming] Inquiring about status

[Phil Steitz <ph...@steitz.com>]

Brett Porter wrote:
>>What about the JNDI Configurator?  Shouldn't that also tie into naming?
> 
> 
> I think we are talking about the reverse here. XmlConfigurator is for populating
> the in memory JNDI from an XML file. I'm guessing that JNDI configurator is for
> commons-configuration to populate Java beans from a JNDI tree?
> 
> I haven't looked at commons-configuration, but there is every chance it also
> loads XML config files and could be used instead.
> 
> I think at this point it is best to see where the release takes it and let
> demand drive the functionality added.
> 
> - Brett

I agree with pretty much all comments on this thread. The code has been 
released and has a lot of good, reusable functionality.  I have not looked 
recently at the tomcat CVS, but up to a few months ago nothing significant 
had changed.  I did fix a couple of small bugs in the code ported from 
tomcat.  I agree with Brett on the XmlConfigurator and relation to 
commons-configuration. It is at conceivable that the reuse could go the 
other way as well -- i.e., we could use commons-configuration to load JNDI 
data from formats other than the tomcat-style XML that the XMLConfigurator 
supports -- but it is probably not a good idea for the dependency to go 
both ways. Adding direct support for properties files and improving the 
XMLConfigurator is proabably the best way to go.

My main problem with the code as it stands is documentation and test cases 
(almost non-existent).  Some refactoring might be in order as well. 
Certainly repackaging, since I don't think I got the core/factory split 
right.  Comments/suggestions for improvement would be greatly appreciated. 
I now have some time to spend on this, so I will get back to it.

Phil

RE: [naming] Inquiring about status

["Noel J. Bergman" <no...@devtech.com>]

> > What about the JNDI Configurator?  Shouldn't that also
> > tie into naming?

> XmlConfigurator is for populating the in memory JNDI from an XML file.

Ah ... so you aren't refering to
http://jakarta.apache.org/commons/configuration/apidocs/org/apache/commons/c
onfiguration/XMLConfiguration.html?  OK, that would explain why we weren't
understanding each other.

> I'm guessing that JNDI configurator is for commons-configuration
> to populate Java beans from a JNDI tree?

See:
http://jakarta.apache.org/commons/configuration/apidocs/org/apache/commons/c
onfiguration/JNDIConfiguration.html.  It implements a Jakarta Commons
configuration based upon a JNDI context root.

You're right in that for naming, we're looking to populate a JNDI context
from a data source for some form, especially the J2EE mandated XML.

	--- Noel

RE: [naming] Inquiring about status

[Brett Porter <br...@apache.org>]

> What about the JNDI Configurator?  Shouldn't that also tie into naming?

I think we are talking about the reverse here. XmlConfigurator is for populating
the in memory JNDI from an XML file. I'm guessing that JNDI configurator is for
commons-configuration to populate Java beans from a JNDI tree?

I haven't looked at commons-configuration, but there is every chance it also
loads XML config files and could be used instead.

I think at this point it is best to see where the release takes it and let
demand drive the functionality added.

- Brett

RE: [naming] Inquiring about status

["Noel J. Bergman" <no...@devtech.com>]

Brett Porter wrote:

> > The Jakarta Commons Configuration folks wanted to
> > know if it was ready for a Release.  Considering
> > the origin of the code, and the fact that we've
> > already released it as part of Tomcat before
> > relocating it, I'd suggest that the Incubator PMC
> > would be willing to allow a Release of that code,
> > if it is ready.
>
> I agree. I've been using it in production for a year now
> (3 years if you count it as part of Tomcat 4 :).

Please note that it is not about code quality.  The justification for it
would be that the ASF has already officially released that code, and has
communities committed to it.

> I think the XmlConfigurator I wrote probably isn't as
> complete as it could be, but is certainly usable.  A
> release would get more feedback from users in this
> regard.

Agreed.  I am thinking that for JAMES v3, we could switch to this for Mailet
configuration, rather than use the Avalon format configuration files.

> > We should also see if there have been any changes to
> > the code in Tomcat. Ideally, Tomcat should get the
> > code from us, and contribute changes back.

> Definitely. However, while the XmlConfigurator is modelled
> on the Tomcat config, I'm sure it doesn't match up. I'd be
> interested if there was a more general way to reuse this
> (I'm not familiar with how Tomcat parses and loads its
> configuration files).

Please feel free to talk with Remy et al about how best to address it.

What about the JNDI Configurator?  Shouldn't that also tie into naming?

	---- Noel

RE: [naming] Inquiring about status

[Brett Porter <br...@apache.org>]

> The Jakarta Commons Configuration folks wanted to know if it was ready for a
> Release.  Considering the origin of the code, and the fact that we've
> already released it as part of Tomcat before relocating it, I'd suggest that
> the Incubator PMC would be willing to allow a Release of that code, if it is
> ready.

I agree. I've been using it in production for a year now (3 years if you count
it as part of Tomcat 4 :).

I think the XmlConfigurator I wrote probably isn't as complete as it could be,
but is certainly usable. A release would get more feedback from users in this
regard.

> 
> We should also see if there have been any changes to the code in Tomcat.
> Ideally, Tomcat should get the code from us, and contribute changes back.
> 

Definitely. However, while the XmlConfigurator is modelled on the Tomcat config,
I'm sure it doesn't match up. I'd be interested if there was a more general way
to reuse this (I'm not familiar with how Tomcat parses and loads its
configuration files).

Cheers,
Brett

RE: [naming] Inquiring about status

["Noel J. Bergman" <no...@devtech.com>]

Alex Karasulu wrote:
> Is there anything left to get the code to a point so
> others can readily use it.  Perhaps use it over at
> Geronimo, on Tomcat or elsewhere like in commons
> configuration?

The Jakarta Commons Configuration folks wanted to know if it was ready for a
Release.  Considering the origin of the code, and the fact that we've
already released it as part of Tomcat before relocating it, I'd suggest that
the Incubator PMC would be willing to allow a Release of that code, if it is
ready.

We should also see if there have been any changes to the code in Tomcat.
Ideally, Tomcat should get the code from us, and contribute changes back.

	--- Noel