David Jencks wrote:
> 
> Hi Mike,
> 
> karaf-full is supposed to be identical to the non-minimal karaf assembly
> from the old style assembly.
> 
> I have no objection to your proposal and I think it would be pretty
> trivial to make such an assembly using the karaf-assembly packaging.
> 
> I'm curious about the environment you are in that would make it easier to
> import an assembly containing 150 bundles 50 of which you don't use than
> the 100 you do use.
> 
> thanks
> david jencks
> 
> On May 3, 2011, at 12:22 PM, mikevan wrote:
> 
>> 
>> mikevan wrote:
>>> 
>>> For folks developing applications to deploy into Karaf on closed
>>> networks,
>>> it is not always feasable to be able to download all the optional
>>> packages
>>> for which we have optional console commands.  I'm thinking web:, http:,
>>> obr:, and the like.
>>> 
>>> I propse we create a new assembly for karaf that will include all of the
>>> optional bundles in the system directory for use in closed-networks. 
>>> After talking about this topic on IRC it seems that many of us
>>> developing
>>> on closed networks have created work-arounds for this.  Because there
>>> are
>>> so many work-arounds, perhaps its time to have a single Karaf-Max
>>> deployment that contains all of the optional bundles for karaf.
>>> 
>>> If it helps, I can write it... :-)
>>> 
>> 
>> After further research, we currently have a karaf-full kar and assembly.
>> However, it doesn't look like karaf-full contains all of the optional
>> dependent bundles. What was supposed to go into karaf-full? Does that
>> suffice the use-case expressed above? If not, would it be appropriate to
>> have a new assembly addressing the karaf-max usecase?
>> 
>> David J, What are you thoughts on this?
>> 
>> 
>> -----
>> Mike Van (aka karafman)
>> Karaf Team (Contributor)
>> --
>> View this message in context:
>> http://karaf.922171.n3.nabble.com/DISCUSS-Karaf-Max-assembly-tp2895460p2895601.html
>> Sent from the Karaf - Dev mailing list archive at Nabble.com.
> 

If we talk in terms of the use-case presented, this would be an assembly for
users who are developing on a closed-network.  A closed-network in this case
would be a network with no external connectivity to the internet.  In the
IRC room, several folks came up with specific types of clients they'd worked
with that had a closed-network as defined above.  Examples presented there
included corporate clients like banks, although many customers with strong
security would fall within that category.

As users become more karaf-savvy, it is certainly forseeable that they would
begin leveraging more of the optional packages that karaf-max would contain. 
For example, when I first started using Karaf, I had no idea what obr was,
but now I use that set of functionality quite a bit, and have provided
patches to increase the functionality of obr.  Another example would be the
pax-web packages which make it easier to deploy servlets into Karaf.  A user
using JMS would not necessarily care about that, however with many
web-services moving to REST, having access to the optional http: and war:
features would make lives easier.

To extend the use-case to make it more understandable, I would assume
karaf-max would be a development-only environment.  After a user has
developed an application using karaf-max, they would likely create a kar out
of the files they have deployed as bundles.  Or, in my case, generate a
local maven repository containing those specific bundles and an
applicatiom-specific features.xml file.  In that case, the actual deployment
would be done using karaf-minimal, and the kar or local maven repository. 

I hope this makes sense.

-----
Mike Van (aka karafman)
Karaf Team (Contributor)
--
View this message in context: 
http://karaf.922171.n3.nabble.com/DISCUSS-Karaf-Max-assembly-tp2895460p2895743.html
Sent from the Karaf - Dev mailing list archive at Nabble.com.

Reply via email to