On Thu,  2 Aug 2001 06:24, Leo Simons wrote:
> whoah! You've been quite busy, haven't you, Pete?

;)

> Could you give me a quick summary of the changes
> you made to phoenix since, say, uh, June?

Well as my time on phoenix was limited and I kept coming up against pain and 
suffering as  a result of javas lack of generic types I decided to dump the 
idea of atlantis (ie generic kernel framework) and just work on our specific 
kernel implementation. I may come back to atlantis in the future or I may 
not. This meant simplification of a lot of our classes (yay!). And removal of 
a lot of duplication.

Also I noticed that the pacakge structure was no longer representative of 
what we wanted to do. As it had evolved classes had been added to packages 
and we had a kind of chaotic arrangement. Thus I adopted the structure that 
is present in cocoon (and my ant proposal) codebase. Basically this places 
one component in each directory of phoenix.components.*. So we have things 
like phoenix.components.manager, phoenix.components.embeddor, 
phoenix.components.frame.

Looking at all my possible use cases I also found that large chunks of old 
facilities never needed to be pluggable. So I merged them into 3 "facilities" 
namely management, frame and configuration (in the future I may readd logging 
- but thats the future). "frame" defines the "execution frame" in which 
blocks operate and contains all support resources.

Configuration and management add other features (guess which ones). Currently 
configuration is setup so that it only stores config data in memory but in 
reality it could store it in any medium (finally!!!!).

The whole system is setup so we can differentiate between "installing" 
(extracting files and registering components) and "deploying" (take extracted 
files and making them go at a particular time). However that is yet to be 
implemented.

I also wanted to make an ant task that verified config files before running 
server but never got around to iut ;)

> Have you done a lot with the JMX and/or Manager code?

nothing.

Cheers,

Pete

*-----------------------------------------------------*
| "Faced with the choice between changing one's mind, |
| and proving that there is no need to do so - almost |
| everyone gets busy on the proof."                   |
|              - John Kenneth Galbraith               |
*-----------------------------------------------------*

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to