On 2/27/06, Andreas Hartmann <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED] wrote:
> > http://lenya.apache.org/1_4/reference/modules/index.html
> > suggests that Modules will be able to add protocols.  Protocols are
> > part of the platform and should not be added dynamically.  Anyone
> > adding a protocol would have a development environment.
> Yes, but this doesn't mean that adding Java code to modules
> is a bad thing, does it?
>
> > Lenya has been usable and very flexible without a development
> > environment.  A binary version of Lenya is very flexible.  I realize
> > that most Committers and some Developers have a development
> > environment, but most Developers do not, and should not, need one.
> >
> > If Modules can contain Java, then:
> > 1. Every Lenya Developer must have a development environment.
> I don't understand this implication. You're not required to
> add Java code to your modules.
Good.

> > 2. We have the dependency issues that started this thread.
> Yes, they have to be solved.
Just have the ClassNotFound Exception open the "Sorry" page and add a
line to the log in any Class that could look for code in a Module.

Each of those Classes needs to be documented to explain how the hook
for Module code works.  Will loading a Module dynamically add to the
information read from WEB-INF/cocoon.xconf?  (That may be the central
place to check for ClassNotFound.)

> > I cannot see how it is good to force Users to install a development
> > environment.
> We're not forcing anyone. We're allowing to use Java.
Good.

> > It increases the complexity and reduces the audience.
> It allows modularization. We have to break up our monolithic
> core into modules, otherwise we'll end up with too many
> connections between components and circular dependencies.
> If we don't allow Java code in modules, people will start to
> implement their stuff using server pages and other suboptimal
> technologies.

I am migrating my Lenya1.2.2 publication to use my interpretation of
Modules.  I moved most of the XMAPs into the Modules, fixed the URLs,
and added a module: protocol.  Tonight's work will be to add template
inheritance to the module: protocol.  Then I will move the rest of the
files into the Modules, and separate parts of the "live" Module into
"authoring" and "xhtml" Modules.

I have not found any function that required or could be improved with
Java, but my Publication could be too simple to need it.

One concern is Cocoon already has a "module:" protocol, which my
version overrides.  I could not find anything in Lenya that used the
Cocoon protocol (and Lenya still works), but this could cause problems
for Cocoon users that are migrating to Lenya.  Opinions?

> > Lenya should become both easier and more powerful.  Modules including
> > Java that needs to be compiled is a step in the wrong direction.
> I disagree.
(Hoping you are not disagreeing with the first sentence)  You have
answered my objections.

Thank you,
solprovider

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

Reply via email to