cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Grzegorz Kossakowski <gkossakow...@apache.org>
Subject Re: pipelineComponent scope troubles
Date Tue, 11 Sep 2007 19:21:03 GMT
Giacomo Pati pisze:
>> Yes, I'm almost sure that's the reason why pipelineComponent scope is freaking.
> 
> Guess what ;-) didn't help removing all map:mount from the sitemap. Exception still exists
and I
> have still no clue what has caused this incompatability with existing code.

Are you sure that there is no map:mount in whole call stack? Is your sitemap managed by Servlet
Service Framework?

> 
> To be honest, I see no reason for me to dig into something I have not a single clue where
to look
> at. But still questioning who has put this backward incompatability into it and how to
migrate from it.

I think it's shared responsibility but the most significant part must be taken by me.
I gave you description of current state of our code to help you get involved but only if you
want
to, of course. Nevertheless, I really, really want to avoid one-man shows as we both know
that it's
not how open source works. I can fix this myself (as soon as time permits) but this will not
help
situation that I and Daniel to some extent will have a any clue about this code.

Implementing this sitemap scope is just a must in order to support old Cocoon code.

In order to help you with migration I must get some details about your current setup and structure
of application.

>> If you want to make a quick test if it's the only problem with current trunk and
work-around NPE
>> mentioned by you just change scope of ObjectModel from "pipelineComponent" to "call"
in this
>> file[1]. Be warned that it's only a quick work-around and implementation of sitemap
scope is the
>> only reliable solution.
> 
> Uhh, this give a even more ugly stacktrace (deep into template, el, and jexl) I wouldn't
even think
> of pasting it into this mail.

It's your turn to be more informative, dear ;-)
Paste this ugly stack trace, it may really help me to understand what's the real cause of
your
troubles. Take into account most (I haven't tested every single one) samples work with current
code
so it's something specific to your application/setup.

Most important: can you reproduce this behaviour when developing some simple test based on
our
archetype?

-- 
Grzegorz Kossakowski
Committer and PMC Member of Apache Cocoon
http://reflectingonthevicissitudes.wordpress.com/

Mime
View raw message