Tim Williams wrote:
I've reverted my changes as requested.
I never requested that :-((
I was trying to understand why this change was necessary.
I'll lay low until I see where
you're going with this. Perhaps I don't understand views as well as I
thought, but I don't see that the output side of them works on xdoc,
but rather the resultant html of document2html.
Did I say that output was XDoc? I didn't mean to, what I meant was the
outpur of document2html is not XHTML2 (its not even html), therefore if
we continue to work with that format we are not using XHTML2 in our
internals.
Either way, I'm deleting my local changes and svn up occasionally to
see where you're going.
No, please don't do that. I never said this change should be reverted, I
asked why it was necessary.
I have vision in my own mind as to where this is going, but that doesn't
mean it is right. Both you and Thorsten are expressing similar concerns,
but I don't understand them. I'm trying to understand.
This commit, in isolation, makes no sense whatsoever. However, I am
aware it is part of a larger set of changes you made locally. I need to
understand what your approach is and if it is in line with Thorstens,
because if it is then it is likely that my approach is wrong.
Ross
--tim
On 9/9/05, Ross Gardler <[EMAIL PROTECTED]> wrote:
Tim Williams wrote:
On 9/9/05, Ross Gardler <[EMAIL PROTECTED]> wrote:
[EMAIL PROTECTED] wrote:
Author: twilliams
Date: Wed Sep 7 21:40:42 2005
New Revision: 279511
URL: http://svn.apache.org/viewcvs?rev=279511&view=rev
Log:
can't return all the html, head, stuff for insertion in the pipeline.
temporary solution to get on with pipeline testing
Why do you want the output to look like document2html output and then be
passed into the existing pipelines?
Because that's how views work right now. I didn't have this grand
redesign scheme in my mind at the time. I thought we were getting
XHTML2 to work with the current views and replacing the **body-*.html
aggregation "part" in the **page pipeline with an xthml2tohtml
equivalent was how I did it.
Views as they are now, work with XDoc. The goal is to make core work
with XHTML2 not XDoc. It is not possible to include views in this work
*and* not do a level of redesign in views (or the old skin system).
Don't we just want XHTML2 -> output plugin = desired output
That's what it is, but views operate on html structured like the
output of document2html. I wasn't redesigning views, just getting
xhtml2 to work with them.
So here you are saying the same as Thorsten.
Meaning we only need an XHTML2_to_HTML stylesheet>
(I recall a conversation Tim and I had on IRC in which he asked me why
we need the XHTML2HTML stylesheet. I seem to remember the answer I gave
was misleading to say the least. Hopefully recent discussions about the
new pipeline will have clarified - if not I'll try agian when I
understand the intent of this commit, I may be missing something)
Ross
I could very well be missing something but like I tried to explain in
the other thread, to get xhtml2 to work with views (or the current
skins for that matter) we need two things:
1) a xslt that outputs html with the same structure of document2html
2) a change in the **page pipeline to call that instead of **body-*.html
Well this is where the "division" needs to be called. In my view it is a
waste of effort refactoring the current skinning system to use XHTML2
and then converting views to get rid of the body-** templates.
In my opinion we should skip the conversion of skins and work directly
on views.
If you revert this commit and do svn up, you will see that the body-*
part of the pipeline is working. All we need to do now is get views to
work with site.xml source files to create the navigation (this is, a
small step).
[NOTE] this work happens in the contracts *not* in the structurer
sitemap so it does not prevent work proceeding with the refactoring of
views.
Even when we do what you suggest and use contracts for everything
instead of aggregation, the implementation of that contract will still
need something similar to this.
Why? I don't think this is the case (see my summary mails on return from
Apachecon and Thorsten expansion of that summary)
Ross
