Having just developed a big project with many, many interdependent
modules, I can confirm that oXygen's validation scenarios are a superb
solution!
John
On Apr 13, 2009, at 5:03 AM, George Cristian Bina wrote:
Hi,
oXygen supports working with modules. We introduced (since version 8
I believe) the concept of a validation scenarios where one can
specify that the main file, in this case docbook.xsl should be
validated instead of the current file, and thus the module gets
validated in the context it is used from and not as a standalone
stylesheet.
For more information please see:
http://www.oxygenxml.com/doc/ug-oxygen/validation-scenario.html
Best Regards,
George
--
George Cristian Bina
<oXygen/> XML Editor, Schema Editor and XSLT Editor/Debugger
http://www.oxygenxml.com
Bergfrid Skaara wrote:
Hi,
I'm using oXygen XML Editor 10.0 and can confirm this behavior there.
My 2 cents is as follows:
The issue is similar to the idref validation errors you get with
modular DocBook XML. The individual fragments will not validate, but
the main document pulling them in, will.
The red validation error indicators in the XSL files has nothing to
do
with the location of the files: simply open one of the stock XSL
files
(except docbook.xsl) in oXygen and you will observe the same red
lines
there when a template calls a named template or references a global
parameter or variable.
In sum, this is a modularization issue and I guess you should see
similar behavior in other validating XML editors.
Best regards,
Bergfrid Skaara
On Fri, Apr 3, 2009 at 1:26 AM, Jennifer Moore
<[email protected]> wrote:
Just realised I'd only sent this to Bob, when I meant to send it
to the
group, in case anyone else had a similar problem...
====
Thanks Bob! That helped - I've made sense of it now.
I was mixing up two kinds of error.
In my new files with custom templates, oXygen was giving me the
warnings
about (other) templates not existing, etc. via its "little red
lines".
I mistakenly thought those were telling me what was stopping the
transformation from working.
But in fact, the transformation was only being stopped by a place
where
I'd still (or again - after all this moving stuff about :-) ) got
the
wrong relative address in the entity tag. As soon as I fixed
that, the
transformation worked again.
oXygen is _still_ giving me little-red-line warnings about missing
templates, but it seems to me now that I can just ignore those, in
that
they're only true within the context of one file by itself. (i.e.
that
file by itself "doesn't know" that by the time it has to do
anything,
I'll have imported docbook.xsl as well.)
Please correct me if I'm wrong in my interpretation there...
Other than that, seems like everything's working OK now - it's just
minorly annoying to be stuck with the irrelevant red lines, which
_usually_ mean something I'd want to pay attention to. (Or maybe
there's a way of telling oXygen not to worry in this context? in
which
case that would be useful...) (mind you this is only oXygen 8.2,
not
the latest, so maybe they _have_ done that in a later one...)
Jennifer
Bob Stayton wrote:
Hi Jennifer,
This is curious. There are almost no constraints on where the
various
stylesheet modules are located. They can be scattered over many
directories, as the DocBook XSL files are. You could even import
with an
http URL.
In general, a relative import address is taken as relative to the
file
that contains the xsl:import element. That keeps things pretty
simple,
fortunately. For each import, you just have to confirm the
location relative
to that file. Once you import the DocBook docbook.xsl file, it
will handle
its own relative imports.
Arranging your customization files outside the distributed files
is good
practice, not foolish, and highly recommended for easy
maintenance and
updates. But clearly something is going wrong with your setup.
I can't
tell what it is from the information so far. Can you provide more
details?
That kind of error means some XSL files are not being read for
some reason.
Are there no other error messages?
Bob Stayton
Sagehill Enterprises
[email protected]
----- Original Message ----- From: "Jennifer Moore"
<[email protected]>
To: <[email protected]>
Sent: Monday, March 30, 2009 2:12 PM
Subject: [docbook-apps] XSL customisation layer: directory
structure
Hello all
(and a belated thanks to Mauritz and Bob for helpful replies to
my last
question, which I seem to have neglected to report back on at
the time -
have got link title attributes working nicely now via XLink.)
I'd been thinking I'd like to learn how to do a custom layer of
XSL, so
as to tweak my HTML output a bit.
I've successfully made an initial file as described in Bob's book
(example 9.1 here:
http://www.sagehill.net/docbookxsl/CustomMethods.html#CustomizationLayer)
and begun to create my own templates, of which I currently have
a couple
in separate files, imported e.g. like this to the main one:
<xsl:import href="emphasis_custom.xsl"/>
Well, I was getting on OK until I got the bright idea that
perhaps it
would be best to move the various custom files into a separate
directory
rather than having them mixed in with the default ones. At that
point
things stopped working. So now I am wondering if in fact that
was a foolish
idea :-)
Is it _possible_ to use different directories, and if so, what
do I need
to do to make the relative addressing work so the templates can
"see" each
other?
Or do the custom files have to be in the same directory as the
docbook
default ones? and in that case, is there any particular naming
convention,
for making the custom files easily identifiable if & when I
upgrade and want
to copy them over?
For background, here's what I did:
First I had set up a subdirectory of the html directory, and
made my "top
level" custom file call docbook like this:
<xsl:import href="../docbook.xsl"/>
It wasn't happy till I corrected the relative addressing for
<!ENTITY % common.entities SYSTEM "../common/entities.ent">
%common.entities;
]>
by making it
<!DOCTYPE xsl:stylesheet [
<!ENTITY % common.entities SYSTEM "../../common/entities.ent">
%common.entities;
]>
Then I got errors like "No template exists named anchor".
Then as an alternative I tried making a custom directory at the
same
level as the html directory, with
<xsl:import href="../html/docbook.xsl"/>
and changing the ENTITY thing above back to what it was before.
But I was still getting error messages of "No template exists
named ...".
It seemed like one template could only "see" another if they
were in the
same directory.
(though oddly enough it doesn't seem to mind having the "main"
custom
layer file in another directory - but then at present that one
isn't calling
any templates, only importing others that do.)
Thanks in advance for clues.
Jennifer
--
www.uncharted-worlds.org/blog/
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
--
www.uncharted-worlds.org/blog/
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
open.org
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
open.org
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
