Thorsten Scherler a écrit :
On Tue, 2005-08-02 at 10:19 +0200, Cyriaque Dupoirieux wrote:
Cyriaque Dupoirieux a écrit :
Ok, I think there is a problem - or at least something different -
with cocoon transformations during views generation process.
If I generate the site with the skin (not the view plug in) I have :
* <div id="credit"></div>
* Because the style sheet does :
* <div id="credit">
o Credits calculation, but I have none...
* </div>
* <div class="clear both"> </div>
* Because the style sheet does :
* <div class="clearboth"> </div>
Why is that happening in skins and not in views?
Maybe it is the sitemap.serializer.xhtml, because I just changed the
serializer to "html" and get:
<div class="test"> </div>
That means we have to look into the xhtml serializer to get our .
No, as I said in my previous mail, the idea of the empty div tag filled
with an is not good.
We have to skip them...
This is not the same behavior :
* First, an empty div tag is kept with this form <div></div>
* Second, the   appears like in my source code which is
what is expected (and not [Â ]).
I think I am the only one to have this problem since I am the only one
to participate to this thread ;-) .
Thx for ignoring me ;-)
Sorry,
Anyway, here is what I think - When we put a hook in an *.fv file :
* There is a css problem if the hook is transformed into an empty
<div/> tag when generating XHTML,
* There is also a problem when we add a blank ( ) because if
the defined style has a background color, the tag <div
...> </div> appears in the page as a box with the background
color
So, the solution would be to skip the hook generation when the content
is empty.
Something like :
|<xsl:choose>
<xsl:when test="normalize-space(.)">
... Generation of div tag and content ...
</xsl:when>
<xsl:otherwise>
... Skip the generation ...
</xsl:otherwise>
</xsl:choose>|
But I don't know exactly where to put this control ?
/home/thorsten/apache/forrest-trunk/whiteboard/plugins/org.apache.forrest.plugin.internal.view/resources/stylesheets/hooksMatcher.xsl
...but it is not that easy because normally you will have
<forrest:hook name="test">
<forrest:contract name="testData"/>
</forrest:hook>
Now what you would have to do is to test whether the *transfomed*
presentation model of the contract (remember a contract is right now
only a xsl viewHelper template) will return something. You may test it
via a xsl:variable but I am not sure.
The hooksMatcher.xsl got imported in the prepare.xhtml.xsl all in the
internal view plugin.
Maybe we can do this in two phases :
* First the prepare.xhtml.sxl (including hooksMatcher.xsl) generates
what he has to generate.
* Second, another sheet clean the generated code by removing empty
div tags ?
Cyriaque,
HTH
salu2
thorsten
Regards,
Cyriaque,
Cyriaque,
Cyriaque Dupoirieux a écrit :
Thorsten Scherler a écrit :
On Fri, 2005-07-29 at 10:25 +0200, Cyriaque Dupoirieux wrote:
Ok, it works.
The only thing is when I edit the XHTML source, I have this :
<div class="clearboth">Â </div>
It seams that   is translated to  ?
Why don't you use   ?
http://www.html-world.de/program/html_sz.php
It is in german, but you will understand. ;-)
Did you try with  ? What is the outcome? I will try after work.
It gives this :
<div class="clearboth" /> !
So I prefer   :-P
You know, I don't really understand why browsers are disturbed by
empty <div/> tags.
But it is a fact, and I have a proof !
Have a look here :
http://cyriaque.dupoirieux.free.fr/testviews1/testDiv.html
Regards,
Cyriaque,
salu2
Cordialement,
Cyriaque,
Thorsten Scherler a écrit :
On Thu, 2005-07-28 at 18:08 +0200, Cyriaque Dupoirieux wrote:
Thorsten Scherler a écrit :
in order to generate <div> </div> if nbsp='true' instead of
<div/> which is not well interpreted by browsers.
But this   is not enough, the generation still gives
<div/>...
¿?
Are you sure? I tried it and is working fine for me.
Actually, I don't understand what happens, if I write a trace to
see if it's the good template which is used, the generation is OK
because of my trace (the div tag is no more empty...)
(It's a little bit like Heisenberg uncertainty principle, the
observer disturb the result of the experiment :-) )
Let it as it is I will continue my investigations, it may come
from something else in my environment...
try a "forrest clean" in your project and build.sh clean in forrest.
As well I recommend to "ant clean" all view plugins.
After a build everything should work fine.
I propose :
<xsl:if test="@nbsp='true'"><span/></xsl:if>
Hmm, I do not like that much because the <span/> do not belong
in there
IMO.
Sure you are right...
By the way, can you try to add something like this in your *.fv
and tell me the result :
<forrest:hook nbsp="true" class="test"/>
<div class="test"> </div>
HTH
salu2
thorsten
Regard,
Cyriaque,
salu2
PS: Sorry, I have lost the thread...
