[xmlc] Re: Re: Re: Accessing the body in a Document

This is a multi-part message in MIME format...

------------=_1276630015-5928-2118
Content-Type: text/plain;charset=utf-8;format="flowed"
Content-Transfer-Encoding: quoted-printable

Hi Franck,

I wrote a bunch of stuff below, but just thought of one thing that might =
be=20
pertinent, so look into this before reading more further below.

Which parser are you using?=C2=A0 Are you using the XML or HTML parser?=C2=
=A0 If HTML,=20
can you try using the XML parser?=C2=A0 Note that the HTML parser won't v=
alidate=20
your markup and may (depending on bugs in NekoHTML) get confused by XHTML=
=20
markup (regardless of whether it validates against the XML parser).=C2=A0=
 You can=20
also try updating to the latest version of NekoHTML[1].=C2=A0 Many bugs h=
ave been=20
fixed in it since the version that shipped with XMLC-2.3.1.


[1] http://nekohtml.sourceforge.net/changes.html

Jake


If you don't make headway after looking into the above question, then rea=
d=20
more below....


Unfortunately, I'm unable to reproduce the issue.=C2=A0 I tried setting u=
p a plain=20
old XMLC project using both the HTML and XHTML DOMs and was able to use=20
getNextSibling() to access the "body" node just fine.

I presume you are using XMLC-2.3.1?=C2=A0 I tried both 2.3.1 and the late=
st SVN=20
head.=C2=A0 In both cases, everything worked just fine.

Note that I place xercesImpl-2.9.1.jar in CATALINA_HOME/lib and=20
xml-apis-1.3.04.jar in CATALINA_HOME/endorsed

It's been a while since I looked at Barracuda, but I presume=20
view.getDocument() (when using XMLC) is returning the XMLObject subclass =
that=20
represents your document.=C2=A0 It's actually a wrapper to the real docum=
ent.=C2=A0=20
Using getOwnerDocument() on any child node would give you the real Docume=
nt=20
object.=C2=A0 You can call xmlObject.getDocument() if the current node re=
presents=20
the document itself (getOwnerDocument() returns null if the current node =
is=20
the Document itself).=C2=A0 Or you can just use XMLCs utility class/metho=
d=20
DOMOps.getDocument(Node) to avoid having to care about where if the hiera=
rchy=20
you are.

In any case, once you have the real owner document, you should just be de=
aling=20
with the plain old Xerces DOM, not XMLC specifics, other than the XHTML=20
specific extension classes (akin to the HTML DOM).=C2=A0 That said, the X=
HTML DOM=20
serves the same purpose as the HTML DOM.=C2=A0 It just provides an [X]HTM=
L-specific=20
API, such as getBody().=C2=A0 All the core DOM implementation details are=
 left to=20
the base Xerces DOM.=C2=A0 It's worth some more investigation, but I can'=
t think of=20
a way that issues with getNextSibling() could possibly be influenced by t=
he=20
XHTML/HTML DOM subclasses?

You might want to just ping the Xerces-user list.=C2=A0 Michael Glavassev=
ich, the=20
lead for Xerces, is pretty responsive and can probably tell you if this i=
s, or=20
has been, an issue with Xerces.


On Tue, 15 Jun 2010 18:35:49 +0200
  Franck Routier <[email protected]> wrote:
> Hi Jake and thanks for your input,
>=20
> - my jars seem to be ok (I have xerces 2.9.1 and xml-apis 1.3.04),
> (although I cannot exclude a classpath problem, but I don't think there
> is one).
> - using html dom, I can access the body element just fine (but my main
> project is using xhtml all over the place)
> - using getElementsByTagNale("body") is not any more lucky (length =3D =
0).
>=20
>=20
> In the object I get on view.getDocument(), I find an attribute called
> fDocument, which is my XHTMLDocumentImpl. The Document contains a
> docElement (XHTMLHtmlElementImpl), that contains a first child
> (XHTMLHeadElementImpl), that has no sibling...
>=20
> But the object that view.getDocument() returns also contains attributes
> like $element_My_Id (I guess these are generated by xmlc, isn't it ?).
> Those elements have a ownerDocument, which is fDocument. And some of
> them have a ownerNode that happens to be a XHTMLBodyElementImpl, and
> that XHTMLBodyElementImpl has the same ownerDocument...
>=20
> So the document structure seems fine, there really must be a bug with
> getNextSibling or something.I'm not sure out to replicate the problem
> outside of my project... if I can do so, I'll post on xerces mailing
> list. Else, I'll just suspend my work on the subject ad wait for xerces
> 2.10.0.
>=20
> Thanks
>Franck
>=20
>=20
>=20
> Le lundi 14 juin 2010 =C3=A0 10:52 -0500, Jacob Kjome a =C3=A9crit :
>> Hi Franck,
>>=20
>> That is strange.  Do you experience the same issue when using the HTML=
 DOM=20
>> rather than the XHTML DOM?  What happens if you use the HTML DOM inter=
face=20
>> method HTMLDocument.getBody()?  How about getElementsByTagName("body")=
 (try=20
>> with and without providing the XHTML namespace)?
>>=20
>> Hmm... I just performed the following search...
>> http://www.google.com/search?q=3Dchild.getNextSibling%28%29+incorrectl=
y+returns+null
>>=20
>> ...and the following results look a bit suspicious...
>>=20
>> http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=3D6333993
>> http://code.google.com/p/android/issues/detail?id=3D779
>>=20
>> When you run your server, try placing xercesImpl-2.9.1.jar in the serv=
er's=20
>> classpath and place the xml-apis.jar that ships with Xerces 2.9.1 in a=
n=20
>> "endorsed" library directory.  Do you get the same results?
>>=20
>> Besides that, it could be related to a recently fixed bug in Xerces de=
aling=20
>> with transferring deferred nodes when using adopNode() (or cloneNode()=
,=20
>>which=20
>> might produce similar behavior and which XMLC does whenever it serves =
up a=20
>> copy of the DOM).  You could try building from the latest Xerces sourc=
e or=20
>> just wait until 2.10.0 is released, which should be any time now accor=
ding=20
>>to=20
>> reports on the Xerces user list.
>>=20
>> https://issues.apache.org/jira/browse/XERCESJ-1450
>>=20
>>=20
>> Jake
>>=20
>> On Mon, 14 Jun 2010 11:54:29 +0200
>>   Franck Routier <[email protected]> wrote:
>> > Hi,
>> >=20
>> > I am playing with components in Barracuda, that I'd like to render i=
n
>> > the body tag of the document (these are script components, currently
>> > rendered in the head, but yslow and google speed suggest it might be
>> > better to render some of them at the end of the document).
>> >=20
>> > So, in my renderer, I access to the Document using the following cod=
e :
>> >=20
>> > Node node =3D view.getNode();
>> > Document doc =3D node.getOwnerDocument();
>> > Element elRoot =3D doc.getDocumentElement();
>> >=20
>> > So here, doc is an ....Impl generated by xmlc, and elRoot correspond=
s to
>> > the htlm tag (in my example I have a XHTMLHtmlElement).
>> > Until now I'm fine.
>> >=20
>> > Node child =3D elRoot.getFirstChild();
>> >=20
>> > --> child is of type HTMLHeadElement. Ok.
>> >=20
>> > child =3D child.getNextSibling();
>> >=20
>> > --> child is null. Argh.
>> >=20
>> > Is there an explaination why my document does not give me access to =
the
>> > body Node ?
>> >=20
>> > Thanks in advance,
>> >=20
>> >Franck
>> >=20
>> >=20
>> >=20
>>=20
>> pi=C3=A8ce jointe document texte brut (message-footer.txt)
>> --
>> You receive this message as a subscriber of the [email protected] mailing l=
ist.
>> To unsubscribe: mailto:[email protected]
>> For general help: mailto:[email protected]?subject=3Dhelp
>> OW2 mailing lists service home page: http://www.ow2.org/wws
>=20
>=20
>=20
>=20


------------=_1276630015-5928-2118
Content-Type: text/plain; charset="UTF-8"; name="message-footer.txt"
Content-Disposition: inline; filename="message-footer.txt"
Content-Transfer-Encoding: quoted-printable


--
You receive this message as a subscriber of the [email protected] mailing list=
.
To unsubscribe: mailto:[email protected]
For general help: mailto:[email protected]?subject=3Dhelp
OW2 mailing lists service home page: http://www.ow2.org/wws

------------=_1276630015-5928-2118--