cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robby Pelssers <>
Subject RE: HTML5 serializer
Date Fri, 06 Jan 2012 16:08:32 GMT
For all I know the serializer does not actually output anything directly.  It hands over the
task to the transformerhandler and this is where the culprit resides.

There is no need to extend the current serializer if  I adapt the current way of working.

XmlSerializer, Html4Serializer and XhtmlSerializer are all just XmlSerializers with  a different
set of properties and the current XMLSerializer class provides static methods to create them.
It makes much more sense to add a static factory method for a html5 serializer  there as well.

The problem is that currently to actually have the transformerhandler output a doctype declaration
you do need to pass I think a doctypepublic property but this cannot be empty from the looks
of it:

    public XMLSerializer setDoctypePublic(String doctypePublic) {
        if (doctypePublic == null || EMPTY.equals(doctypePublic)) {
            throw new SetupException("A doctype-public has to be passed as argument.");

        this.format.put(OutputKeys.DOCTYPE_PUBLIC, doctypePublic);
        return this;

    public XMLSerializer setDoctypeSystem(String doctypeSystem) {
        if (doctypeSystem == null || EMPTY.equals(doctypeSystem)) {
            throw new SetupException("A doctype-system has to be passed as argument.");

        this.format.put(OutputKeys.DOCTYPE_SYSTEM, doctypeSystem);
        return this;


From: Jasha Joachimsthal []
Sent: Friday, January 06, 2012 5:00 PM
Subject: Re: HTML5 serializer

Ok, then create an HTML5Serializer that extends the current Serializer. An other solution
would be to add a boolean that will output differently for html5 but I'd prefer extension
above a number of if statements.


On 6 January 2012 16:56, Robby Pelssers <<>>
I am using Cocoon2.2 but am planning to switch to C3 in the upcoming months.  And  in my mail
I was actually referring to C3.    You are right about what you write but I'd prefer to have
a Serializer which follows the spec so I can just copy the output and validate it without
errors and too many warnings at


From: Jasha Joachimsthal [<>]
Sent: Friday, January 06, 2012 4:51 PM
Subject: Re: HTML5 serializer

Hey Robby,

which Cocoon version are you using for your project? In C2.1 and C2.2 there's not only a XMLSerializer
but also an HTMLSerializer and XHTMLSerializer for their specific needs. So why not create
your own HTML5Serializer?

In HTML5 the specification teams tried to specify what browsers were already doing instead
of making a new theoretical specification. HTML5 should be backwards compatible with previous
(X)HTML versions. This is the reason why some old elements are not deprecated but considered
obsolete (remember marquee, it was so cool on Geocities).
The doctype doesn't really matter, browsers generally ignore the PUBLIC part in the doctype
(apart from some hacks in IE going into quirks mode).
A good presentation about HTML5 is

Jasha Joachimsthal

Europe - Amsterdam - Oosteinde 11, 1017 WT Amsterdam - +31(0)20 522 4466<tel:%2B31%280%2920%20522%204466>
US - Boston - 1 Broadway, Cambridge, MA 02142 - +1 877 414 4776<tel:%2B1%20877%20414%204776>
(toll free)<>
On 6 January 2012 15:48, Robby Pelssers <<>>
Hi all,

I've been looking at how to add a HTML5 serializer to the project.

So far my investigations have led to add following code to org.apache.cocoon.sax.component.XMLSerializer

    public static XMLSerializer createHTML5Serializer() {
        XMLSerializer serializer = new XMLSerializer();


        return serializer;

Using the HTML5 serializer in a test to print the output:

    public void testHTML5Serializer() throws Exception {
        ByteArrayOutputStream baos = new ByteArrayOutputStream();

           new XMLGenerator("<html><head><title>serializer test</title></head><body><p>test</p></body></html>")

        String data = new String(baos.toByteArray());

Would print

<!DOCTYPE html PUBLIC "XSLT-compat">
<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>serializer test</title>

I read a number of articles describing the issues with serializing html5 and so far this was
the best I could come up with which is not 100% conforming due to

*         Non matching doctype although it will not break in the browser  --> should be
<!DOCTYPE html>

*         The charset should be <meta charset="UTF-8"/> according to html5 spec

Does anyone have more knowledge on this subject?


View raw message