new XMLSerializer().serializeToString(document); panics

[mateon1]

<script>new XMLSerializer.serializeToString(document)</script>

This causes a ScriptThread panic, Can't serialize Document node itself

servo/components/script/dom/servoparser/html.rs

Line 253 in fcd6beb

NodeTypeId::Document(_) => panic!("Can't serialize Document node itself"),

Both Chrome and Firefox properly serialize the entire page when serializeToString is called on the document.

[CYBAI]

As @jdm mentioned in #23044 (comment) , I think this will be fixed when servo/html5ever#368 is fixed.

[jdm]

Specifically, this would be handled by https://w3c.github.io/DOM-Parsing/#dfn-xml-serialization-algorithm.

[highfive]

Hi! If you have any questions regarding this issue, feel free to make a comment here, or ask it in the #servo channel in IRC.

If you intend to work on this issue, then add @highfive: assign me to your comment, and I'll assign this to you. 😄

[jdm]

That being said, fixing the panic here is separate from implementing the full algorithm in the html5ever repository. We should implement the document serialization algorithm by checking for a document in the SerializationIterator constructor and pushing all of the immediate child nodes of the document.

[jdm]

If this case is not already tested by tests/wpt/web-platform-tests/domparsing/XMLSerializer-serializeToString.html, then a new testcase should be added.

[5h1rU]

Hello!

would like to contribute here but probably I'll need some assistance since I'm a newbie on the codebase.
@jdm what do you think?

[jdm]

@5h1rU I think you should give it a try and ask questions about anything that is unclear :)

[5h1rU]

@highfive assign me

[highfive]

Hey @5h1rU! Thanks for your interest in working on this issue. It's now assigned to you!

[5h1rU]

After some hours trying to figure out what's going on I have some blockers

• html5ever serializer doesn't have the document serialization algorithm.

Then I change the Document match to:

NodeTypeId::Document(DocumentTypeId::Document) => {
    let doc = n.downcast::<Node>().unwrap();
    // What else?
},

I'm confused about the document serialization algorithm implementation. Not sure where is the best place to put the logic. I tried using the push_node from the SerializationIterator but no luck.

iter.push_node(doc);

@jdm Do you think is possible to use one of the current existent serializer methods?

In conclusion: I don't know what's the next step in order to return the document serialized.

[jdm]

@5h1rU This comment means that we should not modify the match with the panic. However, we should modify the SerializationIterator new method to check for a document node, iterate over its children and push each one of them without pushing the actual document node.