Create HTMLElement instead of pure Element

[krichprollsch]

Our toInterface expect the given Element is always an HTMLElement.
So I think it's better to create always HTMLElement. No b/c we could want to create a non-HTML tag, eg. svg

But I would also like to detect if an Element is really an HTMLElement to avoid crash on toInterface, but I can't figure out one correct solution... done using _dom_element_get_tag_name ✔️

depends on lightpanda-io/libdom#32

[karlseguin]

I think you could stop it from crashing by calling _dom_element_get_tag_name (which I don't think we have exposed in netsurf.zig) instead of _dom_html_element_get_tag_type ?? You could then maybe handle non-HTML tag names (svg, ...??).

[krichprollsch]

So as you suggested, as used _dom_element_get_tag_name + Tag.fromString instead of _dom_html_element_get_tag_type to retrieve the element's Tag.
I think it makes more sense this way: Node.toInterface => Element.toInterface => HTMLElement.toInterfaceFromTag

[krichprollsch]

I pushed more changed to remove netsurf.elementHTMLGetTagType

But i have html/dom/documents/dom-tree-accessors/document.title-09.html crashing again (on c._dom_html_document_create_element_ns).

so still in progress...

[krichprollsch]

@karlseguin WPT tests are passing now.

[krichprollsch]

Hum, Maybe I should keep try Element.toInterface signature unchanged and create a toInterfaceT(comptime T: type, e: *parser.Element) !T { instead to avoid all these changes 🤔

[krichprollsch]

ok here is a cleaner version

[karlseguin]

I know everything in netsurf.zig is inlined, but Andrew's not a fan:

https://ziggit.dev/t/inlining-functions/1341/3

[karlseguin]

Is there a reason to keep the originals? If our document is always an html document, then it will always create an HTMLElement. Only reason I can see to keep both is if we think we'll have a non-html document??

[krichprollsch]

You can create non-html document from scratch in JS and then add elements to them.
If I try to use only HTML element creation, WPT test is crashing on https://github.com/lightpanda-io/wpt/blob/main/html/dom/documents/dom-tree-accessors/document.title-09.html.

It's b/c the document isn't an HTML one.

So depending the document, we want to create HTMLElement or Element...
I could maybe detect the document's type and use the best function...

[krichprollsch]

I'm afraid I can't find a way to detect if a document is an HTML one...
The constructor doesn't set any base doc attribute with a specific value.
https://github.com/lightpanda-io/libdom/blob/master/src/html/html_document.c#L93-L116

[krichprollsch]

ok I did with lightpanda-io/libdom#32

[karlseguin]

LGTM