getElementById support for fakedom

[tobibeer]

Note: This is intended for comparison and @Jermolene has already expressed no desire to merge it.

Uses a basic hash-table applying "the winner takes it all" logic. Specifically, the first element to register an id is the one the id points to, so long as...

• it exists
• the id attribute is not removed from it

[pmario]

IMO this implementation is not according to the standards.

ID is a MUST be unique in the whole document in the html spec: https://html.spec.whatwg.org/multipage/dom.html#the-id-attribute

Also see browser implementations: https://developer.mozilla.org/en-US/docs/Web/HTML/Global_attributes/id and the linked living standard at: https://dom.spec.whatwg.org/#interface-nonelementparentnode which imo defines the behaviour for getElementById()

The getElementById(elementId)
method must return the first element, in tree order, within context object’s descendants, whose ID is elementId, and null if there is no such element otherwise.

So imo the implementation is a bit more complicated, than implemented here, since it basically takes the "starting node" into account and returns the first descendant node, that matches the id string.

There are a lot of links in the standards page, that I didn't copy over, so follow the link

[tobibeer]

ID is a MUST be unique in the whole document in the html spec

It is. Only one element will be returned for a given id. However, I have changed the PR so that a reference for an id, once set, can only be overwritten if...

• the node that registered it first no longer exists
• the id is removed form that node via removeAttribute()

This appears to conform to the standard.

[tobibeer]

Possibly, removing...

1. the id attribute from a node
2. any node having an id

...also requires to reevaluate whether or not other nodes still exists that also have it (wrongfully) assigned.

I think all this is quite superfluous and highly irrelevant for our use cases of rendering static pages.

If you think it's worth the trouble, then I'll add arrays for nodes per id that can be checked for any left-over nodes now wishing to register the id. So, I'll iterate over a stack while shifting away any nodes that no longer exist, e.g.

function getElementById(id) {
    var nodes = elements[id];
    $tw.utils.each(nodes, function(el) {
        if (document.contains(el)) {
            return false;
        } else {
            els.shift();
        }
    }};
    return nodes ? nodes[0] : undefined;
}

[tobibeer]

With respect to my last comment, there would sure be another little bit of functionality not in fakedom: document.contains(node).