[css-view-transitions] Making the callback param non-nullable

[vmpstr]

In css-view-transitions, we define the following, an optional nullable callback with the default value of null:

partial interface Document {
  ViewTransition startViewTransition(optional UpdateCallback? updateCallback = null);
};

In #8960 we resolved to change the callback (in Level 2) to take either a callback or a dictionary of options. Naively, this would have been the following

ViewTransition startViewtransition(optional (UpdateCallback or Dictionary)? param = null)

But that's not possible because of the following restriction:

A nullable type is an IDL type constructed from an existing type (called the inner type) [...] The inner type must not be:

• [...]
• a union type that itself includes a nullable type or has a dictionary type as one of its flattened member types.

A simple solution is to just make these non-nullable:

ViewTransition startViewtransition(optional (UpdateCallback or Dictionary) param)

Additionally, for consistency, it would be nice to make the existing view transition callback non nullable as well, while still remaining optional:

ViewTransition startViewtransition(optional UpdateCallback updateCallback)

I believe the only script visible change here is that this would preclude startViewTransition(null) from being valid. I could be wrong though. If I'm right, then I think that's fine, since that doesn't seem like a correct use of this API.

[css-meeting-bot]

The CSS Working Group just discussed [css-view-transitions] Making the callback param non-nullable, and agreed to the following:

• RESOLVED: Make the parameter non-nullable, but keep it optional

The full IRC log of that discussion

<emilio> noamr: we have a startViewTransition call that takes one param that is marked as optional nullable with default of null
<emilio> ... we resolved to extend this to allow a dictionary
<emilio> ... so naively it'd be an optional nullable union
<emilio> ... but that's not possible in WebIDL
<noamr> it's vmpstr, not noamr
<emilio> vmpstr: so proposal is making it non-nullable
<emilio> ... this only precludes null of being a valid parameter
<emilio> ... it'd still be optional
<emilio> astearns: if we decide on null as a naive way of dealing with the optionality?
<emilio> ... was there a purpose?
<emilio> vmpstr: there wasn't, it was just in the editor's draft since the beginning
<khush> +1
<emilio> q+
<khush> q+
<astearns> emilio: so if null is passed we will start throwing?
<astearns> ack emilio
<emilio> vmpstr: that's right
<emilio> emilio: hopefully nobody is doing that, if you're fine with the compat risk seems fine
<emilio> vmpstr: usage is low enough
<emilio> emilio: sounds good then
<emilio> khush: to add to vmpstr there's no production usecase where you'd do that
<emilio> ... because the callback is where you are supposed to mutate the DOM
<emilio> vmpstr: to clarify it's still optional, it'd be just explicit null that could be problematic
<emilio> RESOLVED: Make the parameter non-nullable, but keep it optional