The :navigating-to() pseudo-class feels awkward

[dbaron]

The initial "What if it could look like this?" example in the route matching explaner uses a :navigating-to() pseudo-class, although this isn't otherwise discussed.

While the power of a :navigating-to() pseudo-class seems like it could be useful in some cases, this example seems unappealing in one way since it looks like a case where dealing with multiple links requires repetitive separate styles for each link.

That said, I see two different options for resolving this, and it's not clear to me if only one is needed or if both are needed:

• a pseudo-class indicating that this link is being navigated because the user activated it
• a pseudo-class indicating that the target of this link is being navigated to, whether via this link or another (in-page, for security/privacy!) mechanism

I'm reasonably confident that in the cases where the above two differ, that some developers would want the first. I'm less sure of whether or not other developers would want the second.

Given that this isn't a fleshed-out feature I expect you're planning further work on this in the future, but I figured I'd write this down in an issue since I was thinking about it as I read the explainer.

[noamr]

Agreed, note that at the top of the explainer I'm linking to issues where the syntax is discussed. My current thinking for CSS is here. All good comments though and lets flesh it out!

[fantasai]

I would separate the questions of "where is this link targetting" (what :navigating-to() seems to be about) from "is this link being navigated to" (the two points dbaron mentions above). The first one I think would make sense to extend :link itself, e.g. :link(<urlpatternargumentthing>). The second one would just be a simple state, like :active but with a different name.

How would this interact with JS navigations? Would the new :active pseudo remain active until the JS is completed or what?

[noamr]

I would separate the questions of "where is this link targetting" (what :navigating-to() seems to be about) from "is this link being navigated to" (the two points dbaron mentions above). The first one I think would make sense to extend :link itself, e.g. :link(<urlpatternargumentthing>). The second one would just be a simple state, like :active but with a different name.

How would this interact with JS navigations? Would the new :active pseudo remain active until the JS is completed or what?

It would follow the navigation API lifecycle, which is well defined, and is also linked to the browser's loading spinner.