Race Condition in Anchor navigation

[cefn]

MDN URL

https://developer.mozilla.org/en-US/docs/Web/Progressive_web_apps/Introduction#re-engageability

What specific section or headline is this issue about?

Any anchored link

What information was incorrect, unhelpful, or incomplete?

I found that navigating directly to an anchor link only sometimes works. I am on Fibre broadband in the UK, but delays loading are still enough to introduce a race condition. My suspicion is that when the anchor is evaluated the content hasn't loaded, yet. This should be addressed by ensuring anchors are (re)evaluated after content load.

An example is as follows. Loading this URL should scroll directly to an inline anchor. However, it manages this only sometimes and more often than not it directly navigates to the head of the page as if there were no anchor.
https://developer.mozilla.org/en-US/docs/Web/Progressive_web_apps/Introduction#re-engageability

By contrast, navigating from links in the contents table of the Introduction page itself (after the page has already loaded) is successful and there is indeed an anchor in the page.

I first encountered this in Chrome by right-clicking from the contents table, and choosing to open an anchored link in a new tab (although I didn't know it was an anchor at the time). When I switched to the tab I found it hadn't navigated to any corresponding content at all.

What did you expect to see?

Navigation to anchors should work

Do you have any supporting links, references, or citations?

No response

Do you have anything more you want to share?

No response

[gaearon]

Not sure if it's the same issue, but I can consistently reproduce this fragment link not working in Chrome:

https://developer.mozilla.org/en-US/docs/Web/HTML/Element/input#methods

In other browsers, it scrolls to the anchor. But Chrome macOS consistently does not.

This feels pretty critical because MDN pages are long, and deep-linking to them is very valuable for external resources.

[caugner]

This sounds like #7168.

@cefn Would you be able to share a screen recording of the issue? That would be awesome. 🙏

[cefn]

I'm not sure there's a whole lot of value sharing a screencast of a behaviour that doesn't happen. Believe me, navigating to a URL with a hash is identical to navigating to the plain URL, which is easy to reproduce.

[caugner]

No worries, closing in favor of #7168.