Let's make this more of a normal algorithm declaration:

To fire a focus event named event at an element target given an element relatedTarget and an optional boolean bubbles (default false), ...

And then update call sites to uniformly be either

[fire a focus event] named foo at bar given baz

or

[fire a focus event] named foo at bar given baz and true

This * means that they belong to the "Window-reflecting body element event handler set" instead of being the normal list at the top of https://whatpr.org/html/10235/webappapis.html#event-handlers-on-elements,-document-objects,-and-window-objects .

Can you confirm that difference is tested? (I'm honestly not sure exactly how you would test, this will take a bit of research. Or maybe change something in the implementation and see if any tests start failing.)

Ok, I see that for example onclick is in the first table and onfocus is in the second one.
I made a quick test page, and it would seem that document.body doesn't do onfocus, but it does to onclick, and that was the only difference that I found.

Does that match up with your understanding?

skimming the chromium code, I'm not sure what the difference is between onclick and onfocus.
They are both listed equally in all these files:

• https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/dom/global_event_handlers.idl
• https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/html/html_attribute_names.json5
• https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/dom/events/event_target.h;l=226-331;drc=82609dbef7832b13135ab4b9cd2f2a5770021677

I think the difference in your test is due to the difference between click being bubbling and focus not.

The example below https://html.spec.whatwg.org/#the-body-element:window-reflecting-body-element-event-handler-set seems to have a possible test case at least for bubbling events.

I bet in Chromium there's some list hidden somewhere for the special ones.

Ah I see, here it is: https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/html/html_body_element.cc;l=115-216;drc=c0c69be50d661fd4952865e0aaf76952d9272f17
These are just for when they're set as HTML attributes rather than element.onfocusin which I am trying to implement in chromium to match safari.
It looks like chromium already implements onfocusin as an HTML attribute, but it is not included in the special body->window logic that onfocus has: https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/html/html_element.cc;l=481-482;drc=040547c1a7d0987ca63e18ca010986113637bf51

So I guess that we should put onfocusin/onfocusout in the list that onfocus is not in? So I should remove that asterisk?

I also made a test page here, but the results only made me confused: https://jsfiddle.net/jarhar/84jLcptn/5/

So I guess that we should put onfocusin/onfocusout in the list that onfocus is not in? So I should remove that asterisk?

Yes, I think that makes the most sense.

For testing, I figured out what we can do.

If you look at https://github.com/web-platform-tests/wpt/blob/master/html/webappapis/scripting/events/resources/event-handler-body.js , it pulls a list of all event handlers from the html.idl file. It then uses the windowReflectingBodyElementEventHandlerSet declaration to test how they should behave.

So, if we decide these should be in the "normal" list, we do no work. Updating the spec will eventually cause html.idl in the WPT repo to update, which will automatically make those tests start testing the right thing.

For your implementation, to verify everything is working correctly, you can modify the html.idl file locally, and make sure the tests still pass. When I do that right now, Chromium does not yet pass, I'd guess because the content attribute reflection isn't implemented. If you can confirm your patch makes such a local change pass, then I'll feel confident about our test coverage.