Re: [whatwg/dom] Meta: a list of side effects that would be impacted by state-preserving move (Issue #1270)

> `moveBefore()` doesn't really introduce that kind of ability, if I understand your point correctly.

The thing that `moveBefore` introduces and that *could* matter here, is that before, moving an `<iframe>` or an `<object>` or any other element embedding another accessible document would unload said document and clear its list of animation frame callbacks. So before, if you moved such an element during an animation frame, I suppose that its document would get removed from the list and the new one might not be part of it yet (though the specs don't seem to really handle this case, so I'm not sure what's really supposed to happen). With `moveBefore` the documents aren't unloaded and so it's important whether the list `docs` is a static list or dynamic, and that could now be observed.

> Yes it seems that Chromium is wrong here. [...]

Interesting that this "bug" can be observed even today, still I think that `moveBefore` through its ability to keep alive the previous docs could introduce new observable discrepancies like this. E.g. I don't think it's possible today to test that the list of `docs` keeps its order during each steps of the task, nor is it clear if it should.


-- 
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/1270#issuecomment-2288619042
You are receiving this because you are subscribed to this thread.

Message ID: <whatwg/dom/issues/1270/2288619042@github.com>