[css-overscroll-1] Should overscroll-behavior change the behavior of a subframe?

[smfr]

https://drafts.csswg.org/css-overscroll

Should overscroll-behavior change the behavior of a (possible cross-origin) subframe? This seems like leakage of CSS through frame boundaries, which isn't something we normally do.

[fred-wang]

I think @smfr has been confused by my (probably incorrect) initial WebKit's patch and testcase at https://bugs.webkit.org/show_bug.cgi?id=176454 where I used something like <iframe style="overscroll-behavior: none">

Reading the current overscroll-behavior spec, overscroll-behavior only applies to scroll containers (https://drafts.csswg.org/css-overflow-3/#scroll-container i.e. have scroll overflow as specified by overflow-x/overflow-y). So I don't think overscroll-behavior applies to iframe.

I think it would be consistent with scroll-behavior here, which applies to scrolling boxes ( https://drafts.csswg.org/cssom-view/#scrolling-box ) instead and says " When this property is specified on the root element, it applies to the viewport instead.".

See also the related #2977 about whether the property should be read on the root element or on the scrolling element. This only matters for quirks mode.

[hiikezoe]

We have encountered a similar issue, in our case the iframe document element has "overscroll-behavior-y: none" but the iframe document in question is not scrollable at all. In such cases whether the overscroll-behavior property should be respected or not? Chrome appears not respecting it, whereas Firefox does as of now.