I was implementing view-transition-types in gecko, and there are two tests that I had to reverse-engineer Chrome to pass:
AFAICT there's nothing in https://drafts.csswg.org/css-view-transitions-2/#the-active-view-transition-type-pseudo that would otherwise prevent those from matching:
- The
ViewTransitionTypeSet can contain -ua- prefixed stuff and none just fine, it's just a set of case-sensitive strings.
- The
:active-view-transition-type() selector doesn't restrict parsing of <custom-ident> to avoid including -ua- and none.
The way this behavior is implemented in Chrome and WebKit is at match time. You never consider a view transition type that starts with -ua- or is an ascii-case-insensitive match for none active. But that's not in the spec.
We have a resolution in #9534 (comment) which seems to indicate that :active-view-transition-type(none) should be invalid, but that's also not what's implemented, what's the intended behavior here?
cc @nt1m @noamr @vmpstr