Re: [css-page-floats] The 'clear' values are backwards

> On Jan 26, 2016, at 11:58 AM, Johannes Wilm <johanneswilm@gmail.com> wrote:
> 
> 
> 
>> On Tue, Jan 26, 2016 at 8:24 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
>> On Tue, Jan 26, 2016 at 11:20 AM, Brad Kemper <brad.kemper@gmail.com> wrote:
>> >> On Jan 26, 2016, at 11:09 AM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
>> >> On Tue, Jan 26, 2016 at 7:40 AM, Brad Kemper <brad.kemper@gmail.com> wrote:
>> >>>> On Jan 25, 2016, at 5:05 PM, Johannes Wilm <johanneswilm@gmail.com> wrote:
>> >>>> After a small conversation offlist: We seem to really have been in agreement. We meant the same thing, we just used different words for it.
>> >>>>
>> >>>> So the conclusion is:
>> >>>>
>> >>>> If we have three floats rights after oneanother where the first float is placed on page one. And the second float has the clear property set so that it will be placed on page two (and not page one), then also float three (which doesn't have the clear property set) will be placed on page 2 (or later if there is not enough space on page 2). Float three will therefore not be placed on page one, even though there would be enough space for it there. We will make sure that the wording ensures that. This is also what inline floats do, so pagefloats are no exceptions.
>> >>>
>> >>> Wouldn't it be all subsequent content that would go to the next page? 'Clear: left' causes all subsequent content to clear the left float.
>> >>>
>> >>> Would 'clear: top' set on a non-float also cause it to move to the top of the next page? Clear:left can be set on any element, not just those with float:left.
>> >>>
>> >>> Is this the same as break-before:always?
>> >>
>> >> Page floats are already disconnected from the flow of in-flow content,
>> >> in a way that inline floats aren't.  Clearing a page float shouldn't
>> >> effectively cause a break.
>> >>
>> >> (This *is* a change from how 'clear' works today, but the change is
>> >> actually more in the way that floats operate when they're "page"
>> >> floats; we'd just be applying that change consistently.)
>> >
>> > So, setting 'clear: top' on a non-float wouldn't affect that element, but would cause subsequent floats to go to the next page/fragmentainer?
>> 
>> I hadn't thought that far ahead, but sure.
> 
> I think this would be useful - but I wonder how we figure out what fragmentainer we target. Should the element also have a float-reference? Do we simply pick the lowest one? Some other way?

Maybe something like 'clear: top column', and if the second value is omitted then pick the lowest (deepest, closest box ancestor) one.