Re: [css3-page] float Rules for Pagination into Varying-Width Pages

(Apologies for the late reply.)

On 20/10/2011 08:47, Peter Moulder wrote:
> On Wed, Oct 19, 2011 at 11:23:22AM -0700, fantasai wrote:
>> On 10/19/2011 07:09 AM, Anton Prowse wrote:
>>> On 20/09/2011 04:31, Peter Moulder wrote:
>>>> On Mon, Sep 19, 2011 at 03:55:17PM -0700, fantasai wrote:
>>>>>
>>>>> However, if we require that a float that was split across pages begin at
>>>>> the top of the page (which I think we should), then that escape hatch is
>>>>> not available on subsequent pages. This could result in either overflow
>>>>> or overlap between floats, which is not normally possible.
>>>>
>>>> I'll note that this proposed change of rules for float widths isn't technically
>>>> necessary: without it, the rules of section 9.5.1 of CSS 2.1 would just mean
>>>> that the second float would be pushed down as far as necessary for it to fit on
>>>> all pages on which it occurs.
>>>>
>>>> Ignoring implementation issues, this would actually be preferable for authors:
>>>> no-one wants a float to overflow off the edge of a page on a subsequent page.
>>>
>>> I agree with this. I wouldn't want floats to be calculated according to dimensions on one page and then look bad on the
>>> continuing page; I imagine that the first impression of authors if they saw that would be that it were a UA bug.
>>>
>>> For me, this issue is even clearer in the case of regions rather than pages. I really don't think we want overlap or overflow.
>>
>> The question then is, if I get a layout like this
>>
>> +----------------------------------+
>> |###########           ############|
>> |###########           ############|
>> |###left####           ####right###|
>> |###float###           ####float###|
>> |###########           ############|
>> |###########           ############|
>> |###########           ############|
>> |###########           ############|
>> +----------------------------------+
>>
>> And on the second page:
>>
>> +-----------------+
>> |###########      |
>> |###left####      |
>> |###float###   A  |
>> |###cont.###      |
>> |###########      |
>> |     ############|
>> |     ###right####|
>> |     ###float####|
>> | B   ###cont.####|
>> |     ############|
>> |      ...        |
>
> There seems to be some confusion as to what proposal is being discussed.
>
> There are I think three approaches that have been discussed in this thread:
>
>    (1) vertically-contiguous floats (i.e. continuations always start at
>        the top of the next page), with float rules modified to overflow or
>        truncate if necessary;
>
>    (2) vertically-contiguous floats, with no modification to float rules
>        (each float is placed as early as possible such that it fits *on
>        all pages*), so no extra overflow.  This is the approach being
>        discussed above, described as "preferable [over option 1] to
>        authors, if we ignore implementation issues".
>
>    (3) vertically-discontiguous floats (i.e. continuations can start lower
>        down the page or even skip a page), with float rules modified such
>        that only one page is considered at a time for whether the float
>        fits.  (No other change to float rules, so no extra overflow.)
>        This is the option that fantasai is depicting above, but
>        is not mentioned in the quoted text.
>
> (Under option 2, the second of the two floats wouldn't start on the first
> page, given that it can't fit in one page, and has no solution in the
> second page if it does start on the first page.)

My understanding matched Peter's.  I thought we were expressing a 
preference between (1) and (2), with my preference being (2) in 
agreement with Peter.  The depicted case (3) seems odd to me; I don't 
think I'd want my float to be vertically-discontiguous.

>> Does text after the float wrap into section A? I think that would make
>> sense if you are using floats as floats (not as a substitution for e.g.
>> flexbox). But it would require some clarification, I think.
>
> I'm not sure I understand the question, but isn't it the same situation
> as in the unpaged case, if we discard the first page?

I would say so.

Cheers,
Anton Prowse
http://dev.moonhenge.net

Received on Monday, 12 December 2011 14:34:09 UTC