Re: [css3-writing-modes] editorial comments on Appendix D: Intrinsic Dimensions

(12/07/31 9:17), fantasai wrote:
> On 07/16/2012 09:01 AM, Kang-Hao (Kenny) Lu wrote:
>>    # For the layout models in CSS2.1, both the min-content extent and
>>    # max-content extent of non-replaced elements are defined as the
>>    # content extent as defined (for horizontal writing modes) in
>>    # CSS2.1§10.6.3 and CSS2.1§17.5.3 for elements with ‘height:
>>    # auto’.
>>
>> Does this mean using 'width: min-content' and 'width: max-content'
>> respectively? Can this be clarified a bit? Also, it would be better if
>> this is part of the definition of the keywords
>> 'min-content'/'max-content', I think.
> 
> "extent" means the logical height, so, yes, it means 'width: min-content'
> and 'width: max-content', but only when in vertical writing mode. :)

I think I didn't described my question clearly. To be precise, my
question is, for the following three examples

data:text/html,<div style="width: 2em;
                           height: auto;
                           border: red solid;">a b c</div>

data:text/html,<div style="width: 2em;
                           height: min-content;
                           border: red solid;">a b c</div>

data:text/html,<div style="width: 2em;
                           height: max-content;
                           border: red solid;">a b c</div>

, do they look the same because 'height: max-content' and 'height:
min-content' are treated as 'height: auto' or do they look different
because 'height: min-content' is the height when <div> is laid out in
'width: min-content' and so on.

My guess after reading the spec was the latter but you seem to indicate
that the former is what's intended (Firefox14 doesn't support
'-moz-max-content/-moz-min-content' on 'height', matching the former
interpretation), but if that's the case:

1. If the spec doesn't just say 'max-content'/'min-content' is treated
as 'auto' on the "block-size property", the sentence should also mention
CSS2.1§10.6.7 ('Auto' heights for block formatting context roots), and
essentially every height calculation algorithm.

2. The sentence seems to indicate that 'max-content extent'/'min-content
extent' are definite size or otherwise why give the "auto height" these
many names? Compared to "max-content/min-content measure", are the
concepts of "max-content/min-content extent" used anywhere?

>> Intrinsic Sizes in Multi-column Layout
>>
>>    # The min-content and max-content sizes of a multi-column element are
>>    # undefined per [CSS3COL].
>>
>> I think this is significantly confusing because you later define
>> 'min-content'/'max-content' using the min/max-content measure of the
>> element's *contents*. My suggestion, if I understand the intention
>> correctly:
>>
>>    | The behavior when 'min-content'/'max-content'/'fit-content' is
>>    | specified on 'width' or 'height' of a multi-column element is
>>    | undefined.
>>
>> Or otherwise, what exactly is this paragraph trying to say?
> 
> Well, you have to understand that the min-content sizes of an element
> are used and need to be defined even when not explicitly specified,
> as they're part of the auto layout algorithms for e.g. tables and floats.
> So keying off specified 'width' or 'height' values in these definitions
> is not useful.

I see. But I think the spec should then define the
'min-content'/'max-content' sizes of the element's *contents*, say,
using a hypothetical non-multi-column element or something. The current
behavior for 'column-width: min-content' and such just look undefined.

By the way, what are the use cases here (for 'column-width')?


Cheers,
Kenny
-- 
Web Specialist, Oupeng Browser, Beijing

Received on Thursday, 9 August 2012 15:44:07 UTC