Re: [css-writing-modes-3] Proposal to modify how inline-block with non-empty block descendants are baseline-aligned

I somehow missed this until Tab mentioned this, sorry about that.

On Fri, Jun 12, 2015 at 9:37 AM, Gérard Talbot <www-style@gtalbot.org>
wrote:

> Hello,
>
> When an inline-block has non-empty block descendants, then
>

Minor details but just want to make sure we're on the same table. So the
topic is when an inline-block has multiple line boxes, not when all
descendants are blocks and it doesn't have any inline descendants, correct?
I originally thought the topic was when all descendants are blocks, but as
I understand more, all what you wrote also applies when I put simple text
in an inline-block and set width to wrap. Your clarification is appreciated.

"
> The baseline of an 'inline-block' is the baseline of its last line box in
> the normal flow (...)
> "
> CSS 2.1, 10.8.1 Leading and half-leading
> http://www.w3.org/TR/CSS21/visudet.html#leading
>
> This was decided *way before* the concept of dominant baseline and before
> central baseline-alignment were created and specified.
>
> Similarly,
>
> "
> The baseline of an 'inline-table' is the baseline of the first row of the
> table.
> "
> CSS 2.1, 10.8.1 Leading and half-leading
> http://www.w3.org/TR/CSS21/visudet.html#leading
>
> But when vertical writing mode spec was created, then we have different
> baselines:
>
> "
> In vertical writing mode, the central baseline is used as the dominant
> baseline when text-orientation is mixed or upright. Otherwise the
> alphabetic baseline is used.
> "
> 4.2 Text Baselines
> http://www.w3.org/TR/css-writing-modes-3/#text-baselines
>
>
> Now, I created a demo and then tested it with an inline-block that has 2
> non-empty block children:
>
>
> http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/vert-align-latin-baseline-inline-block-dhtml.html
>
> First of all, webkit-based browsers not only do not support an unprefixed
> version (when it should!) but they do *not* support the 'mixed' value but
> rather support the old 'vertical-right' value [1]; so that test has been
> adjusted so that it would work for webkit-based browsers.
>
> Steps:
>
> 1- Load
>
>
> http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/vert-align-latin-baseline-inline-block-dhtml.html
>
> into several browsers: say, Firefox 41 nightly build and Chrome 45
>
> 2- Make sure that 'mixed' and 'baseline' radio buttons are checked
>
> Firefox rendering: the inline-block margin box is centered within the line
> box
>
> Chrome rendering: the inline-block is baseline-aligned within the line box
> in accordance with the baseline of its own last line box.
>
> One could say that Chrome complies accurately with the current
> writing-mode spec: no objection here. But, is such rendering best (or
> ideal) for central baseline-alignment? Is such rendering what the spec
> editors wanted for central baseline-alignment?
>

We have two editors ;-) but at least for me, yes. In other words, I did not
think any other options exist, though it's primarily from CSS perspective,
not from use-case perspective.

-------
>
> Steps:
>
> 1- Load
>
> http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/vert-align-latin-baseline-inline-block-dhtml.html
> into several browsers: say, Firefox 41 nightly build and Chrome 45
>
> 2- Make sure that 'sideways-right' and 'baseline' radio buttons are checked
>
> Firefox rendering and Chrome renderings: the inline-block is
> baseline-aligned on the line box in accordance with the baseline of last
> line box.
>
> When 'text-orientation' is 'sideways-right', the spec says the alphabetic
> baseline should be used.
>

Confirmed the difference. Thanks to your interactive demo, it looks like
Firefox handles the central baseline just the same as "middle" if vertical
flow and text-orientation is mixed.


> -------
>
> Proposal:
>
> In vertical writing mode, when the central baseline is used as the
> dominant baseline, then the baseline of an inline-block is halfway between
> its under and over margin edges.
>
> In vertical writing mode, when the central baseline is used as the
> dominant baseline, then the baseline of an inline-table is halfway between
> its under and over margin edges.
>
> Webkit-based browsers would have to be fixed; Firefox would not have any
> modification to do.
>
> -------
>
> When the aplhabetical baseline is used as the dominant baseline, there
> would be no change:
>
> In vertical writing mode, when the aplhabetical baseline is used as the
> dominant baseline, then the baseline of an inline-block is the baseline of
> the last line box.
>
> In vertical writing mode, when the aplhabetical baseline is used as the
> dominant baseline, then the baseline of an inline-table is the baseline of
> the first row of the table.


Use case-wise, I think both use cases exist. I can't answer from top of my
heads which is more common.

>From CSS and implementer perspective, I prefer the current WebKit/Blink
behavior for three reasons:

1. The primary objective of the "central" vertical align is to align the
center of lines. The value "middle" was defined as the center of x-height,
which does not work for CJK use cases. So from my point of view, "central"
aligns characters to characters (from user perspective, sometimes <img> are
characters.)

2. Assuming my understanding of "both use cases exist", "middle" already
serves one, so making "central" the same as "middle" looses another use
case.

3. I prefer less differences when in horizontal and in vertical, unless
it's really needed. Taking the "central" of last line box in horizontal
flow while doing the "middle" of inline-block in vertical flow looks
unnecessary difference to me.

But this part was largely stolen from Steve's line box spec, fantasai wrote
most of text in this section, reviewed by WG members. I'd like to have
other opinions.

fantasai, Steve, could you chime in?

/koji