Re: [CSS3 writing-modes] inline-block-alignment-001 , vertical-alignment-022 , inline-block-alignment-006-rl , text-baseline-006-lr

Le 2014-04-27 18:39, 塩澤 元 a écrit :
> Hi Gérard,

Hello Hajime,

Sorry for the long delay. I have had to deal with personal issues 
(moving to a new apartment) and am still struggling with other issues 
(torn tendom in elbow) these days.

> I divided inline-block-alignment-001.xht into six files.


needs to be updated so that it would have only 1

<div class="control control1">LLLL<span>ABCD</span></div>
<div class="control control2">LLLL<span>LLLLL</span>LLLL<img 
src="support/blue-horiz-line-320x1.png" alt="Image download support must 
be enabled" /></div>

and not 7.

> I created pull request (

I have not used github so far. I'm not even sure how to see how tests 
render with and then without changes... I haven't had time to check this 

> Could
> you review it?
> This is a warm-up commit to tackle with fixing other test.

Some comments on


I wonder why your test has a yellow rectangle. Maybe it's because of me! 
This was probably to visually identify the inline-block. The tester (the 
person taking the test) would probably wonder what that yellow rectangle 
is or means... or what is it supposed to be doing in the test or if it 
is part of the test.

I think the yellow area should be orange unless the yellow area 
(rectangle or square) is actually part of the test, is actually part of 
the test conditions.

If a test file has 7 sub-tests, then it's important to identify 
individually those 7 sub-tests: that's why we use the orange color to 
visually paint each of these 7 sub-tests. But if the test file only has 
1 test, then the orange color is not important and is not part of the 
test. Only vertical alignment of glyphs would matter then inside 1 
individual test.... and only vertical alignment of glyphs is actually 
the purpose of that test.

So, if tests are individual, I am more inclined to believe that tests do 
not need the orange color painting and tests may not even need a blue 
line anyway. If the "L" glyphs are big enough (say, 4em) and close to 
each other, then it should be easy for a tester under normal conditions 
to see if the bottom of "L" glyphs are aligned or not.

Without orange color, yellow color and blue line, the test could be just 
saying "Test passes if the bottom of each 7 'L' glyphs are aligned." or 
"Test passes if the bottom of each 7 'L' glyphs are vertically aligned." 
or "Test passes if the bottom of each 7 'L' glyphs are lined up."

Example given:

When the dominant baseline is alphabetic, then tests do not need to use 
Ahem font...

More later...


> Hajime.
> 2014-04-07 7:30 GMT+09:00 Gérard Talbot <>:
>> Hajime,
>> I did a quick checking of writing-modes tests recently. Here's some
>> feedback
>> 1-
>> submitted/css3-writing-modes/inline-block-alignment-001.xht
>> I believe that at line 78:
>> <div class="control" id="control1">LLLL<span>ABCD</span></div>
>> should be instead
>> <div class="control" id="control1">LLLL<span>LLLLL</span></div>
>> I believe such inline-block-alignment-001.xht test should be broken 
>> into 6
>> smaller tests.
>> For best test checking, you want a test to require a viewport height 
>> of
>> 400px or less. This is for test checking under the test harness.
>> ------------
>> 2-
>> submitted/css3-writing-modes/vertical-alignment-022.xht
>> <img class="line" src="./support/blue-horiz-line_220x1.png" alt="Image
>> download support must be enabled" />
>> is incorrect:
>> it should be
>> blue-horiz-line-220x1.png
>> underscore character should be hyphen character.
>> ------------
>> 3-
>> submitted/css3-writing-modes/inline-block-alignment-006-rl.xht
>> There are 3 orange rectangles, not 2 orange rectangles.
>> ------------
>> 4-
>> submitted/css3-writing-modes/text-baseline-006-lr.xht
>> This test seems incorrect. If "central baseline is used as the 
>> dominant
>> baseline", then all glyphs of different font-size should be "sitting" 
>> on
>> such central baseline; therefore at line 42
>> 'vertical-align: text-top'
>> can not be correct, can not fit into the logic of that test. By 
>> replacing
>> the glyphs in the test with "É" glyphs, I am 99% sure that
>> 'vertical-align: text-top'
>> should be removed.
>> +CC: Koji Ishii
>> Gérard
>> --
>> Web authors' contributions to CSS 2.1 test suite
>> contributions-css21-testsuite.html
>> CSS 2.1 Test suite RC6, March 23rd 2011

Web authors' contributions to CSS 2.1 test suite
CSS 2.1 Test suite RC6, March 23rd 2011

Received on Tuesday, 20 May 2014 17:25:14 UTC