W3C home > Mailing lists > Public > public-css-testsuite@w3.org > February 2013

Re: [RC6] rgb(50%, ..., ...) or rgb(..., 50%, ...) or rgb(..., ..., 50%): fractional value!

From: Gérard Talbot <css21testsuite@gtalbot.org>
Date: Sat, 23 Feb 2013 00:59:49 -0500
Message-ID: <44e774362c1a9ffc4e72fe9afc19f6da.squirrel@ed-sh-cp3.entirelydigital.com>
To: "Rebecca Hauck" <rhauck@adobe.com>, "Arron Eicholz" <arron.eicholz@microsoft.com>
Cc: "Public CSS test suite mailing list" <public-css-testsuite@w3.org>

Le Ven 22 février 2013 21:15, "Gérard Talbot" a écrit :
>
> Le Jeu 10 janvier 2013 15:25, Rebecca Hauck a écrit :
>> Ok, I've cleared up the issues with all of these tests. They're
>> batched
>> in several checkins over the last few days: background-color[1],
>> border-left-color[2], border-right-color[3], border-top-color[4],
>> outline-color[5], color[6].
>>
>> For the background-color tests, I added a second reference file for
>> 049,054,070,075,090,095,110,115. These all use either 1% or 99% rgb()
>> values and now have pngs to match whether a UA rounds up or down.  For
>> the
>> rest of the background-color tests in this batch
>> (052-53,073-74,093-94,113-14), I changed 50% to 40% to eliminate the
>> fractional colors.
>
>
> http://test.csswg.org/source/approved/css2.1/src/backgrounds/background-color-049-020202-ref.xht
>
> http://test.csswg.org/source/approved/css2.1/src/backgrounds/background-color-049-030303-ref.xht
>
>    <meta name="flags" content="image" />
>

I can load the reftests and then, with an advanced text editor, remove
all those <meta name="flags" content="image" /> lines.


>
> is not required, not needed in reftests.
>
> "
> Unlike the format for the test file, there is no metadata except for the
> author credits and optional reference links.
> "
> http://wiki.csswg.org/test/reftest#the-reftest-reference-file
>
> There is another - admittedly small - issue with the
> background-color-nnn tests.
>
>         <div id="test"></div>
>         <br/>
>         <div id="reference">X</div>
>
> is not valid HTML. And in order to replace <br /> in the tests with a
> margin-bottom set on div#test or margin-top set on div#reference, we
> need to supply a defined, specified line-height value. 'line-height:
> normal' can be computed anything from 1.0 to 1.2 but is often between
> 1.10 and 1.20. A line box height of 19.2px can - theoretically speaking
> - be rounded up by some user agents too.
>
> I propose to drop, to remove <br /> in every background-color-nnn tests
> and in their related reftests) and just use a margin-top on the
> div#reference of, say, 20px (or 1.25em).

For consistency reasons, we should probably do what has been done in
other tests: adding a margin-top on the div#reference of 10px.

Eg.
http://test.csswg.org/source/approved/css2.1/src/borders/border-left-color-049.xht

            #reference
            {
                background-color: rgb(1%, 1%, 1%);
                margin-top: 10px;
            }


This can be done with an advanced text editor in a few steps.

And also do this on color-[002-145].xht tests.

>
> I'm checking the other tests...
>
> I wanted to approve and add
> <link rel="reviewer"
> to all the tests and reftests.
>
> Gérard
>
>>
>> For tests 049,052-54,073-75,093-95,113-15 in the
>> border-left/right/top/bottom-/outline-/color tests, I converted to
>> reftests using background-color in the reference elements and ref
>> files.

Okay, I see.

>>
>> Lastly, I updated color-002-145 to remove the pngs from the reference
>> elements to use background-color instead[7]. With the exception of
>> those
>> that I converted to reftests in this suite mentioned above, I kept ref
>> files using pngs.
>>
>> One other notable change that I made to all of the files was based on
>> a
>> discovery from running these in mozilla's and webkit's automated
>> harnesses. I changed the size of the font & img/div elements from 1in
>> to
>> 100px. Using 1in was causing some fuzzing around the edges and none of
>> these tests were pixel perfect. Changing to 100px fixed the problem.
>>
>> I've removed the 'Needs Work' flag from all of these in Shepherd and
>> these
>> are all in either Resubmitted for Review or Accepted status.
>>
>>
>> Please let me know if there are any issues with these changes.
>>

Just the meta-flag line in reftests and the <br> between 2 divs which
could be better replaced with a margin-top: 10px on the appropriate
elements (<div id="reference">).

Gérard


>> Cheers,
>> -Rebecca
>>
>>
>> [1] http://hg.csswg.org/test/rev/e29cb8ff4c58
>> [2] http://hg.csswg.org/test/rev/bd027b72439e
>> [3] http://hg.csswg.org/test/rev/369ab0dae5bc
>> [4] http://hg.csswg.org/test/rev/cbe4225d6c2d
>> [5] http://hg.csswg.org/test/rev/77847b6a522d
>> [6] http://hg.csswg.org/test/rev/5c61040b4f07
>> [7] http://hg.csswg.org/test/rev/e37a9b6098ef
>>
>>


-- 
Contributions to the CSS 2.1 test suite:
http://www.gtalbot.org/BrowserBugsSection/css21testsuite/

CSS 2.1 Test suite RC6, March 23rd 2011:
http://test.csswg.org/suites/css2.1/20110323/html4/toc.html

CSS 2.1 test suite harness:
http://test.csswg.org/harness/

Contributing to to CSS 2.1 test suite:
http://www.gtalbot.org/BrowserBugsSection/css21testsuite/web-authors-contributions-css21-testsuite.html
Received on Saturday, 23 February 2013 06:00:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 23 February 2013 06:00:24 GMT