- From: Håkon Wium Lie <howcome@opera.com>
- Date: Sat, 3 Aug 2013 07:03:01 +0200
- To: =?iso-8859-1?Q?"G\=E9rard\?\= Talbot" <css21testsuite@gtalbot.org>
- Cc: "Public css-testsuite mailing list" <public-css-testsuite@w3.org>
Based on your comments below, I have tried to re-engineer the test. I propose: http://people.opera.com/howcome/2013/tests/multicol/column-fill-auto.html http://people.opera.com/howcome/2013/tests/multicol/column-fill-balance.html (The tests have a build-in reftest, separate reftest files should be extracted once we agree on the tests themselves.) -h&kon Håkon Wium Lie CTO °þe®ª howcome@opera.com http://people.opera.com/howcome Gérard Talbot wrote: > [src] > http://test.csswg.org/source/contributors/opera/submitted/multicol/multicol-fill-002.xht > > [reftest] > http://test.csswg.org/source/contributors/opera/submitted/multicol/multicol-fill-ref.xht > > 1- > The filename should be renamed to multicol-fill-balance-001.xht . That > way, a set of tests testing 'column-fill: balance' would be easier to > find/get. > > 2- > The test in its current version passes in UAs which do not support > multi-column. > > 3- > Computed font-size of multi-column element should be dividable by 5px > without remainer in order to be accurate and reliable across platform. > > 4- > The most important problem with the test is that the test is not truly > creating the appropriate, suitable conditions where 'column-fill: > balance' would create a rendering different from 'column-fill: auto'. If > the test was good, then the test, in my opinion, should fail and fail in > a predictable manner if an UA does not support 'column-fill: balance' > and only support 'column-fill: auto'. Here, the test passes if > 'column-fill: balance' is removed! > > If all the inline content is expected to fill each line and to fill each > column box, then there is no predictable difference to be expected when > setting 'column-fill: balance' or when setting 'column-fill: auto'. > > So, ideally, a test testing 'column-fill: balance' versus 'column-fill: > auto' needs to create not too much inline content so that it can and > will only fill half of all column boxes and then verify that all inline > content fills all column boxes but only half of them (column-fill: > balance) as compared to filling half of all column boxes (column-fill: > auto). > > Imagine something like: > > div > { > column-count: 2; > column-gap: 1em; > height: 6em; > width: 21em; > } > > where digits represents single characters > > <div>1234567890 12 4567 90 123 56 890 123 56789 1 34567 90 12 4567 90 > 123 56 890 123 56</div> > > with 'column-fill: balance': > ------------ ------------ > |1234567890| |1 34567 90| > |12 4567 90| |12 4567 90| > |123 56 890| |123 56 890| > |123 56789 | |123 56 | > | | | | > | | | | > ------------ ------------ > > Same inline content with 'colum-fill: auto': > ------------ ------------ > |1234567890| |123 56 890| > |12 4567 90| |123 56 | > |123 56 890| | | > |123 56789 | | | > |1 34567 90| | | > |12 4567 90| | | > ------------ ------------ > > In conclusion, testing 'column-fill: balance' versus 'column-fill: auto' > requires less content to fill the column boxes, shorter words (versus) > large column boxes. That way, the test provides a leeway where both > column-fill values can "express" their characteristics. > > Gérard > -- > 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, 3 August 2013 05:03:40 UTC