- From: Koji Ishii <kojiishi@gmail.com>
- Date: Sun, 17 May 2015 22:15:58 +0900
- To: Gérard Talbot <css21testsuite@gtalbot.org>
- Cc: Koji Ishii <kojiishi@gluesoft.co.jp>, Public CSS Test suite mailing list <public-css-testsuite@w3.org>
Thank you for reviewing the tests. On Thu, May 14, 2015 at 1:38 PM, Gérard Talbot <css21testsuite@gtalbot.org> wrote: > Koji, > > http://test.csswg.org/source/css-writing-modes-3/orthogonal-parent-shrink-to-fit-001a.html > > line 11: font:25px Ahem; > > Whenever you set a font-size for the Ahem font, you need to set the > line-height as well. Otherwise, line-height is, by default, 'normal' and > 'normal' for IE11 and Webkit-based browsers for the Ahem font is 1.2 while > it is 1.0 for Firefox. > > I suggest > > line 11: font:25px/1 Ahem; > > Ahem Usage > http://testthewebforward.org/docs/test-style-guidelines.html#special-fonts Done, thank you. > line 12: clear:both; > > 'clear' can only apply and can only have a rendering effect in presence of > floats and if there is a float earlier in the source document. So, here, at > line 12, 'clear: both' has no impact, no influence, no relevance. You can > safely remove 'clear: both'. > > 'clear: both' makes sense though in orthogonal-parent-shrink-to-fit-001.html > but not in orthogonal-parent-shrink-to-fit-001?.html > > ---------- > > line 13: margin-top:.2em; > > You can safely remove this declaration since the preceding <p>'s > margin-bottom is 1em and so there will be a margin-collapse of both blocks' > margin. max(1em, 0) or max(1em, 0.2em) is going to be the same anyway. I just keep all variants in sync, this helps me when there were fixes. I changed to use <h3> to make test descriptions more visible, hope this is ok for you. > http://test.csswg.org/source/css-writing-modes-3/orthogonal-parent-shrink-to-fit-001.html > > I think you should not create a "combo" test, a meta-test for all these > orthogonal scenarios and situations. Each of these sub-tests are somewhat > complex (definitely not basic tests) and not easily obvious ... at least to > me... > > I would filename-rename those 1a-1g tests into separate, distinct tests and > not make them part of a "combo" test. Re-read the file name guideline: http://testthewebforward.org/docs/css-naming.html saying: | There may also be a letter affixed after the number, | which can be used to indicate variants of a test. | | For example, float-wrap-001l.xht and float-wrap-001r.xht | might be left and right variants of a float test. This is a test for shrink-to-fit, with different types of children as variants. Isn't this a good fit to use the suffix pattern? /koji
Received on Sunday, 17 May 2015 13:16:25 UTC