- From: Gérard Talbot <css21testsuite@gtalbot.org>
- Date: Sun, 31 May 2015 18:18:24 -0400
- To: Public CSS Test suite mailing list <public-css-testsuite@w3.org>
Hello, [src] http://test.csswg.org/source/css-writing-modes-3/block-flow-direction-009.xht and its reference file: http://test.csswg.org/source/css-writing-modes-3/block-flow-direction-001-ref.xht [nightly-unstable] http://test.csswg.org/suites/css-writing-modes-3_dev/nightly-unstable/html/block-flow-direction-009.htm and its reference file: http://test.csswg.org/suites/css-writing-modes-3_dev/nightly-unstable/html/reference/block-flow-direction-001-ref.htm I have been receiving 2 approximatively identical emails about block-flow-direction-009 test. Basically saying: [ I see, read a big yellow "PASS" word but, in Chrome 40+, it is located on the righthand side of the page when the reference file indicates that the big yellow "PASS" word should be located on the lefthand side of the page. So, is Chrome passing that block-flow-direction-009 test? ] The short answer is ------------------- Chrome passes that block-flow-direction-009 test and Chrome will fail another specific test elsewhere in the Writing Modes test suite. My long answer is ----------------- block-flow-direction-009 test is trying to test only and exclusively if the block flow direction uses a right-to-left direction. block-flow-direction-009 test is *not* trying to test if block-start of writing-mode of containing block is correctly implemented for an abs. pos. element in orthogonal flow. block-flow-direction-009 test's building logic is that the tester (person taking that block-flow-direction-009 test) will be able to read "PASS" if a block flow direction uses a right-to-left direction. Now, *where* that vertical-rl block should be displayed within its containing block is another issue and must be tested in another test; the block-flow-direction-009 test is trying to only checks *how* the vertical-rl block is "built" and rendered. So, both test results (pass or fail) could be considered as valid. Because it is very difficult to isolate the test purpose from other layout issues involved. When a test has 'writing-mode: vertical-rl', then *many* layout css features (line box flow, block flow direction, block start of BFC, precedence algorithms, etc) change altogether. It is impossible to write a simple test (with reduced amount of code) that will isolate 1 and only 1 aspect of the many css features that change. With more constraints and more restraints (which involve more coding, rules, declarations), it could be possible (but longer) to try to isolate 1 aspect at the time. As the author of block-flow-direction-009 test, what was important to do next was to make sure that there is (or will be) another test somewhere, elsewhere testing solely where the vertical-rl block in orthogonal flow should be displayed within its horizontal-tb containing block. And I create 32 tests dedicated at checking that: abs-pos-non-replaced-icb-v* The block-flow-direction-009 test is not ideal because it does not use a pass-fail-conditions sentence to describe expected result. But, as far as tester's ease of checking a test, block-flow-direction-009 test is excellent: if (s)he can read "PASS", then it's a passed test. Easy and fast to do. Again, we have other (32!) tests on absolute positioning where Chrome will be detected as failing. " Adding 32 tests on absolute positioning of child in orthogonal flow having to be positioned within the Initial Containing Block which is in horizontal writing-mode. (...) Chrome 44.0.2391.0 fails 16 tests. " [css-writing-modes-3] 32 abs. pos. in Initial Containing Block tests submitted ( abs-pos-non-replaced-icb-vlr and abs-pos-non-replaced-icb-vrl ) http://lists.w3.org/Archives/Public/public-css-testsuite/2015May/0006.html In conclusion ------------- If you thought that Chrome 40-45 passed block-flow-direction-009, then you are correct. If you thought that, after comparing with reference file, Chrome 40-45 failed block-flow-direction-009, then you are also correct. When Google Chrome developers fix Chrome with regards to block-start of writing-mode of containing block implementation for an abs. pos. element in orthogonal flow, then there will be no hesitation or doubt with regards to Chrome and block-flow-direction-009 test. Gérard -- Test Format Guidelines http://testthewebforward.org/docs/test-format-guidelines.html Test Style Guidelines http://testthewebforward.org/docs/test-style-guidelines.html Test Templates http://testthewebforward.org/docs/test-templates.html CSS Naming Guidelines http://testthewebforward.org/docs/css-naming.html Test Review Checklist http://testthewebforward.org/docs/review-checklist.html CSS Metadata http://testthewebforward.org/docs/css-metadata.html
Received on Sunday, 31 May 2015 22:18:55 UTC