- From: Joanmarie Diggs <jdiggs@igalia.com>
- Date: Mon, 07 Mar 2016 14:04:45 -0500
- To: Joseph Scheuhammer <clown@alum.mit.edu>, Fred Esch <fesch@us.ibm.com>
- Cc: SVG-A11y TF <public-svg-a11y@w3.org>
Hey all. On 03/07/2016 01:17 PM, Joseph Scheuhammer wrote: > No, as far as I can tell. Using Safari 9.0.3, and the Mac > Accessibility > Inspector, for your description example, AXHelp is "the end". On a related note, what you have in the current/shipping Safari is no longer current as of Saturday. The work I've been doing to implement the missing/different-from-core SVG accessibility mappings has landed. [0] And there was much rejoicing. For the specific cases under discussion in this thread, there happens to be no change between shipping and master. For many cases, however, there will be. As a result, verifying the mappings should be done with a nightly build. On a related note, the bulk (~200) of Fred's test cases were included as automated Layout Tests. Actual results of the mappings can be obtained in mere seconds. I still need to re-review all the assertions, but as far as I am aware, there is only one failure I need to fix (test 27 of the roles tests). Of course, as consensus is reached on some of the issues being discussed, we can update WebKit accordingly. (BTW, the "golden hammer" has already been included, with an easy-to-find comment should the group change their mind.) Second-to-lastly, the official results for the two test cases in question: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ============= Name test [1] ============= <svg> <circle id="test22" aria-labelledby="22a 22b" cx="50" cy="100" r="15" data-expected="name: big end"/> <text id="22a" x="40" y="90" aria-label="big">the</text> <text id="22b" x="140" y="90" aria-labelledby="3c">end</text> <text id="22c" x="180" y="90">finish</text> </svg> Mac Results [2]: Expected name: big end AXTitle: AXDescription: big end AXHelp: Linux results [3]: Expected name: big end AXTitle: big end AXDescription: ==================== Description test [4] ==================== <svg> <circle id="test21" aria-describedby="t21a t21b" cx="50" cy="100" r="15" data-expected="description: the end"/> <text id="t21a" x="40" y="90">the</text> <text id="t21b" x="140" y="90" aria-describedby="t2c">end</text> <text id="t21c" x="180" y="90">finish</text> </svg> Mac results [5]: Expected description: the end AXTitle: AXDescription: AXHelp: the end Linux results [6]: Expected description: the end AXTitle: AXDescription: the end ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Lastly, any objections if I make a new page on the SVG A11y wiki containing all the assertions in the format above (i.e. markup, with tests being numbered, and including the actual results for the two platforms)? I find it difficult to visually process the assertions as currently written, let alone keep my/webkit's tests in alignment with Fred's. Thanks! --joanie [0] https://trac.webkit.org/browser/trunk/LayoutTests/accessibility/w3c-svg-roles.html?rev=197616 [1] https://trac.webkit.org/browser/trunk/LayoutTests/accessibility/w3c-svg-name-calculation.html#L133 [2] https://trac.webkit.org/browser/trunk/LayoutTests/platform/mac/accessibility/w3c-svg-name-calculation-expected.txt#L111 [3] https://trac.webkit.org/browser/trunk/LayoutTests/platform/gtk/accessibility/w3c-svg-name-calculation-expected.txt#L90 [4] https://trac.webkit.org/browser/trunk/LayoutTests/accessibility/w3c-svg-description-calculation.html#L144 [5] https://trac.webkit.org/browser/trunk/LayoutTests/platform/mac/accessibility/w3c-svg-description-calculation-expected.txt#L106 [6] https://trac.webkit.org/browser/trunk/LayoutTests/platform/gtk/accessibility/w3c-svg-description-calculation-expected.txt#L86
Received on Monday, 7 March 2016 19:05:18 UTC