Archive index

A11y Slackers Gitter Channel Archive 11th of August 2016

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    @zakim-robot
    Aug 11 02:31
    [michiel] Ardbeg is pretty darn good whisky
  2. zakim-robot
    @zakim-robot
    Aug 11 06:52
    [caesar] Hi guys... could somebody help me identify what's wrong with this scenario? Trying to use a list to create a menu using aria roles: https://codepen.io/anon/pen/bZOYRJ
  3. [caesar] JAWS + IE11 is reading out the link target in addition to the link text; NVDA + Firefox is reading out "menu item" twice for each submenu item; Voiceover is not accessing the submenus at all if I VO+right arrow!
  4. [caesar] What am I missing? Surely something so straightforward shouldn't be so diabolical?
  5. James Nurthen
    @jnurthen
    Aug 11 09:09
    @caesar put role=presentation on the li elements and role=menuitem on the a elements.
  6. Mallory
    @StommePoes
    Aug 11 14:27
    @karlgroves: "Mallory: I’m still confused. Can you go into devtools, expand the necessary node(s) and then trigger the tooltip?" No, because the div that gets added to the DOM is collapsed in the devtools by default. Wouldn't be an issue if I could figure out how to set never-fold-code in that.
  7. All you'd see of the tooltip is <div class="whatever"></div>. When you want to click to open it, the focus is now no longer on the tooltip trigger.
  8. zakim-robot
    @zakim-robot
    Aug 11 14:28
    [karlgroves] Gotcha. How about the idea from Joe Watkins?
  9. Mallory
    @StommePoes
    Aug 11 14:28
    However, I can't find the particular page anymore. I checked Bootstrap, there losing focus doens't close the tip.
  10. If I find that page again (people are always sending it to me so I'll surely see it again), I will try that.
  11. Honestly what I really want in devtools is some ability to do what I need the mouse to do: right-click and Inspect on a specific part of the page opens all that folded HTML to at, or near, where the element is in the DOM. Again if code wasn't folded it would be less of an issue but on dense pages it can still be hard to find something.
  12. er, ability to do that with keyboard. I probably can with FF with caret nav on, but I don't think anyone else has caret nav, plus caret nav's only for text.
  13. @jalbertbowden I personally don't like mailto's because of how often I accidentally click on the link when all I really wanted to do was copy it so I could then go to my mail program and paste it.
  14. But, I may be a clumsy minority.
  15. Mallory
    @StommePoes
    Aug 11 14:34
    @michiel try the supernova. But wait til after you've been paid. It cost around 99 euros.
  16. zakim-robot
    @zakim-robot
    Aug 11 14:41
    [bkardell] anyone in here find this interesting, want to help, or tell me why it's dumb? https://github.com/bkardell/selenium-ide-axe/
  17. [karlgroves] /giphy imma let you finish
    imma let you finish
  18. zakim-robot
    @zakim-robot
    Aug 11 14:47
    [bkardell] @karlgroves: does that mean "yes, I like that as an idea" or "you should use this instead imo"? :)
  19. [karlgroves] Both. ;)
  20. [bkardell] :)
  21. [karlgroves] Yeah. I’m gonna write up a full tutorial on adding the command and using it in an assertion.
  22. [bkardell] nice... I'd be super happy to talk more about how we make those work in IDE because I'd never even thought about how to write a plugin until last night at about 5pm EST.. it works, but prob consulting with someone with deeper knowledge of it might be helpful
  23. zakim-robot
    @zakim-robot
    Aug 11 14:52
    [karlgroves] Its surprisingly easy. Hunting down useful documentation is surprisingly painful.
  24. [bkardell] yeah, i found the same -- one challenge I had was how to deal with async things... I don't know if my solution is ideal... I'd love some feedback from both people familiar with ide extension APIs and from actual potential users
  25. zakim-robot
    @zakim-robot
    Aug 11 19:39
    [alexlande] are there AT impacts of using particular HTML5 input types? tel vs text, for example. Do screen readers treat them different in some way (or refer to them by their type)?
  26. zakim-robot
    @zakim-robot
    Aug 11 20:29
    [marcysutton] It depends on user-agent support and AT support I would think
  27. [marcysutton] (if it wasn't supported it would just be a text field)
  28. [karlgroves] @alexlande: The input types are announced by AT (if supported, as Marcy said) so there’s actually a big benefit to users in terms of predictability. If they know what kind of input it is they will know what is expected WRT interacting with it.
    http://output.jsbin.com/viyumu/1
  29. [karlgroves] If you have a screen reader on your machine, go to that URL and listen.
  30. zakim-robot
    @zakim-robot
    Aug 11 20:45

    [alexlande] That’s perfect, thank you! I tested tel specifically in voice over, but it didn’t get announced as such so I assumed other inputs didn’t have their specific types announced either, but in that page a number of them did.

    Password is announced as “secure edit text,” number as “incrementable edit text,” and so on.

  31. Mallory
    @StommePoes
    Aug 11 21:31
    well one thing about tel
  32. I used to use it on sites
  33. then stopped
  34. on iPads and other appley things I did not own, these were practically turned into links
  35. which was bad because links were given a certain colour with CSS
  36. and backgrounds another
  37. so white text on a dark orange background was fine, until the telephone number in the address you expected to be white because you styled it that way, on iThings turned into a "link" and became orange. Great thing for a customer to find.
  38. And I dunno how many iPad users make phone calls with those things... makes more sense for phones though.
  39. @bkardell "one challenge I had was how to deal with async things..." the little I've done, with one of the expect.js's (what a silly thing, two different projects calling themselves the same name), I do have a done() func on hand to say "this branch of asyncy stuff, stop the test" but I dunno how complex you're going
  40. zakim-robot
    @zakim-robot
    Aug 11 21:36
    [bkardell] looks up expects.js
  41. [bkardell] @StommePoes is this with selenium IDE? If so, share?
  42. Mallory
    @StommePoes
    Aug 11 22:11
    no no, but it's the first unit testing where I saw async stuff in testing.
  43. However, I should ask my QA, he's a Java/Selenium guy.
  44. zakim-robot
    @zakim-robot
    Aug 11 22:13
    [marcysutton] Here's an async example with Selenium Webdriver, but not the IDE https://github.com/marcysutton/axe-webdriverjs-demo/blob/master/spec/test.js#L46 ...you could probably do something similar with browser native promises
  45. Mallory
    @StommePoes
    Aug 11 22:22
    looks pretty promissy, I like.
  46. zakim-robot
    @zakim-robot
    Aug 11 22:45
    [bkardell] @StommePoes please do - I'm sure there is a promise nature in the guts of how IDE allows you to plug in, it's just not documented or even hinted at in any docs or examples I can find
  47. zakim-robot
    @zakim-robot
    Aug 11 23:33
    [caesar] @jnurthen I tried your recommendation but JAWS is still reading out the items twice. Also, my original example was coded per the example given in the W3C ARIA page: https://www.w3.org/TR/wai-aria/roles#tac_example1_header