Meeting minutes
Review agend and next meeting date
<Matt_King> View agenda at https://
<Matt_King> No meeting July 18.
<Matt_King> Next meeting July 25.
<Matt_King> rrsagent , make minutes
jugglinmike: Alex_Flenniken from Bocoup won't be available for review for the foreseeable future
Matt_King: We'll miss him!
Status of Site Updates
Matt_King: I don't think any of the things listed in the agenda are pressing updates
Matt_King: Does anyone object to us waiting for a larger collection before we push them out to the production site?
Jem: When will we consider publishing again?
Matt_King: We have a few patches which are pretty close. I think if we could land the "ListBox" fix, then we should definitely publish a new version
Matt_King: I don't mind waiting until August for that't know if that will necessarily be in July, but I don
Pull request reviews
Feed Example: Convert images to inline SVG
github: w3c/
Matt_King: James provided some good feedback to the author, but James is now on vacation. We would like someone to look at this, in particular, high-contrast mode
Jem: I reviewed it just now, but I didn't review for high-contrast mode
jongund: I can look at high-contrast mode
Matt_King: We'd ask Shirisha and Alyssa to review, previously
Jem: I've assigned you as a review, jongund
Listbox Examples: Update scrolling of listbox item with focus into view when page is magnified
github: w3c/
Matt_King: This is jongund's pull request. jugglinmike had provided performed a review and requested some changes. jongund have you had a chance to respond, yet?
Matt_King: We should include a test for scrolling this into view, and we should probably perform an audit for other patterns which use active descendant
jongund: I can try to write a test for this, but it will probably take me a few weeks. I'm going to be on vacation
Matt_King: Maybe we should raise a separate issue for developing regression tests for scrolling behavior
Matt_King: And then land this without regression tests
Jem: I agree
<Jem> ava info in the apg wiki: https://
Matt_King: Valerie Young might have some feedback for how to write those regression tests because she developed this testing framework
Matt_King: jongund, could you address jugglinmike's other feedback?
jongund: Sure
Matt_King: It'd be great if we could land this one in the next week or so. Then, I would love to push the changes out to production
Landmark Example Pages: remove background colours that cause color contrast failures
github: w3c/
Matt_King: This is a small fix. I think we just need a visual review
Jem: I can do that. I'll assign myself
Matt_King: Thanks!
Add AT Support tables for modal dialog, 1 slider, and navigation menu button
github: w3c/
Matt_King: I'm not going to merge this until I receive a "go ahead" from Vispero, and I'm meeting with them tomorrow
Jem: I approved it yesterday
Reposition AT support tables higher on the page
github: w3c/
Jem: I approved this
Matt_King: Then I'll merge it
Running tests on macOS
Matt_King: I don't know if anyone can run all the test successfully locally, and I don't know if we have a clear path forward for how it can be fixed
Matt_King: If that's correct, then I think we should probably change our readme (at least temporarily) to at least acknowledge this and not expect contributors to run the tests locally and just run them in CI
Matt_King: As discussed recently, this isn't the greatest solution for new contributors, but it's better than where we are now (i.e. misleading documentation)
howard-e: That all sounds right to me. I haven't had a chance to investigate further myself, but I know it's related to a regression to how Node.js version 17 defines "localhost"
howard-e: This leads me to think that some rewrite of the geckodriver file might be needed. But it's too early to say
Matt_King: The only reason we're able to run the tests in CI is because you pinned the CI system to Node.js 16, right?
howard-e: That's right. We can't do that for contributors unless we instruct them on how to use a tool like nvm
howard-e: This problem does not effect Windows; only some Unix environments
howard-e: We know that includes Ubuntu (because that's what the CI uses) and macOS (because that's what I use)
Matt_King: Well, I definitely have timeouts on my Windows machine
Matt_King: I think the logical next step for this issue is to change the readme. Right now, the readme just encourages people to run tests locally. I think we can include a link to the relevant Node.js issue
Matt_King: This was quite a diversion/rabbit-hole for Paul, and I don't want any other new contributors to "fall into" it
New contributing page
github: w3c/
Matt_King: I have this patch in draft. The agenda has a link to the preview
Matt_King: I'm looking for someone else to take a look and giving feedback on whether it includes everything we want
jongund: I can take a look
Matt_King: This was in response to an issue that jongund raised
Matt_King: I copied content from the home page) and cleaned it up a little bit (making it more appropriate for this kind of page
Matt_King: Ideally, it would be the kind of page that points readers to other pages that are helpful when they're getting started
Jem: I was thinking about adding content about "why people should contribute" because right now, it's mostly about the mechanical aspects of contribution
Matt_King: Well, the top of the page has a very brief statement along those lines
Matt_King: If we wanted more and if we wanted to stay true to this style of document, I think that opening statement could include a link--something like "learn more about why contributing to the APG is important..."
Matt_King: So there's a draft pull request; feel free to add commits. I'm happy to co-author content
Is treegrid Example code able to support nested selects
github: w3c/
Matt_King: Stephanie_Leary was trying to use the treeview pattern to make a treeview that included a select in a cell
Matt_King: I replied that adding a select, that could interfere with the navigation in the treeview
Matt_King: I looked around in the code, and I tried to understand the impact of the change on the code. I believe that it will only have negative consequences--not positive
Matt_King: I wrote something up to that effect in my reply, but I would like it if someone could actually test it
Matt_King: I'm pretty sure that the TreeGrid code is very purpose-built for the kind of tree grid we are exemplifying in APG. I don't think it's quite robust enough to support this kind of extension
jongund: I can take a look
Matt_King: You've already signed up for a bunch of things in this meeting; that's very generous!
New Issue Action Planning
Listbox pattern: aria-multiselectable conflicts with the aria spec
github: w3c/
Bryan_Garaventa: I responded to this one
Matt_King: Maybe we actually need to adjust the ARIA spec
Matt_King: Your comment here is spot-on, Bryan_Garaventa. It should work, just in terms of the way it maps
Matt_King: The way the ARIA spec language is written for ARIA multiselectable, it doesn't account for the possibility of using ARIA checked to represent the "selected" state. The spec language is inconsistent with itself: there's language about using "checked" in multiselects
<Jem> +1
Matt_King: So in summary, this is not an APG issue. It's actually an ARIA issue
Jem: I'll add the label "not for practices"
Matt_King: and I'll add a comment to this effect