changelog: Content Change, Update screen reader message to identify a…#6044
changelog: Content Change, Update screen reader message to identify a…#6044SammySteiner merged 5 commits intomainfrom
Conversation
b073f42 to
eaa38a5
Compare
eaa38a5 to
14f376d
Compare
| class: 'usa-button usa-button--big usa-button--full-width margin-bottom-2' %> | ||
| <%= link_to modal.sign_out, | ||
| destroy_user_session_path, | ||
| role: 'link', |
There was a problem hiding this comment.
Was there any change in functionality after adding this role? link_to renders an anchor tag and would have the role explicitly stated on it
There was a problem hiding this comment.
Good catch, it got left in from when I was experimenting with turning it into a button tag.
| <div class="usa-modal-overlay"> | ||
| <%= tag.div( | ||
| role: 'dialog', | ||
| role: 'alertdialog', |
There was a problem hiding this comment.
I am a little concerned about this because this layout is used on all the modals in the app. Is there a way we can use logic to determine if the role should be alertdialog vs dialog?
There was a problem hiding this comment.
Hmm, ok, I added locals parameter to the render and updated the three places where it is used.
| @@ -0,0 +1,248 @@ | |||
| <!DOCTYPE html> | |||
There was a problem hiding this comment.
Hi, apologies I didn't catch this, but this PR ended up checking in some generated HTML into source code, it's noisy, derived data and doesn't belong here.
I've got a commit to clean this up and prevent it in the future: 9603127
|
These changes have the impact of causing the time to be announced every second. Is that expected? Might be caused by the change from Screen.Recording.2022-03-15.at.11.47.46.AM.mov |
So weird, I took a bunch of recordings of the behavior for the acceptance thread with my local deployment and it didn't behave this way. |
…ll options, LG-5857