User tests: Successful: Unsuccessful:
Pull Request resolves #48151.
This PR was drafted with AI assistance (Claude). I reviewed, tested (see below), and am accountable for the change per the Generative AI policy.
layouts/joomla/form/field/subform/repeatable-table.php renders each column header's help text as a role="tooltip" div, revealed visually by :focus/:hover on a preceding icon span (see _form.scss in both the atum and cassiopeia templates: :focus + [role="tooltip"], :hover + [role="tooltip"]). That icon span was both aria-hidden="true" and tabindex="0" — a keyboard user can tab onto it, but a screen reader announces nothing there, and the tooltip text was never programmatically associated with anything (axe-core: aria-hidden-focus, WCAG 4.1.2, serious).
Rather than simply dropping tabindex (which would also silently remove the CSS :focus reveal for sighted keyboard users, since the tooltip is only shown for the preceding sibling of the icon), this follows the pattern already used for this exact icon+tooltip shape elsewhere in core — JGrid::state() (libraries/src/HTML/Helpers/JGrid.php): keep the icon focusable, drop aria-hidden, and let aria-labelledby give the icon its accessible name from the tooltip text. This keeps the existing keyboard-visible-tooltip behavior intact while also making the description announced to screen readers.
Both branches of the file are affected (groupByFieldset and the default per-field branch) and both are fixed identically.
description attribute to a subfield used inside a subform field with layout="joomla.form.field.subform.repeatable-table" (no core form ships one by default; e.g. temporarily add description="test" to the name field inside plugins/editors/tinymce/forms/setoptions.xml's external_plugins subform).Extensions → Plugins → Editor - TinyMCE, External Plugin URLs field) and inspect the column header's icon span (.icon-info-circle).aria-hidden="true" tabindex="0", no aria-labelledby. After: aria-hidden is gone, tabindex="0" remains, aria-labelledby points at the role="tooltip" div's id.:focus rule unaffected), and a screen reader now announces the tooltip text as the icon's accessible name.The icon span is aria-hidden="true" and tabindex="0" at the same time: a keyboard user tabs onto it, but assistive technology announces nothing, and the tooltip text is never programmatically associated with anything. axe-core flags aria-hidden-focus (serious, WCAG 4.1.2).
The icon keeps its tabindex="0" (so the existing CSS :focus reveal still works for sighted keyboard users), aria-hidden is removed, and aria-labelledby gives the icon an accessible name from the tooltip text, so it is announced by assistive technology.
Please select:
Documentation link for guide.joomla.org:
No documentation changes for guide.joomla.org needed
Pull Request link for manual.joomla.org:
No documentation changes for manual.joomla.org needed
🤖 Generated with Claude Code
| Status | New | ⇒ | Pending |
| Category | ⇒ | Layout |
| Title |
|
||||||
I have tested this item 🔴 unsuccessfully on 03b7df9
While I was able to match the BEFORE and AFTER conditions of your PR fix in terms of the Accessibility side of things (gives the icon an accessible name from the tooltip text, so it is announced by assistive technology), unless I misunderstood something, a PR isn't supposed to change the visual look of things, right?
I have tested this item 🔴 unsuccessfully on 03b7df9
While I was able to match the BEFORE and AFTER conditions of your PR fix in terms of the Accessibility side of things (gives the icon an accessible name from the tooltip text, so it is announced by assistive technology), unless I misunderstood something, a PR isn't supposed to change the visual look of things, right?
I have tested this item 🔴 unsuccessfully on 03b7df9
While I was able to match the BEFORE and AFTER conditions of your PR fix in terms of the Accessibility side of things (gives the icon an accessible name from the tooltip text, so it is announced by assistive technology), unless I misunderstood something, a PR isn't supposed to change the visual look of things, right?
Whether, I edit the plugins/editors/tinymce/forms/setoptions.xml file or NOT
There seems to be a missing thin border on the left, and above the Plugin Name and Path cells...
I have tested this item 🔴 unsuccessfully on 03b7df9
While I was able to match the BEFORE and AFTER conditions of your PR fix in terms of the Accessibility side of things (gives the icon an accessible name from the tooltip text, so it is announced by assistive technology), unless I misunderstood something, a PR isn't supposed to change the visual look of things, right?
There seems to be a missing thin border on the left, and above the Plugin Name and Path cells...
and this is whether I edit the plugins/editors/tinymce/forms/setoptions.xml file or NOT, I checked both conditions.
I have tested this item 🔴 unsuccessfully on 03b7df9
While I was able to match the BEFORE and AFTER conditions of your PR fix in terms of the Accessibility side of things (gives the icon an accessible name from the tooltip text, so it is announced by assistive technology), unless I misunderstood something, a PR isn't supposed to change the visual look of things, right?
BEFORE your PR fix after having edited the plugins/editors/tinymce/forms/setoptions.xml file as per your testing sintructions:
AFTER your PR fix applied after having edited the plugins/editors/tinymce/forms/setoptions.xml file as per your testing sintructions:

There seems to be a missing thin border on the left, and above the Plugin Name and Path cells...
when I came across this difference, I also checked with removing the plugins/editors/tinymce/forms/setoptions.xml file changes and the same thing happened to the thin border...
I have tested this item ✅ successfully on 03b7df9
I have tested this item ✅ successfully on 03b7df9
| Labels |
Added:
a11y
bug
PR-6.2-dev
|
||
As the branch is out-of-date with the base branch I had to apply the PR manually - maybe that is why @exlemor saw the border issue. There is nothing in the code of this PR that would touch the border
I've triggered a branch update (and restored the previous human test result). New packages should be available in a few minutes.
As the branch is out-of-date with the base branch I had to apply the PR manually - maybe that is why @exlemor saw the border issue. There is nothing in the code of this PR that would touch the border
I've triggered a branch update (and restored the previous human test results). New packages should be available in a few minutes.
I have tested this item ✅ successfully on 067eadd
I was able to retest with the retriggered branch package and I now had a successful test. Thank you @brianteeman and @richard67!
(Thank you @bcordis for the fix).
I have tested this item ✅ successfully on 067eadd
I was able to retest with the retriggered branch package and I now had a successful test. Thank you @brianteeman and @richard67!
(Thank you @bcordis for the fix).
| Status | Pending | ⇒ | Ready to Commit |
RTC
RTC
| Labels |
Added:
RTC
|
||
| Status | Ready to Commit | ⇒ | Fixed in Code Base |
| Closed_Date | 0000-00-00 00:00:00 | ⇒ | 2026-11-09 15:35:34 |
| Closed_By | ⇒ | drmenzelit |
Thank you
This pull request has been automatically rebased to 6.2-dev. Joomla! 5.4 is now in security only mode.