I have tested this item 🔴 unsuccessfully on 3fbc5c3

The label is now doubled
The accessibility is now brokenn

In which form is that?

every form

I can't reproduce your results.

I tested it twice

hmm looks good to me.

without the patch

with the patch

How can we reproduce the issue you mention? Or I'm missing something here?

all i did was apply the PR with patchtester

after

before

hmm i have now looked into the screenshots more deeply. I have marked the differences, i found on mine. Does the additional for="jform_link_titles" break accessibility? Maybe I just not getting what the issue is as it seams to me that the label is there before and after two times, on both my and your screenshots.

after

before

I'll revert the related change.

We can't even use hiddenLabel attribute to actually hide labels anymore 🤷‍♂ .

Test again please.

Tried again with a pure git checkout instead of patchtester - dont get the double anymore but there is an empty div now

after

before

Otherwise its ok

Out of scope of this PR. Can happen with other fields because main layout doesn't check for empty input/label.

As its a change in behaviour created by this PR I dont see how its out of scope

This behavior is not specific to this field. Any field that returns an empty string in getLabel() or getInput() methods is going to cause empty divs to be rendered , e.g.

<field
	name="whatever"
	type="note"
/>

Outputs:

<div class="control-group">
	<div class="control-label"></div>
	<div class="controls">
			</div>
	</div>

If it was my call, I'd just drop all these workarounds and have #25085 reverted. It was done with good intentions but with wrong assumptions that the labels can only be hidden for cosmetic purposes.

Please fix conflicts.

I have tested this item ✅ successfully on bd7ff1b

RTC