Update accname/name/comp_labelledby.html with aria-labeledby [sic] tests #43698
Conversation
There was a problem hiding this comment.
This is a great investigation, Rahim. Thanks...
Hopefully stating the obvious that the non-standard spelling tests with different results should not land until the WG has had a chance to compare the results and come up with a general rule.
Some options:
- You could wait on the WG discussion in this PR (that might take a while), so the PR could get downstream conflicts (probably okay).
- You could commit the ones where the implementations agree, and file a second PR with the others.
- You could break the non-standard spelling tests into a standalone file: less likely to conflict, but more likely to get out of sync with the tests for the primary spelling.
If the WG decides this should be an allowed alternate spelling and otherwise behave the same as aria-labelledby, I think this PR is okay to land. However, it may be that the ARIA WG would not want it to conflict with some tests, such as the host language labeling mechanisms like label/for and alt. I don't know what the outcome will be.
|
@cookiecrook Got it, I'll hold onto this for the time being pending WG discussion and consensus. |
|
I wonder, in what ways does it matter if some browsers alias the "incorrect" spelling? It's probably okay that there's a test that points this specific use case out, mostly because it's a quirk (comparatively). JMO, though. I don't think I want spec to openly support both, but rather just know what a browser will do in either case, for this specific use case. |
|
@MelSumner FWIW, I agree. Fortunately, all browsers appear to prefer |
There was a problem hiding this comment.
The last commit that added the ordered attribute changes is a good one long term, but will cause new failures in the Interop 2024 Focus Area. Those 3 new failing tests can either go in a tentative file, or in a later PR after Interop 2024 ends.
[Update: Actually I misread... Those are bad test expectations b/c it's accurate to use the correct spelling regardless of the order in which they appear.]
…nvalid misspelling.
|
I believe this PR should land given the changes were approved by the ARIA WG: w3c/aria#2371 |
|
@spectranaut wrote:
Agreed. My objection above was to a test expectation that has since been removed from the PR. |
|
This appears to have changed the Interop 2024 results. I am not seeing a corresponding test change proposal in https://github.com/web-platform-tests/interop/issues?q=is%3Aissue%20label%3Atest-change-proposal%20 so I think the change should either be reverted, or the new tests added to another file, until Interop 2024 is complete. |
|
Apologies. I thought Interop 2024 was done. For those of us more on the outside of the Interop admin process, when does it end? |
|
No worries :) It's a bit confusing and we should invest more in automation to help people avoid making unintentional changes. The current plan is to freeze the dashboard on or before next Thursday 6th. |
|
PR is up at @jgraham Once the passing CI results are in, please r+ and merge at will. |
|
I've merged in the partial revert, and plan to re-revert that sometime next month. |
|
Thanks, it's much appreciated :) |
WPT investigation for this issue: