Update: fix inconsistency around aria-setsize=-1

[scottaohara]

closes #1759
closes #2346
also closes w3c/core-aam#149

This PR adjusts the requirements for aria-setsize to fix the inconsistency between the ARIA and CORE AAM specs regarding the use of -1 as a value.

Test, Documentation and Implementation tracking

Once this PR has been reviewed and has consensus from the working group, tests should be written and issues should be opened on browsers. Add N/A and check when not applicable.

• "author MUST" tests:
• "user agent MUST" tests:
• Browser implementations (link to issue or commit):
  • WebKit: https://bugs.webkit.org/show_bug.cgi?id=292844
  • Gecko: https://bugzilla.mozilla.org/show_bug.cgi?id=1358473
  • Blink: https://issues.chromium.org/issues/420965168
• Does this need AT implementations?
• Related APG Issue/PR:
• MDN Issue/PR:

---

Preview | Diff

[netlify]

✅ Deploy Preview for wai-aria ready!

Name	Link
🔨 Latest commit	32e0608
🔍 Latest deploy log	https://app.netlify.com/sites/wai-aria/deploys/67069cb8cf16df0009437f36
😎 Deploy Preview	https://deploy-preview-2341--wai-aria.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[MelSumner]

This seems to resolve the confusion, thank you!

[scottaohara]

talking with @aleventhal about this some more, we should also add in what to do about if someone uses the empty string, or puts in literal "garbage" text strings.

So this can probably be
positive full numbers works as presently spec'd
0 and any negative number == indeterminate
anything else == undefined and thus ignore.

[aleventhal]

So this can probably be positive full numbers works as presently spec'd 0 and any negative number == indeterminate anything else == undefined and thus ignore.

I know this is going to happen in a follow-up, but it's probably good if it's not too strict, e.g. spaces before and after the number are ok. I was going to do the following:

1. Empty string or not present = undefined and thus ignore
2. Pass it to a non-strict int conversion function, and comes back < 1, treat as follows:
   • aria-posinset or aria-level -> 1
   • aria-setsize -> -1 (indeterminate)

[rahimabdi]

@aleventhal @scottaohara Would this be important enough to flag a console message?

[aleventhal]

setsize=-1 is really rare in my experience. I've never actually seen an example with it.
I'm not sure what/when the console message would be, but my quick answer based on prevalence is that it's not worth any code in the browser engine.

[spectranaut]

@adampage to investigate the status of browsers w.r.t. this change and file issues if necessary

[spectranaut]

Discussed in today's meeting: https://www.w3.org/2025/05/15-aria-minutes.html#ffff

[adampage]

Hey all, I’ve finished investigating this and have published some results.

Some fun trivia:

• Webkit reported a set size of 184346744073798551615 for some aria-setsize="-1 tests specimens.

• article was the only role that exhibited no issues, seemingly on account of ignoring aria-setsize or aria-posinset altogether. Windows + Firefox + Narrator was the one exception.

• The main browser repair techniques for calculating set size seem to be 1) reporting the actual number of DOM elements within the group, or 2) accepting the highest value reported by the author via aria-posinset.

• Aside from repairing or misreporting the set size, many browser + AT combos also repaired or misreported each item’s stated position within the set.

The tl;dr is that every browser exhibits issues with both aria-setsize="0" and aria-setsize="-1" in relation to this PR. menuitem, option, and radio had the most.

cc: @scottaohara @spectranaut

[adampage]

The PR checklist now has issue links for Webkit, Gecko, and Blink cc: @rahimabdi, @aleventhal, @jcsteh

[rahimabdi]

An incredibly thorough investigation, thank you @adampage!

[spectranaut]

Discussed briefly in today's meeting: https://www.w3.org/2025/05/29-aria-minutes.html#26c1