fix: role for aside element when nested within sectioning content

[jlp-craigmorten]

Changes

Relates to #25

Updates the role returned for a aside element to be either:

1. complementary when the element is named;
2. complementary when the element is not nested within sectioning content;
3. generic otherwise.

This detail is currently missing from the HTML-ARIA spec, but is describe in the sibling HTML-AAM spec:

• https://www.w3.org/TR/html-aam-1.0/#el-aside-ancestorbodymain - when scoped to body or main (i.e. not sectioning content)
• https://www.w3.org/TR/html-aam-1.0/#el-aside - when scoped to sectioning content

How to Review

Mirrors the existing functionality and tests for combination of the img element which has similar named behaviours, and the footer which has similar scoping/nesting behaviours.

Undecided whether would be cleaner to scope the calculateAccessibleName() check within the new getAsideRole() - perhaps feels a bit awkward to have a getAsideRole() that holds partial logic for getting the role, the rest living elsewhere 🤔 wdyt?

Starting to see patterns arising, e.g. attributes['aria-label'] || attributes['aria-labelledby'] repeated, but perhaps still borderline rule or 3 to warrant refactoring at this stage?

Checklist

• Unit tests updated

[changeset-bot]

🦋 Changeset detected

Latest commit: 9d45089

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
html-aria	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[drwpow]

Undecided whether would be cleaner to scope the calculateAccessibleName() check within the new getAsideRole() - perhaps feels a bit awkward to have a getAsideRole() that holds partial logic for getting the role, the rest living elsewhere 🤔 wdyt?

I don’t mind this. I see this as more of a “private” API that can remain undocumented. I think just to be safe I’d err on the side of not exporting it from the entry so we could always refactor later if needed.

Starting to see patterns arising, e.g. attributes['aria-label'] || attributes['aria-labelledby'] repeated, but perhaps still borderline rule or 3 to warrant refactoring at this stage?

Ha I usually wait till way past 3 to refactor, until it’s fairly obvious.