fix(attributes): preserve non-standard boolean attribute values like hidden=until-found

[xxiaoxiong]

Problem

When using hidden="until-found" (an HTML Living Standard feature), Cheerio normalizes the attribute value back to just hidden:

const $ = cheerio.load('<div id="test" hidden="until-found"></div>');
$('#test').attr('hidden'); // Returns "hidden", but should return "until-found"

This affects any boolean HTML attribute that receives a non-standard value.

Root cause

In src/api/attributes.ts, the attribute getter checks if the attribute name matches a boolean attribute regex, and if so, always returns the attribute name instead of the actual value:

return !xmlMode && rboolean.test(name) ? name : elem.attribs[name];

Changes

Modified the getter to check the actual attribute value. If the value is empty or matches the attribute name, the boolean normalization is preserved. If the attribute has a non-standard value (e.g., "until-found" for hidden), the actual value is returned.

Closes

Closes #5073

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b482443ed8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Normalize boolean values case-insensitively

For parsed HTML such as <input checked="CHECKED">, boolean attributes whose value is the attribute name in a different case are still standard boolean syntax, but this strict comparison now treats them as custom values and returns "CHECKED" instead of preserving Cheerio's normalized "checked" behavior. Since rboolean already matches names case-insensitively, this should compare the value to the attribute name case-insensitively before falling through to return the raw value.

Useful? React with 👍 / 👎.