.changeset/small-words-show.md (1)

5-5: Tighten the changeset copy, add examples, and fix the link.

• Use past tense, proper naming (“eslint‑plugin‑react”), and end sentences with a period.
• Add tiny valid/invalid examples.
• Point to the correct rule page; the current slug likely collides with the CSS rule page. Update once the docs page exists.

Apply this diff:

-Added the new rule [`no-unknown-property`](https://biomejs.dev/linter/rules/no-unknown-property/)  from eslint react
+#### Lint rules.
+
+Added the new nursery rule [`noUnknownProperty`](<update-with-correct-docs-url>) ported from `eslint-plugin-react`.
+
+Examples:
+
+```jsx
+/* valid */
+<App class="foo" /><div data-foo="bar" />
+```
+
+```jsx
+/* invalid */
+<div class="foo" /><div data-testID="bar" />
+```

As per coding guidelines.

crates/biome_js_analyze/tests/specs/nursery/noUnknownProperty/ignore/valid.jsx (1)

1-6: Nice focused “ignore” case.

Consider adding a sibling negative case (property not on the ignore list) to prove the rule still fires.

crates/biome_js_analyze/tests/specs/nursery/noUnknownProperty/lowercase/invalid.jsx (1)

1-4: Confirm parity on custom components.

Do we intend requireDataLowercase to apply to custom components like <App />? ESLint’s rule often limits DOM‑prop checks to host elements. If we want parity, scope this to intrinsic elements; if we want stricter Biome behaviour, keep as is and document it.

crates/biome_js_analyze/tests/specs/nursery/noUnknownProperty/valid.jsx (1)

1-204: Big omnibus “valid” test – consider splitting by theme.

Smaller themed fixtures (events, SVG, boolean attrs, data/aria, legacy attrs) make diffs and failures easier to triage. Also keep the React 19 popover block gated in a dedicated future test.

Quick sweep for potential tag‑specific mismatches (e.g. allowFullScreen on non‑iframe) would be good to double‑check against the rule’s allowed‑tags map.

crates/biome_js_analyze/tests/specs/nursery/noUnknownProperty/invalid.jsx (2)

21-38: Add popover/onBeforeToggle cases to lock React 19 support

Please add invalid/valid cases for popover API:

• Attributes: popover, popovertarget, popovertargetaction.
• Events: onBeforeToggle/onToggle on allowed tags.
  This prevents regressions and aligns with your commented code in the rule.

---

1-3: Consider a companion valid case for custom elements

Add a valid fixture using a custom element (e.g., ) to assert the rule skips custom elements and tags with an is attribute. This documents intent.

crates/biome_js_analyze/src/lint/nursery/no_unknown_property.rs (5)

1065-1077: Avoid avoidable allocations for attribute names

name/value_token().text_trimmed().to_string() and the manual namespace concatenation allocate on every attribute. Use TokenText and Cow to allocate only for namespaced cases.

@@
-        let node_name = match node.name().ok()? {
-            AnyJsxAttributeName::JsxName(name) => {
-                name.value_token().ok()?.text_trimmed().to_string()
-            }
-            AnyJsxAttributeName::JsxNamespaceName(name) => {
-                let namespace = name.namespace().ok()?.value_token().ok()?;
-                let name = &name.name().ok()?.value_token().ok()?;
-                // There could be better way, but i couldn't extract namespaced attributes
-                // For e.g xlink:href
-                // without manually concatenating with ':'
-                namespace.text_trimmed().to_string() + ":" + name.text_trimmed()
-            }
-        };
+        use std::borrow::Cow;
+        let node_name: Cow<'_, str> = match node.name().ok()? {
+            AnyJsxAttributeName::JsxName(name) => {
+                Cow::Borrowed(name.value_token().ok()?.text_trimmed())
+            }
+            AnyJsxAttributeName::JsxNamespaceName(name) => {
+                let ns = name.namespace().ok()?.value_token().ok()?.text_trimmed();
+                let local = name.name().ok()?.value_token().ok()?.text_trimmed();
+                Cow::Owned(format!("{ns}:{local}"))
+            }
+        };
@@
-        if options.ignore.contains(&node_name) {
+        if options.ignore.iter().any(|s| s == &*node_name) {
             return None;
         }
@@
-        let name = normalize_attribute_case(&node_name);
+        let name = normalize_attribute_case(&node_name);

Also applies to: 1079-1081

---

1009-1014: Simplify tag_name_has_dot signature

This always returns Some(..). Make it return bool and drop ? at call sites; avoids needless Option churn.

-fn tag_name_has_dot(node: &AnyJsxElement) -> Option<bool> {
-    Some(matches!(
-        node.name().ok()?,
-        AnyJsxElementName::JsxMemberName(_)
-    ))
-}
+fn tag_name_has_dot(node: &AnyJsxElement) -> bool {
+    matches!(node.name().ok(), Ok(AnyJsxElementName::JsxMemberName(_)))
+}
@@
-        if tag_name_has_dot(&element)? {
+        if tag_name_has_dot(&element) {
             return None;
         }

---

1167-1183: Minor copy tweak

The diagnostic says “for React components”, but this rule validates DOM elements after gating out components. Suggest “in JSX” to avoid confusion.

-                    "Use '"{standard_name}"' instead of '"{name}"' for React components."
+                    "Use '"{standard_name}"' instead of '"{name}"' in JSX."

---

1484-1510: Use case‑insensitive check for data‑ only after the prefix*

Current has_uppercase(name) flags any uppercase anywhere, which is fine, but the option intent is about the part after data-. Consider restricting the scan to the suffix to keep the message crisp for odd inputs.

-            if options.require_data_lowercase && has_uppercase(&name) {
+            if options.require_data_lowercase {
+                let suffix = &name["data-".len()..];
+                if has_uppercase(suffix) {
                     return Some(NoUnknownPropertyDiagnostic::DataLowercaseRequired {
                         name: name.to_string(),
                         lowercase_name: name.to_lowercase(),
                     });
-            }
+                }
+            }

---

226-231: Keys are sorted — guard test is defensive, not critical

The ATTRIBUTE_TAGS_MAP keys are currently sorted and the code is correct. However, the suggestion to add a guard test is reasonable: without one, future edits could accidentally break sort order without triggering a compile-time error, and binary_search_by_key would silently fail at runtime.

Consider adding a simple unit test at the end of the file to assert the map remains sorted:

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn attribute_tags_map_is_sorted() {
        let keys: Vec<_> = ATTRIBUTE_TAGS_MAP.iter().map(|(k, _)| k).collect();
        let mut sorted = keys.clone();
        sorted.sort();
        assert_eq!(keys, sorted, "ATTRIBUTE_TAGS_MAP keys must remain sorted for binary_search_by_key");
    }
}

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_analyze/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:20:19.139Z
Learning: Applies to crates/biome_analyze/crates/biome_rule_options/lib/**/*.rs : For options types, derive Serialize, Deserialize, Deserializable (and JsonSchema under the schema feature) and use #[serde(rename_all="camelCase", deny_unknown_fields, default)] with skip_serializing_if where appropriate

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_formatter/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:22:15.851Z
Learning: Applies to crates/biome_formatter/tests/specs/**/options.json : Use options.json files colocated with test inputs to override formatting options for all files in that folder

Learnt from: dyc3
PR: biomejs/biome#7081
File: packages/@biomejs/biome/configuration_schema.json:7765-7781
Timestamp: 2025-08-05T14:43:29.581Z
Learning: The file `packages/biomejs/biome/configuration_schema.json` is auto-generated and should not be manually edited or reviewed for schema issues; any changes should be made at the code generation source.

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_analyze/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:20:19.139Z
Learning: Applies to crates/biome_analyze/**/tests/specs/**/*.{js,jsx,ts,tsx} : Place snapshot test cases under tests/specs/<group>/<ruleName>/ with files typically prefixed by invalid/ or valid/

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_analyze/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:20:19.139Z
Learning: A rule’s diagnostic should explain what the error is, why it’s triggered, and what to do (ideally via a code action or a note)

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_diagnostics/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:21:24.116Z
Learning: A diagnostic must explain why something went wrong, using log advices and links to documentation when helpful

crates/biome_js_analyze/src/lint/nursery/no_unknown_property.rs (1)

713-714: Event casing: add onBeforeToggle; drop ontoggle; prefer camelCase suggestions

Current list includes lowercase "onbeforetoggle" and "ontoggle", but misses the canonical React props "onBeforeToggle" and (already present) "onToggle". As written, ontoggle/onbeforetoggle will be accepted and won’t auto‑fix to camelCase.

-    "onbeforetoggle",
+    "onBeforeToggle",
@@
-    "ontoggle",
+    // Intentionally omit lowercase "ontoggle" so it maps to "onToggle"

Add tests asserting:

• <div onbeforetoggle={() => {}} /> → diagnostic suggesting onBeforeToggle
• <div ontoggle={() => {}} /> → diagnostic suggesting onToggle

This keeps behaviour aligned with React 19’s events and improves auto‑fix.

Also applies to: 847-848

crates/biome_js_analyze/src/lint/nursery/no_unknown_property.rs (3)

1065-1077: Avoid unnecessary allocations for attribute names; make ignore check case‑insensitive

You allocate for every attribute (to_string, manual concat). We can avoid allocations in the common case and only allocate for namespaced attributes, while matching ignore without allocating.

+use std::borrow::Cow;
@@
-        let node_name = match node.name().ok()? {
-            AnyJsxAttributeName::JsxName(name) => {
-                name.value_token().ok()?.text_trimmed().to_string()
-            }
-            AnyJsxAttributeName::JsxNamespaceName(name) => {
-                let namespace = name.namespace().ok()?.value_token().ok()?;
-                let name = &name.name().ok()?.value_token().ok()?;
-                // There could be better way, but i couldn't extract namespaced attributes
-                // For e.g xlink:href
-                // without manually concatenating with ':'
-                namespace.text_trimmed().to_string() + ":" + name.text_trimmed()
-            }
-        };
+        let node_name_text: Cow<'_, str> = match node.name().ok()? {
+            AnyJsxAttributeName::JsxName(name) => {
+                Cow::Borrowed(name.value_token().ok()?.text_trimmed())
+            }
+            AnyJsxAttributeName::JsxNamespaceName(name) => {
+                let ns = name.namespace().ok()?.value_token().ok()?.text_trimmed();
+                let nm = name.name().ok()?.value_token().ok()?.text_trimmed();
+                Cow::Owned(format!("{ns}:{nm}"))
+            }
+        };
@@
-        if options.ignore.contains(&node_name) {
+        if options
+            .ignore
+            .iter()
+            .any(|s| s.eq_ignore_ascii_case(node_name_text.as_ref()))
+        {
             return None;
         }
@@
-        let name = normalize_attribute_case(&node_name);
+        let name = normalize_attribute_case(node_name_text.as_ref());

This removes hot‑path allocations and makes ignore friendlier. If ignore is meant to be case‑sensitive, drop eq_ignore_ascii_case.

Also applies to: 1079-1084

---

972-973: Redundant use inside function

use biome_string_case::StrOnlyExtension; is already imported at the module level. Remove the inner use.

-    use biome_string_case::StrOnlyExtension;

---

1169-1177: Offer quick‑fixes in diagnostics (rename)

You already tell users what to do; wire in code actions to auto‑rename to standard_name and lower‑case data-* attributes.

Happy to sketch the .action plumbing if useful.

Also applies to: 1205-1213

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_analyze/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:20:19.139Z
Learning: A rule’s diagnostic should explain what the error is, why it’s triggered, and what to do (ideally via a code action or a note)

Learnt from: CR
PR: biomejs/biome#0
File: crates/biome_diagnostics/CONTRIBUTING.md:0-0
Timestamp: 2025-10-15T09:21:24.116Z
Learning: A diagnostic must explain why something went wrong, using log advices and links to documentation when helpful