fix(rome_js_analyze): fix a false positive for noDuplicateCase (#4709)

Author: Conaclos

CHANGELOG.md

@@ -174,6 +174,8 @@ if no error diagnostics are emitted.
 
 - Fix [`noInvalidConstructorSuper`](https://docs.rome.tools/lint/rules/noinvalidconstructorsuper/) rule that erroneously reported generic parents [#4624](https://github.com/rome/tools/issues/4624).
 
+-  Fix [`noDuplicateCase`](https://docs.rome.tools/lint/rules/noDuplicateCase/) rule that erroneously reported as equals the strings literals `"'"` and `'"'` [#4706](https://github.com/rome/tools/issues/4706).
+
 - Relax [`noBannedTypes`](https://docs.rome.tools/lint/rules/nobannedtypes/) and improve documentation
 
   The rule no longer reports a user type that reuses a banned type name.

crates/rome_js_analyze/src/analyzers/style/use_enum_initializers.rs

@@ -4,7 +4,9 @@ use rome_analyze::{declare_rule, ActionCategory, Ast, Rule, RuleDiagnostic};
 use rome_console::markup;
 use rome_diagnostics::Applicability;
 use rome_js_factory::make;
-use rome_js_syntax::{AnyJsExpression, AnyJsLiteralExpression, JsSyntaxKind, TsEnumMember};
+use rome_js_syntax::{
+    inner_text, AnyJsExpression, AnyJsLiteralExpression, JsSyntaxKind, TsEnumMember,
+};
 use rome_rowan::{AstNode, BatchMutationExt, Direction};
 
 declare_rule! {
@@ -118,8 +120,7 @@ impl Rule for UseEnumInitializers {
                 }
                 AnyJsLiteralExpression::JsStringLiteralExpression(expr) => {
                     let prev_enum_delim_val = expr.value_token().ok()?;
-                    let prev_enum_delim_val = prev_enum_delim_val.text();
-                    let prev_enum_val = &prev_enum_delim_val[1..(prev_enum_delim_val.len() - 1)];
+                    let prev_enum_val = inner_text(&prev_enum_delim_val);
                     if prev_name.text() == prev_enum_val {
                         let enum_name = enum_member.name().ok()?.text();
                         Some(AnyJsLiteralExpression::JsStringLiteralExpression(

crates/rome_js_analyze/src/semantic_analyzers/nursery/use_naming_convention.rs

@@ -17,7 +17,7 @@ use rome_deserialize::{
 use rome_diagnostics::Applicability;
 use rome_js_semantic::CanBeImportedExported;
 use rome_js_syntax::{
-    binding_ext::AnyJsBindingDeclaration, AnyJsClassMember, AnyJsObjectMember,
+    binding_ext::AnyJsBindingDeclaration, inner_text, AnyJsClassMember, AnyJsObjectMember,
     AnyJsVariableDeclaration, AnyTsTypeMember, JsIdentifierBinding, JsLiteralExportName,
     JsLiteralMemberName, JsPrivateClassMemberName, JsSyntaxKind, JsSyntaxToken,
     JsVariableDeclarator, JsVariableKind, TsEnumMember, TsIdentifierBinding, TsTypeParameterName,
@@ -290,10 +290,7 @@ impl Rule for UseNamingConvention {
             return None;
         }
         let name_token = node.name_token().ok()?;
-        let mut name = name_token.text_trimmed();
-        if name_token.kind() == JsSyntaxKind::JS_STRING_LITERAL {
-            name = &name[1..name.len() - 1];
-        }
+        let name = inner_text(&name_token);
         if !is_js_ident(name) {
             // ignore non-identifier strings
             return None;
@@ -323,10 +320,7 @@ impl Rule for UseNamingConvention {
             suggested_name,
         } = state;
         let name_token = ctx.query().name_token().ok()?;
-        let mut name = name_token.text_trimmed();
-        if name_token.kind() == JsSyntaxKind::JS_STRING_LITERAL {
-            name = &name[1..name.len() - 1];
-        }
+        let name = inner_text(&name_token);
         let trimmed_name = trim_underscore_dollar(name);
         let allowed_cases = element.allowed_cases(ctx.options());
         let allowed_case_names = allowed_cases

crates/rome_js_analyze/src/utils.rs

@@ -1,6 +1,6 @@
 use rome_js_factory::make;
 use rome_js_syntax::{
-    AnyJsStatement, JsLanguage, JsModuleItemList, JsStatementList, JsSyntaxNode, JsSyntaxToken,
+    inner_text, AnyJsStatement, JsLanguage, JsModuleItemList, JsStatementList, JsSyntaxNode,
     JsVariableDeclaration, JsVariableDeclarator, JsVariableDeclaratorList, JsVariableStatement, T,
 };
 use rome_rowan::{AstNode, AstSeparatedList, BatchMutation, Direction, WalkEvent};
@@ -225,7 +225,7 @@ pub(crate) fn is_node_equal(a_node: &JsSyntaxNode, b_node: &JsSyntaxNode) -> boo
             (None, Some(_)) | (Some(_), None) => return false,
             // both are tokens
             (Some(a), Some(b)) => {
-                if !is_token_text_equal(a, b) {
+                if inner_text(a) != inner_text(b) {
                     return false;
                 }
                 continue;
@@ -236,18 +236,6 @@ pub(crate) fn is_node_equal(a_node: &JsSyntaxNode, b_node: &JsSyntaxNode) -> boo
     true
 }
 
-/// Verify that tokens' inner text are equal
-fn is_token_text_equal(a: &JsSyntaxToken, b: &JsSyntaxToken) -> bool {
-    static QUOTES: [char; 2] = ['"', '\''];
-
-    a.token_text_trimmed()
-        .trim_start_matches(QUOTES)
-        .trim_end_matches(QUOTES)
-        == b.token_text_trimmed()
-            .trim_start_matches(QUOTES)
-            .trim_end_matches(QUOTES)
-}
-
 #[test]
 fn ok_to_camel_case() {
     assert_eq!(to_camel_case("camelCase"), Cow::Borrowed("camelCase"));

crates/rome_js_analyze/tests/specs/suspicious/noDuplicateCase/valid.js

@@ -113,3 +113,9 @@ switch (a) {
 	case toString:
 		break;
 }
+switch (a) {
+  case "'":
+	return ''';
+  case '"':
+	return '"';
+}

crates/rome_js_analyze/tests/specs/suspicious/noDuplicateCase/valid.js.snap

@@ -119,7 +119,12 @@ switch (a) {
 	case toString:
 		break;
 }
-
+switch (a) {
+  case "'":
+	return ''';
+  case '"':
+	return '"';
+}
 ```
 
 

crates/rome_js_syntax/src/lib.rs

@@ -272,3 +272,30 @@ impl OperatorPrecedence {
         matches!(self, OperatorPrecedence::Exponential)
     }
 }
+
+/// Similar to `JsSyntaxToken::text_trimmed` with the difference that delimiters of string literals are trimmed.
+///
+/// ## Examples
+///
+/// ```
+/// use rome_js_syntax::{JsSyntaxKind, JsSyntaxToken, inner_text};
+///
+/// let a = JsSyntaxToken::new_detached(JsSyntaxKind::JS_STRING_LITERAL, "'inner_text'", [], []);
+/// let b = JsSyntaxToken::new_detached(JsSyntaxKind::JS_STRING_LITERAL, "\"inner_text\"", [], []);
+/// assert_eq!(inner_text(&a), inner_text(&b));
+///
+/// let a = JsSyntaxToken::new_detached(JsSyntaxKind::LET_KW, "let", [], []);
+/// let b = JsSyntaxToken::new_detached(JsSyntaxKind::LET_KW, "let", [], []);
+/// assert_eq!(inner_text(&a), inner_text(&b));
+///
+/// let a = JsSyntaxToken::new_detached(JsSyntaxKind::LET_KW, "let", [], []);
+/// let b = JsSyntaxToken::new_detached(JsSyntaxKind::CONST_KW, "const", [], []);
+/// assert!(inner_text(&a) != inner_text(&b));
+/// ```
+pub fn inner_text(token: &JsSyntaxToken) -> &str {
+    let mut result = token.text_trimmed();
+    if token.kind() == JsSyntaxKind::JS_STRING_LITERAL {
+        result = &result[1..result.len() - 1];
+    }
+    result
+}