feat: handle escaped unicode

biomejs · Nov 27, 2023 · 38493ce · 38493ce
1 parent adcb3b4
commit 38493ce
Show file tree

Hide file tree

Showing 5 changed files with 434 additions and 8 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/crates/biome_js_analyze/Cargo.toml b/crates/biome_js_analyze/Cargo.toml
@@ -27,6 +27,7 @@ biome_rowan            = { workspace = true }
 bpaf.workspace         = true
 lazy_static            = { workspace = true }
 natord                 = "1.0.9"
+regex                  = "1.5.5"
 roaring                = "0.10.1"
 rustc-hash             = { workspace = true }
 schemars               = { workspace = true, optional = true }

diff --git a/crates/biome_js_analyze/src/semantic_analyzers/nursery/no_misleading_character_class.rs b/crates/biome_js_analyze/src/semantic_analyzers/nursery/no_misleading_character_class.rs
@@ -4,6 +4,7 @@ use biome_console::markup;
 use biome_diagnostics::Applicability;
 use biome_js_syntax::{JsRegexLiteralExpression, JsSyntaxKind, JsSyntaxToken};
 use biome_rowan::{AstNode, BatchMutationExt, TextRange};
+use regex::Regex;
 
 declare_rule! {
     /// Disallow characters which are made with multiple code points in character class syntax
@@ -88,37 +89,36 @@ impl Rule for NoMisleadingCharacterClass {
             return None;
         }
 
-        let l = pattern.text();
-
+        let l = replace_escaped_unicode(pattern.text());
         let has_u_flag = flags.text().contains('u');
-        if !has_u_flag && has_surrogate_pair(l) {
+        if !has_u_flag && has_surrogate_pair(&l) {
             return Some(RuleState {
                 range,
                 message: Message::SurrogatePairWithoutUFlag,
             });
         }
 
-        if has_combining_class_or_vs16(l) {
+        if has_combining_class_or_vs16(&l) {
             return Some(RuleState {
                 range,
                 message: Message::CombiningClassOrVs16,
             });
         }
-        if has_regional_indicator_symbol(l) {
+        if has_regional_indicator_symbol(&l) {
             return Some(RuleState {
                 range,
                 message: Message::RegionalIndicatorSymbol,
             });
         }
 
-        if has_emoji_modifier(l) {
+        if has_emoji_modifier(&l) {
             return Some(RuleState {
                 range,
                 message: Message::EmojiModifier,
             });
         }
 
-        if zwj(l) {
+        if zwj(&l) {
             return Some(RuleState {
                 range,
                 message: Message::JoinedCharSequence,
@@ -232,3 +232,14 @@ fn zwj(chars: &str) -> bool {
 fn has_surrogate_pair(s: &str) -> bool {
     s.chars().any(|c| c as u32 > 0xFFFF)
 }
+
+fn replace_escaped_unicode(input: &str) -> String {
+    let re = Regex::new(r"\\u\{([0-9a-fA-F]+)\}").unwrap();
+    re.replace_all(input, |caps: &regex::Captures| {
+        u32::from_str_radix(&caps[1], 16)
+            .ok()
+            .and_then(char::from_u32)
+            .map_or_else(String::new, |c| c.to_string())
+    })
+    .into_owned()
+}
diff --git a/crates/biome_js_analyze/tests/specs/nursery/noMisleadingCharacterClass/invalid.js b/crates/biome_js_analyze/tests/specs/nursery/noMisleadingCharacterClass/invalid.js
@@ -3,4 +3,32 @@
 /^[👶🏻]$/u;
 /^[🇯🇵]$/u;
 /^[👨‍👩‍👦]$/u;
-/^[👍]$/;
+/^[👍]$/;
+
+/[👍]/;
+/[\uD83D\uDC4D]/;
+/[👍]\\a/;
+/(?<=[👍])/;
+/[Á]/;
+/[Á]/u;
+/[\u0041\u0301]/;
+/[\u0041\u0301]/u;
+/[\u{41}\u{301}]/u;
+/[❇️]/;
+/[❇️]/u;
+/[\u2747\uFE0F]/;
+/[\u2747\uFE0F]/u;
+/[\u{2747}\u{FE0F}]/u;
+/[👶🏻]/;
+/[👶🏻]/u;
+/[\uD83D\uDC76\uD83C\uDFFB]/u;
+/[\u{1F476}\u{1F3FB}]/u;
+/[🇯🇵]/;
+/[🇯🇵]/i;
+/[🇯🇵]/u;
+/[\uD83C\uDDEF\uD83C\uDDF5]/u;
+/[\u{1F1EF}\u{1F1F5}]/u;
+/[👨‍👩‍👦]/;
+/[👨‍👩‍👦]/u;
+/[\uD83D\uDC68\u200D\uD83D\uDC69\u200D\uD83D\uDC66]/u;
+/[\u{1F468}\u{200D}\u{1F469}\u{200D}\u{1F466}]/u;