feat:: add restricted glob

Cargo.toml

@@ -178,6 +178,7 @@ crossbeam          = "0.8.4"
 dashmap            = "6.1.0"
 enumflags2         = "0.7.10"
 getrandom          = "0.2.15"
+globset            = "0.4.15"
 ignore             = "0.4.23"
 indexmap           = { version = "2.6.0", features = ["serde"] }
 insta              = "1.40.0"

crates/biome_js_analyze/Cargo.toml

@@ -28,6 +28,7 @@ biome_suppression        = { workspace = true }
 biome_unicode_table      = { workspace = true }
 bitvec                   = "1.0.1"
 enumflags2               = { workspace = true }
+globset                  = { workspace = true }
 natord                   = { workspace = true }
 regex                    = { workspace = true }
 roaring                  = "0.10.6"

crates/biome_js_analyze/src/assists/source/organize_imports.rs

@@ -2,13 +2,15 @@ use biome_analyze::{
     context::RuleContext, declare_source_rule, ActionCategory, Ast, FixKind, Rule, SourceActionKind,
 };
 use biome_console::markup;
+use biome_deserialize::Deserializable;
+use biome_deserialize_macros::Deserializable;
 use biome_js_syntax::JsModule;
 use biome_rowan::BatchMutationExt;
 
-use crate::JsRuleAction;
+use crate::{utils::restricted_glob::RestrictedGlob, JsRuleAction};
 
-pub mod util;
 pub mod legacy;
+pub mod util;
 
 declare_source_rule! {
     /// Provides a whole-source code action to sort the imports in the file
@@ -47,7 +49,7 @@ impl Rule for OrganizeImports {
     type Query = Ast<JsModule>;
     type State = State;
     type Signals = Option<Self::State>;
-    type Options = ();
+    type Options = Options;
 
     fn run(ctx: &RuleContext<Self>) -> Option<Self::State> {
         let root = ctx.query();
@@ -76,3 +78,47 @@ pub enum State {
     Legacy(legacy::ImportGroups),
     Modern,
 }
+
+#[derive(Clone, Debug, Default, serde::Deserialize, Deserializable, serde::Serialize)]
+#[cfg_attr(feature = "schema", derive(schemars::JsonSchema))]
+#[serde(rename_all = "camelCase", deny_unknown_fields, default)]
+pub struct Options {
+    legacy: bool,
+    import_groups: Box<[ImportGroup]>,
+}
+
+#[derive(Clone, Debug, serde::Deserialize, serde::Serialize)]
+#[cfg_attr(feature = "schema", derive(schemars::JsonSchema))]
+#[serde(untagged)]
+pub enum ImportGroup {
+    Predefined(PredefinedImportGroup),
+    Custom(RestrictedGlob),
+}
+impl Deserializable for ImportGroup {
+    fn deserialize(
+        value: &impl biome_deserialize::DeserializableValue,
+        name: &str,
+        diagnostics: &mut Vec<biome_deserialize::DeserializationDiagnostic>,
+    ) -> Option<Self> {
+        Some(
+            if let Some(predefined) = Deserializable::deserialize(value, name, diagnostics) {
+                ImportGroup::Predefined(predefined)
+            } else {
+                ImportGroup::Custom(Deserializable::deserialize(value, name, diagnostics)?)
+            },
+        )
+    }
+}
+
+#[derive(Clone, Debug, serde::Deserialize, Deserializable, Eq, PartialEq, serde::Serialize)]
+#[cfg_attr(feature = "schema", derive(schemars::JsonSchema))]
+pub enum PredefinedImportGroup {
+    #[serde(rename = ":blank-line:")]
+    BlankLine,
+    #[serde(rename = ":bun:")]
+    Bun,
+    #[serde(rename = ":node:")]
+    Node,
+    #[serde(rename = ":types:")]
+    Types,
+}

crates/biome_js_analyze/src/utils.rs

@@ -4,6 +4,7 @@ use std::iter;
 
 pub mod batch;
 pub mod rename;
+pub mod restricted_glob;
 pub mod restricted_regex;
 #[cfg(test)]
 pub mod tests;

crates/biome_js_analyze/src/utils/restricted_glob.rs

@@ -0,0 +1,154 @@
+use biome_deserialize_macros::Deserializable;
+
+/// A restricted glov pattern only supports the following syntaxes:
+///
+/// - star `*` that matches zero or more character inside a path segment
+/// - globstar `**` that matches zero or more path segments
+/// - Use `\*` to escape `*`
+/// - `?`, `[`, `]`, `{`, and `}` must be escaped using `\`.
+///   These characters are reserved for future use.
+/// - `!` must be escaped if it is the first characrter of the pattern
+///
+/// A path segment is delimited by path separator `/` or the start/end of the path.
+#[derive(Clone, Debug, Deserializable, serde::Deserialize, serde::Serialize)]
+#[serde(try_from = "String", into = "String")]
+pub struct RestrictedGlob(globset::GlobMatcher);
+
+impl std::ops::Deref for RestrictedGlob {
+    type Target = globset::GlobMatcher;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+impl std::fmt::Display for RestrictedGlob {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        let repr = self.0.glob().to_string();
+        f.write_str(&repr)
+    }
+}
+
+impl From<RestrictedGlob> for String {
+    fn from(value: RestrictedGlob) -> Self {
+        value.to_string()
+    }
+}
+
+impl std::str::FromStr for RestrictedGlob {
+    type Err = globset::ErrorKind;
+
+    fn from_str(value: &str) -> Result<Self, Self::Err> {
+        is_restricted_glob(value)?;
+        let mut glob_builder = globset::GlobBuilder::new(value);
+        // Allow escaping with `\` on all platforms.
+        glob_builder.backslash_escape(true);
+        // Only `**` can match `/`
+        glob_builder.literal_separator(true);
+        match glob_builder.build() {
+            Ok(glob) => Ok(RestrictedGlob(glob.compile_matcher())),
+            Err(error) => Err(error.kind().clone()),
+        }
+    }
+}
+
+impl TryFrom<String> for RestrictedGlob {
+    type Error = globset::ErrorKind;
+
+    fn try_from(value: String) -> Result<Self, Self::Error> {
+        value.parse()
+    }
+}
+
+#[cfg(feature = "schemars")]
+impl schemars::JsonSchema for RestrictedGlob {
+    fn schema_name() -> String {
+        "Regex".to_string()
+    }
+
+    fn json_schema(gen: &mut schemars::gen::SchemaGenerator) -> schemars::schema::Schema {
+        String::json_schema(gen)
+    }
+}
+
+/// Returns an error if `pattern` doesn't follow the restricted glob syntax.
+fn is_restricted_glob(pattern: &str) -> Result<(), globset::ErrorKind> {
+    let mut it = pattern.bytes().enumerate();
+    while let Some((i, c)) = it.next() {
+        match c {
+            b'!' if i == 0 => {
+                return Err(globset::ErrorKind::Regex(
+                    r"Negated globs `!` are not supported. Use `\!` to escape the character."
+                        .to_string(),
+                ));
+            }
+            b'\\' => {
+                // Accept a restrictive set of escape sequence
+                if let Some((_, c)) = it.next() {
+                    if !matches!(c, b'!' | b'*' | b'?' | b'{' | b'}' | b'[' | b']' | b'\\') {
+                        // SAFETY: safe because of the match
+                        let c = unsafe { char::from_u32_unchecked(c as u32) };
+                        // Escape sequences https://docs.rs/regex/latest/regex/#escape-sequences
+                        // and Perl char classes https://docs.rs/regex/latest/regex/#perl-character-classes-unicode-friendly
+                        return Err(globset::ErrorKind::Regex(format!(
+                            "Escape sequence \\{c} is not supported."
+                        )));
+                    }
+                } else {
+                    return Err(globset::ErrorKind::DanglingEscape);
+                }
+            }
+            b'?' => {
+                return Err(globset::ErrorKind::Regex(
+                    r"`?` matcher is not supported. Use `\?` to escape the character.".to_string(),
+                ));
+            }
+            b'[' | b']' => {
+                return Err(globset::ErrorKind::Regex(
+                    r"Character class `[]` are not supported. Use `\[` and `\]` to escape the characters."
+                        .to_string(),
+                ));
+            }
+            b'{' | b'}' => {
+                return Err(globset::ErrorKind::Regex(
+                    r"Alternates `{}` are not supported. Use `\{` and `\}` to escape the characters.".to_string(),
+                ));
+            }
+            _ => {}
+        }
+    }
+    Ok(())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_is_restricted_glob() {
+        assert!(is_restricted_glob("!*.js").is_err());
+        assert!(is_restricted_glob("*.[jt]s").is_err());
+        assert!(is_restricted_glob("*.{js,ts}").is_err());
+        assert!(is_restricted_glob("?*.js").is_err());
+        assert!(is_restricted_glob(r"\").is_err());
+        assert!(is_restricted_glob("!").is_err());
+
+        assert!(is_restricted_glob("*.js").is_ok());
+        assert!(is_restricted_glob("**/*.js").is_ok());
+        assert!(is_restricted_glob(r"\*").is_ok());
+        assert!(is_restricted_glob(r"\!").is_ok());
+    }
+
+    #[test]
+    fn test_restricted_regex() {
+        assert!(!"*.js"
+            .parse::<RestrictedGlob>()
+            .unwrap()
+            .is_match("file/path.js"));
+
+        assert!("**/*.js"
+            .parse::<RestrictedGlob>()
+            .unwrap()
+            .is_match("file/path.js"));
+    }
+}

crates/biome_js_analyze/src/utils/restricted_regex.rs

@@ -90,7 +90,7 @@ impl PartialEq for RestrictedRegex {
     }
 }
 
-/// Rteurns an error if `pattern` doesn't follow the restricted regular expression syntax.
+/// Returns an error if `pattern` doesn't follow the restricted regular expression syntax.
 fn is_restricted_regex(pattern: &str) -> Result<(), regex::Error> {
     let mut it = pattern.bytes();
     let mut is_in_char_class = false;
@@ -218,7 +218,7 @@ mod tests {
     use super::*;
 
     #[test]
-    fn test() {
+    fn test_is_restricted_regex() {
         assert!(is_restricted_regex("^a").is_err());
         assert!(is_restricted_regex("a$").is_err());
         assert!(is_restricted_regex(r"\").is_err());