Add cache-dir to command-line and pyproject.toml (#1351)

README.md

@@ -318,6 +318,8 @@ Options:
           Show violations with source code
       --respect-gitignore
           Respect file exclusions via `.gitignore` and other standard ignore files
+      --force-exclude
+          Enforce exclusions, even for paths passed to Ruff directly on the command-line
       --show-files
           See the files Ruff will be run against with the current settings
       --show-settings
@@ -336,6 +338,8 @@ Options:
           The name of the file when passing it through stdin
       --explain <EXPLAIN>
           Explain a rule
+      --cache-dir <CACHE_DIR>
+          Path to the cache directory
   -h, --help
           Print help information
   -V, --version
@@ -1595,6 +1599,31 @@ allowed-confusables = ["−", "ρ", "∗"]
 
 ---
 
+#### [`cache-dir`](#cache-dir)
+
+A path to the cache directory.
+
+By default, Ruff stores cache results in a `.ruff_cache` directory in the current
+project root.
+
+However, Ruff will also respect the `RUFF_CACHE_DIR` environment variable, which takes
+precedence over that default.
+
+This setting will override even the `RUFF_CACHE_DIR` environment variable, if set.
+
+**Default value**: `.ruff_cache`
+
+**Type**: `PathBuf`
+
+**Example usage**:
+
+```toml
+[tool.ruff]
+cache-dir = "~/.cache/ruff"
+```
+
+---
+
 #### [`dummy-variable-rgx`](#dummy-variable-rgx)
 
 A regular expression used to identify "dummy" variables, or those which should be

flake8_to_ruff/src/converter.rs

@@ -315,6 +315,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -369,6 +370,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -423,6 +425,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -477,6 +480,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -531,6 +535,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -629,6 +634,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,
@@ -684,6 +690,7 @@ mod tests {
             src: None,
             target_version: None,
             unfixable: None,
+            cache_dir: None,
             flake8_annotations: None,
             flake8_bugbear: None,
             flake8_errmsg: None,

src/cache.rs

@@ -3,7 +3,7 @@ use std::fs;
 use std::fs::{create_dir_all, File, Metadata};
 use std::hash::{Hash, Hasher};
 use std::io::Write;
-use std::path::Path;
+use std::path::{Path, PathBuf};
 
 use anyhow::Result;
 use filetime::FileTime;
@@ -36,8 +36,12 @@ struct CheckResult {
     messages: Vec<Message>,
 }
 
-fn cache_dir() -> &'static Path {
-    Path::new(CACHE_DIR.as_ref().map_or(".ruff_cache", String::as_str))
+/// Return the cache directory for a given project root. Defers to the
+/// `RUFF_CACHE_DIR` environment variable, if set.
+pub fn cache_dir(project_root: &Path) -> PathBuf {
+    CACHE_DIR
+        .as_ref()
+        .map_or_else(|| project_root.join(".ruff_cache"), PathBuf::from)
 }
 
 fn content_dir() -> &'static Path {
@@ -53,10 +57,8 @@ fn cache_key<P: AsRef<Path>>(path: P, settings: &Settings, autofix: fixer::Mode)
     hasher.finish()
 }
 
-/// Initialize the cache directory.
-pub fn init() -> Result<()> {
-    let path = cache_dir();
-
+/// Initialize the cache at the specified `Path`.
+pub fn init(path: &Path) -> Result<()> {
     // Create the cache directories.
     create_dir_all(path.join(content_dir()))?;
 
@@ -75,15 +77,15 @@ pub fn init() -> Result<()> {
     Ok(())
 }
 
-fn write_sync(key: u64, value: &[u8]) -> Result<(), std::io::Error> {
+fn write_sync(cache_dir: &Path, key: u64, value: &[u8]) -> Result<(), std::io::Error> {
     fs::write(
-        cache_dir().join(content_dir()).join(format!("{key:x}")),
+        cache_dir.join(content_dir()).join(format!("{key:x}")),
         value,
     )
 }
 
-fn read_sync(key: u64) -> Result<Vec<u8>, std::io::Error> {
-    fs::read(cache_dir().join(content_dir()).join(format!("{key:x}")))
+fn read_sync(cache_dir: &Path, key: u64) -> Result<Vec<u8>, std::io::Error> {
+    fs::read(cache_dir.join(content_dir()).join(format!("{key:x}")))
 }
 
 /// Get a value from the cache.
@@ -98,7 +100,7 @@ pub fn get<P: AsRef<Path>>(
         return None;
     };
 
-    let encoded = read_sync(cache_key(path, settings, autofix)).ok()?;
+    let encoded = read_sync(&settings.cache_dir, cache_key(path, settings, autofix)).ok()?;
     let (mtime, messages) = match bincode::deserialize::<CheckResult>(&encoded[..]) {
         Ok(CheckResult {
             metadata: CacheMetadata { mtime },
@@ -135,6 +137,7 @@ pub fn set<P: AsRef<Path>>(
         messages,
     };
     if let Err(e) = write_sync(
+        &settings.cache_dir,
         cache_key(path, settings, autofix),
         &bincode::serialize(&check_result).unwrap(),
     ) {

src/cli.rs

@@ -133,6 +133,9 @@ pub struct Cli {
     /// Generate shell completion
     #[arg(long, hide = true, value_name = "SHELL")]
     pub generate_shell_completion: Option<clap_complete_command::Shell>,
+    /// Path to the cache directory.
+    #[arg(long)]
+    pub cache_dir: Option<PathBuf>,
 }
 
 impl Cli {
@@ -180,6 +183,7 @@ impl Cli {
                 fix: resolve_bool_arg(self.fix, self.no_fix),
                 format: self.format,
                 force_exclude: resolve_bool_arg(self.force_exclude, self.no_force_exclude),
+                cache_dir: self.cache_dir,
             },
         )
     }
@@ -238,6 +242,7 @@ pub struct Overrides {
     pub fix: Option<bool>,
     pub format: Option<SerializationFormat>,
     pub force_exclude: Option<bool>,
+    pub cache_dir: Option<PathBuf>,
 }
 
 /// Map the CLI settings to a `LogLevel`.

src/commands.rs

@@ -20,7 +20,7 @@ use crate::message::Message;
 use crate::resolver::{FileDiscovery, PyprojectDiscovery};
 use crate::settings::flags;
 use crate::settings::types::SerializationFormat;
-use crate::{packages, resolver};
+use crate::{cache, packages, resolver};
 
 /// Run the linter over a collection of files.
 pub fn run(
@@ -47,6 +47,30 @@ pub fn run(
             .collect::<Vec<_>>(),
     );
 
+    // Initialize the cache.
+    if matches!(cache, flags::Cache::Enabled) {
+        match &pyproject_strategy {
+            PyprojectDiscovery::Fixed(settings) => {
+                if let Err(e) = cache::init(&settings.cache_dir) {
+                    error!(
+                        "Failed to initialize cache at {}: {e:?}",
+                        settings.cache_dir.to_string_lossy()
+                    );
+                }
+            }
+            PyprojectDiscovery::Hierarchical(default) => {
+                for settings in std::iter::once(default).chain(resolver.iter()) {
+                    if let Err(e) = cache::init(&settings.cache_dir) {
+                        error!(
+                            "Failed to initialize cache at {}: {e:?}",
+                            settings.cache_dir.to_string_lossy()
+                        );
+                    }
+                }
+            }
+        }
+    };
+
     let start = Instant::now();
     let mut diagnostics: Diagnostics = par_iter(&paths)
         .map(|entry| {

src/main.rs

@@ -18,6 +18,7 @@ use std::sync::mpsc::channel;
 
 use ::ruff::autofix::fixer;
 use ::ruff::cli::{extract_log_level, Cli, Overrides};
+use ::ruff::commands;
 use ::ruff::logging::{set_up_logging, LogLevel};
 use ::ruff::printer::Printer;
 use ::ruff::resolver::{resolve_settings, FileDiscovery, PyprojectDiscovery, Relativity};
@@ -26,7 +27,6 @@ use ::ruff::settings::types::SerializationFormat;
 use ::ruff::settings::{pyproject, Settings};
 #[cfg(feature = "update-informer")]
 use ::ruff::updates;
-use ::ruff::{cache, commands};
 use anyhow::Result;
 use clap::{CommandFactory, Parser};
 use colored::Colorize;
@@ -123,6 +123,7 @@ fn inner_main() -> Result<ExitCode> {
     } else {
         fixer::Mode::None
     };
+    let cache = !cli.no_cache;
 
     if let Some(code) = cli.explain {
         commands::explain(&code, &format)?;
@@ -137,13 +138,6 @@ fn inner_main() -> Result<ExitCode> {
         return Ok(ExitCode::SUCCESS);
     }
 
-    // Initialize the cache.
-    let mut cache_enabled: bool = !cli.no_cache;
-    if cache_enabled && cache::init().is_err() {
-        eprintln!("Unable to initialize cache; disabling...");
-        cache_enabled = false;
-    }
-
     let printer = Printer::new(&format, &log_level);
     if cli.watch {
         if matches!(autofix, fixer::Mode::Generate | fixer::Mode::Apply) {
@@ -168,7 +162,7 @@ fn inner_main() -> Result<ExitCode> {
             &pyproject_strategy,
             &file_strategy,
             &overrides,
-            cache_enabled.into(),
+            cache.into(),
             fixer::Mode::None,
         )?;
         printer.write_continuously(&messages)?;
@@ -198,7 +192,7 @@ fn inner_main() -> Result<ExitCode> {
                             &pyproject_strategy,
                             &file_strategy,
                             &overrides,
-                            cache_enabled.into(),
+                            cache.into(),
                             fixer::Mode::None,
                         )?;
                         printer.write_continuously(&messages)?;
@@ -237,7 +231,7 @@ fn inner_main() -> Result<ExitCode> {
                 &pyproject_strategy,
                 &file_strategy,
                 &overrides,
-                cache_enabled.into(),
+                cache.into(),
                 autofix,
             )?
         };

src/resolver.rs

@@ -86,6 +86,11 @@ impl Resolver {
                 .unwrap_or(default),
         }
     }
+
+    /// Return an iterator over the resolved `Settings` in this `Resolver`.
+    pub fn iter(&self) -> impl Iterator<Item = &Settings> {
+        self.settings.values()
+    }
 }
 
 /// Recursively resolve a `Configuration` from a `pyproject.toml` file at the

src/settings/configuration.rs

@@ -46,6 +46,7 @@ pub struct Configuration {
     pub src: Option<Vec<PathBuf>>,
     pub target_version: Option<PythonVersion>,
     pub unfixable: Option<Vec<CheckCodePrefix>>,
+    pub cache_dir: Option<PathBuf>,
     // Plugins
     pub flake8_annotations: Option<flake8_annotations::settings::Options>,
     pub flake8_bugbear: Option<flake8_bugbear::settings::Options>,
@@ -130,6 +131,14 @@ impl Configuration {
                 .transpose()?,
             target_version: options.target_version,
             unfixable: options.unfixable,
+            cache_dir: options
+                .cache_dir
+                .map(|dir| {
+                    let dir = shellexpand::full(&dir);
+                    dir.map(|dir| PathBuf::from(dir.as_ref()))
+                })
+                .transpose()
+                .map_err(|e| anyhow!("Invalid `cache-dir` value: {e}"))?,
             // Plugins
             flake8_annotations: options.flake8_annotations,
             flake8_bugbear: options.flake8_bugbear,
@@ -184,6 +193,7 @@ impl Configuration {
             src: self.src.or(config.src),
             target_version: self.target_version.or(config.target_version),
             unfixable: self.unfixable.or(config.unfixable),
+            cache_dir: self.cache_dir.or(config.cache_dir),
             // Plugins
             flake8_annotations: self.flake8_annotations.or(config.flake8_annotations),
             flake8_bugbear: self.flake8_bugbear.or(config.flake8_bugbear),
@@ -254,6 +264,9 @@ impl Configuration {
         if let Some(unfixable) = overrides.unfixable {
             self.unfixable = Some(unfixable);
         }
+        if let Some(cache_dir) = overrides.cache_dir {
+            self.cache_dir = Some(cache_dir);
+        }
         // Special-case: `extend_ignore` and `extend_select` are parallel arrays, so
         // push an empty array if only one of the two is provided.
         match (overrides.extend_ignore, overrides.extend_select) {

src/settings/mod.rs

@@ -13,6 +13,7 @@ use path_absolutize::path_dedot;
 use regex::Regex;
 use rustc_hash::FxHashSet;
 
+use crate::cache::cache_dir;
 use crate::checks::CheckCode;
 use crate::checks_gen::{CheckCodePrefix, SuffixLength, CATEGORIES};
 use crate::settings::configuration::Configuration;
@@ -49,6 +50,7 @@ pub struct Settings {
     pub show_source: bool,
     pub src: Vec<PathBuf>,
     pub target_version: PythonVersion,
+    pub cache_dir: PathBuf,
     // Plugins
     pub flake8_annotations: flake8_annotations::settings::Settings,
     pub flake8_bugbear: flake8_bugbear::settings::Settings,
@@ -140,6 +142,7 @@ impl Settings {
                 .unwrap_or_else(|| vec![project_root.to_path_buf()]),
             target_version: config.target_version.unwrap_or(PythonVersion::Py310),
             show_source: config.show_source.unwrap_or_default(),
+            cache_dir: config.cache_dir.unwrap_or_else(|| cache_dir(project_root)),
             // Plugins
             flake8_annotations: config
                 .flake8_annotations
@@ -209,6 +212,7 @@ impl Settings {
             show_source: false,
             src: vec![path_dedot::CWD.clone()],
             target_version: PythonVersion::Py310,
+            cache_dir: cache_dir(path_dedot::CWD.as_path()),
             flake8_annotations: flake8_annotations::settings::Settings::default(),
             flake8_bugbear: flake8_bugbear::settings::Settings::default(),
             flake8_errmsg: flake8_errmsg::settings::Settings::default(),
@@ -242,6 +246,7 @@ impl Settings {
             show_source: false,
             src: vec![path_dedot::CWD.clone()],
             target_version: PythonVersion::Py310,
+            cache_dir: cache_dir(path_dedot::CWD.as_path()),
             flake8_annotations: flake8_annotations::settings::Settings::default(),
             flake8_bugbear: flake8_bugbear::settings::Settings::default(),
             flake8_errmsg: flake8_errmsg::settings::Settings::default(),

src/settings/options.rs

@@ -321,6 +321,23 @@ pub struct Options {
         "#
     )]
     pub unfixable: Option<Vec<CheckCodePrefix>>,
+    #[option(
+        doc = r#"
+            A path to the cache directory.
+
+            By default, Ruff stores cache results in a `.ruff_cache` directory in the current
+            project root.
+
+            However, Ruff will also respect the `RUFF_CACHE_DIR` environment variable, which takes
+            precedence over that default.
+
+            This setting will override even the `RUFF_CACHE_DIR` environment variable, if set.
+        "#,
+        default = ".ruff_cache",
+        value_type = "PathBuf",
+        example = r#"cache-dir = "~/.cache/ruff""#
+    )]
+    pub cache_dir: Option<String>,
     // Plugins
     #[option_group]
     pub flake8_annotations: Option<flake8_annotations::settings::Options>,

src/settings/pyproject.rs

@@ -141,6 +141,7 @@ mod tests {
                     src: None,
                     target_version: None,
                     unfixable: None,
+                    cache_dir: None,
                     flake8_annotations: None,
                     flake8_bugbear: None,
                     flake8_errmsg: None,
@@ -189,6 +190,7 @@ line-length = 79
                     src: None,
                     target_version: None,
                     unfixable: None,
+                    cache_dir: None,
                     flake8_annotations: None,
                     flake8_bugbear: None,
                     flake8_errmsg: None,
@@ -237,6 +239,7 @@ exclude = ["foo.py"]
                     src: None,
                     target_version: None,
                     unfixable: None,
+                    cache_dir: None,
                     flake8_annotations: None,
                     flake8_errmsg: None,
                     flake8_bugbear: None,
@@ -285,6 +288,7 @@ select = ["E501"]
                     src: None,
                     target_version: None,
                     unfixable: None,
+                    cache_dir: None,
                     flake8_annotations: None,
                     flake8_bugbear: None,
                     flake8_errmsg: None,
@@ -334,6 +338,7 @@ ignore = ["E501"]
                     src: None,
                     target_version: None,
                     unfixable: None,
+                    cache_dir: None,
                     flake8_annotations: None,
                     flake8_bugbear: None,
                     flake8_errmsg: None,
@@ -415,6 +420,7 @@ other-attribute = 1
                 format: None,
                 force_exclude: None,
                 unfixable: None,
+                cache_dir: None,
                 per_file_ignores: Some(FxHashMap::from_iter([(
                     "__init__.py".to_string(),
                     vec![CheckCodePrefix::F401]