Migrate SyntaxMapping impl to new system

Most existing builtins are removed but not yet ported, so the test
`user_can_override_builtin_mappings` should fail. It should pass once the old
rules have been ported.

CHANGELOG.md

@@ -16,6 +16,8 @@
 - Update Arch Linux package URL in README files #2779 (@brunobell)
 - Update and improve `zsh` completion, see #2772 (@okapia)
 - More extensible syntax mapping mechanism #2755 (@cyqsimon)
+  - [BREAKING] Precedence order of user-defined syntax mappings (using `-m/--map-syntax`) has been reversed
+    - `-m '*.foo:alpha' -m '*.foo:bravo'` now maps `bar.foo` to `alpha`
 
 ## Syntaxes
 
@@ -23,6 +25,10 @@
 
 ## `bat` as a library
 
+- Changes to `syntax_mapping::SyntaxMapping` #2755 (@cyqsimon)
+  - `SyntaxMapping::get_syntax_for` is now correctly public
+  - [BREAKING] `SyntaxMapping::{empty,builtin}` are removed; use `SyntaxMapping::new` instead
+  - [BREAKING] `SyntaxMapping::mappings` is replaced by `SyntaxMapping::{builtin,custom,all}_mappings`
 
 # v0.24.0
 

src/assets.rs

@@ -441,7 +441,7 @@ mod tests {
         fn new() -> Self {
             SyntaxDetectionTest {
                 assets: HighlightingAssets::from_binary(),
-                syntax_mapping: SyntaxMapping::builtin(),
+                syntax_mapping: SyntaxMapping::new(),
                 temp_dir: TempDir::new().expect("creation of temporary directory"),
             }
         }

src/bin/bat/app.rs

@@ -121,7 +121,7 @@ impl App {
             _ => unreachable!("other values for --paging are not allowed"),
         };
 
-        let mut syntax_mapping = SyntaxMapping::builtin();
+        let mut syntax_mapping = SyntaxMapping::new();
 
         if let Some(values) = self.matches.get_many::<String>("ignored-suffix") {
             for suffix in values {

src/bin/bat/main.rs

@@ -79,7 +79,7 @@ fn run_cache_subcommand(
 }
 
 fn get_syntax_mapping_to_paths<'a>(
-    mappings: &[(GlobMatcher, MappingTarget<'a>)],
+    mappings: &[(&GlobMatcher, &MappingTarget<'a>)],
 ) -> HashMap<&'a str, Vec<String>> {
     let mut map = HashMap::new();
     for mapping in mappings {
@@ -123,7 +123,7 @@ pub fn get_languages(config: &Config, cache_dir: &Path) -> Result<String> {
 
     languages.sort_by_key(|lang| lang.name.to_uppercase());
 
-    let configured_languages = get_syntax_mapping_to_paths(config.syntax_mapping.mappings());
+    let configured_languages = get_syntax_mapping_to_paths(&config.syntax_mapping.all_mappings());
 
     for lang in &mut languages {
         if let Some(additional_paths) = configured_languages.get(lang.name.as_str()) {

src/syntax_mapping.rs

@@ -3,6 +3,7 @@ use std::path::Path;
 use globset::{Candidate, GlobBuilder, GlobMatcher};
 
 use crate::error::Result;
+use builtin::BUILTIN_MAPPINGS;
 use ignored_suffixes::IgnoredSuffixes;
 
 mod builtin;
@@ -39,201 +40,58 @@ pub enum MappingTarget<'a> {
 
 #[derive(Debug, Clone, Default)]
 pub struct SyntaxMapping<'a> {
-    mappings: Vec<(GlobMatcher, MappingTarget<'a>)>,
+    /// User-defined mappings at run time.
+    custom_mappings: Vec<(GlobMatcher, MappingTarget<'a>)>,
     pub(crate) ignored_suffixes: IgnoredSuffixes<'a>,
 }
 
 impl<'a> SyntaxMapping<'a> {
-    pub fn empty() -> SyntaxMapping<'a> {
+    pub fn new() -> SyntaxMapping<'a> {
         Default::default()
     }
 
-    pub fn builtin() -> SyntaxMapping<'a> {
-        let mut mapping = Self::empty();
-        mapping.insert("*.h", MappingTarget::MapTo("C++")).unwrap();
-        mapping
-            .insert(".clang-format", MappingTarget::MapTo("YAML"))
-            .unwrap();
-        mapping.insert("*.fs", MappingTarget::MapTo("F#")).unwrap();
-        mapping
-            .insert("build", MappingTarget::MapToUnknown)
-            .unwrap();
-        mapping
-            .insert("**/.ssh/config", MappingTarget::MapTo("SSH Config"))
-            .unwrap();
-        mapping
-            .insert(
-                "**/bat/config",
-                MappingTarget::MapTo("Bourne Again Shell (bash)"),
-            )
-            .unwrap();
-        mapping
-            .insert(
-                "/etc/profile",
-                MappingTarget::MapTo("Bourne Again Shell (bash)"),
-            )
-            .unwrap();
-        mapping
-            .insert(
-                "os-release",
-                MappingTarget::MapTo("Bourne Again Shell (bash)"),
-            )
-            .unwrap();
-        mapping
-            .insert("*.pac", MappingTarget::MapTo("JavaScript (Babel)"))
-            .unwrap();
-        mapping
-            .insert("fish_history", MappingTarget::MapTo("YAML"))
-            .unwrap();
-
-        for glob in ["*.jsonl", "*.sarif"] {
-            mapping.insert(glob, MappingTarget::MapTo("JSON")).unwrap();
-        }
-
-        // See #2151, https://nmap.org/book/nse-language.html
-        mapping
-            .insert("*.nse", MappingTarget::MapTo("Lua"))
-            .unwrap();
-
-        // See #1008
-        mapping
-            .insert("rails", MappingTarget::MapToUnknown)
-            .unwrap();
-
-        mapping
-            .insert("Containerfile", MappingTarget::MapTo("Dockerfile"))
-            .unwrap();
-
-        mapping
-            .insert("*.ksh", MappingTarget::MapTo("Bourne Again Shell (bash)"))
-            .unwrap();
-
-        // Nginx and Apache syntax files both want to style all ".conf" files
-        // see #1131 and #1137
-        mapping
-            .insert("*.conf", MappingTarget::MapExtensionToUnknown)
-            .unwrap();
-
-        for glob in &[
-            "/etc/nginx/**/*.conf",
-            "/etc/nginx/sites-*/**/*",
-            "nginx.conf",
-            "mime.types",
-        ] {
-            mapping.insert(glob, MappingTarget::MapTo("nginx")).unwrap();
-        }
-
-        for glob in &[
-            "/etc/apache2/**/*.conf",
-            "/etc/apache2/sites-*/**/*",
-            "httpd.conf",
-        ] {
-            mapping
-                .insert(glob, MappingTarget::MapTo("Apache Conf"))
-                .unwrap();
-        }
-
-        for glob in &[
-            "**/systemd/**/*.conf",
-            "**/systemd/**/*.example",
-            "*.automount",
-            "*.device",
-            "*.dnssd",
-            "*.link",
-            "*.mount",
-            "*.netdev",
-            "*.network",
-            "*.nspawn",
-            "*.path",
-            "*.service",
-            "*.scope",
-            "*.slice",
-            "*.socket",
-            "*.swap",
-            "*.target",
-            "*.timer",
-        ] {
-            mapping.insert(glob, MappingTarget::MapTo("INI")).unwrap();
-        }
-
-        // unix mail spool
-        for glob in &["/var/spool/mail/*", "/var/mail/*"] {
-            mapping.insert(glob, MappingTarget::MapTo("Email")).unwrap()
-        }
-
-        // pacman hooks
-        mapping
-            .insert("*.hook", MappingTarget::MapTo("INI"))
-            .unwrap();
-
-        mapping
-            .insert("*.ron", MappingTarget::MapTo("Rust"))
-            .unwrap();
-
-        // Global git config files rooted in `$XDG_CONFIG_HOME/git/` or `$HOME/.config/git/`
-        // See e.g. https://git-scm.com/docs/git-config#FILES
-        match (
-            std::env::var_os("XDG_CONFIG_HOME").filter(|val| !val.is_empty()),
-            std::env::var_os("HOME")
-                .filter(|val| !val.is_empty())
-                .map(|home| Path::new(&home).join(".config")),
-        ) {
-            (Some(xdg_config_home), Some(default_config_home))
-                if xdg_config_home == default_config_home => {
-                insert_git_config_global(&mut mapping, &xdg_config_home)
-            }
-            (Some(xdg_config_home), Some(default_config_home)) /* else guard */ => {
-                insert_git_config_global(&mut mapping, &xdg_config_home);
-                insert_git_config_global(&mut mapping, &default_config_home)
-            }
-            (Some(config_home), None) => insert_git_config_global(&mut mapping, &config_home),
-            (None, Some(config_home)) => insert_git_config_global(&mut mapping, &config_home),
-            (None, None) => (),
-        };
-
-        fn insert_git_config_global(mapping: &mut SyntaxMapping, config_home: impl AsRef<Path>) {
-            let git_config_path = config_home.as_ref().join("git");
-
-            mapping
-                .insert(
-                    &git_config_path.join("config").to_string_lossy(),
-                    MappingTarget::MapTo("Git Config"),
-                )
-                .ok();
-
-            mapping
-                .insert(
-                    &git_config_path.join("ignore").to_string_lossy(),
-                    MappingTarget::MapTo("Git Ignore"),
-                )
-                .ok();
-
-            mapping
-                .insert(
-                    &git_config_path.join("attributes").to_string_lossy(),
-                    MappingTarget::MapTo("Git Attributes"),
-                )
-                .ok();
-        }
-
-        mapping
-    }
-
     pub fn insert(&mut self, from: &str, to: MappingTarget<'a>) -> Result<()> {
         let matcher = make_glob_matcher(from)?;
-        self.mappings.push((matcher, to));
+        self.custom_mappings.push((matcher, to));
         Ok(())
     }
 
-    pub fn mappings(&self) -> &[(GlobMatcher, MappingTarget<'a>)] {
+    /// Returns all mappings. User-defined mappings are listed before builtin
-        &self.mappings
+    /// mappings; mappings in front have higher precedence.
+    ///
+    /// Note that this function ignores builtin mappings that are invalid under
+    /// the current environment (i.e. their rules require an environment
+    /// variable that is unset).
+    pub fn all_mappings(&self) -> Vec<(&GlobMatcher, &MappingTarget<'a>)> {
+        self.custom_mappings()
+            .iter()
+            .map(|(matcher, target)| (matcher, target)) // as_ref
+            .chain(self.builtin_mappings())
+            .collect()
+    }
+
+    /// Returns all valid builtin mappings. Mappings in front have higher
+    /// precedence.
+    ///
+    /// If a mapping rule requires an environment variable that is unset, it
+    /// will be ignored.
+    pub fn builtin_mappings(&self) -> Vec<(&'static GlobMatcher, &'static MappingTarget<'static>)> {
+        BUILTIN_MAPPINGS
+            .iter()
+            .filter_map(|(matcher, target)| matcher.as_ref().map(|glob| (glob, target)))
+            .collect()
+    }
+
+    /// Returns all user-defined mappings.
+    pub fn custom_mappings(&self) -> &[(GlobMatcher, MappingTarget<'a>)] {
+        &self.custom_mappings
     }
 
     pub fn get_syntax_for(&self, path: impl AsRef<Path>) -> Option<MappingTarget<'a>> {
         // Try matching on the file name as-is.
         let candidate = Candidate::new(&path);
         let candidate_filename = path.as_ref().file_name().map(Candidate::new);
-        for (ref glob, ref syntax) in self.mappings.iter().rev() {
+        for (glob, syntax) in self.all_mappings().into_iter() {
             if glob.is_match_candidate(&candidate)
                 || candidate_filename
                     .as_ref()
@@ -261,7 +119,7 @@ mod tests {
     use super::*;
     #[test]
     fn basic() {
-        let mut map = SyntaxMapping::empty();
+        let mut map = SyntaxMapping::new();
         map.insert("/path/to/Cargo.lock", MappingTarget::MapTo("TOML"))
             .ok();
         map.insert("/path/to/.ignore", MappingTarget::MapTo("Git Ignore"))
@@ -281,7 +139,7 @@ mod tests {
 
     #[test]
     fn user_can_override_builtin_mappings() {
-        let mut map = SyntaxMapping::builtin();
+        let mut map = SyntaxMapping::new();
 
         assert_eq!(
             map.get_syntax_for("/etc/profile"),
@@ -297,35 +155,11 @@ mod tests {
 
     #[test]
     fn builtin_mappings() {
-        let map = SyntaxMapping::builtin();
+        let map = SyntaxMapping::new();
 
         assert_eq!(
             map.get_syntax_for("/path/to/build"),
             Some(MappingTarget::MapToUnknown)
         );
     }
-
-    #[test]
-    /// verifies that SyntaxMapping::builtin() doesn't repeat `Glob`-based keys
-    fn no_duplicate_builtin_keys() {
-        let mappings = SyntaxMapping::builtin().mappings;
-        for i in 0..mappings.len() {
-            let tail = mappings[i + 1..].into_iter();
-            let (dupl, _): (Vec<_>, Vec<_>) =
-                tail.partition(|item| item.0.glob() == mappings[i].0.glob());
-
-            // emit repeats on failure
-            assert_eq!(
-                dupl.len(),
-                0,
-                "Glob pattern `{}` mapped to multiple: {:?}",
-                mappings[i].0.glob().glob(),
-                {
-                    let (_, mut dupl_targets): (Vec<GlobMatcher>, Vec<MappingTarget>) =
-                        dupl.into_iter().cloned().unzip();
-                    dupl_targets.push(mappings[i].1)
-                },
-            )
-        }
-    }
 }