77dbbc7 · helix · git.iko.soy

git.iko.soy/helix

Detect filetype from shebang line (#1001)

commit77dbbc73f9c9b6599bc39b18625285685fe2e4b1 authorath3 <45574139+ath3@users.noreply.github.com> dateMon 8 Nov 2021, 16:19 (5 years ago) parent29e6849

5 files changed+65−2

compare with master

unifiedsplit

book/src/guides/adding_languages.md+2−1 helix-core/src/indent.rs+1 helix-core/src/syntax.rs+24 helix-view/src/document.rs+3−1 languages.toml+35

book/src/guides/adding_languages.md+2−1

@@ -33,10 +33,11 @@ These are the available keys and descriptions for the file.

3333| scope | A string like `source.js` that identifies the language. Currently, we strive to match the scope names used by popular TextMate grammars and by the Linguist library. Usually `source.<name>` or `text.<name>` in case of markup languages |

3434| injection-regex | regex pattern that will be tested against a language name in order to determine whether this language should be used for a potential [language injection][treesitter-language-injection] site. |

3535| file-types | The filetypes of the language, for example `["yml", "yaml"]` |

36+| shebangs | The interpreters from the shebang line, for example `["sh", "bash"]` |

3637| roots | A set of marker files to look for when trying to find the workspace root. For example `Cargo.lock`, `yarn.lock` |

3738| auto-format | Whether to autoformat this language when saving |

3839| comment-token | The token to use as a comment-token |

39−| indent | The indent to use. Has sub keys `tab-width` and `unit` |

40+| indent | The indent to use. Has sub keys `tab-width` and `unit` |

4041| config | Language server configuration |

4142

4243## Queries

helix-core/src/indent.rs+1

@@ -450,6 +450,7 @@ where

450450 language: vec![LanguageConfiguration {

451451 scope: "source.rust".to_string(),

452452 file_types: vec!["rs".to_string()],

453+ shebangs: vec![],

453454 language_id: "Rust".to_string(),

454455 highlight_config: OnceCell::new(),

455456 config: None,

helix-core/src/syntax.rs+24

@@ -14,6 +14,8 @@ use std::{

1414 cell::RefCell,

1515 collections::{HashMap, HashSet},

1616 fmt,

17+ fs::File,

18+ io::Read,

1719 path::Path,

1820 sync::Arc,

1921};

@@ -52,6 +54,7 @@ pub struct LanguageConfiguration {

5254 pub language_id: String,

5355 pub scope: String, // source.rust

5456 pub file_types: Vec<String>, // filename ends_with? <Gemfile, rb, etc>

57+ pub shebangs: Vec<String>, // interpreter(s) associated with language

5558 pub roots: Vec<String>, // these indicate project roots <.git, Cargo.toml>

5659 pub comment_token: Option<String>,

5760

@@ -254,6 +257,7 @@ pub struct Loader {

254257 // highlight_names ?

255258 language_configs: Vec<Arc<LanguageConfiguration>>,

256259 language_config_ids_by_file_type: HashMap<String, usize>, // Vec<usize>

260+ language_config_ids_by_shebang: HashMap<String, usize>,

257261}

258262

259263impl Loader {

@@ -261,6 +265,7 @@ impl Loader {

261265 let mut loader = Self {

262266 language_configs: Vec::new(),

263267 language_config_ids_by_file_type: HashMap::new(),

268+ language_config_ids_by_shebang: HashMap::new(),

264269 };

265270

266271 for config in config.language {

@@ -273,6 +278,11 @@ impl Loader {

273278 .language_config_ids_by_file_type

274279 .insert(file_type.clone(), language_id);

275280 }

281+ for shebang in &config.shebangs {

282+ loader

283+ .language_config_ids_by_shebang

284+ .insert(shebang.clone(), language_id);

285+ }

276286

277287 loader.language_configs.push(Arc::new(config));

278288 }

@@ -298,6 +308,20 @@ impl Loader {

298308 // TODO: content_regex handling conflict resolution

299309 }

300310

311+ pub fn language_config_for_shebang(&self, path: &Path) -> Option<Arc<LanguageConfiguration>> {

312+ // Read the first 128 bytes of the file. If its a shebang line, try to find the language

313+ let file = File::open(path).ok()?;

314+ let mut buf = String::with_capacity(128);

315+ file.take(128).read_to_string(&mut buf).ok()?;

316+ static SHEBANG_REGEX: Lazy<Regex> =

317+ Lazy::new(|| Regex::new(r"^#!\s*(?:\S*[/\\](?:env\s+)?)?([^\s\.\d]+)").unwrap());

318+ let configuration_id = SHEBANG_REGEX

319+ .captures(&buf)

320+ .and_then(|cap| self.language_config_ids_by_shebang.get(&cap[1]));

321+

322+ configuration_id.and_then(|&id| self.language_configs.get(id).cloned())

323+ }

324+

301325 pub fn language_config_for_scope(&self, scope: &str) -> Option<Arc<LanguageConfiguration>> {

302326 self.language_configs

303327 .iter()

helix-view/src/document.rs+3−1

@@ -494,7 +494,9 @@ impl Document {

494494 /// Detect the programming language based on the file type.

495495 pub fn detect_language(&mut self, theme: Option<&Theme>, config_loader: &syntax::Loader) {

496496 if let Some(path) = &self.path {

497− let language_config = config_loader.language_config_for_file_name(path);

497+ let language_config = config_loader

498+ .language_config_for_file_name(path)

499+ .or_else(|| config_loader.language_config_for_shebang(path));

498500 self.set_language(theme, language_config);

499501 }

500502 }

languages.toml+35

@@ -3,6 +3,7 @@ name = "rust"

33scope = "source.rust"

44injection-regex = "rust"

55file-types = ["rs"]

6+shebangs = []

67roots = []

78auto-format = true

89comment-token = "//"

@@ -17,6 +18,7 @@ name = "toml"

1718scope = "source.toml"

1819injection-regex = "toml"

1920file-types = ["toml"]

21+shebangs = []

2022roots = []

2123comment-token = "#"

2224

@@ -27,6 +29,7 @@ name = "protobuf"

2729scope = "source.proto"

2830injection-regex = "protobuf"

2931file-types = ["proto"]

32+shebangs = []

3033roots = []

3134comment-token = "//"

3235

@@ -37,6 +40,7 @@ name = "elixir"

3740scope = "source.elixir"

3841injection-regex = "elixir"

3942file-types = ["ex", "exs"]

43+shebangs = []

4044roots = []

4145comment-token = "#"

4246

@@ -48,6 +52,7 @@ name = "mint"

4852scope = "source.mint"

4953injection-regex = "mint"

5054file-types = ["mint"]

55+shebangs = []

5156roots = []

5257comment-token = "//"

5358

@@ -59,6 +64,7 @@ name = "json"

5964scope = "source.json"

6065injection-regex = "json"

6166file-types = ["json"]

67+shebangs = []

6268roots = []

6369

6470indent = { tab-width = 2, unit = " " }

@@ -68,6 +74,7 @@ name = "c"

6874scope = "source.c"

6975injection-regex = "c"

7076file-types = ["c"] # TODO: ["h"]

77+shebangs = []

7178roots = []

7279comment-token = "//"

7380

@@ -79,6 +86,7 @@ name = "cpp"

7986scope = "source.cpp"

8087injection-regex = "cpp"

8188file-types = ["cc", "hh", "cpp", "hpp", "h", "ipp", "tpp", "cxx", "hxx", "ixx", "txx", "ino"]

89+shebangs = []

8290roots = []

8391comment-token = "//"

8492

@@ -90,6 +98,7 @@ name = "c-sharp"

9098scope = "source.csharp"

9199injection-regex = "c-?sharp"

92100file-types = ["cs"]

101+shebangs = []

93102roots = []

94103comment-token = "//"

95104

@@ -100,6 +109,7 @@ name = "go"

100109scope = "source.go"

101110injection-regex = "go"

102111file-types = ["go"]

112+shebangs = []

103113roots = ["Gopkg.toml", "go.mod"]

104114auto-format = true

105115comment-token = "//"

@@ -113,6 +123,7 @@ name = "javascript"

113123scope = "source.js"

114124injection-regex = "^(js|javascript)$"

115125file-types = ["js", "mjs"]

126+shebangs = []

116127roots = []

117128comment-token = "//"

118129# TODO: highlights-jsx, highlights-params

@@ -124,6 +135,7 @@ name = "typescript"

124135scope = "source.ts"

125136injection-regex = "^(ts|typescript)$"

126137file-types = ["ts"]

138+shebangs = []

127139roots = []

128140# TODO: highlights-jsx, highlights-params

129141

@@ -135,6 +147,7 @@ name = "tsx"

135147scope = "source.tsx"

136148injection-regex = "^(tsx)$" # |typescript

137149file-types = ["tsx"]

150+shebangs = []

138151roots = []

139152# TODO: highlights-jsx, highlights-params

140153

@@ -146,6 +159,7 @@ name = "css"

146159scope = "source.css"

147160injection-regex = "css"

148161file-types = ["css"]

162+shebangs = []

149163roots = []

150164

151165indent = { tab-width = 2, unit = " " }

@@ -155,6 +169,7 @@ name = "html"

155169scope = "text.html.basic"

156170injection-regex = "html"

157171file-types = ["html"]

172+shebangs = []

158173roots = []

159174

160175indent = { tab-width = 2, unit = " " }

@@ -164,6 +179,7 @@ name = "python"

164179scope = "source.python"

165180injection-regex = "python"

166181file-types = ["py"]

182+shebangs = ["python"]

167183roots = []

168184comment-token = "#"

169185

@@ -176,6 +192,7 @@ name = "nix"

176192scope = "source.nix"

177193injection-regex = "nix"

178194file-types = ["nix"]

195+shebangs = []

179196roots = []

180197comment-token = "#"

181198

@@ -187,6 +204,7 @@ name = "ruby"

187204scope = "source.ruby"

188205injection-regex = "ruby"

189206file-types = ["rb"]

207+shebangs = ["ruby"]

190208roots = []

191209comment-token = "#"

192210

@@ -198,6 +216,7 @@ name = "bash"

198216scope = "source.bash"

199217injection-regex = "bash"

200218file-types = ["sh", "bash"]

219+shebangs = ["sh", "bash", "dash"]

201220roots = []

202221comment-token = "#"

203222

@@ -209,6 +228,7 @@ name = "php"

209228scope = "source.php"

210229injection-regex = "php"

211230file-types = ["php"]

231+shebangs = ["php"]

212232roots = []

213233

214234indent = { tab-width = 4, unit = " " }

@@ -218,6 +238,7 @@ name = "latex"

218238scope = "source.tex"

219239injection-regex = "tex"

220240file-types = ["tex"]

241+shebangs = []

221242roots = []

222243comment-token = "%"

223244

@@ -228,6 +249,7 @@ name = "julia"

228249scope = "source.julia"

229250injection-regex = "julia"

230251file-types = ["jl"]

252+shebangs = []

231253roots = []

232254comment-token = "#"

233255language-server = { command = "julia", args = [

@@ -253,6 +275,7 @@ name = "java"

253275scope = "source.java"

254276injection-regex = "java"

255277file-types = ["java"]

278+shebangs = []

256279roots = []

257280indent = { tab-width = 4, unit = " " }

258281

@@ -261,6 +284,7 @@ name = "ledger"

261284scope = "source.ledger"

262285injection-regex = "ledger"

263286file-types = ["ldg", "ledger", "journal"]

287+shebangs = []

264288roots = []

265289comment-token = ";"

266290indent = { tab-width = 4, unit = " " }

@@ -270,6 +294,7 @@ name = "ocaml"

270294scope = "source.ocaml"

271295injection-regex = "ocaml"

272296file-types = ["ml"]

297+shebangs = []

273298roots = []

274299comment-token = "(**)"

275300indent = { tab-width = 2, unit = " " }

@@ -278,6 +303,7 @@ indent = { tab-width = 2, unit = " " }

278303name = "ocaml-interface"

279304scope = "source.ocaml.interface"

280305file-types = ["mli"]

306+shebangs = []

281307roots = []

282308comment-token = "(**)"

283309indent = { tab-width = 2, unit = " "}

@@ -286,6 +312,7 @@ indent = { tab-width = 2, unit = " "}

286312name = "lua"

287313scope = "source.lua"

288314file-types = ["lua"]

315+shebangs = []

289316roots = []

290317comment-token = "--"

291318indent = { tab-width = 2, unit = " " }

@@ -295,6 +322,7 @@ name = "svelte"

295322scope = "source.svelte"

296323injection-regex = "svelte"

297324file-types = ["svelte"]

325+shebangs = []

298326roots = []

299327indent = { tab-width = 2, unit = " " }

300328language-server = { command = "svelteserver", args = ["--stdio"] }

@@ -305,6 +333,7 @@ name = "vue"

305333scope = "source.vue"

306334injection-regex = "vue"

307335file-types = ["vue"]

336+shebangs = []

308337roots = []

309338indent = { tab-width = 2, unit = " " }

310339

@@ -312,6 +341,7 @@ indent = { tab-width = 2, unit = " " }

312341name = "yaml"

313342scope = "source.yaml"

314343file-types = ["yml", "yaml"]

344+shebangs = []

315345roots = []

316346comment-token = "#"

317347indent = { tab-width = 2, unit = " " }

@@ -331,6 +361,7 @@ name = "zig"

331361scope = "source.zig"

332362injection-regex = "zig"

333363file-types = ["zig"]

364+shebangs = []

334365roots = ["build.zig"]

335366auto-format = true

336367comment-token = "//"

@@ -343,6 +374,7 @@ name = "prolog"

343374scope = "source.prolog"

344375roots = []

345376file-types = ["pl", "prolog"]

377+shebangs = ["swipl"]

346378comment-token = "%"

347379

348380language-server = { command = "swipl", args = [

@@ -354,6 +386,7 @@ language-server = { command = "swipl", args = [

354386name = "tsq"

355387scope = "source.tsq"

356388file-types = ["scm"]

389+shebangs = []

357390roots = []

358391comment-token = ";"

359392indent = { tab-width = 2, unit = " " }

@@ -362,6 +395,7 @@ indent = { tab-width = 2, unit = " " }

362395name = "cmake"

363396scope = "source.cmake"

364397file-types = ["cmake", "CMakeLists.txt"]

398+shebangs = []

365399roots = []

366400comment-token = "#"

367401indent = { tab-width = 2, unit = " " }

@@ -371,6 +405,7 @@ language-server = { command = "cmake-language-server" }

371405name = "perl"

372406scope = "source.perl"

373407file-types = ["pl", "pm"]

408+shebangs = ["perl"]

374409roots = []

375410comment-token = "#"

376411indent = { tab-width = 2, unit = " " }