Add Chapel Lexer

alecthomas · Jun 23, 2022 · f941d46 · f941d46
1 parent 298b727
commit f941d46
Show file tree

Hide file tree

Showing 4 changed files with 112 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -38,7 +38,7 @@ Prefix | Language
 :----: | --------
 A | ABAP, ABNF, ActionScript, ActionScript 3, Ada, Angular2, ANTLR, ApacheConf, APL, AppleScript, Arduino, Awk
 B | Ballerina, Base Makefile, Bash, Batchfile, BibTeX, Bicep, BlitzBasic, BNF, Brainfuck
-C | C, C#, C++, Caddyfile, Caddyfile Directives, Cap'n Proto, Cassandra CQL, Ceylon, CFEngine3, cfstatement, ChaiScript, Cheetah, Clojure, CMake, COBOL, CoffeeScript, Common Lisp, Coq, Crystal, CSS, Cython
+C | C, C#, C++, Caddyfile, Caddyfile Directives, Cap'n Proto, Cassandra CQL, Ceylon, CFEngine3, cfstatement, ChaiScript, Chapel, Cheetah, Clojure, CMake, COBOL, CoffeeScript, Common Lisp, Coq, Crystal, CSS, Cython
 D | D, Dart, Diff, Django/Jinja, Docker, DTD, Dylan
 E | EBNF, Elixir, Elm, EmacsLisp, Erlang
 F | Factor, Fish, Forth, Fortran, FSharp

diff --git a/lexers/chapel.go b/lexers/chapel.go
@@ -0,0 +1,62 @@
+package lexers
+
+import (
+	. "github.com/alecthomas/chroma/v2" // nolint
+)
+
+// Chapel lexer.
+var Chapel = Register(MustNewLexer(
+	&Config{
+		Name:      "Chapel",
+		Aliases:   []string{"chapel", "chpl"},
+		Filenames: []string{"*.chpl"},
+		MimeTypes: []string{},
+	},
+	func() Rules {
+		return Rules{
+			"root": {
+				{`\n`, TextWhitespace, nil},
+				{`\s+`, TextWhitespace, nil},
+				{`\\\n`, Text, nil},
+				{`//(.*?)\n`, CommentSingle, nil},
+				{`/(\\\n)?[*](.|\n)*?[*](\\\n)?/`, CommentMultiline, nil},
+				{Words(``, `\b`, `config`, `const`, `in`, `inout`, `out`, `param`, `ref`, `type`, `var`), KeywordDeclaration, nil},
+				{Words(``, `\b`, `false`, `nil`, `none`, `true`), KeywordConstant, nil},
+				{Words(``, `\b`, `bool`, `bytes`, `complex`, `imag`, `int`, `locale`, `nothing`, `opaque`, `range`, `real`, `string`, `uint`, `void`), KeywordType, nil},
+				{Words(``, `\b`, `atomic`, `single`, `sync`, `borrowed`, `owned`, `shared`, `unmanaged`, `align`, `as`, `begin`, `break`, `by`, `catch`, `cobegin`, `coforall`, `continue`, `defer`, `delete`, `dmapped`, `do`, `domain`, `else`, `enum`, `except`, `export`, `extern`, `for`, `forall`, `foreach`, `forwarding`, `if`, `implements`, `import`, `index`, `init`, `inline`, `label`, `lambda`, `let`, `lifetime`, `local`, `new`, `noinit`, `on`, `only`, `otherwise`, `override`, `pragma`, `primitive`, `private`, `prototype`, `public`, `reduce`, `require`, `return`, `scan`, `select`, `serial`, `sparse`, `subdomain`, `then`, `this`, `throw`, `throws`, `try`, `use`, `when`, `where`, `while`, `with`, `yield`, `zip`), Keyword, nil},
+				{`(iter)(\s+)`, ByGroups(Keyword, TextWhitespace), Push("procname")},
+				{`(proc)(\s+)`, ByGroups(Keyword, TextWhitespace), Push("procname")},
+				{`(operator)(\s+)`, ByGroups(Keyword, TextWhitespace), Push("procname")},
+				{`(class|interface|module|record|union)(\s+)`, ByGroups(Keyword, TextWhitespace), Push("classname")},
+				{`\d+i`, LiteralNumber, nil},
+				{`\d+\.\d*([Ee][-+]\d+)?i`, LiteralNumber, nil},
+				{`\.\d+([Ee][-+]\d+)?i`, LiteralNumber, nil},
+				{`\d+[Ee][-+]\d+i`, LiteralNumber, nil},
+				{`(\d*\.\d+)([eE][+-]?[0-9]+)?i?`, LiteralNumberFloat, nil},
+				{`\d+[eE][+-]?[0-9]+i?`, LiteralNumberFloat, nil},
+				{`0[bB][01]+`, LiteralNumberBin, nil},
+				{`0[xX][0-9a-fA-F]+`, LiteralNumberHex, nil},
+				{`0[oO][0-7]+`, LiteralNumberOct, nil},
+				{`[0-9]+`, LiteralNumberInteger, nil},
+				{`"(\\\\|\\"|[^"])*"`, LiteralString, nil},
+				{`'(\\\\|\\'|[^'])*'`, LiteralString, nil},
+				{`(=|\+=|-=|\*=|/=|\*\*=|%=|&=|\|=|\^=|&&=|\|\|=|<<=|>>=|<=>|<~>|\.\.|by|#|\.\.\.|&&|\|\||!|&|\||\^|~|<<|>>|==|!=|<=|>=|<|>|[+\-*/%]|\*\*)`, Operator, nil},
+				{`[:;,.?()\[\]{}]`, Punctuation, nil},
+				{`[a-zA-Z_][\w$]*`, NameOther, nil},
+			},
+			"classname": {
+				{`[a-zA-Z_][\w$]*`, NameClass, Pop(1)},
+			},
+			"procname": {
+				{`([a-zA-Z_][.\w$]*|\~[a-zA-Z_][.\w$]*|[+*/!~%<>=&^|\-:]{1,2})`, NameFunction, Pop(1)},
+				{`\(`, Punctuation, Push("receivertype")},
+				{`\)+\.`, Punctuation, nil},
+			},
+			"receivertype": {
+				{Words(``, `\b`, `atomic`, `single`, `sync`, `borrowed`, `owned`, `shared`, `unmanaged`), Keyword, nil},
+				{Words(``, `\b`, `bool`, `bytes`, `complex`, `imag`, `int`, `locale`, `nothing`, `opaque`, `range`, `real`, `string`, `uint`, `void`), KeywordType, nil},
+				{`[^()]*`, NameOther, Pop(1)},
+			},
+		}
+	},
+))
diff --git a/lexers/testdata/chapel.actual b/lexers/testdata/chapel.actual
@@ -0,0 +1,6 @@
+config const numMessages = 100;
+
+forall msg in 1..numMessages do
+  writeln("Hello, world! (from iteration ", msg, " of ", numMessages, ")");
+
+// Oh, and a comment for good measure.
diff --git a/lexers/testdata/chapel.expected b/lexers/testdata/chapel.expected
@@ -0,0 +1,43 @@
+[
+  {"type":"KeywordDeclaration","value":"config"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"KeywordDeclaration","value":"const"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"NameOther","value":"numMessages"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"Operator","value":"="},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"LiteralNumberInteger","value":"100"},
+  {"type":"Punctuation","value":";"},
+  {"type":"TextWhitespace","value":"\n\n"},
+  {"type":"Keyword","value":"forall"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"NameOther","value":"msg"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"KeywordDeclaration","value":"in"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"LiteralNumberInteger","value":"1"},
+  {"type":"Operator","value":".."},
+  {"type":"NameOther","value":"numMessages"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"Keyword","value":"do"},
+  {"type":"TextWhitespace","value":"\n  "},
+  {"type":"NameOther","value":"writeln"},
+  {"type":"Punctuation","value":"("},
+  {"type":"LiteralString","value":"\"Hello, world! (from iteration \""},
+  {"type":"Punctuation","value":","},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"NameOther","value":"msg"},
+  {"type":"Punctuation","value":","},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"LiteralString","value":"\" of \""},
+  {"type":"Punctuation","value":","},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"NameOther","value":"numMessages"},
+  {"type":"Punctuation","value":","},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"LiteralString","value":"\")\""},
+  {"type":"Punctuation","value":");"},
+  {"type":"TextWhitespace","value":"\n\n"},
+  {"type":"CommentSingle","value":"// Oh, and a comment for good measure.\n"}
+]