Syntax highlighting for comments

CHANGELOG.md

@@ -6,6 +6,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/)
 and this project adheres to [Semantic Versioning](http://semver.org/).
 
 ## [Unreleased]
+### Added
+- Syntax highlighting for comments (`#`) ([#212](https://github.com/cucumber/language-service/pull/212))
+
 ### Fixed
 - (Ruby) Support `And` and `But` step definition annotations ([#211](https://github.com/cucumber/language-service/pull/211))
 

src/service/getGherkinSemanticTokens.ts

@@ -14,11 +14,12 @@ export const semanticTokenTypes: SemanticTokenTypes[] = [
   SemanticTokenTypes.type, // @tags and DocString ```type
   SemanticTokenTypes.variable, // step <placeholder>
   SemanticTokenTypes.property, // examples table header row
+  SemanticTokenTypes.comment, // # comments
 ]
 
 const indexByType = Object.fromEntries(semanticTokenTypes.map((type, index) => [type, index]))
 
-// https://microsoft.github.io/language-server-protocol/specifications/specification-3-17/#textDocument_semanticTokens
+// https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#textDocument_semanticTokens
 export function getGherkinSemanticTokens(
   gherkinSource: string,
   expressions: readonly Expression[]
@@ -170,9 +171,45 @@ export function getGherkinSemanticTokens(
       inExamples = false
       return arr
     },
+    comment(comment, arr) {
+      return makeLocationToken(comment.location, comment.text, SemanticTokenTypes.comment, arr)
+    },
   })
 
+  // Order tokens by positive relative line numbers. Required as table rows
+  // are processed before comments. Without this ordering, table rows can
+  // have negative relative line numbers, which is not supported by the
+  // Language Server Protocol.
+  const orderedData = orderRelativeTokenPositions(data)
+
   return {
-    data,
+    data: orderedData,
+  }
+}
+
+function orderRelativeTokenPositions(data: number[]): number[] {
+  // Create tokens and convert relative line numbers to absolute
+  const tokens = []
+  let line = 0
+  for (let tokenStartIndex = 0; tokenStartIndex < data.length; tokenStartIndex += 5) {
+    line += data[tokenStartIndex]
+    tokens.push([
+      line,
+      data[tokenStartIndex + 1],
+      data[tokenStartIndex + 2],
+      data[tokenStartIndex + 3],
+      data[tokenStartIndex + 4],
+    ])
   }
+
+  // Sort tokens by line number
+  tokens.sort((a, b) => a[0] - b[0])
+
+  // Convert absolute line numbers back to relative
+  for (let i = tokens.length - 1; i > 0; i--) {
+    tokens[i][0] -= tokens[i - 1][0]
+  }
+
+  // Flatten the array
+  return tokens.flat()
 }

test/service/getGherkinSemanticTokens.test.ts

@@ -58,6 +58,7 @@ Feature: a
     ])
     const actual = tokenize(gherkinSource, semanticTokens.data)
     const expected: TokenWithType[] = [
+      ['# some comment', SemanticTokenTypes.comment],
       ['@foo', SemanticTokenTypes.type],
       ['@bar', SemanticTokenTypes.type],
       ['Feature', SemanticTokenTypes.keyword],