sugar-high 1.3.0 → 2.0.0

package/lib/presets/lang/javascript.js

@@ -0,0 +1,6 @@
+// @ts-check
+import { tokenize as tokenizeJavaScript } from './javascript-runtime.js'
+
+// JavaScript includes JSX and preserves the lightweight TypeScript heuristic used by the default
+// highlighter. Use the TypeScript preset to force TS/TSX behavior for ambiguous snippets.
+export const tokenize = tokenizeJavaScript

package/lib/presets/lang/json.js

@@ -1,4 +1,8 @@
 // @ts-check
 
+import { onCommentEnd, onCommentStart } from './clike-base.js'
+
 export const keywords = new Set(['true', 'false', 'null'])
 export const quotedKeys = true
+// JSONC is treated as the comment-tolerant JSON dialect rather than a separate language.
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/kotlin.js

@@ -0,0 +1,5 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './clike-base.js'
+export const keywords = new Set(['as','break','by','catch','class','companion','const','constructor','continue','data','do','else','enum','false','finally','for','fun','get','if','import','in','infix','init','interface','internal','is','lateinit','noinline','null','object','open','operator','out','override','package','private','protected','public','reified','return','sealed','set','suspend','tailrec','this','throw','true','try','typealias','val','var','vararg','when','where','while'])
+export const typeKeywords = new Set(['Any','Boolean','Byte','Char','Double','Float','Int','Long','Nothing','Short','String','Unit'])
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/markdown.js

@@ -0,0 +1,15 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './plain-base.js'
+
+export const keywords = new Set([])
+
+export const annotateLine = (line) => {
+  let annotation = ''
+  if (/^#{1,6}\s/.test(line.value)) annotation = 'markdown-heading'
+  else if (/^\s*>/.test(line.value)) annotation = 'markdown-quote'
+  else if (/^\s*(?:[-*+] |\d+[.)] )/.test(line.value)) annotation = 'markdown-list'
+  else if (/^\s*```/.test(line.value)) annotation = 'markdown-fence'
+  if (annotation) line.annotations.push(annotation)
+}
+
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/php.js

@@ -0,0 +1,5 @@
+// @ts-check
+export const keywords = new Set(['abstract','and','array','as','break','callable','case','catch','class','clone','const','continue','declare','default','do','echo','else','elseif','empty','enddeclare','endfor','endforeach','endif','endswitch','endwhile','enum','eval','exit','extends','false','final','finally','fn','for','foreach','from','function','global','goto','if','implements','include','include_once','instanceof','insteadof','interface','isset','list','match','namespace','new','null','or','print','private','protected','public','readonly','require','require_once','return','static','switch','throw','trait','true','try','unset','use','var','while','xor','yield'])
+export const typeKeywords = new Set(['bool','float','int','iterable','mixed','never','object','string','void'])
+export const onCommentStart = (curr, next) => curr === '#' ? 1 : curr + next === '//' ? 1 : curr + next === '/*' ? 2 : 0
+export const onCommentEnd = (prev, curr) => curr === '\n' ? 1 : prev + curr === '*/' ? 2 : 0

package/lib/presets/lang/plain-base.js

@@ -0,0 +1,4 @@
+// @ts-check
+
+export const onCommentStart = () => 0
+export const onCommentEnd = () => 0

package/lib/presets/lang/powershell.js

@@ -0,0 +1,5 @@
+// @ts-check
+export const caseInsensitive = true
+export const keywords = new Set(['begin','break','catch','class','continue','data','define','do','dynamicparam','else','elseif','end','enum','exit','filter','finally','for','foreach','from','function','if','in','param','process','return','switch','throw','trap','try','until','using','while'])
+export const onCommentStart = (curr, next) => curr === '#' ? 1 : curr + next === '<#' ? 2 : 0
+export const onCommentEnd = (prev, curr) => curr === '\n' ? 1 : prev + curr === '#>' ? 2 : 0

package/lib/presets/lang/shell.js

@@ -0,0 +1,10 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './hash-comment-base.js'
+
+export const keywords = new Set([
+  'case', 'coproc', 'do', 'done', 'elif', 'else', 'esac', 'export', 'fi', 'for',
+  'function', 'if', 'in', 'local', 'readonly', 'return', 'select', 'then', 'time',
+  'until', 'while',
+])
+
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/sql.js

@@ -0,0 +1,31 @@
+// @ts-check
+
+export const keywords = new Set([
+  'add', 'all', 'alter', 'and', 'as', 'asc', 'between', 'by', 'case', 'check',
+  'column', 'constraint', 'create', 'cross', 'database', 'default', 'delete', 'desc',
+  'distinct', 'drop', 'else', 'end', 'exists', 'foreign', 'from', 'full', 'group',
+  'having', 'in', 'index', 'inner', 'insert', 'into', 'is', 'join', 'key', 'left',
+  'like', 'limit', 'not', 'null', 'offset', 'on', 'or', 'order', 'outer', 'primary',
+  'references', 'right', 'select', 'set', 'table', 'then', 'union', 'unique', 'update',
+  'values', 'view', 'when', 'where', 'with',
+])
+
+export const typeKeywords = new Set([
+  'bigint', 'binary', 'bit', 'blob', 'boolean', 'char', 'date', 'datetime', 'decimal',
+  'double', 'float', 'int', 'integer', 'interval', 'json', 'numeric', 'real', 'smallint',
+  'text', 'time', 'timestamp', 'uuid', 'varchar',
+])
+
+export const caseInsensitive = true
+
+export const onCommentStart = (currentChar, nextChar) => {
+  const pair = currentChar + nextChar
+  if (pair === '--') return 1
+  if (pair === '/*') return 2
+  return 0
+}
+
+export const onCommentEnd = (prevChar, currChar) => {
+  if (currChar === '\n') return 1
+  return prevChar + currChar === '*/' ? 2 : 0
+}

package/lib/presets/lang/swift.js

@@ -0,0 +1,5 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './clike-base.js'
+export const keywords = new Set(['as','associatedtype','break','case','catch','class','continue','convenience','default','defer','deinit','didSet','do','dynamic','else','enum','extension','fallthrough','false','fileprivate','final','for','func','get','guard','if','import','in','indirect','infix','init','inout','internal','is','lazy','let','mutating','nil','nonmutating','open','operator','override','precedencegroup','private','protocol','public','repeat','required','rethrows','return','self','set','some','static','struct','subscript','super','switch','throw','throws','true','try','typealias','unowned','var','weak','where','while','willSet'])
+export const typeKeywords = new Set(['Any','Bool','Character','Double','Float','Int','Never','String','UInt','Void'])
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/toml.js

@@ -0,0 +1,5 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './hash-comment-base.js'
+export const keywords = new Set(['false','true'])
+export const quotedKeys = true
+export { onCommentEnd, onCommentStart }

package/lib/presets/lang/typescript.js

@@ -0,0 +1,7 @@
+// @ts-check
+import { tokenize as tokenizeJavaScript } from './javascript-runtime.js'
+
+export const tokenize = (code, options) => tokenizeJavaScript(code, {
+  ...options,
+  typescript: true,
+})

package/lib/presets/lang/yaml.js

@@ -0,0 +1,10 @@
+// @ts-check
+import { onCommentEnd, onCommentStart } from './hash-comment-base.js'
+
+export const keywords = new Set([
+  'false', 'False', 'FALSE', 'no', 'No', 'NO', 'null', 'Null', 'NULL',
+  'off', 'Off', 'OFF', 'on', 'On', 'ON', 'true', 'True', 'TRUE', 'yes', 'Yes', 'YES',
+])
+
+export const quotedKeys = true
+export { onCommentEnd, onCommentStart }

package/lib/shared.js

@@ -0,0 +1,149 @@
+// @ts-check
+
+const TokenTypes = /** @type {const} */ ('identifier keyword string class property entity jsxliterals sign comment break space'.split(' '))
+const [
+  T_IDENTIFIER, T_KEYWORD, T_STRING, T_CLASS, T_PROPERTY, T_ENTITY,
+  T_JSX_LITERALS, T_SIGN, T_COMMENT, T_BREAK, T_SPACE,
+] = TokenTypes.map((_, index) => index)
+
+const SugarHigh = /** @type {const} */ ({
+  TokenTypes,
+  TokenMap: new Map(TokenTypes.map((type, index) => [type, index])),
+})
+
+/** @param {string} value @param {Array<[number, string]>} tokens */
+function assemble(value, tokens) {
+  const lines = []
+  let lineIndex = 0
+  /** @type {Array<[number, string]>} */
+  const lineTokens = []
+  let lastWasBreak = false
+
+  /** @param {Array<[number, string]>} tokens */
+  function flushLine(tokens) {
+    lines.push({
+      index: lineIndex++,
+      value: tokens.map(([, tokenValue]) => tokenValue).join(''),
+      tokens: tokens.map(([type, tokenValue]) => ({
+        type: TokenTypes[type],
+        value: tokenValue,
+      })),
+      annotations: [],
+    })
+  }
+
+  for (let index = 0; index < tokens.length; index++) {
+    const token = tokens[index]
+    const [type, value] = token
+    if (type !== T_BREAK) {
+      if (value.includes('\n')) {
+        const values = value.split('\n')
+        for (let part = 0; part < values.length; part++) {
+          lineTokens.push([type, values[part]])
+          if (part < values.length - 1) {
+            flushLine(lineTokens)
+            lineTokens.length = 0
+          }
+        }
+      } else {
+        lineTokens.push(token)
+      }
+      lastWasBreak = false
+    } else {
+      if (lastWasBreak) flushLine([])
+      else {
+        flushLine(lineTokens)
+        lineTokens.length = 0
+      }
+      if (index === tokens.length - 1) flushLine([])
+      lastWasBreak = true
+    }
+  }
+
+  if (lineTokens.length) flushLine(lineTokens)
+  return { value, lines }
+}
+
+/**
+ * @param {import('./core.js').ParsedCode} parsed
+ * @param {import('./core.js').DisplayOptions | undefined} options
+ */
+function generate(parsed, options) {
+  const cx = options?.cx
+  const mark = options?.mark
+  const markLine = options?.markLine
+
+  return parsed.lines.map((parsedLine) => {
+    const line = {
+      index: parsedLine.index,
+      value: parsedLine.value,
+      tokens: parsedLine.tokens,
+      annotations: parsedLine.annotations,
+      className: `sh__line${parsedLine.annotations.map(annotation => ` sh__line--${annotation}`).join('')}`,
+      style: {},
+      properties: {},
+    }
+    markLine?.(line)
+
+    return {
+      type: 'element',
+      tagName: 'span',
+      children: parsedLine.tokens.map(({ type, value }) => {
+        const extraClassName = cx?.[type]
+        const token = {
+          type,
+          value,
+          className: `sh__token--${type}${extraClassName ? ` ${extraClassName}` : ''}`,
+          style: { color: `var(--sh-${type})` },
+          properties: {},
+        }
+        mark?.(token)
+        return {
+          type: 'element',
+          tokenType: token.type,
+          tagName: 'span',
+          children: [{ type: 'text', value: token.value }],
+          properties: {
+            ...token.properties,
+            className: token.className,
+            style: token.style,
+          },
+        }
+      }),
+      properties: {
+        ...line.properties,
+        className: line.className,
+        style: line.style,
+      },
+    }
+  })
+}
+
+const entities = { '&': '&amp;', '<': '&lt;', '>': '&gt;', '"': '&quot;', "'": '&#039;' }
+/** @param {string} value */
+const encode = (value) => value.replace(/[&<>"']/g, character => entities[character])
+
+/** @param {Record<string, any>} values */
+function attributes(values) {
+  const style = Object.entries(values.style || {})
+    .map(([key, value]) => `${key.replace(/[A-Z]/g, match => `-${match.toLowerCase()}`)}:${value}`).join(';')
+  const properties = Object.entries(values)
+    .filter(([key, value]) => /^[\w:-]+$/.test(key) && key !== 'className' && key !== 'style' && value !== false && value != null)
+    .map(([key, value]) => value === true ? key : `${key}="${encode(String(value))}"`).join(' ')
+  return `class="${encode(values.className || '')}"${style ? ` style="${encode(style)}"` : ''}${properties ? ` ${properties}` : ''}`
+}
+
+/** @param {Array<any>} lines */
+function toHtml(lines) {
+  return lines.map(line => {
+    const children = line.children.map(token => {
+      return `<${token.tagName} ${attributes(token.properties)}>${encode(token.children[0].value)}</${token.tagName}>`
+    }).join('')
+    return `<${line.tagName} ${attributes(line.properties)}>${children}</${line.tagName}>`
+  }).join('\n')
+}
+
+export {
+  assemble, encode, generate, SugarHigh, toHtml, TokenTypes, T_BREAK, T_CLASS, T_COMMENT, T_ENTITY, T_IDENTIFIER,
+  T_JSX_LITERALS, T_KEYWORD, T_PROPERTY, T_SIGN, T_SPACE, T_STRING,
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "sugar-high",
-  "version": "1.3.0",
+  "version": "2.0.0",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/huozhi/sugar-high.git",
@@ -14,23 +14,28 @@
       "types": "./lib/index.d.ts",
       "default": "./lib/index.js"
     },
-    "./presets": {
-      "types": "./lib/presets/index.d.ts",
-      "default": "./lib/presets/index.js"
+    "./core": {
+      "types": "./lib/core.d.ts",
+      "default": "./lib/core.js"
+    },
+    "./lang": {
+      "types": "./lib/lang.d.ts",
+      "default": "./lib/lang.js"
     }
   },
-  "description": "Super lightweight JSX syntax highlighter",
+  "description": "Lightweight, zero-dependency syntax highlighting for popular programming languages",
   "files": [
     "lib"
   ],
   "license": "MIT",
-  "scripts": {
-    "test": "vitest",
-    "build": "echo 'package requires no build'"
-  },
   "devDependencies": {
     "@types/node": "22.12.0",
     "typescript": "6.0.2",
     "vitest": "^3.0.2"
+  },
+  "scripts": {
+    "test": "vitest",
+    "build": "echo 'package requires no build'",
+    "benchmark": "node scripts/benchmark.mjs"
   }
-}
+}

package/lib/presets/index.d.ts

@@ -1,18 +0,0 @@
-type LanguageConfig = {
-  keywords: Set<string>
-  typeKeywords?: Set<string>
-  onCommentStart?(curr: string, next: string): 0 | 1 | 2
-  onCommentEnd?(prev: string, curr: string): 0 | 1 | 2
-  onQuote?(curr: string, i: number, code: string): number | null | undefined
-  quotedKeys?: boolean
-  lineClassName?(line: string, index: number): string | null | undefined
-}
-
-export const css: LanguageConfig
-export const rust: LanguageConfig
-export const python: LanguageConfig
-export const c: LanguageConfig
-export const go: LanguageConfig
-export const java: LanguageConfig
-export const diff: LanguageConfig
-export const json: LanguageConfig

package/lib/presets/index.js

@@ -1,8 +0,0 @@
-export * as css from './lang/css.js'
-export * as rust from './lang/rust.js'
-export * as python from './lang/python.js'
-export * as c from './lang/c.js'
-export * as go from './lang/go.js'
-export * as java from './lang/java.js'
-export * as diff from './lang/diff.js'
-export * as json from './lang/json.js'