postcss-normalize-string 9.0.3 → 9.0.5

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "postcss-normalize-string",
-  "version": "9.0.3",
+  "version": "9.0.5",
   "description": "Normalize wrapping quotes for CSS string literals.",
   "exports": {
     ".": {
@@ -33,7 +33,7 @@
     "url": "git+https://github.com/cssnano/cssnano.git"
   },
   "dependencies": {
-    "postcss-value-parser": "^4.2.0"
+    "@csstools/css-tokenizer": "^4.0.2"
   },
   "bugs": {
     "url": "https://github.com/cssnano/cssnano/issues"

package/src/index.js

@@ -1,233 +1,311 @@
-import valueParser from 'postcss-value-parser';
-
-/*
- * Constants (parser usage)
- */
+import { tokenize, TokenType } from '@csstools/css-tokenizer';
 
 const SINGLE_QUOTE = "'".charCodeAt(0);
 const DOUBLE_QUOTE = '"'.charCodeAt(0);
 const BACKSLASH = '\\'.charCodeAt(0);
 const NEWLINE = '\n'.charCodeAt(0);
-const SPACE = ' '.charCodeAt(0);
 const FEED = '\f'.charCodeAt(0);
-const TAB = '\t'.charCodeAt(0);
 const CR = '\r'.charCodeAt(0);
+const SPACE = ' '.charCodeAt(0);
+const TAB = '\t'.charCodeAt(0);
 
-const WORD_END = /[ \n\t\r\f'"\\]/g;
-
-/*
- * Constants (node type strings)
+/**
+ * @param {number} code
+ * @return {boolean}
  */
+function isHexDigit(code) {
+  return (
+    (code >= 0x30 && code <= 0x39) ||
+    (code >= 0x41 && code <= 0x46) ||
+    (code >= 0x61 && code <= 0x66)
+  );
+}
 
-const C_STRING = 'string';
-const C_ESCAPED_SINGLE_QUOTE = 'escapedSingleQuote';
-const C_ESCAPED_DOUBLE_QUOTE = 'escapedDoubleQuote';
-const C_SINGLE_QUOTE = 'singleQuote';
-const C_DOUBLE_QUOTE = 'doubleQuote';
-const C_NEWLINE = 'newline';
-const C_SINGLE = 'single';
+/**
+ * Return the index just past the hex digits of the escape whose backslash is at `pos`.
+ * CSS Syntax 3 allows one to six hex digits.
+ *
+ * @param {string} inner
+ * @param {number} pos
+ * @return {number}
+ */
+function getHexDigitsEnd(inner, pos) {
+  const limit = Math.min(pos + 7, inner.length);
+  let end = pos + 2;
+  while (end < limit && isHexDigit(inner.charCodeAt(end))) {
+    end++;
+  }
+  return end;
+}
 
-/*
- * Literals
+/**
+ * Return the length of the whitespace at `pos` (CRLF counts as one newline), or 0.
+ *
+ * @param {string} inner
+ * @param {number} pos
+ * @return {number}
  */
+function getWhitespaceLength(inner, pos) {
+  const code = inner.charCodeAt(pos);
+  if (code === CR) {
+    return inner.charCodeAt(pos + 1) === NEWLINE ? 2 : 1;
+  }
+  return code === SPACE || code === TAB || code === NEWLINE || code === FEED
+    ? 1
+    : 0;
+}
 
-const L_SINGLE_QUOTE = `'`;
-const L_DOUBLE_QUOTE = `"`;
-const L_NEWLINE = `\\\n`;
+/**
+ * Return the space that keeps an open hex escape from absorbing the code point at `pos`.
+ *
+ * @param {string} inner
+ * @param {number} pos
+ * @param {boolean} acceptsMoreDigits
+ * @return {string}
+ */
+function getHexEscapeDelimiter(inner, pos, acceptsMoreDigits) {
+  return getWhitespaceLength(inner, pos) > 0 ||
+    (acceptsMoreDigits && isHexDigit(inner.charCodeAt(pos)))
+    ? ' '
+    : '';
+}
 
-/*
- * Parser nodes
+/**
+ * Verify whether a string token has matching delimiters and balanced trailing escapes.
+ *
+ * @param {string} value
+ * @return {boolean}
  */
+function isClosedString(value) {
+  const len = value.length;
+  if (len < 2 || value.charCodeAt(len - 1) !== value.charCodeAt(0)) {
+    return false;
+  }
 
-const T_ESCAPED_SINGLE_QUOTE = { type: C_ESCAPED_SINGLE_QUOTE, value: `\\'` };
-const T_ESCAPED_DOUBLE_QUOTE = { type: C_ESCAPED_DOUBLE_QUOTE, value: `\\"` };
-const T_SINGLE_QUOTE = { type: C_SINGLE_QUOTE, value: L_SINGLE_QUOTE };
-const T_DOUBLE_QUOTE = { type: C_DOUBLE_QUOTE, value: L_DOUBLE_QUOTE };
-const T_NEWLINE = { type: C_NEWLINE, value: L_NEWLINE };
+  let backslashes = 0;
+  for (
+    let index = len - 2;
+    index >= 0 && value.charCodeAt(index) === BACKSLASH;
+    index--
+  ) {
+    backslashes++;
+  }
+
+  return backslashes % 2 === 0;
+}
 
-/** @typedef {typeof T_ESCAPED_SINGLE_QUOTE | typeof T_ESCAPED_DOUBLE_QUOTE | typeof T_SINGLE_QUOTE | typeof T_NEWLINE} StringAstNode */
 /**
- * @typedef {{nodes: StringAstNode[],
- *            types: {escapedSingleQuote: number, escapedDoubleQuote: number, singleQuote: number, doubleQuote: number},
- *            quotes: boolean}} StringAst
+ * Return the length of an escaped line continuation, or 0 if not a newline escape.
+ *
+ * @param {string} inner
+ * @param {number} pos
+ * @param {number} len
+ * @return {number}
  */
+function getEscapedNewlineLength(inner, pos, len) {
+  const next = inner.charCodeAt(pos + 1);
+  if (next === NEWLINE || next === FEED) {
+    return 2;
+  }
+  if (next === CR) {
+    return pos + 2 < len && inner.charCodeAt(pos + 2) === NEWLINE ? 3 : 2;
+  }
+  return 0;
+}
 
 /**
- * @param {StringAst} ast
- * @return {string}
+ * Scan string contents to count quotes and detect escaped newlines or redundant escapes.
+ *
+ * @param {string} inner
+ * @param {string} currentQuote
+ * @return {{ singleCount: number, doubleCount: number, hasEscapedNewline: boolean, hasRedundantEscape: boolean }}
  */
-function stringify(ast) {
-  let str = '';
-  for (const { value } of ast.nodes) {
-    // Collapse multiple line strings automatically
-    if (value !== L_NEWLINE) {
-      str += value;
+function scanQuotes(inner, currentQuote) {
+  const len = inner.length;
+  let singleCount = 0;
+  let doubleCount = 0;
+  let hasEscapedNewline = false;
+  let hasRedundantEscape = false;
+  let pos = 0;
+
+  while (pos < len) {
+    const code = inner.charCodeAt(pos);
+    if (code !== BACKSLASH) {
+      if (code === SINGLE_QUOTE) {
+        singleCount++;
+      } else if (code === DOUBLE_QUOTE) {
+        doubleCount++;
+      }
+      pos++;
+      continue;
+    }
+    const newlineLen = getEscapedNewlineLength(inner, pos, len);
+    if (newlineLen > 0) {
+      hasEscapedNewline = true;
+      pos += newlineLen;
+      continue;
     }
+    const next = inner.charCodeAt(pos + 1);
+    if (next === SINGLE_QUOTE) {
+      singleCount++;
+      hasRedundantEscape ||= currentQuote === '"';
+    } else if (next === DOUBLE_QUOTE) {
+      doubleCount++;
+      hasRedundantEscape ||= currentQuote === "'";
+    }
+    pos += 2;
+  }
+
+  return { singleCount, doubleCount, hasEscapedNewline, hasRedundantEscape };
+}
+
+/**
+ * Determine the optimal wrapping quote delimiter.
+ *
+ * @param {number} singleCount
+ * @param {number} doubleCount
+ * @param {'single' | 'double'} preferredQuote
+ * @param {string} currentQuote
+ * @return {string}
+ */
+function selectTargetQuote(
+  singleCount,
+  doubleCount,
+  preferredQuote,
+  currentQuote
+) {
+  if (singleCount > doubleCount) {
+    return '"';
+  }
+  if (doubleCount > singleCount) {
+    return "'";
+  }
+  if (singleCount === 0 && doubleCount === 0) {
+    return preferredQuote === 'single' ? "'" : '"';
   }
-  return str;
+  return currentQuote;
 }
 
 /**
- * @param {string} str
- * @return {StringAst}
+ * Reconstruct string with the target quote, collapsing newlines and updating escapes.
+ *
+ * @param {string} inner
+ * @param {string} targetQuote
+ * @return {string}
  */
-function parse(str) {
-  let code, next, value;
+function reconstructString(inner, targetQuote) {
+  const len = inner.length;
+  const targetCode = targetQuote.charCodeAt(0);
+  const otherQuote = targetQuote === "'" ? '"' : "'";
+  const otherCode = otherQuote.charCodeAt(0);
+  const escapedTargetQuote = '\\' + targetQuote;
+  const chunks = [targetQuote];
+  let cursor = 0;
   let pos = 0;
-  const len = str.length;
-
-  /** @type StringAst */
-  const ast = {
-    nodes: [],
-    types: {
-      escapedSingleQuote: 0,
-      escapedDoubleQuote: 0,
-      singleQuote: 0,
-      doubleQuote: 0,
-    },
-    quotes: false,
-  };
+  // A hex escape without a whitespace terminator would absorb a following hex
+  // digit or whitespace once the line continuations after it are removed.
+  let openHexEnd = -1;
+  let openHexIsShort = false;
 
   while (pos < len) {
-    code = str.charCodeAt(pos);
-
-    switch (code) {
-      case SPACE:
-      case TAB:
-      case CR:
-      case FEED:
-        next = pos;
-
-        do {
-          next += 1;
-          code = str.charCodeAt(next);
-        } while (
-          code === SPACE ||
-          code === NEWLINE ||
-          code === TAB ||
-          code === CR ||
-          code === FEED
-        );
-
-        ast.nodes.push({
-          type: 'space',
-          value: str.slice(pos, next),
-        });
-        pos = next - 1;
-        break;
-      case SINGLE_QUOTE:
-        ast.nodes.push(T_SINGLE_QUOTE);
-        ast.types[C_SINGLE_QUOTE]++;
-        ast.quotes = true;
-        break;
-      case DOUBLE_QUOTE:
-        ast.nodes.push(T_DOUBLE_QUOTE);
-        ast.types[C_DOUBLE_QUOTE]++;
-        ast.quotes = true;
-        break;
-      case BACKSLASH:
-        next = pos + 1;
-
-        if (str.charCodeAt(next) === SINGLE_QUOTE) {
-          ast.nodes.push(T_ESCAPED_SINGLE_QUOTE);
-          ast.types[C_ESCAPED_SINGLE_QUOTE]++;
-          ast.quotes = true;
-          pos = next;
-          break;
-        } else if (str.charCodeAt(next) === DOUBLE_QUOTE) {
-          ast.nodes.push(T_ESCAPED_DOUBLE_QUOTE);
-          ast.types[C_ESCAPED_DOUBLE_QUOTE]++;
-          ast.quotes = true;
-          pos = next;
-          break;
-        } else if (str.charCodeAt(next) === NEWLINE) {
-          ast.nodes.push(T_NEWLINE);
-          pos = next;
-          break;
+    const code = inner.charCodeAt(pos);
+    if (code === BACKSLASH) {
+      const newlineLen = getEscapedNewlineLength(inner, pos, len);
+      if (newlineLen > 0) {
+        if (pos > cursor) {
+          chunks.push(inner.slice(cursor, pos));
         }
-      /*
-       * We need to fall through here to handle the token as
-       * a whole word. The missing 'break' is intentional.
-       */
-      // oxlint-disable-next-line no-fallthrough
-      default:
-        WORD_END.lastIndex = pos + 1;
-        WORD_END.test(str);
-
-        if (WORD_END.lastIndex === 0) {
-          next = len - 1;
-        } else {
-          next = WORD_END.lastIndex - 2;
+        const followsOpenHex = pos === openHexEnd;
+        pos += newlineLen;
+        cursor = pos;
+        if (followsOpenHex) {
+          openHexEnd = pos;
+          chunks.push(getHexEscapeDelimiter(inner, pos, openHexIsShort));
         }
+        continue;
+      }
 
-        value = str.slice(pos, next + 1);
+      const next = inner.charCodeAt(pos + 1);
+      if (next === otherCode) {
+        if (pos > cursor) {
+          chunks.push(inner.slice(cursor, pos));
+        }
+        chunks.push(otherQuote);
+        pos += 2;
+        cursor = pos;
+        continue;
+      }
 
-        ast.nodes.push({
-          type: C_STRING,
-          value,
-        });
+      if (isHexDigit(next)) {
+        const digitsEnd = getHexDigitsEnd(inner, pos);
+        const terminatorLen = getWhitespaceLength(inner, digitsEnd);
+        if (terminatorLen === 0) {
+          openHexEnd = digitsEnd;
+          openHexIsShort = digitsEnd - pos < 7;
+        }
+        pos = digitsEnd + terminatorLen;
+        continue;
+      }
+
+      pos += 2;
+      continue;
+    }
 
-        pos = next;
+    if (code === targetCode) {
+      if (pos > cursor) {
+        chunks.push(inner.slice(cursor, pos));
+      }
+      chunks.push(escapedTargetQuote);
+      pos++;
+      cursor = pos;
+      continue;
     }
+
     pos++;
   }
 
-  return ast;
+  if (cursor < len) {
+    chunks.push(inner.slice(cursor));
+  }
+  chunks.push(targetQuote);
+  return chunks.join('');
 }
 
 /**
- * @param {valueParser.StringNode} node
- * @param {StringAst} ast
- * @return {void}
+ * Normalizes quotes and line continuations in a single closed CSS string token.
+ *
+ * @param {string} raw
+ * @param {'single' | 'double'} preferredQuote
+ * @return {string}
  */
-function changeWrappingQuotes(node, ast) {
-  const { types } = ast;
+function normalizeString(raw, preferredQuote) {
+  const currentQuote = raw[0];
+  const targetDelimiter = preferredQuote === 'single' ? "'" : '"';
+  const inner = raw.slice(1, -1);
 
-  if (types[C_SINGLE_QUOTE] || types[C_DOUBLE_QUOTE]) {
-    return;
+  if (currentQuote === targetDelimiter && !inner.includes('\\')) {
+    return raw;
   }
 
-  if (
-    node.quote === L_SINGLE_QUOTE &&
-    types[C_ESCAPED_SINGLE_QUOTE] > 0 &&
-    !types[C_ESCAPED_DOUBLE_QUOTE]
-  ) {
-    node.quote = L_DOUBLE_QUOTE;
-  }
+  const { singleCount, doubleCount, hasEscapedNewline, hasRedundantEscape } =
+    scanQuotes(inner, currentQuote);
+
+  const targetQuote = selectTargetQuote(
+    singleCount,
+    doubleCount,
+    preferredQuote,
+    currentQuote
+  );
 
   if (
-    node.quote === L_DOUBLE_QUOTE &&
-    types[C_ESCAPED_DOUBLE_QUOTE] > 0 &&
-    !types[C_ESCAPED_SINGLE_QUOTE]
+    targetQuote === currentQuote &&
+    !hasEscapedNewline &&
+    !hasRedundantEscape
   ) {
-    node.quote = L_SINGLE_QUOTE;
+    return raw;
   }
 
-  ast.nodes = changeChildQuotes(ast.nodes, node.quote);
-}
-/**
- * @param {StringAstNode[]} childNodes
- * @param {string} parentQuote
- * @return {StringAstNode[]}
- */
-function changeChildQuotes(childNodes, parentQuote) {
-  const updatedChildren = [];
-  for (const child of childNodes) {
-    if (
-      child.type === C_ESCAPED_DOUBLE_QUOTE &&
-      parentQuote === L_SINGLE_QUOTE
-    ) {
-      updatedChildren.push(T_DOUBLE_QUOTE);
-    } else if (
-      child.type === C_ESCAPED_SINGLE_QUOTE &&
-      parentQuote === L_DOUBLE_QUOTE
-    ) {
-      updatedChildren.push(T_SINGLE_QUOTE);
-    } else {
-      updatedChildren.push(child);
-    }
-  }
-  return updatedChildren;
+  return reconstructString(inner, targetQuote);
 }
 
 /**
@@ -236,29 +314,18 @@ function changeChildQuotes(childNodes, parentQuote) {
  * @return {string}
  */
 function normalize(value, preferredQuote) {
-  if (!value || !value.length) {
-    return value;
+  const chunks = [];
+  let cursor = 0;
+  for (const [type, raw, start, end] of tokenize({ css: value })) {
+    if (type !== TokenType.String || !isClosedString(raw)) continue;
+    const normalized = normalizeString(raw, preferredQuote);
+    if (normalized === raw) continue;
+    chunks.push(value.slice(cursor, start), normalized);
+    cursor = end + 1;
   }
-
-  return valueParser(value)
-    .walk((child) => {
-      if (child.type !== C_STRING) {
-        return;
-      }
-
-      const ast = parse(child.value);
-
-      if (ast.quotes) {
-        changeWrappingQuotes(child, ast);
-      } else if (preferredQuote === C_SINGLE) {
-        child.quote = L_SINGLE_QUOTE;
-      } else {
-        child.quote = L_DOUBLE_QUOTE;
-      }
-
-      child.value = stringify(ast);
-    })
-    .toString();
+  if (cursor === 0) return value;
+  chunks.push(value.slice(cursor));
+  return chunks.join('');
 }
 
 /**
@@ -268,28 +335,30 @@ function normalize(value, preferredQuote) {
  * @return {string}
  */
 function minify(original, cache, preferredQuote) {
-  const key = original + '|' + preferredQuote;
-  if (cache.has(key)) {
-    return /** @type {string} */ (cache.get(key));
+  if (!original || (!original.includes("'") && !original.includes('"'))) {
+    return original;
+  }
+  if (cache.has(original)) {
+    return /** @type {string} */ (cache.get(original));
   }
   const newValue = normalize(original, preferredQuote);
-  cache.set(key, newValue);
+  cache.set(original, newValue);
   return newValue;
 }
 
+/** @param {import('postcss').Declaration} decl @param {string} value */
+function assignValue(decl, value) {
+  decl.value = value;
+  if (decl.raws.value?.raw) decl.raws.value = { raw: value, value };
+}
+
 /** @typedef {{preferredQuote?: 'double' | 'single'}} Options */
 /**
  * @param {Options} opts
  * @return {import('postcss').Plugin}
  */
 function pluginCreator(opts) {
-  const { preferredQuote } = Object.assign(
-    {},
-    {
-      preferredQuote: 'double',
-    },
-    opts
-  );
+  const { preferredQuote = 'double' } = opts ?? {};
 
   return {
     postcssPlugin: 'postcss-normalize-string',
@@ -305,11 +374,19 @@ function pluginCreator(opts) {
           case 'rule':
             node.selector = minify(node.selector, cache, preferredQuote);
             break;
-          case 'decl':
-            node.value = minify(node.value, cache, preferredQuote);
+          case 'decl': {
+            const rawValue = node.raws.value;
+            const value =
+              rawValue?.value === node.value
+                ? (rawValue.raw ?? node.value)
+                : node.value;
+            assignValue(node, minify(value, cache, preferredQuote));
             break;
+          }
           case 'atrule':
-            node.params = minify(node.params, cache, preferredQuote);
+            if (node.name.toLowerCase() !== 'charset') {
+              node.params = minify(node.params, cache, preferredQuote);
+            }
             break;
         }
       });

package/types/index.d.ts

@@ -1,30 +1,3 @@
-declare const T_ESCAPED_SINGLE_QUOTE: {
-    type: string;
-    value: string;
-};
-declare const T_ESCAPED_DOUBLE_QUOTE: {
-    type: string;
-    value: string;
-};
-declare const T_SINGLE_QUOTE: {
-    type: string;
-    value: string;
-};
-declare const T_NEWLINE: {
-    type: string;
-    value: string;
-};
-export type StringAstNode = typeof T_ESCAPED_SINGLE_QUOTE | typeof T_ESCAPED_DOUBLE_QUOTE | typeof T_SINGLE_QUOTE | typeof T_NEWLINE;
-export type StringAst = {
-    nodes: StringAstNode[];
-    types: {
-        escapedSingleQuote: number;
-        escapedDoubleQuote: number;
-        singleQuote: number;
-        doubleQuote: number;
-    };
-    quotes: boolean;
-};
 export type Options = {
     preferredQuote?: 'double' | 'single';
 };

package/types/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.js"],"names":[],"mappings":"AAyCA,QAAA,MAAM,sBAAsB;;;CAAiD,CAAC;AAC9E,QAAA,MAAM,sBAAsB;;;CAAiD,CAAC;AAC9E,QAAA,MAAM,cAAc;;;CAAkD,CAAC;AAEvE,QAAA,MAAM,SAAS;;;CAAwC,CAAC;AAEpD,YAAoH,aAAa,GAAvH,OAAO,sBAAsB,GAAG,OAAO,sBAAsB,GAAG,OAAO,cAAc,GAAG,OAAO,SAAS,CAAe;AAElI,YAE6B,SAAS,GAF5B;IAAC,KAAK,EAAE,aAAa,EAAE,CAAC;IACvB,KAAK,EAAE;QAAC,kBAAkB,EAAE,MAAM,CAAC;QAAC,kBAAkB,EAAE,MAAM,CAAC;QAAC,WAAW,EAAE,MAAM,CAAC;QAAC,WAAW,EAAE,MAAM,CAAA;KAAC,CAAC;IAC1G,MAAM,EAAE,OAAO,CAAA;CAAC,CAAW;AAoOrC,YAAkD,OAAO,GAA/C;IAAC,cAAc,CAAC,EAAE,QAAQ,GAAG,QAAQ,CAAA;CAAC,CAAS;AAA7D,gEAAgE;AAChE;;;GAGG;AACH,iBAAS,aAAa,CAAC,IAAI,EAHhB,OAGgB,GAFf,OAAO,SAAS,EAAE,MAAM,CAmCnC;kBAjCQ,aAAa;iBAkCX,IAAI;;AAEf,QAAA,MAAM,aAAa,sBAAgB,CAAC;AAEpC,OAAO,EAAE,aAAa,IAAI,OAAO,EAAE,aAAa,IAAI,gBAAgB,EAAE,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.js"],"names":[],"mappings":"AAkWI,YAAkD,OAAO,GAA/C;IAAC,cAAc,CAAC,EAAE,QAAQ,GAAG,QAAQ,CAAA;CAAC,CAAS;AAA7D,gEAAgE;AAChE;;;GAGG;AACH,iBAAS,aAAa,CAAC,IAAI,EAHhB,OAGgB,GAFf,OAAO,SAAS,EAAE,MAAM,CAqCnC;kBAnCQ,aAAa;iBAoCX,IAAI;;AAEf,QAAA,MAAM,aAAa,sBAAgB,CAAC;AAEpC,OAAO,EAAE,aAAa,IAAI,OAAO,EAAE,aAAa,IAAI,gBAAgB,EAAE,CAAC"}