printer-notifications/node_modules/css-tree/cjs/utils/string.cjs

100 lines
3.2 KiB
JavaScript
Raw Normal View History

2023-11-13 21:10:04 +00:00
'use strict';
const charCodeDefinitions = require('../tokenizer/char-code-definitions.cjs');
const utils = require('../tokenizer/utils.cjs');
const REVERSE_SOLIDUS = 0x005c; // U+005C REVERSE SOLIDUS (\)
const QUOTATION_MARK = 0x0022; // "
const APOSTROPHE = 0x0027; // '
function decode(str) {
const len = str.length;
const firstChar = str.charCodeAt(0);
const start = firstChar === QUOTATION_MARK || firstChar === APOSTROPHE ? 1 : 0;
const end = start === 1 && len > 1 && str.charCodeAt(len - 1) === firstChar ? len - 2 : len - 1;
let decoded = '';
for (let i = start; i <= end; i++) {
let code = str.charCodeAt(i);
if (code === REVERSE_SOLIDUS) {
// special case at the ending
if (i === end) {
// if the next input code point is EOF, do nothing
// otherwise include last quote as escaped
if (i !== len - 1) {
decoded = str.substr(i + 1);
}
break;
}
code = str.charCodeAt(++i);
// consume escaped
if (charCodeDefinitions.isValidEscape(REVERSE_SOLIDUS, code)) {
const escapeStart = i - 1;
const escapeEnd = utils.consumeEscaped(str, escapeStart);
i = escapeEnd - 1;
decoded += utils.decodeEscaped(str.substring(escapeStart + 1, escapeEnd));
} else {
// \r\n
if (code === 0x000d && str.charCodeAt(i + 1) === 0x000a) {
i++;
}
}
} else {
decoded += str[i];
}
}
return decoded;
}
// https://drafts.csswg.org/cssom/#serialize-a-string
// § 2.1. Common Serializing Idioms
function encode(str, apostrophe) {
const quote = apostrophe ? '\'' : '"';
const quoteCode = apostrophe ? APOSTROPHE : QUOTATION_MARK;
let encoded = '';
let wsBeforeHexIsNeeded = false;
for (let i = 0; i < str.length; i++) {
const code = str.charCodeAt(i);
// If the character is NULL (U+0000), then the REPLACEMENT CHARACTER (U+FFFD).
if (code === 0x0000) {
encoded += '\uFFFD';
continue;
}
// If the character is in the range [\1-\1f] (U+0001 to U+001F) or is U+007F,
// the character escaped as code point.
// Note: Do not compare with 0x0001 since 0x0000 is precessed before
if (code <= 0x001f || code === 0x007F) {
encoded += '\\' + code.toString(16);
wsBeforeHexIsNeeded = true;
continue;
}
// If the character is '"' (U+0022) or "\" (U+005C), the escaped character.
if (code === quoteCode || code === REVERSE_SOLIDUS) {
encoded += '\\' + str.charAt(i);
wsBeforeHexIsNeeded = false;
} else {
if (wsBeforeHexIsNeeded && (charCodeDefinitions.isHexDigit(code) || charCodeDefinitions.isWhiteSpace(code))) {
encoded += ' ';
}
// Otherwise, the character itself.
encoded += str.charAt(i);
wsBeforeHexIsNeeded = false;
}
}
return quote + encoded + quote;
}
exports.decode = decode;
exports.encode = encode;