refactor: no extra work for CSS unescaping

This commit is contained in:
alexander.akait 2024-11-29 21:39:14 +03:00
parent f4092a6059
commit 8edbc7ce2a
2 changed files with 157 additions and 150 deletions

View File

@ -99,6 +99,136 @@ const normalizeUrl = (str, isString) => {
return str; return str;
}; };
// eslint-disable-next-line no-useless-escape
const regexSingleEscape = /[ -,.\/:-@[\]\^`{-~]/;
const regexExcessiveSpaces =
/(^|\\+)?(\\[A-F0-9]{1,6})\u0020(?![a-fA-F0-9\u0020])/g;
/**
* @param {string} str string
* @returns {string} escaped identifier
*/
const escapeIdentifier = str => {
let output = "";
let counter = 0;
while (counter < str.length) {
const character = str.charAt(counter++);
let value;
if (/[\t\n\f\r\u000B]/.test(character)) {
const codePoint = character.charCodeAt(0);
value = `\\${codePoint.toString(16).toUpperCase()} `;
} else if (character === "\\" || regexSingleEscape.test(character)) {
value = `\\${character}`;
} else {
value = character;
}
output += value;
}
const firstChar = str.charAt(0);
if (/^-[-\d]/.test(output)) {
output = `\\-${output.slice(1)}`;
} else if (/\d/.test(firstChar)) {
output = `\\3${firstChar} ${output.slice(1)}`;
}
// Remove spaces after `\HEX` escapes that are not followed by a hex digit,
// since theyre redundant. Note that this is only possible if the escape
// sequence isnt preceded by an odd number of backslashes.
output = output.replace(regexExcessiveSpaces, ($0, $1, $2) => {
if ($1 && $1.length % 2) {
// Its not safe to remove the space, so dont.
return $0;
}
// Strip the space.
return ($1 || "") + $2;
});
return output;
};
const CONTAINS_ESCAPE = /\\/;
/**
* @param {string} str string
* @returns {[string, number] | undefined} hex
*/
const gobbleHex = str => {
const lower = str.toLowerCase();
let hex = "";
let spaceTerminated = false;
for (let i = 0; i < 6 && lower[i] !== undefined; i++) {
const code = lower.charCodeAt(i);
// check to see if we are dealing with a valid hex char [a-f|0-9]
const valid = (code >= 97 && code <= 102) || (code >= 48 && code <= 57);
// https://drafts.csswg.org/css-syntax/#consume-escaped-code-point
spaceTerminated = code === 32;
if (!valid) break;
hex += lower[i];
}
if (hex.length === 0) return undefined;
const codePoint = Number.parseInt(hex, 16);
const isSurrogate = codePoint >= 0xd800 && codePoint <= 0xdfff;
// Add special case for
// "If this number is zero, or is for a surrogate, or is greater than the maximum allowed code point"
// https://drafts.csswg.org/css-syntax/#maximum-allowed-code-point
if (isSurrogate || codePoint === 0x0000 || codePoint > 0x10ffff) {
return ["\uFFFD", hex.length + (spaceTerminated ? 1 : 0)];
}
return [
String.fromCodePoint(codePoint),
hex.length + (spaceTerminated ? 1 : 0)
];
};
/**
* @param {string} str string
* @returns {string} unescaped string
*/
const unescapeIdentifier = str => {
const needToProcess = CONTAINS_ESCAPE.test(str);
if (!needToProcess) return str;
let ret = "";
for (let i = 0; i < str.length; i++) {
if (str[i] === "\\") {
const gobbled = gobbleHex(str.slice(i + 1, i + 7));
if (gobbled !== undefined) {
ret += gobbled[0];
i += gobbled[1];
continue;
}
// Retain a pair of \\ if double escaped `\\\\`
// https://github.com/postcss/postcss-selector-parser/commit/268c9a7656fb53f543dc620aa5b73a30ec3ff20e
if (str[i + 1] === "\\") {
ret += "\\";
i += 1;
continue;
}
// if \\ is at the end of the string retain it
// https://github.com/postcss/postcss-selector-parser/commit/01a6b346e3612ce1ab20219acc26abdc259ccefb
if (str.length === i + 1) {
ret += str[i];
}
continue;
}
ret += str[i];
}
return ret;
};
class LocConverter { class LocConverter {
/** /**
* @param {string} input input * @param {string} input input
@ -482,7 +612,7 @@ class CssParser extends Parser {
// CSS Variable // CSS Variable
const { line: sl, column: sc } = locConverter.get(propertyNameStart); const { line: sl, column: sc } = locConverter.get(propertyNameStart);
const { line: el, column: ec } = locConverter.get(propertyNameEnd); const { line: el, column: ec } = locConverter.get(propertyNameEnd);
const name = propertyName.slice(2); const name = unescapeIdentifier(propertyName.slice(2));
const dep = new CssLocalIdentifierDependency( const dep = new CssLocalIdentifierDependency(
name, name,
[propertyNameStart, propertyNameEnd], [propertyNameStart, propertyNameEnd],
@ -490,9 +620,7 @@ class CssParser extends Parser {
); );
dep.setLoc(sl, sc, el, ec); dep.setLoc(sl, sc, el, ec);
module.addDependency(dep); module.addDependency(dep);
declaredCssVariables.add( declaredCssVariables.add(name);
CssSelfLocalIdentifierDependency.unescapeIdentifier(name)
);
} else if ( } else if (
OPTIONALLY_VENDOR_PREFIXED_ANIMATION_PROPERTY.test(propertyName) OPTIONALLY_VENDOR_PREFIXED_ANIMATION_PROPERTY.test(propertyName)
) { ) {
@ -507,9 +635,11 @@ class CssParser extends Parser {
if (inAnimationProperty && lastIdentifier) { if (inAnimationProperty && lastIdentifier) {
const { line: sl, column: sc } = locConverter.get(lastIdentifier[0]); const { line: sl, column: sc } = locConverter.get(lastIdentifier[0]);
const { line: el, column: ec } = locConverter.get(lastIdentifier[1]); const { line: el, column: ec } = locConverter.get(lastIdentifier[1]);
const name = lastIdentifier[2] const name = unescapeIdentifier(
lastIdentifier[2]
? input.slice(lastIdentifier[0], lastIdentifier[1]) ? input.slice(lastIdentifier[0], lastIdentifier[1])
: input.slice(lastIdentifier[0] + 1, lastIdentifier[1] - 1); : input.slice(lastIdentifier[0] + 1, lastIdentifier[1] - 1)
);
const dep = new CssSelfLocalIdentifierDependency(name, [ const dep = new CssSelfLocalIdentifierDependency(name, [
lastIdentifier[0], lastIdentifier[0],
lastIdentifier[1] lastIdentifier[1]
@ -882,10 +1012,11 @@ class CssParser extends Parser {
end end
); );
if (!ident) return end; if (!ident) return end;
const name = const name = unescapeIdentifier(
ident[2] === true ident[2] === true
? input.slice(ident[0], ident[1]) ? input.slice(ident[0], ident[1])
: input.slice(ident[0] + 1, ident[1] - 1); : input.slice(ident[0] + 1, ident[1] - 1)
);
const { line: sl, column: sc } = locConverter.get(ident[0]); const { line: sl, column: sc } = locConverter.get(ident[0]);
const { line: el, column: ec } = locConverter.get(ident[1]); const { line: el, column: ec } = locConverter.get(ident[1]);
const dep = new CssLocalIdentifierDependency(name, [ const dep = new CssLocalIdentifierDependency(name, [
@ -900,10 +1031,8 @@ class CssParser extends Parser {
if (!ident) return end; if (!ident) return end;
let name = input.slice(ident[0], ident[1]); let name = input.slice(ident[0], ident[1]);
if (!name.startsWith("--") || name.length < 3) return end; if (!name.startsWith("--") || name.length < 3) return end;
name = name.slice(2); name = unescapeIdentifier(name.slice(2));
declaredCssVariables.add( declaredCssVariables.add(name);
CssSelfLocalIdentifierDependency.unescapeIdentifier(name)
);
const { line: sl, column: sc } = locConverter.get(ident[0]); const { line: sl, column: sc } = locConverter.get(ident[0]);
const { line: el, column: ec } = locConverter.get(ident[1]); const { line: el, column: ec } = locConverter.get(ident[1]);
const dep = new CssLocalIdentifierDependency( const dep = new CssLocalIdentifierDependency(
@ -996,7 +1125,7 @@ class CssParser extends Parser {
end end
); );
if (!ident) return end; if (!ident) return end;
const name = input.slice(ident[0], ident[1]); const name = unescapeIdentifier(input.slice(ident[0], ident[1]));
const dep = new CssLocalIdentifierDependency(name, [ const dep = new CssLocalIdentifierDependency(name, [
ident[0], ident[0],
ident[1] ident[1]
@ -1013,7 +1142,7 @@ class CssParser extends Parser {
hash: (input, start, end, isID) => { hash: (input, start, end, isID) => {
if (isNextRulePrelude && isLocalMode() && isID) { if (isNextRulePrelude && isLocalMode() && isID) {
const valueStart = start + 1; const valueStart = start + 1;
const name = input.slice(valueStart, end); const name = unescapeIdentifier(input.slice(valueStart, end));
const dep = new CssLocalIdentifierDependency(name, [valueStart, end]); const dep = new CssLocalIdentifierDependency(name, [valueStart, end]);
const { line: sl, column: sc } = locConverter.get(start); const { line: sl, column: sc } = locConverter.get(start);
const { line: el, column: ec } = locConverter.get(end); const { line: el, column: ec } = locConverter.get(end);
@ -1258,12 +1387,15 @@ class CssParser extends Parser {
if (name === "var") { if (name === "var") {
const customIdent = walkCssTokens.eatIdentSequence(input, end); const customIdent = walkCssTokens.eatIdentSequence(input, end);
if (!customIdent) return end; if (!customIdent) return end;
const name = input.slice(customIdent[0], customIdent[1]); let name = input.slice(customIdent[0], customIdent[1]);
// A custom property is any property whose name starts with two dashes (U+002D HYPHEN-MINUS), like --foo. // A custom property is any property whose name starts with two dashes (U+002D HYPHEN-MINUS), like --foo.
// The <custom-property-name> production corresponds to this: // The <custom-property-name> production corresponds to this:
// its defined as any <dashed-ident> (a valid identifier that starts with two dashes), // its defined as any <dashed-ident> (a valid identifier that starts with two dashes),
// except -- itself, which is reserved for future use by CSS. // except -- itself, which is reserved for future use by CSS.
if (!name.startsWith("--") || name.length < 3) return end; if (!name.startsWith("--") || name.length < 3) return end;
name = unescapeIdentifier(
input.slice(customIdent[0] + 2, customIdent[1])
);
const afterCustomIdent = walkCssTokens.eatWhitespaceAndComments( const afterCustomIdent = walkCssTokens.eatWhitespaceAndComments(
input, input,
customIdent[1] customIdent[1]
@ -1301,7 +1433,7 @@ class CssParser extends Parser {
} else if (from[2] === false) { } else if (from[2] === false) {
const dep = new CssIcssImportDependency( const dep = new CssIcssImportDependency(
path.slice(1, -1), path.slice(1, -1),
name.slice(2), name,
[customIdent[0], from[1] - 1] [customIdent[0], from[1] - 1]
); );
const { line: sl, column: sc } = locConverter.get( const { line: sl, column: sc } = locConverter.get(
@ -1321,7 +1453,7 @@ class CssParser extends Parser {
customIdent[1] customIdent[1]
); );
const dep = new CssSelfLocalIdentifierDependency( const dep = new CssSelfLocalIdentifierDependency(
name.slice(2), name,
[customIdent[0], customIdent[1]], [customIdent[0], customIdent[1]],
"--", "--",
declaredCssVariables declaredCssVariables
@ -1466,3 +1598,5 @@ class CssParser extends Parser {
} }
module.exports = CssParser; module.exports = CssParser;
module.exports.escapeIdentifier = escapeIdentifier;
module.exports.unescapeIdentifier = unescapeIdentifier;

View File

@ -9,6 +9,7 @@ const { cssExportConvention } = require("../util/conventions");
const createHash = require("../util/createHash"); const createHash = require("../util/createHash");
const { makePathsRelative } = require("../util/identifier"); const { makePathsRelative } = require("../util/identifier");
const makeSerializable = require("../util/makeSerializable"); const makeSerializable = require("../util/makeSerializable");
const memoize = require("../util/memoize");
const NullDependency = require("./NullDependency"); const NullDependency = require("./NullDependency");
/** @typedef {import("webpack-sources").ReplaceSource} ReplaceSource */ /** @typedef {import("webpack-sources").ReplaceSource} ReplaceSource */
@ -30,6 +31,8 @@ const NullDependency = require("./NullDependency");
/** @typedef {import("../util/Hash")} Hash */ /** @typedef {import("../util/Hash")} Hash */
/** @typedef {import("../util/createHash").Algorithm} Algorithm */ /** @typedef {import("../util/createHash").Algorithm} Algorithm */
const getCssParser = memoize(() => require("../css/CssParser"));
/** /**
* @param {string} local css local * @param {string} local css local
* @param {CssModule} module module * @param {CssModule} module module
@ -78,50 +81,6 @@ const getLocalIdent = (local, module, chunkGraph, runtimeTemplate) => {
.replace(/^((-?[0-9])|--)/, "_$1"); .replace(/^((-?[0-9])|--)/, "_$1");
}; };
const CONTAINS_ESCAPE = /\\/;
/**
* @param {string} str string
* @returns {[string, number] | undefined} hex
*/
const gobbleHex = str => {
const lower = str.toLowerCase();
let hex = "";
let spaceTerminated = false;
for (let i = 0; i < 6 && lower[i] !== undefined; i++) {
const code = lower.charCodeAt(i);
// check to see if we are dealing with a valid hex char [a-f|0-9]
const valid = (code >= 97 && code <= 102) || (code >= 48 && code <= 57);
// https://drafts.csswg.org/css-syntax/#consume-escaped-code-point
spaceTerminated = code === 32;
if (!valid) break;
hex += lower[i];
}
if (hex.length === 0) return undefined;
const codePoint = Number.parseInt(hex, 16);
const isSurrogate = codePoint >= 0xd800 && codePoint <= 0xdfff;
// Add special case for
// "If this number is zero, or is for a surrogate, or is greater than the maximum allowed code point"
// https://drafts.csswg.org/css-syntax/#maximum-allowed-code-point
if (isSurrogate || codePoint === 0x0000 || codePoint > 0x10ffff) {
return ["\uFFFD", hex.length + (spaceTerminated ? 1 : 0)];
}
return [
String.fromCodePoint(codePoint),
hex.length + (spaceTerminated ? 1 : 0)
];
};
// eslint-disable-next-line no-useless-escape
const regexSingleEscape = /[ -,.\/:-@[\]\^`{-~]/;
const regexExcessiveSpaces =
/(^|\\+)?(\\[A-F0-9]{1,6})\u0020(?![a-fA-F0-9\u0020])/g;
class CssLocalIdentifierDependency extends NullDependency { class CssLocalIdentifierDependency extends NullDependency {
/** /**
* @param {string} name name * @param {string} name name
@ -130,99 +89,13 @@ class CssLocalIdentifierDependency extends NullDependency {
*/ */
constructor(name, range, prefix = "") { constructor(name, range, prefix = "") {
super(); super();
this.name = CssLocalIdentifierDependency.unescapeIdentifier(name); this.name = name;
this.range = range; this.range = range;
this.prefix = prefix; this.prefix = prefix;
this._conventionNames = undefined; this._conventionNames = undefined;
this._hashUpdate = undefined; this._hashUpdate = undefined;
} }
/**
* @param {string} str string
* @returns {string} unescaped string
*/
static unescapeIdentifier(str) {
const needToProcess = CONTAINS_ESCAPE.test(str);
if (!needToProcess) return str;
let ret = "";
for (let i = 0; i < str.length; i++) {
if (str[i] === "\\") {
const gobbled = gobbleHex(str.slice(i + 1, i + 7));
if (gobbled !== undefined) {
ret += gobbled[0];
i += gobbled[1];
continue;
}
// Retain a pair of \\ if double escaped `\\\\`
// https://github.com/postcss/postcss-selector-parser/commit/268c9a7656fb53f543dc620aa5b73a30ec3ff20e
if (str[i + 1] === "\\") {
ret += "\\";
i += 1;
continue;
}
// if \\ is at the end of the string retain it
// https://github.com/postcss/postcss-selector-parser/commit/01a6b346e3612ce1ab20219acc26abdc259ccefb
if (str.length === i + 1) {
ret += str[i];
}
continue;
}
ret += str[i];
}
return ret;
}
/**
* @param {string} str string
* @returns {string} escaped identifier
*/
static escapeIdentifier(str) {
let output = "";
let counter = 0;
while (counter < str.length) {
const character = str.charAt(counter++);
let value;
if (/[\t\n\f\r\u000B]/.test(character)) {
const codePoint = character.charCodeAt(0);
value = `\\${codePoint.toString(16).toUpperCase()} `;
} else if (character === "\\" || regexSingleEscape.test(character)) {
value = `\\${character}`;
} else {
value = character;
}
output += value;
}
const firstChar = str.charAt(0);
if (/^-[-\d]/.test(output)) {
output = `\\-${output.slice(1)}`;
} else if (/\d/.test(firstChar)) {
output = `\\3${firstChar} ${output.slice(1)}`;
}
// Remove spaces after `\HEX` escapes that are not followed by a hex digit,
// since theyre redundant. Note that this is only possible if the escape
// sequence isnt preceded by an odd number of backslashes.
output = output.replace(regexExcessiveSpaces, ($0, $1, $2) => {
if ($1 && $1.length % 2) {
// Its not safe to remove the space, so dont.
return $0;
}
// Strip the space.
return ($1 || "") + $2;
});
return output;
}
get type() { get type() {
return "css local identifier"; return "css local identifier";
} }
@ -325,7 +198,7 @@ CssLocalIdentifierDependency.Template = class CssLocalIdentifierDependencyTempla
return ( return (
dep.prefix + dep.prefix +
CssLocalIdentifierDependency.escapeIdentifier( getCssParser().escapeIdentifier(
getLocalIdent(local, module, chunkGraph, runtimeTemplate) getLocalIdent(local, module, chunkGraph, runtimeTemplate)
) )
); );