From cfdb1dfe59b33bf7441b8a8e4fc58d75e4f54cee Mon Sep 17 00:00:00 2001
From: Ryan Wilson-Perkin <ryanwilsonperkin@gmail.com>
Date: Fri, 24 Feb 2023 16:17:15 -0500
Subject: [PATCH] Improve performance of hashRegExp lookup

For applications with a very large number of assets, the cost of
invoking a single regular expression with many many values in a group
becomes very high.

By changing to a list of regular expressions (with helper methods for
maintaining the original design) we can get a large performance
improvement.
---
 lib/optimize/RealContentHashPlugin.js | 58 ++++++++++++++++++++-------
 1 file changed, 44 insertions(+), 14 deletions(-)

diff --git a/lib/optimize/RealContentHashPlugin.js b/lib/optimize/RealContentHashPlugin.js
index 7ab9c46fb8b..ba058b753a2 100644
--- a/lib/optimize/RealContentHashPlugin.js
+++ b/lib/optimize/RealContentHashPlugin.js
@@ -178,10 +178,43 @@ class RealContentHashPlugin {
 						}
 					}
 					if (hashToAssets.size === 0) return;
-					const hashRegExp = new RegExp(
-						Array.from(hashToAssets.keys(), quoteMeta).join("|"),
-						"g"
+					const hashRegExps = Array.from(hashToAssets.keys(), quoteMeta).map(
+						hash => new RegExp(hash, "g")
 					);
+
+					/**
+					 * @param {string} str string to be matched against all hashRegExps
+					 * @returns {string[] | null} matches found
+					 */
+					const hashMatch = str => {
+						/** @type {string[]} */
+						const results = [];
+						for (const hashRegExp of hashRegExps) {
+							const matches = str.match(hashRegExp);
+							if (matches) {
+								matches.forEach(match => results.push(match));
+							}
+						}
+						if (results.length) {
+							return results;
+						} else {
+							return null;
+						}
+					};
+
+					/**
+					 * @param {string} str string to be replaced with all hashRegExps
+					 * @param {function(string): string} fn replacement function to use when a hash is found
+					 * @returns {string} replaced content
+					 */
+					const hashReplace = (str, fn) => {
+						let result = str;
+						for (const hashRegExp of hashRegExps) {
+							result = result.replace(hashRegExp, fn);
+						}
+						return result;
+					};
+
 					await Promise.all(
 						assetsWithInfo.map(async asset => {
 							const { name, source, content, hashes } = asset;
@@ -198,7 +231,7 @@ class RealContentHashPlugin {
 								await cacheAnalyse.providePromise(name, etag, () => {
 									const referencedHashes = new Set();
 									let ownHashes = new Set();
-									const inContent = content.match(hashRegExp);
+									const inContent = hashMatch(content);
 									if (inContent) {
 										for (const hash of inContent) {
 											if (hashes.has(hash)) {
@@ -298,7 +331,7 @@ ${referencingAssets
 									identifier,
 									etag,
 									() => {
-										const newContent = asset.content.replace(hashRegExp, hash =>
+										const newContent = hashReplace(asset.content, hash =>
 											hashToNewHash.get(hash)
 										);
 										return new RawSource(newContent);
@@ -323,15 +356,12 @@ ${referencingAssets
 									identifier,
 									etag,
 									() => {
-										const newContent = asset.content.replace(
-											hashRegExp,
-											hash => {
-												if (asset.ownHashes.has(hash)) {
-													return "";
-												}
-												return hashToNewHash.get(hash);
+										const newContent = hashReplace(asset.content, hash => {
+											if (asset.ownHashes.has(hash)) {
+												return "";
 											}
-										);
+											return hashToNewHash.get(hash);
+										});
 										return new RawSource(newContent);
 									}
 								);
@@ -374,7 +404,7 @@ ${referencingAssets
 					await Promise.all(
 						assetsWithInfo.map(async asset => {
 							await computeNewContent(asset);
-							const newName = asset.name.replace(hashRegExp, hash =>
+							const newName = hashReplace(asset.name, hash =>
 								hashToNewHash.get(hash)
 							);