From 66b0a3697dc4f026960d53daa888cddd9913a242 Mon Sep 17 00:00:00 2001
From: SukkaW <isukkaw@gmail.com>
Date: Sat, 12 Oct 2024 12:37:45 +0800
Subject: [PATCH] Chore: a hash collision test

---
 Build/lib/string-hash.ts              | 11 +++++++
 Build/validate-hash-collision-test.ts | 46 +++++++++++++++++++++++++++
 package.json                          |  1 +
 pnpm-lock.yaml                        |  8 +++++
 4 files changed, 66 insertions(+)
 create mode 100644 Build/validate-hash-collision-test.ts

diff --git a/Build/lib/string-hash.ts b/Build/lib/string-hash.ts
index 8084201e..355856d3 100644
--- a/Build/lib/string-hash.ts
+++ b/Build/lib/string-hash.ts
@@ -43,4 +43,15 @@ export function fnv1a52(str: string) {
   );
 }
 
+export function fnv1a(s: string) {
+  let h = 0x81_1C_9D_C5;
+
+  for (let i = 0, l = s.length; i < l; i++) {
+    h ^= s.charCodeAt(i);
+    h += (h << 1) + (h << 4) + (h << 7) + (h << 8) + (h << 24);
+  }
+
+  return (h >>> 0);
+}
+
 export const stringHash = (payload: string) => fnv1a52(payload).toString(36) + payload.length.toString(36);
diff --git a/Build/validate-hash-collision-test.ts b/Build/validate-hash-collision-test.ts
new file mode 100644
index 00000000..fbadb1de
--- /dev/null
+++ b/Build/validate-hash-collision-test.ts
@@ -0,0 +1,46 @@
+/* eslint-disable no-await-in-loop -- no concurrent */
+import { fdir as Fdir } from 'fdir';
+import { OUTPUT_SURGE_DIR } from './constants/dir';
+import path from 'node:path';
+import { readFileIntoProcessedArray } from './lib/fetch-text-by-line';
+import { xxhash3 } from 'hash-wasm';
+
+(async () => {
+  const hashMap = new Map<string, Set<string>>();
+
+  const runHash = async (inputs: string[]) => {
+    for (const input of inputs) {
+      const hash = await xxhash3(input);
+      if (!hashMap.has(hash)) {
+        hashMap.set(hash, new Set());
+      }
+      hashMap.get(hash)!.add(input);
+    }
+  };
+
+  const files = await new Fdir()
+    .withRelativePaths()
+    .crawl(OUTPUT_SURGE_DIR)
+    .withPromise();
+
+  for (const file of files) {
+    const fullpath = path.join(OUTPUT_SURGE_DIR, file);
+    if (file.startsWith('domainset' + path.sep)) {
+      await runHash((await readFileIntoProcessedArray(fullpath)).map(i => (i[0] === '.' ? i.slice(1) : i)));
+    } else if (file.startsWith('non_ip' + path.sep)) {
+      await runHash((await readFileIntoProcessedArray(fullpath)).map(i => i.split(',')[1]));
+    }
+  }
+
+  console.log(hashMap.size);
+  let collision = 0;
+  hashMap.forEach((v, k) => {
+    if (v.size > 1) {
+      collision++;
+      console.log(k, '=>', v);
+    }
+  });
+  if (collision === 0) {
+    console.log(hashMap);
+  }
+})();
diff --git a/package.json b/package.json
index 4ac8ba3b..e3a25a4e 100644
--- a/package.json
+++ b/package.json
@@ -31,6 +31,7 @@
     "fast-cidr-tools": "^0.3.1",
     "fdir": "^6.4.0",
     "foxact": "^0.2.38",
+    "hash-wasm": "^4.11.0",
     "json-stringify-pretty-compact": "^3.0.0",
     "make-fetch-happen": "^14.0.1",
     "mnemonist": "^0.39.8",
diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index 171d9783..219b0dbf 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -44,6 +44,9 @@ importers:
       foxact:
         specifier: ^0.2.38
         version: 0.2.38
+      hash-wasm:
+        specifier: ^4.11.0
+        version: 4.11.0
       json-stringify-pretty-compact:
         specifier: ^3.0.0
         version: 3.0.0
@@ -1162,6 +1165,9 @@ packages:
     resolution: {integrity: sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==}
     engines: {node: '>=8'}
 
+  hash-wasm@4.11.0:
+    resolution: {integrity: sha512-HVusNXlVqHe0fzIzdQOGolnFN6mX/fqcrSAOcTBXdvzrXVHwTz11vXeKRmkR5gTuwVpvHZEIyKoePDvuAR+XwQ==}
+
   hasown@2.0.2:
     resolution: {integrity: sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==}
     engines: {node: '>= 0.4'}
@@ -2983,6 +2989,8 @@ snapshots:
 
   has-flag@4.0.0: {}
 
+  hash-wasm@4.11.0: {}
+
   hasown@2.0.2:
     dependencies:
       function-bind: 1.1.2