Feat: introduce make-fetch-happen (#44)

Build/build-speedtest-domainset.ts

@@ -3,10 +3,9 @@ import path from 'node:path';
 import { Sema } from 'async-sema';
 import { getHostname } from 'tldts-experimental';
 import { task } from './trace';
-import { fetchWithRetry } from './lib/fetch-retry';
+import { $fetch } from './lib/make-fetch-happen';
 import { SHARED_DESCRIPTION } from './lib/constants';
 import { readFileIntoProcessedArray } from './lib/fetch-text-by-line';
-import { TTL, deserializeArray, fsFetchCache, serializeArray, createCacheKey } from './lib/cache-filesystem';
 
 import { DomainsetOutput } from './lib/create-file';
 import { OUTPUT_SURGE_DIR } from './constants/dir';
@@ -137,20 +136,10 @@ const PREDEFINE_DOMAINS = [
 ];
 
 const s = new Sema(2);
-const cacheKey = createCacheKey(__filename);
 
-const latestTopUserAgentsPromise = fsFetchCache.applyWithHttp304<string[]>(
-  'https://cdn.jsdelivr.net/npm/top-user-agents@latest/src/desktop.json',
-  cacheKey('https://cdn.jsdelivr.net/npm/top-user-agents@latest/src/desktop.json'),
-  async (res) => {
-    const userAgents = await (res.json() as Promise<string[]>);
-    return userAgents.filter(ua => ua.startsWith('Mozilla/5.0 '));
-  },
-  {
-    serializer: serializeArray,
-    deserializer: deserializeArray
-  }
-);
+const latestTopUserAgentsPromise = $fetch('https://cdn.jsdelivr.net/npm/top-user-agents@latest/src/desktop.json')
+  .then(res => res.json())
+  .then((userAgents: string[]) => userAgents.filter(ua => ua.startsWith('Mozilla/5.0 ')));
 
 const querySpeedtestApi = async (keyword: string): Promise<Array<string | null>> => {
   const topUserAgents = await latestTopUserAgentsPromise;
@@ -160,45 +149,34 @@ const querySpeedtestApi = async (keyword: string): Promise<Array<string | null>>
   try {
     const randomUserAgent = topUserAgents[Math.floor(Math.random() * topUserAgents.length)];
 
-    return await fsFetchCache.apply(
-      cacheKey(url),
-      () => s.acquire().then(() => fetchWithRetry(url, {
-        headers: {
-          dnt: '1',
-          Referer: 'https://www.speedtest.net/',
-          accept: 'application/json, text/plain, */*',
-          'User-Agent': randomUserAgent,
-          'Accept-Language': 'en-US,en;q=0.9',
-          ...(randomUserAgent.includes('Chrome')
-            ? {
-              'Sec-Ch-Ua-Mobile': '?0',
-              'Sec-Fetch-Dest': 'empty',
-              'Sec-Fetch-Mode': 'cors',
-              'Sec-Fetch-Site': 'same-origin',
-              'Sec-Gpc': '1'
-            }
-            : {})
-        },
-        signal: AbortSignal.timeout(1000 * 60),
-        retry: {
-          retries: 2
-        }
-      })).then(r => r.json() as any).then((data: Array<{ url: string, host: string }>) => data.reduce<string[]>(
-        (prev, cur) => {
-          const line = cur.host || cur.url;
-          const hn = getHostname(line, { detectIp: false, validateHostname: true });
-          if (hn) {
-            prev.push(hn);
+    return await s.acquire().then(() => $fetch(url, {
+      headers: {
+        dnt: '1',
+        Referer: 'https://www.speedtest.net/',
+        accept: 'application/json, text/plain, */*',
+        'User-Agent': randomUserAgent,
+        'Accept-Language': 'en-US,en;q=0.9',
+        ...(randomUserAgent.includes('Chrome')
+          ? {
+            'Sec-Ch-Ua-Mobile': '?0',
+            'Sec-Fetch-Dest': 'empty',
+            'Sec-Fetch-Mode': 'cors',
+            'Sec-Fetch-Site': 'same-origin',
+            'Sec-Gpc': '1'
           }
-          return prev;
-        }, []
-      )).finally(() => s.release()),
-      {
-        ttl: TTL.ONE_WEEK(),
-        serializer: serializeArray,
-        deserializer: deserializeArray
-      }
-    );
+          : {})
+      },
+      timeout: 1000 * 60
+    })).then(r => r.json() as any).then((data: Array<{ url: string, host: string }>) => data.reduce<string[]>(
+      (prev, cur) => {
+        const line = cur.host || cur.url;
+        const hn = getHostname(line, { detectIp: false, validateHostname: true });
+        if (hn) {
+          prev.push(hn);
+        }
+        return prev;
+      }, []
+    )).finally(() => s.release());
   } catch (e) {
     console.error(e);
     return [];

Build/build-telegram-cidr.ts

@@ -1,5 +1,4 @@
 // @ts-check
-import { defaultRequestInit, fetchWithRetry } from './lib/fetch-retry';
 import { createReadlineInterfaceFromResponse } from './lib/fetch-text-by-line';
 import { isProbablyIpv4, isProbablyIpv6 } from './lib/is-fast-ip';
 import { processLine } from './lib/process-line';
@@ -7,9 +6,10 @@ import { task } from './trace';
 import { SHARED_DESCRIPTION } from './lib/constants';
 import { createMemoizedPromise } from './lib/memo-promise';
 import { RulesetOutput } from './lib/create-file';
+import { $fetch } from './lib/make-fetch-happen';
 
 export const getTelegramCIDRPromise = createMemoizedPromise(async () => {
-  const resp = await fetchWithRetry('https://core.telegram.org/resources/cidr.txt', defaultRequestInit);
+  const resp = await $fetch('https://core.telegram.org/resources/cidr.txt');
   const lastModified = resp.headers.get('last-modified');
   const date = lastModified ? new Date(lastModified) : new Date();
 

Build/download-mock-assets.ts

@@ -1,11 +1,10 @@
 import { task } from './trace';
 import path from 'node:path';
 import fs from 'node:fs';
-import { Readable } from 'node:stream';
 import { pipeline } from 'node:stream/promises';
-import { fetchWithRetry } from './lib/fetch-retry';
 import { OUTPUT_MOCK_DIR } from './constants/dir';
 import { mkdirp } from './lib/misc';
+import { $fetch } from './lib/make-fetch-happen';
 
 const ASSETS_LIST = {
   'www-google-analytics-com_ga.js': 'https://raw.githubusercontent.com/AdguardTeam/Scriptlets/master/dist/redirect-files/google-analytics-ga.js',
@@ -18,7 +17,7 @@ const ASSETS_LIST = {
 export const downloadMockAssets = task(require.main === module, __filename)((span) => Promise.all(Object.entries(ASSETS_LIST).map(
   ([filename, url]) => span
     .traceChildAsync(url, async () => {
-      const res = await fetchWithRetry(url);
+      const res = await $fetch(url);
       if (!res.body) {
         throw new Error(`Empty body from ${url}`);
       }
@@ -27,7 +26,7 @@ export const downloadMockAssets = task(require.main === module, __filename)((spa
       const src = path.join(OUTPUT_MOCK_DIR, filename);
 
       return pipeline(
-        Readable.fromWeb(res.body),
+        res.body,
         fs.createWriteStream(src, 'utf-8')
       );
     })

Build/download-previous-build.ts

@@ -1,13 +1,12 @@
 import path from 'node:path';
 import fs from 'node:fs';
 import { pipeline } from 'node:stream/promises';
+import picocolors from 'picocolors';
 import { task } from './trace';
-import { defaultRequestInit, fetchWithRetry } from './lib/fetch-retry';
 import { extract as tarExtract } from 'tar-fs';
 import type { Headers as TarEntryHeaders } from 'tar-fs';
 import zlib from 'node:zlib';
-import { Readable } from 'node:stream';
-import picocolors from 'picocolors';
+import { $fetch } from './lib/make-fetch-happen';
 
 const GITHUB_CODELOAD_URL = 'https://codeload.github.com/sukkalab/ruleset.skk.moe/tar.gz/master';
 const GITLAB_CODELOAD_URL = 'https://gitlab.com/SukkaW/ruleset.skk.moe/-/archive/master/ruleset.skk.moe-master.tar.gz';
@@ -21,13 +20,7 @@ export const downloadPreviousBuild = task(require.main === module, __filename)(a
   }
 
   const tarGzUrl = await span.traceChildAsync('get tar.gz url', async () => {
-    const resp = await fetchWithRetry(GITHUB_CODELOAD_URL, {
-      ...defaultRequestInit,
-      method: 'HEAD',
-      retry: {
-        retryOnNon2xx: false
-      }
-    });
+    const resp = await $fetch(GITHUB_CODELOAD_URL, { method: 'HEAD' });
     if (resp.status !== 200) {
       console.warn('Download previous build from GitHub failed! Status:', resp.status);
       console.warn('Switch to GitLab');
@@ -37,19 +30,12 @@ export const downloadPreviousBuild = task(require.main === module, __filename)(a
   });
 
   return span.traceChildAsync('download & extract previoud build', async () => {
-    const resp = await fetchWithRetry(tarGzUrl, {
+    const resp = await $fetch(tarGzUrl, {
       headers: {
         'User-Agent': 'curl/8.9.1',
         // https://github.com/unjs/giget/issues/97
         // https://gitlab.com/gitlab-org/gitlab/-/commit/50c11f278d18fe1f3fb12eb595067216bb58ade2
         'sec-fetch-mode': 'same-origin'
-      },
-      // https://github.com/unjs/giget/issues/97
-      // https://gitlab.com/gitlab-org/gitlab/-/commit/50c11f278d18fe1f3fb12eb595067216bb58ade2
-
-      mode: 'same-origin',
-      retry: {
-        retryOnNon2xx: false
       }
     });
 
@@ -79,7 +65,7 @@ export const downloadPreviousBuild = task(require.main === module, __filename)(a
     );
 
     return pipeline(
-      Readable.fromWeb(resp.body),
+      resp.body,
       gunzip,
       extract
     );

Build/lib/convert-clash-meta-mrs.ts

@@ -1,14 +1,13 @@
 import path from 'node:path';
 import fs from 'node:fs';
 import fsp from 'node:fs/promises';
-import { Readable } from 'node:stream';
 import { pipeline } from 'node:stream/promises';
 import zlib from 'node:zlib';
 import process from 'node:process';
 
 import { async as ezspawn } from '@jsdevtools/ez-spawn';
 import { mkdirp } from './misc';
-import { fetchWithRetry } from './fetch-retry';
+import { $fetch } from './make-fetch-happen';
 
 const mihomoBinaryDir = path.join(__dirname, '../../node_modules/.cache/mihomo');
 const mihomoBinaryPath = path.join(mihomoBinaryDir, 'mihomo');
@@ -33,7 +32,7 @@ const ensureMihomoBinary = async () => {
       throw new Error(`Unsupported platform: ${process.platform} ${process.arch}`);
     }
 
-    const res = await fetchWithRetry(downloadUrl);
+    const res = await $fetch(downloadUrl);
 
     if (!res.ok || !res.body) {
       throw new Error(`Failed to download mihomo binary: ${res.statusText}`);
@@ -42,7 +41,7 @@ const ensureMihomoBinary = async () => {
     const gunzip = zlib.createGunzip();
 
     await pipeline(
-      Readable.fromWeb(res.body),
+      res.body,
       gunzip,
       writeStream
     );

Build/lib/fetch-text-by-line.ts

@@ -1,12 +1,13 @@
 import fs from 'node:fs';
 import { Readable } from 'node:stream';
-import { fetchWithRetry, defaultRequestInit } from './fetch-retry';
 import type { FileHandle } from 'node:fs/promises';
 
 import { TextLineStream } from './text-line-transform-stream';
 import type { ReadableStream } from 'node:stream/web';
 import { TextDecoderStream } from 'node:stream/web';
 import { processLine } from './process-line';
+import { $fetch } from './make-fetch-happen';
+import type { NodeFetchResponse } from './make-fetch-happen';
 
 const getReadableStream = (file: string | FileHandle): ReadableStream => {
   if (typeof file === 'string') {
@@ -20,7 +21,7 @@ export const readFileByLine: ((file: string | FileHandle) => AsyncIterable<strin
   .pipeThrough(new TextDecoderStream())
   .pipeThrough(new TextLineStream());
 
-const ensureResponseBody = (resp: Response) => {
+const ensureResponseBody = <T extends Response | NodeFetchResponse>(resp: T): NonNullable<T['body']> => {
   if (!resp.body) {
     throw new Error('Failed to fetch remote text');
   }
@@ -30,12 +31,20 @@ const ensureResponseBody = (resp: Response) => {
   return resp.body;
 };
 
-export const createReadlineInterfaceFromResponse: ((resp: Response) => AsyncIterable<string>) = (resp) => ensureResponseBody(resp)
-  .pipeThrough(new TextDecoderStream())
-  .pipeThrough(new TextLineStream());
+export const createReadlineInterfaceFromResponse: ((resp: Response | NodeFetchResponse) => AsyncIterable<string>) = (resp) => {
+  const stream = ensureResponseBody(resp);
 
-export function fetchRemoteTextByLine(url: string | URL) {
-  return fetchWithRetry(url, defaultRequestInit).then(createReadlineInterfaceFromResponse);
+  const webStream: ReadableStream<Uint8Array> = 'getReader' in stream
+    ? stream
+    : Readable.toWeb(new Readable().wrap(stream)) as any;
+
+  return webStream
+    .pipeThrough(new TextDecoderStream())
+    .pipeThrough(new TextLineStream());
+};
+
+export function fetchRemoteTextByLine(url: string) {
+  return $fetch(url).then(createReadlineInterfaceFromResponse);
 }
 
 export async function readFileIntoProcessedArray(file: string | FileHandle) {

Build/lib/make-fetch-happen.ts

@@ -0,0 +1,23 @@
+import path from 'node:path';
+import fs from 'node:fs';
+import makeFetchHappen from 'make-fetch-happen';
+// eslint-disable-next-line @typescript-eslint/no-restricted-imports -- type only
+export type { Response as NodeFetchResponse } from 'node-fetch';
+
+const cachePath = path.resolve(__dirname, '../../.cache/__make_fetch_happen__');
+fs.mkdirSync(cachePath, { recursive: true });
+
+export const $fetch = makeFetchHappen.defaults({
+  cachePath,
+  maxSockets: 32, /**
+   * They said 15 is a good default that prevents knocking out others' routers,
+   * I disagree. 32 is a good number.
+   */
+  headers: {
+    'User-Agent': 'curl/8.9.1 (https://github.com/SukkaW/Surge)'
+  },
+  retry: {
+    retries: 5,
+    randomize: true
+  }
+});

Build/lib/parse-dnsmasq.ts

@@ -1,6 +1,7 @@
 import { createReadlineInterfaceFromResponse } from './fetch-text-by-line';
 import { parse as tldtsParse } from 'tldts';
-import { fetchWithRetry, defaultRequestInit } from './fetch-retry';
+import { $fetch } from './make-fetch-happen';
+import type { NodeFetchResponse } from './make-fetch-happen';
 
 const isDomainLoose = (domain: string): boolean => {
   const { isIcann, isPrivate, isIp } = tldtsParse(domain);
@@ -14,7 +15,7 @@ export const extractDomainsFromFelixDnsmasq = (line: string): string | null => {
   return null;
 };
 
-export const parseFelixDnsmasqFromResp = async (resp: Response): Promise<string[]> => {
+export const parseFelixDnsmasqFromResp = async (resp: Response | NodeFetchResponse): Promise<string[]> => {
   const results: string[] = [];
 
   for await (const line of createReadlineInterfaceFromResponse(resp)) {
@@ -27,7 +28,7 @@ export const parseFelixDnsmasqFromResp = async (resp: Response): Promise<string[
   return results;
 };
 
-export const parseFelixDnsmasq = async (url: string | URL): Promise<string[]> => {
-  const resp = await fetchWithRetry(url, defaultRequestInit);
+export const parseFelixDnsmasq = async (url: string): Promise<string[]> => {
+  const resp = await $fetch(url);
   return parseFelixDnsmasqFromResp(resp);
 };

Build/validate-domestic.ts

@@ -5,14 +5,14 @@ import path from 'node:path';
 import { processLine } from './lib/process-line';
 import { parseFelixDnsmasq } from './lib/parse-dnsmasq';
 import { SOURCE_DIR } from './constants/dir';
-import { fetchWithRetry } from './lib/fetch-retry';
+import { $fetch } from './lib/make-fetch-happen';
 
 export const parseDomesticList = async () => {
   const trie = createTrie(await parseFelixDnsmasq('https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/accelerated-domains.china.conf'));
 
   const top5000 = new Set<string>();
 
-  const res = await (await fetchWithRetry('https://radar.cloudflare.com/charts/LargerTopDomainsTable/attachment?id=1077&top=10000', {
+  const res = await (await $fetch('https://radar.cloudflare.com/charts/LargerTopDomainsTable/attachment?id=1077&top=10000', {
     headers: {
       accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7',
       'accept-language': 'en-US,en;q=0.9,zh-CN;q=0.8,zh;q=0.7,zh-TW;q=0.6,es;q=0.5',

Build/validate-gfwlist.ts

@@ -6,13 +6,13 @@ import { parse } from 'csv-parse/sync';
 import { readFileByLine } from './lib/fetch-text-by-line';
 import path from 'node:path';
 import { SOURCE_DIR } from './constants/dir';
-import { fetchWithRetry } from './lib/fetch-retry';
+import { $fetch } from './lib/make-fetch-happen';
 
 export const parseGfwList = async () => {
   const whiteSet = new Set<string>();
   const blackSet = new Set<string>();
 
-  const text = await (await fetchWithRetry('https://raw.githubusercontent.com/gfwlist/gfwlist/master/gfwlist.txt')).text();
+  const text = await (await $fetch('https://raw.githubusercontent.com/gfwlist/gfwlist/master/gfwlist.txt')).text();
   for (const l of atob(text).split('\n')) {
     const line = processLine(l);
     if (!line) continue;
@@ -55,13 +55,13 @@ export const parseGfwList = async () => {
       continue;
     }
   }
-  for (const l of (await (await fetchWithRetry('https://raw.githubusercontent.com/Loyalsoldier/cn-blocked-domain/release/domains.txt')).text()).split('\n')) {
+  for (const l of (await (await $fetch('https://raw.githubusercontent.com/Loyalsoldier/cn-blocked-domain/release/domains.txt')).text()).split('\n')) {
     blackSet.add(l);
   }
 
   const top500Gfwed = new Set<string>();
 
-  const res = await (await fetchWithRetry('https://radar.cloudflare.com/charts/LargerTopDomainsTable/attachment?id=1077&top=10000', {
+  const res = await (await $fetch('https://radar.cloudflare.com/charts/LargerTopDomainsTable/attachment?id=1077&top=10000', {
     headers: {
       accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7',
       'accept-language': 'en-US,en;q=0.9,zh-CN;q=0.8,zh;q=0.7,zh-TW;q=0.6,es;q=0.5',

package.json

@@ -32,6 +32,7 @@
     "fdir": "^6.4.0",
     "foxact": "^0.2.38",
     "json-stringify-pretty-compact": "^3.0.0",
+    "make-fetch-happen": "^14.0.1",
     "mnemonist": "^0.39.8",
     "picocolors": "^1.1.0",
     "punycode": "^2.3.1",
@@ -48,7 +49,9 @@
     "@types/async-retry": "^1.4.9",
     "@types/better-sqlite3": "^7.6.11",
     "@types/chai": "^4.3.20",
+    "@types/make-fetch-happen": "^10.0.4",
     "@types/mocha": "^10.0.9",
+    "@types/node-fetch": "2",
     "@types/punycode": "^2.1.4",
     "@types/tar-fs": "^2.0.4",
     "@types/tar-stream": "^3.1.3",