build-microsoft-cdn.ts 3.3 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374
  1. import { task } from './trace';
  2. import { SHARED_DESCRIPTION } from './constants/description';
  3. import { once } from 'foxts/once';
  4. import { RulesetOutput } from './lib/rules/ruleset';
  5. import Worktank from 'worktank';
  6. const pool = new Worktank({
  7. pool: {
  8. name: 'build-internal-reverse-chn-cidr',
  9. size: 1 // The number of workers to keep in the pool, if more workers are needed they will be spawned up to this limit
  10. },
  11. worker: {
  12. autoAbort: 10000,
  13. autoTerminate: 30000, // The interval of milliseconds at which to check if the pool can be automatically terminated, to free up resources, workers will be spawned up again if needed
  14. autoInstantiate: true,
  15. methods: {
  16. // eslint-disable-next-line object-shorthand -- workertank
  17. getMicrosoftCdnRuleset: async function (importMetaUrl: string): Promise<[domains: string[], domainSuffixes: string[]]> {
  18. // TODO: createRequire is a temporary workaround for https://github.com/nodejs/node/issues/51956
  19. const { default: module } = await import('node:module');
  20. const __require = module.createRequire(importMetaUrl);
  21. const { HostnameSmolTrie } = __require('./lib/trie');
  22. const { PROBE_DOMAINS, DOMAINS, DOMAIN_SUFFIXES, BLACKLIST } = __require('./constants/microsoft-cdn') as typeof import('./constants/microsoft-cdn');
  23. const { fetchRemoteTextByLine } = __require('./lib/fetch-text-by-line') as typeof import('./lib/fetch-text-by-line');
  24. const { appendArrayInPlace } = __require('foxts/append-array-in-place') as typeof import('foxts/append-array-in-place');
  25. const { extractDomainsFromFelixDnsmasq } = __require('./lib/parse-dnsmasq') as typeof import('./lib/parse-dnsmasq');
  26. const trie = new HostnameSmolTrie();
  27. for await (const line of await fetchRemoteTextByLine('https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/accelerated-domains.china.conf')) {
  28. const domain = extractDomainsFromFelixDnsmasq(line);
  29. if (domain) {
  30. trie.add(domain);
  31. }
  32. }
  33. // remove blacklist domain from trie, to prevent them from being included in the later dump
  34. BLACKLIST.forEach(black => trie.whitelist(black));
  35. const domains: string[] = DOMAINS;
  36. const domainSuffixes = appendArrayInPlace(PROBE_DOMAINS.flatMap(domain => trie.find(domain)), DOMAIN_SUFFIXES);
  37. return [domains, domainSuffixes] as const;
  38. }
  39. }
  40. }
  41. });
  42. export const getMicrosoftCdnRulesetPromise = once<Promise<[domains: string[], domainSuffixes: string[]]>>(async () => {
  43. const res = await pool.exec(
  44. 'getMicrosoftCdnRuleset',
  45. [import.meta.url]
  46. );
  47. pool.terminate();
  48. return res;
  49. });
  50. export const buildMicrosoftCdn = task(require.main === module, __filename)(async (span) => {
  51. const [domains, domainSuffixes] = await span.traceChildPromise('get microsoft cdn domains', getMicrosoftCdnRulesetPromise());
  52. return new RulesetOutput(span, 'microsoft_cdn', 'non_ip')
  53. .withTitle('Sukka\'s Ruleset - Microsoft CDN')
  54. .appendDescription(SHARED_DESCRIPTION)
  55. .appendDescription(
  56. '',
  57. 'This file contains Microsoft\'s domains using their China mainland CDN servers.'
  58. )
  59. .appendDataSource('https://github.com/felixonmars/dnsmasq-china-list')
  60. .bulkAddDomain(domains)
  61. .bulkAddDomainSuffix(domainSuffixes)
  62. .write();
  63. });