build-microsoft-cdn.ts 3.3 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071
  1. import { task } from './trace';
  2. import { SHARED_DESCRIPTION } from './constants/description';
  3. import { RulesetOutput } from './lib/rules/ruleset';
  4. import Worktank from 'worktank';
  5. import { RULES } from './constants/microsoft-cdn';
  6. import { wait } from 'foxts/wait';
  7. const pool = new Worktank({
  8. pool: {
  9. name: 'get-microsoft-cdn',
  10. size: 1 // The number of workers to keep in the pool, if more workers are needed they will be spawned up to this limit
  11. },
  12. worker: {
  13. autoAbort: 10000,
  14. 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
  15. autoInstantiate: true,
  16. methods: {
  17. // eslint-disable-next-line object-shorthand -- workertank
  18. getMicrosoftCdnRuleset: async function (__filename: string): Promise<[domains: string[], domainSuffixes: string[]]> {
  19. // TODO: createRequire is a temporary workaround for https://github.com/nodejs/node/issues/51956
  20. const { default: module } = await import('node:module');
  21. const __require = module.createRequire(__filename);
  22. const { HostnameSmolTrie } = __require('./lib/trie');
  23. const { PROBE_DOMAINS, DOMAINS, DOMAIN_SUFFIXES, BLACKLIST } = __require('./constants/microsoft-cdn') as typeof import('./constants/microsoft-cdn');
  24. const { fetchRemoteTextByLine } = __require('./lib/fetch-text-by-line') as typeof import('./lib/fetch-text-by-line');
  25. const { appendArrayInPlace } = __require('foxts/append-array-in-place') as typeof import('foxts/append-array-in-place');
  26. const { extractDomainsFromFelixDnsmasq } = __require('./lib/parse-dnsmasq') as typeof import('./lib/parse-dnsmasq');
  27. const trie = new HostnameSmolTrie();
  28. for await (const line of await fetchRemoteTextByLine('https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/accelerated-domains.china.conf')) {
  29. const domain = extractDomainsFromFelixDnsmasq(line);
  30. if (domain) {
  31. trie.add(domain);
  32. }
  33. }
  34. // remove blacklist domain from trie, to prevent them from being included in the later dump
  35. BLACKLIST.forEach(black => trie.whitelist(black));
  36. const domains: string[] = DOMAINS;
  37. const domainSuffixes = appendArrayInPlace(PROBE_DOMAINS.flatMap(domain => trie.find(domain)), DOMAIN_SUFFIXES);
  38. return [domains, domainSuffixes] as const;
  39. }
  40. }
  41. }
  42. });
  43. export const getMicrosoftCdnRulesetPromise = wait(0).then(() => pool.exec(
  44. 'getMicrosoftCdnRuleset',
  45. [__filename]
  46. )).finally(() => pool.terminate());
  47. export const buildMicrosoftCdn = task(require.main === module, __filename)(async (span) => {
  48. const [domains, domainSuffixes] = await span.traceChildPromise('get microsoft cdn domains', getMicrosoftCdnRulesetPromise);
  49. return new RulesetOutput(span, 'microsoft_cdn', 'non_ip')
  50. .withTitle('Sukka\'s Ruleset - Microsoft CDN')
  51. .appendDescription(SHARED_DESCRIPTION)
  52. .appendDescription(
  53. '',
  54. 'This file contains Microsoft\'s domains using their China mainland CDN servers.'
  55. )
  56. .addFromRuleset(RULES)
  57. .appendDataSource('https://github.com/felixonmars/dnsmasq-china-list')
  58. .bulkAddDomain(domains)
  59. .bulkAddDomainSuffix(domainSuffixes)
  60. .write();
  61. });