build-microsoft-cdn.ts 2.3 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162
  1. import { task } from './trace';
  2. import { fetchRemoteTextByLine } from './lib/fetch-text-by-line';
  3. import { HostnameSmolTrie } from './lib/trie';
  4. import { SHARED_DESCRIPTION } from './constants/description';
  5. import { createMemoizedPromise } from './lib/memo-promise';
  6. import { extractDomainsFromFelixDnsmasq } from './lib/parse-dnsmasq';
  7. import { RulesetOutput } from './lib/rules/ruleset';
  8. import { appendArrayInPlace } from './lib/append-array-in-place';
  9. const PROBE_DOMAINS = ['.microsoft.com', '.windows.net', '.windows.com', '.windowsupdate.com', '.windowssearch.com', '.office.net'];
  10. const DOMAINS = [
  11. 'res.cdn.office.net',
  12. 'res-1.cdn.office.net',
  13. 'statics.teams.cdn.office.net'
  14. ];
  15. const DOMAIN_SUFFIXES = ['download.prss.microsoft.com'];
  16. const BLACKLIST = [
  17. 'www.microsoft.com',
  18. 'windowsupdate.com'
  19. ];
  20. export const getMicrosoftCdnRulesetPromise = createMemoizedPromise<[domains: string[], domainSuffixes: string[]]>(async () => {
  21. // First trie is to find the microsoft domains that matches probe domains
  22. const trie = new HostnameSmolTrie();
  23. for await (const line of await fetchRemoteTextByLine('https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/accelerated-domains.china.conf')) {
  24. const domain = extractDomainsFromFelixDnsmasq(line);
  25. if (domain) {
  26. trie.add(domain);
  27. }
  28. }
  29. // remove blacklist domain from trie, to prevent them from being included in the later dump
  30. BLACKLIST.forEach(black => trie.whitelist(black));
  31. const domains: string[] = DOMAINS;
  32. const domainSuffixes = appendArrayInPlace(PROBE_DOMAINS.flatMap(domain => trie.find(domain)), DOMAIN_SUFFIXES);
  33. return [domains, domainSuffixes] as const;
  34. });
  35. export const buildMicrosoftCdn = task(require.main === module, __filename)(async (span) => {
  36. const description = [
  37. ...SHARED_DESCRIPTION,
  38. '',
  39. 'This file contains Microsoft\'s domains using their China mainland CDN servers.',
  40. '',
  41. 'Data from:',
  42. ' - https://github.com/felixonmars/dnsmasq-china-list'
  43. ];
  44. const [domains, domainSuffixes] = await span.traceChildPromise('get microsoft cdn domains', getMicrosoftCdnRulesetPromise());
  45. return new RulesetOutput(span, 'microsoft_cdn', 'non_ip')
  46. .withTitle('Sukka\'s Ruleset - Microsoft CDN')
  47. .withDescription(description)
  48. .bulkAddDomain(domains)
  49. .bulkAddDomainSuffix(domainSuffixes)
  50. .write();
  51. });