@noodleseed/one 0.93.0 → 0.94.0

package/node_modules/tldts-core/src/factory.ts

@@ -0,0 +1,199 @@
+/**
+ * Implement a factory allowing to plug different implementations of suffix
+ * lookup (e.g.: using a trie or the packed hashes datastructures). This is used
+ * and exposed in `tldts.ts` and `tldts-experimental.ts` bundle entrypoints.
+ */
+
+import getDomain from './domain';
+import getDomainWithoutSuffix from './domain-without-suffix';
+import extractHostname, {
+  extractedHostnameValidated,
+} from './extract-hostname';
+import isIp from './is-ip';
+import isSpecialUse from './is-special-use';
+import isValidHostname from './is-valid';
+import { IPublicSuffix, ISuffixLookupOptions } from './lookup/interface';
+import { IOptions, setDefaults } from './options';
+import getSubdomain from './subdomain';
+
+export interface IResult {
+  // `hostname` is either a registered name (including but not limited to a
+  // hostname), or an IP address, directly extracted from the input URL. IPv4
+  // addresses are in dot-decimal notation. IPv6 is returned without its
+  // surrounding brackets; both bracketed (in URLs, e.g. `http://[::1]/`) and
+  // bare unbracketed (e.g. `2a01:e35::1`) IPv6 literals are accepted.
+  hostname: string | null;
+
+  // Is `hostname` an IP? (IPv4 or IPv6)
+  isIp: boolean | null;
+
+  // `hostname` split between subdomain, domain and its public suffix (if any)
+  subdomain: string | null;
+  domain: string | null;
+  publicSuffix: string | null;
+  domainWithoutSuffix: string | null;
+
+  // Specifies if `publicSuffix` comes from the ICANN or PRIVATE section of the list
+  isIcann: boolean | null;
+  isPrivate: boolean | null;
+
+  // Is `hostname` a special-use domain from the IANA registry (RFC 6761 et al.:
+  // e.g. `localhost`, `*.test`, `*.local`, `*.onion`, `home.arpa`)? `isIcann`/
+  // `isPrivate` do not identify these (most are not in the Public Suffix List;
+  // the few that are appear as ordinary ICANN suffixes). `null` unless the
+  // `detectSpecialUse` option is enabled (see is-special-use.ts).
+  isSpecialUse: boolean | null;
+}
+
+export function getEmptyResult(): IResult {
+  return {
+    domain: null,
+    domainWithoutSuffix: null,
+    hostname: null,
+    isIcann: null,
+    isIp: null,
+    isPrivate: null,
+    isSpecialUse: null,
+    publicSuffix: null,
+    subdomain: null,
+  };
+}
+
+export function resetResult(result: IResult): void {
+  result.domain = null;
+  result.domainWithoutSuffix = null;
+  result.hostname = null;
+  result.isIcann = null;
+  result.isIp = null;
+  result.isPrivate = null;
+  result.isSpecialUse = null;
+  result.publicSuffix = null;
+  result.subdomain = null;
+}
+
+// Flags representing steps in the `parse` function. They are used to implement
+// an early stop mechanism (simulating some form of laziness) to avoid doing
+// more work than necessary to perform a given action (e.g.: we don't need to
+// extract the domain and subdomain if we are only interested in public suffix).
+export const enum FLAG {
+  HOSTNAME,
+  IS_VALID,
+  PUBLIC_SUFFIX,
+  DOMAIN,
+  SUB_DOMAIN,
+  ALL,
+}
+
+export function parseImpl(
+  url: string,
+  step: FLAG,
+  suffixLookup: (
+    _1: string,
+    _2: ISuffixLookupOptions,
+    _3: IPublicSuffix,
+  ) => void,
+  partialOptions: Partial<IOptions> | undefined,
+  result: IResult,
+): IResult {
+  const options: IOptions = /*@__INLINE__*/ setDefaults(partialOptions);
+
+  // Very fast approximate check to make sure `url` is a string. This is needed
+  // because the library will not necessarily be used in a typed setup and
+  // values of arbitrary types might be given as argument.
+  if (typeof url !== 'string') {
+    return result;
+  }
+
+  // Extract hostname from `url` only if needed. This can be made optional
+  // using `options.extractHostname`. This option will typically be used
+  // whenever we are sure the inputs to `parse` are already hostnames and not
+  // arbitrary URLs.
+  //
+  // `mixedInput` allows to specify if we expect a mix of URLs and hostnames
+  // as input. If only hostnames are expected then `extractHostname` can be
+  // set to `false` to speed-up parsing. If only URLs are expected then
+  // `mixedInputs` can be set to `false`. The `mixedInputs` is only a hint
+  // and will not change the behavior of the library.
+  // Whether `url` itself was already a valid hostname (only computed on the
+  // mixedInputs path). Lets us skip the post-extraction validation below when
+  // extractHostname returned `url` unchanged (same reference).
+  let urlIsValid = false;
+  if (!options.extractHostname) {
+    result.hostname = url;
+  } else if (options.mixedInputs) {
+    urlIsValid = isValidHostname(url);
+    result.hostname = extractHostname(
+      url,
+      urlIsValid,
+      options.validateHostname,
+    );
+  } else {
+    result.hostname = extractHostname(url, false, options.validateHostname);
+  }
+
+  // Check if `hostname` is a valid ip address
+  if (options.detectIp && result.hostname !== null) {
+    result.isIp = isIp(result.hostname);
+    if (result.isIp) {
+      return result;
+    }
+  }
+
+  // Perform hostname validation if enabled. If hostname is not valid, no need to
+  // go further as there will be no valid domain or sub-domain. This validation
+  // is applied before any early returns to ensure consistent behavior across
+  // all API methods including getHostname().
+  if (
+    options.validateHostname &&
+    options.extractHostname &&
+    result.hostname !== null &&
+    // Skip the re-scan when `url` was already validated and extractHostname
+    // returned it unchanged (same reference => identical string, still valid).
+    !(urlIsValid && result.hostname === url) &&
+    // Skip the re-scan when extractHostname already validated the host inline
+    // (a confirmed-valid simple authority — see extract-hostname.ts).
+    !extractedHostnameValidated &&
+    !isValidHostname(result.hostname)
+  ) {
+    result.hostname = null;
+    return result;
+  }
+
+  if (step === FLAG.HOSTNAME || result.hostname === null) {
+    return result;
+  }
+
+  // Flag special-use domains, only when opted in (`detectSpecialUse`) and only
+  // for the full `parse()` result (FLAG.ALL). Computed here, before the
+  // public-suffix/domain early-returns below, so single-label names like
+  // `localhost` (which have no registrable domain) are still flagged.
+  if (step === FLAG.ALL && options.detectSpecialUse) {
+    result.isSpecialUse = isSpecialUse(result.hostname);
+  }
+
+  // Extract public suffix
+  suffixLookup(result.hostname, options, result);
+  if (step === FLAG.PUBLIC_SUFFIX || result.publicSuffix === null) {
+    return result;
+  }
+
+  // Extract domain
+  result.domain = getDomain(result.publicSuffix, result.hostname, options);
+  if (step === FLAG.DOMAIN || result.domain === null) {
+    return result;
+  }
+
+  // Extract subdomain
+  result.subdomain = getSubdomain(result.hostname, result.domain);
+  if (step === FLAG.SUB_DOMAIN) {
+    return result;
+  }
+
+  // Extract domain without suffix
+  result.domainWithoutSuffix = getDomainWithoutSuffix(
+    result.domain,
+    result.publicSuffix,
+  );
+
+  return result;
+}

package/node_modules/tldts-core/src/is-ip.ts

@@ -0,0 +1,87 @@
+/**
+ * Check if a hostname is an IP. You should be aware that this only works
+ * because `hostname` is already garanteed to be a valid hostname!
+ */
+function isProbablyIpv4(hostname: string): boolean {
+  // Cannot be shorted than 1.1.1.1
+  if (hostname.length < 7) {
+    return false;
+  }
+
+  // Cannot be longer than: 255.255.255.255
+  if (hostname.length > 15) {
+    return false;
+  }
+
+  let numberOfDots = 0;
+
+  for (let i = 0; i < hostname.length; i += 1) {
+    const code = hostname.charCodeAt(i);
+
+    if (code === 46 /* '.' */) {
+      numberOfDots += 1;
+    } else if (code < 48 /* '0' */ || code > 57 /* '9' */) {
+      return false;
+    }
+  }
+
+  return (
+    numberOfDots === 3 &&
+    hostname.charCodeAt(0) !== 46 /* '.' */ &&
+    hostname.charCodeAt(hostname.length - 1) !== 46 /* '.' */
+  );
+}
+
+/**
+ * Similar to isProbablyIpv4.
+ */
+function isProbablyIpv6(hostname: string): boolean {
+  if (hostname.length < 3) {
+    return false;
+  }
+
+  let start = hostname.startsWith('[') ? 1 : 0;
+  let end = hostname.length;
+
+  if (hostname[end - 1] === ']') {
+    end -= 1;
+  }
+
+  // We only consider the maximum size of a normal IPV6. Note that this will
+  // fail on so-called "IPv4 mapped IPv6 addresses" but this is a corner-case
+  // and a proper validation library should be used for these.
+  if (end - start > 39) {
+    return false;
+  }
+
+  let hasColon = false;
+
+  for (; start < end; start += 1) {
+    const code = hostname.charCodeAt(start);
+
+    if (code === 58 /* ':' */) {
+      hasColon = true;
+    } else if (
+      !(
+        (
+          (code >= 48 && code <= 57) || // 0-9
+          (code >= 97 && code <= 102) || // a-f
+          (code >= 65 && code <= 70)
+        ) // A-F (RFC 4291 §2.2: an IPv6 hextet is hex digits only)
+      )
+    ) {
+      return false;
+    }
+  }
+
+  return hasColon;
+}
+
+/**
+ * Check if `hostname` is *probably* a valid ip addr (either ipv6 or ipv4).
+ * This *will not* work on any string. We need `hostname` to be a valid
+ * hostname.
+ */
+export default function isIp(hostname: string): boolean {
+  return isProbablyIpv6(hostname) || isProbablyIpv4(hostname);
+}

package/node_modules/tldts-core/src/is-special-use.ts

@@ -0,0 +1,65 @@
+/**
+ * Special-use domain names from the IANA "Special-Use Domain Names" registry:
+ * the authoritative list, created by RFC 6761 and maintained as new RFCs add to
+ * it: https://www.iana.org/assignments/special-use-domain-names/
+ * Snapshot: 2026-05-24. (RFC 6761 is not obsoleted; draft-hoffman-rfc6761bis
+ * proposes to retire its prose but keep this registry, so the registry is the
+ * source of truth; re-sync this list against it.)
+ *
+ * These names never correspond to a public registration, yet neither
+ * `isIcann` nor `isPrivate` marks one as special-use: most are absent from the
+ * Public Suffix List (so `a.test` looks like a registrable domain), and the
+ * few that are listed (`onion`, `home.arpa`) appear there as ordinary ICANN
+ * suffixes. `isSpecialUse` is the single signal that covers them all.
+ *
+ * Per the registry and RFC 6761 ("and any names falling within these domains"),
+ * the designation covers each listed name AND all of its sub-domains. DNS labels
+ * are case-insensitive (RFC 4343); `hostname` is expected to be already
+ * lower-cased and trailing-dot-stripped, as produced by `extractHostname`, the
+ * same normalization the Public-Suffix-List lookup relies on.
+ *
+ * Two groups of registry entries are intentionally excluded: the numeric
+ * reverse-DNS delegation zones (`10.in-addr.arpa`, the `*.ip6.arpa` ranges, …),
+ * which are reverse-DNS PTR zones rather than hostnames and whose parents
+ * (`in-addr.arpa`/`ip6.arpa`) are already in the Public Suffix List; and the
+ * deprecated `eap-noob.arpa` entry.
+ */
+const SPECIAL_USE_DOMAINS: readonly string[] = [
+  'test', // RFC 6761
+  'localhost', // RFC 6761
+  'invalid', // RFC 6761
+  'example', // RFC 6761
+  'example.com', // RFC 6761
+  'example.net', // RFC 6761
+  'example.org', // RFC 6761
+  'local', // RFC 6762 (mDNS)
+  'onion', // RFC 7686 (Tor)
+  'alt', // RFC 9476
+  'home.arpa', // RFC 8375
+  'ipv4only.arpa', // RFC 8880
+  'resolver.arpa', // RFC 9462
+  'service.arpa', // RFC 9665
+  '6tisch.arpa', // RFC 9031
+  'eap.arpa', // RFC 9965
+];
+
+/**
+ * Return `true` if `hostname` is, or is a sub-domain of, a special-use domain
+ * (see the registry note above). Expects an already-normalized `hostname`.
+ */
+export default function isSpecialUse(hostname: string): boolean {
+  for (const name of SPECIAL_USE_DOMAINS) {
+    // Match on a label boundary: `hostname` is either exactly `name` or ends
+    // with `.name` (so `latest` is not matched by `test`, nor `myexample.com`
+    // by `example.com`).
+    if (
+      hostname.endsWith(name) &&
+      (hostname.length === name.length ||
+        hostname.charCodeAt(hostname.length - name.length - 1) === 46) /* '.' */
+    ) {
+      return true;
+    }
+  }
+
+  return false;
+}

package/node_modules/tldts-core/src/is-valid.ts

@@ -0,0 +1,89 @@
+/**
+ * Implements fast shallow verification of hostnames. This does not perform a
+ * struct check on the content of labels (classes of Unicode characters, etc.)
+ * but instead check that the structure is valid (number of labels, length of
+ * labels, etc.).
+ *
+ * If you need stricter validation, consider using an external library.
+ */
+
+// KEEP IN SYNC with `extract-hostname.ts` `isValidHostnameChar` + its inline
+// scan/verdict, which duplicate these structural rules to validate during
+// extraction (a perf fusion). That copy additionally accepts A-Z (the host is
+// not yet lowercased there) and folds in '-' / '_'. Any change to the accepted
+// character set or the label/length rules here must be mirrored there.
+function isValidAscii(code: number): boolean {
+  return (
+    (code >= 97 && code <= 122) || (code >= 48 && code <= 57) || code > 127
+  );
+}
+
+/**
+ * Check if a hostname string is valid. It's usually a preliminary check before
+ * trying to use getDomain or anything else.
+ *
+ * Beware: it does not check if the TLD exists.
+ */
+export default function (hostname: string): boolean {
+  if (hostname.length > 255) {
+    return false;
+  }
+
+  if (hostname.length === 0) {
+    return false;
+  }
+
+  if (
+    /*@__INLINE__*/ !isValidAscii(hostname.charCodeAt(0)) &&
+    hostname.charCodeAt(0) !== 46 && // '.' (dot)
+    hostname.charCodeAt(0) !== 95 // '_' (underscore)
+  ) {
+    return false;
+  }
+
+  // Validate hostname according to RFC
+  let lastDotIndex = -1;
+  let lastCharCode = -1;
+  const len = hostname.length;
+
+  for (let i = 0; i < len; i += 1) {
+    const code = hostname.charCodeAt(i);
+    if (code === 46 /* '.' */) {
+      if (
+        // Check that previous label is < 63 bytes long (64 = 63 + '.')
+        i - lastDotIndex > 64 ||
+        // Check that previous character was not already a '.'
+        lastCharCode === 46 ||
+        // Check that the previous label does not end with '-' (RFC 1035 §2.3.1 LDH).
+        // '_' is intentionally NOT restricted: DNS allows any octet (RFC 2181 §11) and
+        // WHATWG URL does not treat '_' as a forbidden host code point.
+        lastCharCode === 45
+      ) {
+        return false;
+      }
+
+      lastDotIndex = i;
+    } else if (
+      // A forbidden character in the label...
+      !(/*@__INLINE__*/ (isValidAscii(code) || code === 45 || code === 95)) ||
+      // ...or a '-' starting a label (the byte right after a '.'). A label must
+      // not begin with a hyphen (RFC 1034 §3.5 / RFC 1035 §2.3.1 LDH, as amended
+      // by RFC 1123 §2.1; cf. UTS #46 CheckHyphens). The first label is covered by
+      // the leading-character guard above; mirrors the trailing-'-' rule below.
+      (code === 45 && lastCharCode === 46)
+    ) {
+      return false;
+    }
+
+    lastCharCode = code;
+  }
+
+  return (
+    // Check that last label is shorter than 63 chars
+    len - lastDotIndex - 1 <= 63 &&
+    // Check that the last character is an allowed trailing label character.
+    // Since we already checked that the char is a valid hostname character,
+    // we only need to check that it's different from '-'.
+    lastCharCode !== 45
+  );
+}

package/node_modules/tldts-core/src/lookup/fast-path.ts

@@ -0,0 +1,80 @@
+import { IPublicSuffix, ISuffixLookupOptions } from './interface';
+
+export default function (
+  hostname: string,
+  options: ISuffixLookupOptions,
+  out: IPublicSuffix,
+): boolean {
+  // Fast path for very popular suffixes; this allows to by-pass lookup
+  // completely as well as any extra allocation or string manipulation.
+  if (!options.allowPrivateDomains && hostname.length > 3) {
+    const last: number = hostname.length - 1;
+    const c3: number = hostname.charCodeAt(last);
+    const c2: number = hostname.charCodeAt(last - 1);
+    const c1: number = hostname.charCodeAt(last - 2);
+    const c0: number = hostname.charCodeAt(last - 3);
+
+    if (
+      c3 === 109 /* 'm' */ &&
+      c2 === 111 /* 'o' */ &&
+      c1 === 99 /* 'c' */ &&
+      c0 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'com';
+      return true;
+    } else if (
+      c3 === 103 /* 'g' */ &&
+      c2 === 114 /* 'r' */ &&
+      c1 === 111 /* 'o' */ &&
+      c0 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'org';
+      return true;
+    } else if (
+      c3 === 117 /* 'u' */ &&
+      c2 === 100 /* 'd' */ &&
+      c1 === 101 /* 'e' */ &&
+      c0 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'edu';
+      return true;
+    } else if (
+      c3 === 118 /* 'v' */ &&
+      c2 === 111 /* 'o' */ &&
+      c1 === 103 /* 'g' */ &&
+      c0 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'gov';
+      return true;
+    } else if (
+      c3 === 116 /* 't' */ &&
+      c2 === 101 /* 'e' */ &&
+      c1 === 110 /* 'n' */ &&
+      c0 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'net';
+      return true;
+    } else if (
+      c3 === 101 /* 'e' */ &&
+      c2 === 100 /* 'd' */ &&
+      c1 === 46 /* '.' */
+    ) {
+      out.isIcann = true;
+      out.isPrivate = false;
+      out.publicSuffix = 'de';
+      return true;
+    }
+  }
+
+  return false;
+}

package/node_modules/tldts-core/src/lookup/interface.ts

@@ -0,0 +1,10 @@
+export interface IPublicSuffix {
+  isIcann: boolean | null;
+  isPrivate: boolean | null;
+  publicSuffix: string | null;
+}
+
+export interface ISuffixLookupOptions {
+  allowIcannDomains: boolean;
+  allowPrivateDomains: boolean;
+}

package/node_modules/tldts-core/src/options.ts

@@ -0,0 +1,45 @@
+export interface IOptions {
+  allowIcannDomains: boolean;
+  allowPrivateDomains: boolean;
+  detectIp: boolean;
+  // Detect RFC 6761 / IANA special-use domains and expose the result as
+  // `isSpecialUse` on `parse()`. Off by default so the common path stays
+  // allocation-free with no extra work; enable it to populate the field.
+  detectSpecialUse: boolean;
+  extractHostname: boolean;
+  mixedInputs: boolean;
+  validHosts: string[] | null;
+  validateHostname: boolean;
+}
+
+function setDefaultsImpl({
+  allowIcannDomains = true,
+  allowPrivateDomains = false,
+  detectIp = true,
+  detectSpecialUse = false,
+  extractHostname = true,
+  mixedInputs = true,
+  validHosts = null,
+  validateHostname = true,
+}: Partial<IOptions>): IOptions {
+  return {
+    allowIcannDomains,
+    allowPrivateDomains,
+    detectIp,
+    detectSpecialUse,
+    extractHostname,
+    mixedInputs,
+    validHosts,
+    validateHostname,
+  };
+}
+
+const DEFAULT_OPTIONS = /*@__INLINE__*/ setDefaultsImpl({});
+
+export function setDefaults(options?: Partial<IOptions>): IOptions {
+  if (options === undefined) {
+    return DEFAULT_OPTIONS;
+  }
+
+  return /*@__INLINE__*/ setDefaultsImpl(options);
+}

package/node_modules/tldts-core/src/subdomain.ts

@@ -0,0 +1,11 @@
+/**
+ * Returns the subdomain of a hostname string
+ */
+export default function getSubdomain(hostname: string, domain: string): string {
+  // If `hostname` and `domain` are the same, then there is no sub-domain
+  if (domain.length === hostname.length) {
+    return '';
+  }
+
+  return hostname.slice(0, -domain.length - 1);
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@noodleseed/one",
-  "version": "0.93.0",
+  "version": "0.94.0",
   "private": false,
   "description": "Noodle CLI by Noodle Seed — author, run, and deploy declarative MCP servers. Embedding the assistant in your own web app is @noodleseed/assistant.",
   "license": "Apache-2.0",
@@ -174,6 +174,8 @@
     "statuses",
     "stripe",
     "tailwindcss",
+    "tldts",
+    "tldts-core",
     "toidentifier",
     "tslib",
     "typescript",