TLD Parsing code from Gecko

gistfile1.cpp

// Finds the base domain for a host, with requested number of additional parts.
// This will fail, generating an error, if the host is an IPv4/IPv6 address,
// if more subdomain parts are requested than are available, or if the hostname
// includes characters that are not valid in a URL. Normalization is performed
// on the host string and the result will be in UTF8.
nsresult
nsEffectiveTLDService::GetBaseDomainInternal(nsCString  &aHostname,
                                             int32_t    aAdditionalParts,
                                             nsACString &aBaseDomain)
{
  if (aHostname.IsEmpty())
    return NS_ERROR_INSUFFICIENT_DOMAIN_LEVELS;

  // chomp any trailing dot, and keep track of it for later
  bool trailingDot = aHostname.Last() == '.';
  if (trailingDot)
    aHostname.Truncate(aHostname.Length() - 1);

  // check the edge cases of the host being '.' or having a second trailing '.',
  // since subsequent checks won't catch it.
  if (aHostname.IsEmpty() || aHostname.Last() == '.')
    return NS_ERROR_INVALID_ARG;

  // Check if we're dealing with an IPv4/IPv6 hostname, and return
  PRNetAddr addr;
  PRStatus result = PR_StringToNetAddr(aHostname.get(), &addr);
  if (result == PR_SUCCESS)
    return NS_ERROR_HOST_IS_IP_ADDRESS;

  // Walk up the domain tree, most specific to least specific,
  // looking for matches at each level.  Note that a given level may
  // have multiple attributes (e.g. IsWild() and IsNormal()).
  const char *prevDomain = nullptr;
  const char *currDomain = aHostname.get();
  const char *nextDot = strchr(currDomain, '.');
  const char *end = currDomain + aHostname.Length();
  const char *eTLD = currDomain;
  while (1) {
    // sanity check the string we're about to look up: it should not begin with
    // a '.'; this would mean the hostname began with a '.' or had an
    // embedded '..' sequence.
    if (*currDomain == '.')
      return NS_ERROR_INVALID_ARG;

    // perform the hash lookup.
    nsDomainEntry *entry = mHash.GetEntry(currDomain);
    if (entry) {
      if (entry->IsWild() && prevDomain) {
        // wildcard rules imply an eTLD one level inferior to the match.
        eTLD = prevDomain;
        break;

      } else if (entry->IsNormal() || !nextDot) {
        // specific match, or we've hit the top domain level
        eTLD = currDomain;
        break;

      } else if (entry->IsException()) {
        // exception rules imply an eTLD one level superior to the match.
        eTLD = nextDot + 1;
        break;
      }
    }

    if (!nextDot) {
      // we've hit the top domain level; use it by default.
      eTLD = currDomain;
      break;
    }

    prevDomain = currDomain;
    currDomain = nextDot + 1;
    nextDot = strchr(currDomain, '.');
  }

  const char *begin, *iter;
  if (aAdditionalParts < 0) {
    NS_ASSERTION(aAdditionalParts == -1,
                 "aAdditionalParts can't be negative and different from -1");

    for (iter = aHostname.get(); iter != eTLD && *iter != '.'; iter++);

    if (iter != eTLD) {
      iter++;
    }
    if (iter != eTLD) {
      aAdditionalParts = 0;
    }
  } else {
    // count off the number of requested domains.
    begin = aHostname.get();
    iter = eTLD;

    while (1) {
      if (iter == begin)
        break;

      if (*(--iter) == '.' && aAdditionalParts-- == 0) {
        ++iter;
        ++aAdditionalParts;
        break;
      }
    }
  }

  if (aAdditionalParts != 0)
    return NS_ERROR_INSUFFICIENT_DOMAIN_LEVELS;

  aBaseDomain = Substring(iter, end);
  // add on the trailing dot, if applicable
  if (trailingDot)
    aBaseDomain.Append('.');

  return NS_OK;
}