path: root/src/lib/utils/parsing.cpp

/*
* Various string utils and parsing functions
* (C) 1999-2007,2013,2014,2015,2018 Jack Lloyd
* (C) 2015 Simon Warta (Kullo GmbH)
* (C) 2017 René Korthaus, Rohde & Schwarz Cybersecurity
*
* Botan is released under the Simplified BSD License (see license.txt)
*/

#include <botan/parsing.h>
#include <botan/exceptn.h>
#include <botan/charset.h>
#include <botan/loadstor.h>
#include <algorithm>
#include <cctype>
#include <limits>
#include <set>

namespace Botan {

uint16_t to_uint16(const std::string& str)
   {
   const uint32_t x = to_u32bit(str);

   if(x >> 16)
      throw Invalid_Argument("Integer value exceeds 16 bit range");

   return static_cast<uint16_t>(x);
   }

uint32_t to_u32bit(const std::string& str)
   {
   // std::stoul is not strict enough. Ensure that str is digit only [0-9]*
   for(const char chr : str)
      {
      if(chr < '0' || chr > '9')
         {
         std::string chrAsString(1, chr);
         throw Invalid_Argument("String contains non-digit char: " + chrAsString);
         }
      }

   const unsigned long int x = std::stoul(str);

   if(sizeof(unsigned long int) > 4)
      {
      // x might be uint64
      if (x > std::numeric_limits<uint32_t>::max())
         {
         throw Invalid_Argument("Integer value of " + str + " exceeds 32 bit range");
         }
      }

   return static_cast<uint32_t>(x);
   }

/*
* Convert a string into a time duration
*/
uint32_t timespec_to_u32bit(const std::string& timespec)
   {
   if(timespec.empty())
      return 0;

   const char suffix = timespec[timespec.size()-1];
   std::string value = timespec.substr(0, timespec.size()-1);

   uint32_t scale = 1;

   if(Charset::is_digit(suffix))
      value += suffix;
   else if(suffix == 's')
      scale = 1;
   else if(suffix == 'm')
      scale = 60;
   else if(suffix == 'h')
      scale = 60 * 60;
   else if(suffix == 'd')
      scale = 24 * 60 * 60;
   else if(suffix == 'y')
      scale = 365 * 24 * 60 * 60;
   else
      throw Decoding_Error("timespec_to_u32bit: Bad input " + timespec);

   return scale * to_u32bit(value);
   }

/*
* Parse a SCAN-style algorithm name
*/
std::vector<std::string> parse_algorithm_name(const std::string& namex)
   {
   if(namex.find('(') == std::string::npos &&
      namex.find(')') == std::string::npos)
      return std::vector<std::string>(1, namex);

   std::string name = namex, substring;
   std::vector<std::string> elems;
   size_t level = 0;

   elems.push_back(name.substr(0, name.find('(')));
   name = name.substr(name.find('('));

   for(auto i = name.begin(); i != name.end(); ++i)
      {
      char c = *i;

      if(c == '(')
         ++level;
      if(c == ')')
         {
         if(level == 1 && i == name.end() - 1)
            {
            if(elems.size() == 1)
               elems.push_back(substring.substr(1));
            else
               elems.push_back(substring);
            return elems;
            }

         if(level == 0 || (level == 1 && i != name.end() - 1))
            throw Invalid_Algorithm_Name(namex);
         --level;
         }

      if(c == ',' && level == 1)
         {
         if(elems.size() == 1)
            elems.push_back(substring.substr(1));
         else
            elems.push_back(substring);
         substring.clear();
         }
      else
         substring += c;
      }

   if(!substring.empty())
      throw Invalid_Algorithm_Name(namex);

   return elems;
   }

std::vector<std::string> split_on(const std::string& str, char delim)
   {
   return split_on_pred(str, [delim](char c) { return c == delim; });
   }

std::vector<std::string> split_on_pred(const std::string& str,
                                       std::function<bool (char)> pred)
   {
   std::vector<std::string> elems;
   if(str.empty()) return elems;

   std::string substr;
   for(auto i = str.begin(); i != str.end(); ++i)
      {
      if(pred(*i))
         {
         if(!substr.empty())
            elems.push_back(substr);
         substr.clear();
         }
      else
         substr += *i;
      }

   if(substr.empty())
      throw Invalid_Argument("Unable to split string: " + str);
   elems.push_back(substr);

   return elems;
   }

/*
* Join a string
*/
std::string string_join(const std::vector<std::string>& strs, char delim)
   {
   std::string out = "";

   for(size_t i = 0; i != strs.size(); ++i)
      {
      if(i != 0)
         out += delim;
      out += strs[i];
      }

   return out;
   }

/*
* Parse an ASN.1 OID string
*/
std::vector<uint32_t> parse_asn1_oid(const std::string& oid)
   {
   std::string substring;
   std::vector<uint32_t> oid_elems;

   for(auto i = oid.begin(); i != oid.end(); ++i)
      {
      char c = *i;

      if(c == '.')
         {
         if(substring.empty())
            throw Invalid_OID(oid);
         oid_elems.push_back(to_u32bit(substring));
         substring.clear();
         }
      else
         substring += c;
      }

   if(substring.empty())
      throw Invalid_OID(oid);
   oid_elems.push_back(to_u32bit(substring));

   if(oid_elems.size() < 2)
      throw Invalid_OID(oid);

   return oid_elems;
   }

/*
* X.500 String Comparison
*/
bool x500_name_cmp(const std::string& name1, const std::string& name2)
   {
   auto p1 = name1.begin();
   auto p2 = name2.begin();

   while((p1 != name1.end()) && Charset::is_space(*p1)) ++p1;
   while((p2 != name2.end()) && Charset::is_space(*p2)) ++p2;

   while(p1 != name1.end() && p2 != name2.end())
      {
      if(Charset::is_space(*p1))
         {
         if(!Charset::is_space(*p2))
            return false;

         while((p1 != name1.end()) && Charset::is_space(*p1)) ++p1;
         while((p2 != name2.end()) && Charset::is_space(*p2)) ++p2;

         if(p1 == name1.end() && p2 == name2.end())
            return true;
         if(p1 == name1.end() || p2 == name2.end())
            return false;
         }

      if(!Charset::caseless_cmp(*p1, *p2))
         return false;
      ++p1;
      ++p2;
      }

   while((p1 != name1.end()) && Charset::is_space(*p1)) ++p1;
   while((p2 != name2.end()) && Charset::is_space(*p2)) ++p2;

   if((p1 != name1.end()) || (p2 != name2.end()))
      return false;
   return true;
   }

/*
* Convert a decimal-dotted string to binary IP
*/
uint32_t string_to_ipv4(const std::string& str)
   {
   std::vector<std::string> parts = split_on(str, '.');

   if(parts.size() != 4)
      throw Decoding_Error("Invalid IP string " + str);

   uint32_t ip = 0;

   for(auto part = parts.begin(); part != parts.end(); ++part)
      {
      uint32_t octet = to_u32bit(*part);

      if(octet > 255)
         throw Decoding_Error("Invalid IP string " + str);

      ip = (ip << 8) | (octet & 0xFF);
      }

   return ip;
   }

/*
* Convert an IP address to decimal-dotted string
*/
std::string ipv4_to_string(uint32_t ip)
   {
   std::string str;

   for(size_t i = 0; i != sizeof(ip); ++i)
      {
      if(i)
         str += ".";
      str += std::to_string(get_byte(i, ip));
      }

   return str;
   }

std::string erase_chars(const std::string& str, const std::set<char>& chars)
   {
   std::string out;

   for(auto c: str)
      if(chars.count(c) == 0)
         out += c;

   return out;
   }

std::string replace_chars(const std::string& str,
                          const std::set<char>& chars,
                          char to_char)
   {
   std::string out = str;

   for(size_t i = 0; i != out.size(); ++i)
      if(chars.count(out[i]))
         out[i] = to_char;

   return out;
   }

std::string replace_char(const std::string& str, char from_char, char to_char)
   {
   std::string out = str;

   for(size_t i = 0; i != out.size(); ++i)
      if(out[i] == from_char)
         out[i] = to_char;

   return out;
   }

namespace {

std::string tolower_string(const std::string& in)
   {
   std::string s = in;
   for(size_t i = 0; i != s.size(); ++i)
      {
      if(std::isalpha(static_cast<unsigned char>(s[i])))
         s[i] = std::tolower(static_cast<unsigned char>(s[i]));
      }
   return s;
   }

}

bool host_wildcard_match(const std::string& issued_, const std::string& host_)
   {
   const std::string issued = tolower_string(issued_);
   const std::string host = tolower_string(host_);

   if(host.empty() || issued.empty())
      return false;

   /*
   If there are embedded nulls in your issued name
   Well I feel bad for you son
   */
   if(std::count(issued.begin(), issued.end(), char(0)) > 0)
      return false;

   // If more than one wildcard, then issued name is invalid
   const size_t stars = std::count(issued.begin(), issued.end(), '*');
   if(stars > 1)
      return false;

   // '*' is not a valid character in DNS names so should not appear on the host side
   if(std::count(host.begin(), host.end(), '*') != 0)
      return false;

   // Similarly a DNS name can't end in .
   if(host[host.size() - 1] == '.')
      return false;

   // And a host can't have an empty name component, so reject that
   if(host.find("..") != std::string::npos)
      return false;

   // Exact match: accept
   if(issued == host)
      {
      return true;
      }

   /*
   Otherwise it might be a wildcard

   If the issued size is strictly longer than the hostname size it
   couldn't possibly be a match, even if the issued value is a
   wildcard. The only exception is when the wildcard ends up empty
   (eg www.example.com matches www*.example.com)
   */
   if(issued.size() > host.size() + 1)
      {
      return false;
      }

   // If no * at all then not a wildcard, and so not a match
   if(stars != 1)
      {
      return false;
      }

   /*
   Now walk through the issued string, making sure every character
   matches. When we come to the (singular) '*', jump forward in the
   hostname by the corresponding amount. We know exactly how much
   space the wildcard takes because it must be exactly `len(host) -
   len(issued) + 1 chars`.

   We also verify that the '*' comes in the leftmost component, and
   doesn't skip over any '.' in the hostname.
   */
   size_t dots_seen = 0;
   size_t host_idx = 0;

   for(size_t i = 0; i != issued.size(); ++i)
      {
      dots_seen += (issued[i] == '.');

      if(issued[i] == '*')
         {
         // Fail: wildcard can only come in leftmost component
         if(dots_seen > 0)
            {
            return false;
            }

         /*
         Since there is only one * we know the tail of the issued and
         hostname must be an exact match. In this case advance host_idx
         to match.
         */
         const size_t advance = (host.size() - issued.size() + 1);

         if(host_idx + advance > host.size()) // shouldn't happen
            return false;

         // Can't be any intervening .s that we would have skipped
         if(std::count(host.begin() + host_idx,
                       host.begin() + host_idx + advance, '.') != 0)
            return false;

         host_idx += advance;
         }
      else
         {
         if(issued[i] != host[host_idx])
            {
            return false;
            }

         host_idx += 1;
         }
      }

   // Wildcard issued name must have at least 3 components
   if(dots_seen < 2)
      {
      return false;
      }

   return true;
   }

}