regex/doc/partial_matches.qbk

[/ 
  Copyright 2006-2007 John Maddock.
  Distributed under the Boost Software License, Version 1.0.
  (See accompanying file LICENSE_1_0.txt or copy at
  http://www.boost.org/LICENSE_1_0.txt).
]

[section:partial_matches Partial Matches]

The [match_flag_type] `match_partial` can be passed to the following algorithms: 
[regex_match], [regex_search], and [regex_grep], and used with the 
iterator [regex_iterator]. When used it indicates that partial as 
well as full matches should be found. A partial match is one that 
matched one or more characters at the end of the text input, but 
did not match all of the regular expression (although it may have done 
so had more input been available). Partial matches are typically used 
when either validating data input (checking each character as it is 
entered on the keyboard), or when searching texts that are either too long 
to load into memory (or even into a memory mapped file), or are of 
indeterminate length (for example the source may be a socket or similar). 
Partial and full matches can be differentiated as shown in the following 
table (the variable M represents an instance of [match_results] as filled in 
by [regex_match], [regex_search] or [regex_grep]):

[table
[[ ][Result][M\[0\].matched][M\[0\].first][M\[0\].second]]
[[No match][False][Undefined][Undefined][Undefined]]
[[Partial match][True][False][Start of partial match.][End of partial match (end of text).]]
[[Full match][True][True][Start of full match.][End of full match.]]
]

Be aware that using partial matches can sometimes result in somewhat 
imperfect behavior:

* There are some expressions, such as ".\*abc" that will always produce a partial match.  This problem can be reduced by careful construction of the regular expressions used, or by setting flags like match_not_dot_newline so that expressions like .\* can't match past line boundaries.
* Boost.Regex currently prefers leftmost matches to full matches, so for example matching "abc|b" against "ab" produces a partial match against the "ab" rather than a full match against "b".  It's more efficient to work this way, but may not be the behavior you want in all situations.

The following example tests to see whether the text could be a valid 
credit card number, as the user presses a key, the character entered 
would be added to the string being built up, and passed to `is_possible_card_number`. 
If this returns true then the text could be a valid card number, so the 
user interface's OK button would be enabled. If it returns false, then 
this is not yet a valid card number, but could be with more input, so 
the user interface would disable the OK button. Finally, if the procedure 
throws an exception the input could never become a valid number, and the 
inputted character must be discarded, and a suitable error indication 
displayed to the user.

   #include <string>
   #include <iostream>
   #include <boost/regex.hpp>

   boost::regex e("(\\d{3,4})[- ]?(\\d{4})[- ]?(\\d{4})[- ]?(\\d{4})");

   bool is_possible_card_number(const std::string& input)
   {
      //
      // return false for partial match, true for full match, or throw for
      // impossible match based on what we have so far...
      boost::match_results<std::string::const_iterator> what;
      if(0 == boost::regex_match(input, what, e, boost::match_default | boost::match_partial))
      {
         // the input so far could not possibly be valid so reject it:
         throw std::runtime_error(
            "Invalid data entered - this could not possibly be a valid card number");
      }
      // OK so far so good, but have we finished?
      if(what[0].matched)
      {
         // excellent, we have a result:
         return true;
      }
      // what we have so far is only a partial match...
      return false;
   }

In the following example, text input is taken from a stream containing an 
unknown amount of text; this example simply counts the number of html tags 
encountered in the stream. The text is loaded into a buffer and searched a 
part at a time, if a partial match was encountered, then the partial match 
gets searched a second time as the start of the next batch of text:

   #include <iostream>
   #include <fstream>
   #include <sstream>
   #include <string>
   #include <boost/regex.hpp>

   // match some kind of html tag:
   boost::regex e("<[^>]*>");
   // count how many:
   unsigned int tags = 0;

   void search(std::istream& is)
   {
      // buffer we'll be searching in:
      char buf[4096];
      // saved position of end of partial match:
      const char* next_pos = buf + sizeof(buf);
      // flag to indicate whether there is more input to come:
      bool have_more = true;

      while(have_more)
      {
         // how much do we copy forward from last try:
         unsigned leftover = (buf + sizeof(buf)) - next_pos;
         // and how much is left to fill:
         unsigned size = next_pos - buf;
         // copy forward whatever we have left:
         std::memmove(buf, next_pos, leftover);
         // fill the rest from the stream:
         is.read(buf + leftover, size);
         unsigned read = is.gcount();
         // check to see if we've run out of text:
         have_more = read == size;
         // reset next_pos:
         next_pos = buf + sizeof(buf);
         // and then iterate:
         boost::cregex_iterator a(
            buf, 
            buf + read + leftover, 
            e, 
            boost::match_default | boost::match_partial); 
         boost::cregex_iterator b;

         while(a != b)
         {
            if((*a)[0].matched == false)
            {
               // Partial match, save position and break:
               next_pos = (*a)[0].first;
               break;
            }
            else
            {
               // full match:
               ++tags;
            }
            
            // move to next match:
            ++a;
         }
      }
   }

[endsect]