1 /*
2 *
3 * Copyright (c) 2003
4 * John Maddock
5 *
6 * Use, modification and distribution are subject to the
7 * Boost Software License, Version 1.0. (See accompanying file
8 * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
9 *
10 */
11
12 /*
13 * LOCATION: see http://www.boost.org for most recent version.
14 * FILE regex_token_iterator_example_2.cpp
15 * VERSION see <boost/version.hpp>
16 * DESCRIPTION: regex_token_iterator example: spit out linked URL's.
17 */
18
19
20 #include <boost/regex.hpp>
21 #include <fstream>
22 #include <iostream>
23 #include <iterator>
24
25 boost::regex e("<\\s*A\\s+[^>]*href\\s*=\\s*\"([^\"]*)\"",
26 boost::regex::normal | boost::regbase::icase);
27
load_file(std::string & s,std::istream & is)28 void load_file(std::string& s, std::istream& is)
29 {
30 s.erase();
31 if(is.bad()) return;
32 //
33 // attempt to grow string buffer to match file size,
34 // this doesn't always work...
35 s.reserve(static_cast<std::string::size_type>(is.rdbuf()->in_avail()));
36 char c;
37 while(is.get(c))
38 {
39 // use logarithmic growth stategy, in case
40 // in_avail (above) returned zero:
41 if(s.capacity() == s.size())
42 s.reserve(s.capacity() * 3);
43 s.append(1, c);
44 }
45 }
46
main(int argc,char ** argv)47 int main(int argc, char** argv)
48 {
49 std::string s;
50 int i;
51 for(i = 1; i < argc; ++i)
52 {
53 std::cout << "Findings URL's in " << argv[i] << ":" << std::endl;
54 s.erase();
55 std::ifstream is(argv[i]);
56 load_file(s, is);
57 is.close();
58 boost::sregex_token_iterator i(s.begin(), s.end(), e, 1);
59 boost::sregex_token_iterator j;
60 while(i != j)
61 {
62 std::cout << *i++ << std::endl;
63 }
64 }
65 //
66 // alternative method:
67 // test the array-literal constructor, and split out the whole
68 // match as well as $1....
69 //
70 for(i = 1; i < argc; ++i)
71 {
72 std::cout << "Findings URL's in " << argv[i] << ":" << std::endl;
73 s.erase();
74 std::ifstream is(argv[i]);
75 load_file(s, is);
76 is.close();
77 const int subs[] = {1, 0,};
78 boost::sregex_token_iterator i(s.begin(), s.end(), e, subs);
79 boost::sregex_token_iterator j;
80 while(i != j)
81 {
82 std::cout << *i++ << std::endl;
83 }
84 }
85
86 return 0;
87 }
88
89
90