Unexpected results in matching a regex

[Dave]

#define BOOST_REGEX_DYN_LINK

#include <iostream>
#include <boost\regex.hpp>

using namespace boost;
using namespace std;

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

 

[Andre Kostur]

#define BOOST_REGEX_DYN_LINK

#include <iostream>
#include <boost\regex.hpp>

using namespace boost;
using namespace std;

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

You're asking in the wrong place.... that's not Standard C++. You'll be
much better off asking in the boost user's mailing lists and such.

 

[Victor Bazarov]

#define BOOST_REGEX_DYN_LINK

#include <iostream>
#include <boost\regex.hpp>

Should change it to

#include <boost/regex.hpp>

and _never_ again use backslashes in the header name.

using namespace boost;
using namespace std;

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

Oh, did you try Boost online forum? What did they say?

V

 

[Dave]

#define BOOST_REGEX_DYN_LINK

#include <iostream>
#include <boost\regex.hpp>

Should change it to

#include <boost/regex.hpp>

and _never_ again use backslashes in the header name.

using namespace boost;
using namespace std;

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

Oh, did you try Boost online forum? What did they say?

V

Actually, yes I did post to Boost, but my message has not yet appeared. Your
comment on backslashes contributes absolutely nothing to this thread.

 

[Victor Bazarov]

#define BOOST_REGEX_DYN_LINK

#include <iostream>
#include <boost\regex.hpp>

Should change it to

#include <boost/regex.hpp>

and _never_ again use backslashes in the header name.

using namespace boost;
using namespace std;

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

Oh, did you try Boost online forum? What did they say?

V

Actually, yes I did post to Boost, but my message has not yet appeared.

So, you're growing impatient, is that it? Generally, Boost functionality
is OT here. If you asked about std::tr1, you'd be closer to the topic of
this newsgroup.

Have you tried other tools, like Perl or 'sed' to see if the expression is
a match? No, I didn't. I just think that "\x2Aqfw" is not a match for
"\x2A..". You probably wanted to add another dot there.

> Your
comment on backslashes contributes absolutely nothing to this thread.

Not true, it promotes healthy coding practices and that is always
a GOOD THING(tm).

V

 

[Pete Becker]

int main()
{
regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

// I would expect this to match, and it does.
cout << regex_match("\xF8p\x2Aqf\x0DPFFF", reg_exp) << endl;

// I would expect this to *not* match due to the 'w'.
// Why does it match?
cout << regex_match("\xF8p\x2Aqfw\x0DPFFF", reg_exp) << endl;
}

I can't speak for the Boost implementation, but our TR1 implementation
gets the same result. It's a little clearer with a slight rewriting of
the regular expression. Yours was:

regex reg_exp("\xF8.\x2A..\x0D(P|V)[0-9A-F]{3}");

In the ASCII encoding that's equivalent to:

regex reg_exp("\xF8.*..\x0D(P|V)[0-9A-F]{3}");


And that embedded ".*.." will eat up two or more characters.