//#define SPIRIT_DEBUG
#include "lexer.hpp"
#include <iostream>

using namespace spirit;
using namespace std;

char to_char(wchar_t c)
{
    if (spirit::impl::isgraph_(char(c)) || spirit::impl::isspace_(char(c)))
	    return char(c);
    else
        return '.';
}

ostream& operator<<(ostream& out, wchar_t const* c)
{
    while (*c)
	    out << to_char(*c++);
    return out;
}

template <typename LexerT, typename CharT>
void test_match(LexerT& lexer, const CharT* str)
{
    int token;
    const CharT* end = str;
    while (*end)
        ++end;

    cout << "testing string: " << str << endl;
    while ((token = lexer.next_token(str, end)) != -1 && str != end)
    {
        cout << "matched token #" << token << endl;
    }
    if (token == -1)
    {
        cout << "lexing failed at: " << str << endl;
    }
    else
    {
        cout << "matched token #" << token << endl;
    }
}

int main(int , char** )
{
    /*
	g.parse("[[:blank:]]+");
	g.parse("[[:blank:]]*");
	g.parse("[^[:blank:]\\n]");
	g.parse("\\r?\\n");
	g.parse("([[:alpha:]_][[:alnum:]_-]*)");
	g.parse("[^[:alpha:]_*\\n]+");
	g.parse("(\\\\([^\\n]|[0-7]{1,3}|x[[:xdigit:]]{1,2}))");
	g.parse("(\"[:\"[[:alpha:]]+\":]\")");
	node* n = g.parse("a*b*c*");
	n = g.parse("a*b*c*|c*b*a*");
	n = g.parse("(a|b)*abb+c?");

    */

    //this is pathological! Generates ~65,000 dfa states
    //dfa_t dfa = g.createdfa("(a|b)*a(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)(a|b)");
    //
    //g.register_regex("(a|b)*abb+c?(ab(c|d)*)?");
    //g.register_regex("abcd");
    //g.register_regex("dabcd");
    //g.register_regex("abcde");
    //g.register_regex("(abcd)|(dabcd)");
    //g.register_regex("[\\a-zA-\\x5a][a-zA-Z0-9_]*"); // identifier
    //g.register_regex("[0-9]*"); // integer

    lexer<> lex;
    lex.register_regex("(a|b)*abb+c?(ab(c|d)*)?", 99);
    lex.register_regex("[\\a-zA-\\x5a][a-zA-Z0-9_]*", 100); // identifier
    lex.register_regex("[0-9]*", 200); // integer
    test_match(lex, "aaabb");
    test_match(lex, "aaabbd");
    test_match(lex, "aaabbabd");
    test_match(lex, "aaabbde");
    test_match(lex, "baaabcabd");
    test_match(lex, "baaabcabde");
    test_match(lex, "abc123");
    test_match(lex, "1234");
    test_match(lex, "12a45");

    lexer<> lex2;
    lex2.register_regex("do", 1);
    lex2.register_regex("do1", 2);
    lex2.register_regex("do_more", 3);
    lex2.register_regex("_more", 4);
    lex2.register_regex("some_more", 5);
    lex2.register_regex("some_", 6);
    test_match(lex2, "do");
    test_match(lex2, "do1");
    test_match(lex2, "do_more");
    test_match(lex2, "_more");
    test_match(lex2, "some_more");
    test_match(lex2, "some_");
    test_match(lex2, "dodo1");
    test_match(lex2, "do_more_more");
    test_match(lex2, "some_more_moresome__more");

#define WCHAR_TESTS
#ifdef WCHAR_TESTS
	lexer<wchar_t const *> wlex;
    //wg.register_regex(L"(a|b)*abb+c?(ab(c|d)*)?");
    //wg.register_regex(L"d*a*b*c*d*");
    //wg.register_regex(L"[^\x01020304-\x05060708]+");
    //wg.register_regex(L"[^a-z]+");
    wlex.register_regex(L"[a-z]+", 2);

    test_match(wlex, L"aaabb");
    test_match(wlex, L"abc123");
    test_match(wlex, L"1234");
    test_match(wlex, L"12a45");
    test_match(wlex, L"abc");
    
// only compile this if the compiler has 4-byte wide chars.
#if defined(__GNUC__) && defined(linux)
	lexer<wchar_t const *> wlex2;
    //wlex2.register_regex(L"[^\\x07020304-\\x08060708]*", 1);
    wlex2.register_regex(L"[\\x07020304-\\x08060708]*", 1);
    
    test_match(wlex2, L"\x01020304");
    test_match(wlex2, L"\x01020303");
    test_match(wlex2, L"\x01020305");
    test_match(wlex2, L"\x02020305");
    test_match(wlex2, L"\x05060707");
    test_match(wlex2, L"\x05060708");
    test_match(wlex2, L"\x05060709");
    test_match(wlex2, L"\x07020303");
    test_match(wlex2, L"\x07020304");
    test_match(wlex2, L"\x07020305");
    test_match(wlex2, L"\x07120305");
    test_match(wlex2, L"\x08060707");
    test_match(wlex2, L"\x08060708");
    test_match(wlex2, L"\x08060709");
#endif
#ifdef SPIRIT_DEBUG
    wlex.dump(cout);
#endif

#endif

}