#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include "regexParser.h"

static NodePtr charClass(RegexInfoPtr);
static void token(RegexInfoPtr);
static NodePtr regexAtom(RegexInfoPtr);
NodePtr regex(RegexInfoPtr);

/**
 * Create a node of regex parse tree.
 *     tokenType
 *     regexPosition(state)
 *     stateTransitionTable
 */

static
NodePtr allocateNode() {
    NodePtr n = NEW(Node);
    n->cc = NULL;
    n->left = NULL;
    n->right = NULL;
    return n;
}

static
NodePtr createNode(RegexInfoPtr ri,unsigned char type,CharClassPtr cc, NodePtr left, NodePtr right) {
    NodePtr n = allocateNode();

    n->tokenType = type;
    n->cc = cc;
    n->left = left;
    n->right = right;
    n->nodeNumber = ri->nodeNumber;
    ri->nodeNumber++;

    return n;
}

CharClassPtr createCharClassWord(RegexInfoPtr ri) {
    CharClassPtr cc = NEW(CharClass);
    cc->type = 'a';
    cc->cond.w.word = ri->tokenValue;
    cc->cond.w.length = ri->ptr - ri->tokenValue;
    cc->nextState.bitContainer = 0;
    return cc;
}

CharClassPtr charClassMerge(CharClassPtr cc, unsigned char begin, unsigned char end, CharClassPtr next) {
    CharClassPtr cc1 = NEW(CharClass);
    return cc1;
}

// <charClass> ::= '['<literal>'-'<literal>']'
static
NodePtr charClass(RegexInfoPtr ri) {
    CharClassPtr cc = NEW(CharClass);
    NodePtr n = createNode(ri,'c',cc,0,0);
    cc->type = 'r';
    cc->nextState.bitContainer = 0;
    RangeListPtr rangeList = &cc->cond.range;
    rangeList->begin = (unsigned long)*ri->ptr;
    rangeList->end = (unsigned long)*ri->ptr;
    rangeList->next = NULL;

    for (ri->ptr++; *ri->ptr && *ri->ptr != ']'; ri->ptr++) {
        if (*ri->ptr == '-') {
            rangeList->end = (unsigned long)*(ri->ptr + 1);
            ri->ptr++;
            continue;
        }
        if (ri->ptr[0] == 0 || ri->ptr[0] == ']') break;
        if (ri->ptr[0] == rangeList->end + 1) {
            rangeList->end = (unsigned long)*ri->ptr;
            continue;
        }
        rangeList->next = NEW(RangeList);
        rangeList = rangeList->next;
        rangeList->begin = (unsigned long)*ri->ptr;
        rangeList->end = (unsigned long)*ri->ptr;
        rangeList->next = NULL;
    }
    // TODO literal support
    // merge rangeList here
    if (*ri->ptr) ri->ptr++;
    token(ri);
    return n;
}

// <literal> ::= [a-z][A-Z][0-9]
static
NodePtr literal(RegexInfoPtr ri) {
    CharClassPtr cc = createCharClassWord(ri);
    token(ri);
    NodePtr n = createNode(ri,'a',cc,0,0);
    return n;
}

static
void token(RegexInfoPtr ri) {
    while (ri->ptr[0] != '\0') {
        if (ri->ptr[0] == '('){
            ri->ptr++;
            ri->tokenType = '(';
            ri->tokenValue = NULL;
            return;
        } else if (ri->ptr[0] == ')') {
            ri->ptr++;
            ri->tokenType = ')';
            ri->tokenValue = ri->ptr;
            return;
        } else if (ri->ptr[0] == ']') {
            ri->ptr++;
            ri->tokenType = ']';
            ri->tokenValue = ri->ptr;
            return;
        } else if (ri->ptr[0] == '|'){
            ri->ptr++;
            ri->tokenType = '|';
            ri->tokenValue = NULL;
            return;
        } else if (ri->ptr[0] == '*'){
            ri->ptr++;
            ri->tokenType = '*';
            ri->tokenValue = NULL;
            return;
        } else if (ri->ptr[0] == '\\'){
            // need more proccesing 
            /*
                \277
                \0xa5
                \[
                \\
                \utf-8 etc...
            */
        } else if (ri->ptr[0] == '[') {
            ri->ptr++;
            ri->tokenType = 'c';
            ri->tokenValue = ri->ptr;
            return;
        } else {
            ri->tokenType = 'a';
            ri->tokenValue = ri->ptr;
            if (isalnum(ri->ptr[0])) {
                ri->ptr++;
            }
            return;
        }
    }
    ri->tokenType = 0;
    ri->tokenValue = NULL;
    return;
}

// <regexAtom> ::= <literal>|<charClass>|<group>
static
NodePtr regexAtom(RegexInfoPtr ri) {

    NodePtr n = NULL;
    if (ri->tokenType == 'c') n = charClass(ri);
    else if (ri->tokenType == 'a') n = literal(ri);
    else if (ri->tokenType == '(') {
        n = regex(ri);
        if (ri->tokenType != ')') {
            // error
        }
        token(ri);
    }
    if (ri->tokenType == '*') {
        n = createNode(ri,'*',0,n,0);
        token(ri);
    }

    return n;
}

// <regex> ::= <regexAtom> | <regexAtom>'*'<regex> | <regexAtom>'|'<regex> | <regexAtom><regexAtom>'*' | <regexAtom><regex>
NodePtr regex(RegexInfoPtr ri) {
    token(ri);
    NodePtr n = regexAtom(ri);
    while (ri->tokenType) {
        if (ri->tokenType == '*') {
            n = createNode(ri,'*',0,n,0);
            token(ri);
            return n;
        } else if (ri->tokenType == '|') {
            n = createNode(ri,'|',0,n,0);
            NodePtr n1 = regex(ri);
            n->right = n1;
        } else if (ri->tokenType == ')') {
            return n;
        } else {
            n = createNode(ri,'+',0,n,0);
            NodePtr n1 = regexAtom(ri);
            n->right = n1;
        }
    }
    return n;
}