path: root/parser.cpp

#include <cstdio>
#include <cctype>
#include <sstream>
#include "parser.h"
#include "exc.h"
#include "consts.h"
#include "builtin.h"

using std::stringstream;

static char buff[TOKEN_BUFF_SIZE];
static FrameObj *parse_stack[PARSE_STACK_SIZE];
extern Pair *empty_list;

Tokenizor::Tokenizor() : stream(stdin), buff_ptr(buff), escaping(false) {}
void Tokenizor::set_stream(FILE *_stream) {
    stream = _stream;
}

#define IS_NEWLINE(ch) \
    ((ch) == '\n')
#define IS_QUOTE(ch) \
    ((ch) == '\"')
#define IS_SLASH(ch) \
    ((ch) == '\\')
#define IS_BRACKET(ch) \
    ((ch) == '(' || (ch) == ')') 
#define IS_SPACE(ch) \
    ((ch) == ' ' || (ch) == '\t' || IS_NEWLINE(ch))
#define IS_COMMENT(ch) \
    ((ch) == ';')
#define IS_LITERAL(ch) \
    ((ch) == '\'')
#define IS_DELIMITER(ch) \
    (IS_BRACKET(ch) || IS_SPACE(ch) ||  \
     IS_COMMENT(ch) || IS_QUOTE(ch))

#define POP \
    do { \
        *buff_ptr = 0; \
        ret = string(buff); \
        buff_ptr = buff; \
    } while (0)
#define TOP (*(buff_ptr - 1))

void str_to_lower(string &str) {
    size_t len = str.length();
    for (int i = 0; i < len; i++)
        if ('A' <= str[i] && str[i] <= 'Z')
            str[i] -= 'A' - 'a';
}

bool Tokenizor::get_token(string &ret) {
    char ch;
    bool flag = false;
    while (fread(&ch, 1, 1, stream))
    {
        if (escaping)
        {
            escaping = false;
            switch (ch)
            {
                case '\\': *buff_ptr++ = '\\'; break;
                case '\"': *buff_ptr++ = '\"'; break;
                case 'n': *buff_ptr++ = '\n'; break;
                case 't': *buff_ptr++ = '\t'; break;
                default: {
                             buff_ptr = buff;
                             throw TokenError(string("") + ch, 
                                     PAR_ERR_ILLEGAL_CHAR_IN_ESC);
                         }
            }
        }
        else
        {
            bool in_quote = buff_ptr != buff && IS_QUOTE(*buff);
            if (buff_ptr != buff)
            {
                if (buff_ptr - buff == 1 && IS_LITERAL(TOP))
                {
                    POP;
                    flag = true;
                }
                else if ((IS_BRACKET(TOP) || IS_DELIMITER(ch)))
                {
                    if (IS_COMMENT(*buff))
                    {
                        if (IS_NEWLINE(ch)) buff_ptr = buff;
                        else buff_ptr = buff + 1;
                    }
                    else if (!in_quote) // not in a double-quote
                    {
                        if (!(buff_ptr - buff == 1 && ch == '(' && TOP ==