CPlayground/src/lexer.h

#ifndef LEXER_H
#define LEXER_H

#include <stdlib.h>

#include "./token.h"
#include "./str.h"

struct lexer {
	struct str in;
	size_t pos;
	size_t rpos;
	char c;
};

enum lexer_err_code {
	LEXER_ERR_OK = 0,
	LEXER_ERR_INVALID_PP_IDENT,
	LEXER_ERR_INVALID_PP_INCLUDE
};

struct lexer_err {
	enum lexer_err_code code;
	const char *name;
};

struct lexer lexer_create(struct str in);
void lexer_read_char(struct lexer *l);
char lexer_peek_char(struct lexer *l);

struct token lexer_next_token(struct lexer *l, struct lexer_err *err);

struct str lexer_read_ident(struct lexer *l);
struct str lexer_read_str_lit(struct lexer *l);
struct str lexer_read_int_lit(struct lexer *l);
struct str lexer_read_until_or(struct lexer *l, char tgt, char limit,
			       bool *reached_limit);

enum token_type_enum lexer_lookup_ident(struct str ident);
void lexer_skip_whitespace(struct lexer *l);

struct token lexer_handle_pp(struct lexer *l, struct lexer_err *err);
enum token_type_enum _lexer_lookup_pp(struct str ident);

struct lexer_err lexer_err_create(enum lexer_err_code code);

bool _lexer_is_letter(char c);
bool _lexer_is_number(char c);
void _lexer_set_err(struct lexer_err *err, enum lexer_err_code code);

#if defined(IMP) || defined(LEXER_IMP)

struct lexer 
lexer_create(struct str in)
{
	struct lexer l = {0};
	l.in = in;
	lexer_read_char(&l);
	return l;
}

void 
lexer_read_char(struct lexer *l)
{
	if ( l->rpos >= l->in.size ) {
		l->c = '\0';
	} else {
		l->c = l->in.data[l->rpos];
	}
	l->pos = l->rpos;
	++l->rpos;
}

char 
lexer_peek_char(struct lexer *l)
{
	if ( l->rpos >= l->in.size ) {
		return '\0';
	}
	return l->in.data[l->rpos];
}

struct token 
lexer_next_token(struct lexer *l, struct lexer_err *err)
{
#define _LEXER_CUR_CHAR str_slice(l->in, l->pos, l->pos+1)
#define _LEXER_2_CHAR str_slice(l->in, l->pos, l->pos+2)

	struct token t = TOKEN_ILLEGAL;

	lexer_skip_whitespace(l);

	switch ( l->c ) {
	case '=':
		if ( lexer_peek_char(l) == '=' ) {
			t = token_create(TT_EQ, _LEXER_2_CHAR);
			lexer_read_char(l);
			goto aft_swt;
		}
		t = token_create(TT_ASSIGN, _LEXER_CUR_CHAR);
		break;
	case ';':  t = token_create(TT_SEMICOLON, _LEXER_CUR_CHAR); break;
	case '(':  t = token_create(TT_LPAREN,    _LEXER_CUR_CHAR); break;
	case ')':  t = token_create(TT_RPAREN,    _LEXER_CUR_CHAR); break;
	case '!':
		if ( lexer_peek_char(l) == '=' ) {
			t = token_create(TT_NOT_EQ, _LEXER_2_CHAR);
			lexer_read_char(l);
			goto aft_swt;
		}
		t = token_create(TT_BANG,      _LEXER_CUR_CHAR);
		break;
	case '/':  t = token_create(TT_SLASH,     _LEXER_CUR_CHAR); break;
	case ',':  t = token_create(TT_COMMA,     _LEXER_CUR_CHAR); break;
	case '*':  t = token_create(TT_ASTERISK,  _LEXER_CUR_CHAR); break;
	case '-':  t = token_create(TT_DASH,      _LEXER_CUR_CHAR); break;
	case '.':  t = token_create(TT_DOT,       _LEXER_CUR_CHAR); break;
	case '+':  t = token_create(TT_PLUS,      _LEXER_CUR_CHAR); break;
	case '{':  t = token_create(TT_LBRACE,    _LEXER_CUR_CHAR); break;
	case '}':  t = token_create(TT_RBRACE,    _LEXER_CUR_CHAR); break;
	case '<':  t = token_create(TT_LABRACKET, _LEXER_CUR_CHAR); break;
	case '>':  t = token_create(TT_RABRACKET, _LEXER_CUR_CHAR); break;
	case '\'': t = token_create(TT_SQUOTE,    _LEXER_CUR_CHAR); break;
	case '#':
		lexer_read_char(l);
		if ( _lexer_is_letter(l->c) ) {
			t = lexer_handle_pp(l, err);
			if ( t.typ.code == TT_ILLEGAL ) {
				goto ret_invalid;
			}
			goto ret_ok;
		}
		_lexer_set_err(err, LEXER_ERR_INVALID_PP_IDENT);
		goto ret_invalid;
		break;
	case '"':
		t = token_create(TT_STR_LIT, lexer_read_str_lit(l));
		goto ret_ok;
		break;
	case '\0': t = token_create(TT_EOF, STR_EMPTY); break;
	default: 
		if ( _lexer_is_letter(l->c) ) {
			struct str ident = lexer_read_ident(l);
			t = token_create(lexer_lookup_ident(ident), ident);
			goto ret_ok;
		}
		if ( _lexer_is_number(l->c) ) {
			t = token_create(TT_INT_LIT, lexer_read_int_lit(l));
			goto ret_ok;
		}
		break;
	}
aft_swt:
	lexer_read_char(l);

ret_ok:
	_lexer_set_err(err, LEXER_ERR_OK);
	return t;
ret_invalid:
	return TOKEN_ILLEGAL;

#undef _LEXER_2_CHAR
#undef _LEXER_CUR_CHAR
}

struct str 
lexer_read_ident(struct lexer *l)
{
	size_t pos = l->pos;
	while ( _lexer_is_letter(l->c) ) {
		lexer_read_char(l);
	}
	return str_slice(l->in, pos, l->pos);
}

struct str
lexer_read_str_lit(struct lexer *l)
{
	size_t pos = l->pos;
	lexer_read_char(l);
loop:
	switch ( l->c ) {
	case '"':
	case '\0':
		break;

	default:
		lexer_read_char(l);
		goto loop;
		break;
	}

	lexer_read_char(l);
	return str_slice(l->in, pos+1, l->pos-1);
}

struct str
lexer_read_int_lit(struct lexer *l)
{
	size_t pos = l->pos;
	while ( _lexer_is_number(l->c) ) {
		lexer_read_char(l);
	}
	return str_slice(l->in, pos, l->pos);
	
}

struct str
lexer_read_until_or(struct lexer *l, char tgt, char limit, bool *reached_limit)
{
	size_t pos = l->pos;
loop:
	if ( l->c == tgt ) {
		goto ret_ok;
	}
	if ( l->c == limit || l->c == '\0' ) {
		goto ret_err;
	}
	lexer_read_char(l);
	goto loop;

ret_ok:
	if ( reached_limit != NULL ) {
		*reached_limit = false;
	}
	lexer_read_char(l);
	return str_slice(l->in, pos, l->pos-1);
ret_err:
	if ( reached_limit != NULL ) {
		*reached_limit = true;
	}
	return STR_EMPTY;
}

enum token_type_enum
lexer_lookup_ident(struct str ident)
{
	if ( ident.size < 2 ) {
		return TT_IDENT;
	}

	switch ( ident.data[0] ) {
	case 'i':
		if ( str_eq_cstr(ident, "if", 2) ) {
			return TT_IF;
		}
		if ( str_eq_cstr(ident, "int", 3) ) {
			return TT_TYPE;
		}
		break;
	case 's':
		if ( str_eq_cstr(ident, "switch", 6) ) {
			return TT_SWITCH;
		}
		if ( str_eq_cstr(ident, "struct", 6) ) {
			return TT_SWITCH;
		}
		break;
	case 'c':
		if ( str_eq_cstr(ident, "char", 4) ) {
			return TT_TYPE;
		}
		if ( str_eq_cstr(ident, "case", 4) ) {
			return TT_CASE;
		}
		if ( str_eq_cstr(ident, "const", 5) ) {
			return TT_CONST;
		}
		if ( str_eq_cstr(ident, "continue", 8) ) {
			return TT_CONST;
		}
		break;
	case 'r':
		if ( str_eq_cstr(ident, "return", 6) ) {
			return TT_RETURN;
		}
		break;
	case 'e':
		if ( str_eq_cstr(ident, "else", 4) ) {
			return TT_ELSE;
		}
		if ( str_eq_cstr(ident, "enum", 4) ) {
			return TT_ENUM;
		}
		break;
	case 'f':
		if ( str_eq_cstr(ident, "for", 3) ) {
			return TT_FOR;
		}
		break;
	case 'b':
		if ( str_eq_cstr(ident, "break", 5) ) {
			return TT_BREAK;
		}
		break;
	case 'd':
		if ( str_eq_cstr(ident, "do", 2) ) {
			return TT_WHILE;
		}
		break;
	case 'w':
		if ( str_eq_cstr(ident, "while", 5) ) {
			return TT_WHILE;
		}
		break;
	}


	return TT_IDENT;
}


void 
lexer_skip_whitespace(struct lexer *l)
{
loop:
	switch ( l->c ) {
	case ' ':
	case '\t':
	case '\r':
	case '\n':
		lexer_read_char(l);
		goto loop;
	}

	return;
}


struct token 
lexer_handle_pp(struct lexer *l, struct lexer_err *err)
{
	struct token t = TOKEN_ILLEGAL;
	struct str ident = lexer_read_ident(l);
	enum token_type_enum tt = _lexer_lookup_pp(ident);
	bool reached_limit = false;

	if ( tt == TT_ILLEGAL ) {
		_lexer_set_err(err, LEXER_ERR_INVALID_PP_IDENT);
		goto ret_illegal;
	}

	lexer_skip_whitespace(l);

	switch ( tt ) {
	case TT_PP_INCLUDE:
		switch ( l->c ) {
		case '<':
			lexer_read_char(l);
			ident = lexer_read_until_or(l, '>', '\n', &reached_limit);
			if ( reached_limit == true ) {
				_lexer_set_err(err, LEXER_ERR_INVALID_PP_INCLUDE);
				goto ret_illegal;
			}
			t = token_create(tt, ident);
			goto ret_ok;
			break;
		case '"':
			lexer_read_char(l);
			ident = lexer_read_until_or(l, '"', '\n', &reached_limit);
			if ( reached_limit == true ) {
				_lexer_set_err(err, LEXER_ERR_INVALID_PP_INCLUDE);
				goto ret_illegal;
			}
			t = token_create(tt, ident);
			goto ret_ok;
			break;
		default: goto ret_illegal; break;
		}
		goto ret_ok;
		break;

	case TT_PP_DEFINE:
		ident = lexer_read_until_or(l, '\n', '\\', &reached_limit);
		if ( reached_limit == true ) {
			goto ret_illegal;	
		}

		t = token_create(tt, ident);
		goto ret_ok;
		break;
	case TT_PP_IFNDEF: goto ret_illegal; break;
	case TT_PP_IFDEF: goto ret_illegal; break;
	case TT_PP_IF: goto ret_illegal; break;

	default: goto ret_illegal; break;
	}

ret_ok:
	_lexer_set_err(err, LEXER_ERR_OK);
	return t;

ret_illegal:
	return TOKEN_ILLEGAL;
}

enum token_type_enum
_lexer_lookup_pp(struct str ident)
{
	if ( ident.size < 2 ) {
		return TT_ILLEGAL;
	}

	switch ( ident.size ) {
	case 7:
		if ( str_eq_cstr(ident, "include", 7) ) {
			return TT_PP_INCLUDE;
		}
		break;
	case 6:
		if ( str_eq_cstr(ident, "define", 6) ) {
			return TT_PP_DEFINE;
		}
		if ( str_eq_cstr(ident, "ifndef", 6) ) {
			return TT_PP_IFNDEF;
		}
		break;
	case 5:
		if ( str_eq_cstr(ident, "ifdef", 5) ) {
			return TT_PP_IFDEF;
		}
		break;
	case 2:
		if ( str_eq_cstr(ident, "if", 2) ) {
			return TT_PP_IF;
		}
		break;
	}

	return TT_ILLEGAL;
}

struct lexer_err
lexer_err_create(enum lexer_err_code code)
{
#define _LEXER_ERR_CASE(it) case it: le.name = #it; break;
	struct lexer_err le = {0};
	le.code = code;

	switch ( code ) {
	_LEXER_ERR_CASE(LEXER_ERR_OK);
	_LEXER_ERR_CASE(LEXER_ERR_INVALID_PP_IDENT);
	_LEXER_ERR_CASE(LEXER_ERR_INVALID_PP_INCLUDE);
	}

	return le;
#undef _LEXER_ERR_CASE
}

bool
_lexer_is_letter(char c)
{
	return ( c >= 0x41 && c <= 0x5A ) \
	       || ( c >= 0x61 && c <= 0x7A ) \
	       || c == 0x5F;
}

bool
_lexer_is_number(char c)
{
	return ( c >= '0' && c <= '9' );
}

void
_lexer_set_err(struct lexer_err *err, enum lexer_err_code code)
{
	if ( err == NULL ) {
		return;
	}

	*err = lexer_err_create(code);
}

#endif /* defined(IMP) || defined(LEXER_IMP) */

#endif /* LEXER_H */