1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/jitcraft/Scanner.cpp Mon Jun 25 02:57:28 2012 -0700
1.3 @@ -0,0 +1,227 @@
1.4 +/*
1.5 + * Copyright (C) 2012 David Anderson
1.6 + *
1.7 + * This file is part of JITCraft.
1.8 + *
1.9 + * JITCraft is free software: you can redistribute it and/or modify it under
1.10 + * the terms of the GNU General Public License as published by the Free
1.11 + * Software Foundation, either version 3 of the License, or (at your option)
1.12 + * any later version.
1.13 + *
1.14 + * Foobar is distributed in the hope that it will be useful, but WITHOUT ANY
1.15 + * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
1.16 + * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
1.17 + *
1.18 + * You should have received a copy of the GNU General Public License along with
1.19 + * JITCraft. If not, see http://www.gnu.org/licenses/.
1.20 + */
1.21 +#include <assert.h>
1.22 +#include <ctype.h>
1.23 +#include "Scanner.h"
1.24 +
1.25 +using namespace jc;
1.26 +
1.27 +const char *jc::TokenNames[] =
1.28 +{
1.29 +#define _(name, str) str,
1.30 + TOKENMAP(_)
1.31 +#undef _
1.32 + NULL
1.33 +};
1.34 +
1.35 +Scanner::Scanner(const char *stream, size_t length)
1.36 + : stream_(stream),
1.37 + length_(length),
1.38 + pushed_(false),
1.39 + saved_('\0'),
1.40 + pos_(stream_)
1.41 +{
1.42 +}
1.43 +
1.44 +char
1.45 +Scanner::getChar()
1.46 +{
1.47 + if (saved_ != '\0') {
1.48 + char temp = saved_;
1.49 + saved_ = '\0';
1.50 + return temp;
1.51 + }
1.52 + if (pos_ >= stream_ + length_)
1.53 + return '\0';
1.54 + return *pos_++;
1.55 +}
1.56 +
1.57 +void
1.58 +Scanner::putChar(char c)
1.59 +{
1.60 + assert(saved_ == '\0');
1.61 + saved_ = c;
1.62 +}
1.63 +
1.64 +void
1.65 +Scanner::eatSingleLine()
1.66 +{
1.67 + char c;
1.68 + while ((c = getChar()) != '\0') {
1.69 + if (c == '\n' || c == '\r')
1.70 + break;
1.71 + }
1.72 +}
1.73 +
1.74 +bool
1.75 +Scanner::eatComment()
1.76 +{
1.77 + char c;
1.78 +
1.79 + while ((c = getChar()) != '\0') {
1.80 + while (c == '*') {
1.81 + c = getChar();
1.82 + if (c == '\0')
1.83 + return false;
1.84 + if (c == '/')
1.85 + return true;
1.86 + }
1.87 + }
1.88 +
1.89 + return false;
1.90 +}
1.91 +
1.92 +TokenKind
1.93 +Scanner::integerLiteral(char first)
1.94 +{
1.95 + literal_ = first;
1.96 +
1.97 + for (;;) {
1.98 + char c = getChar();
1.99 + if (c < '0' || c > '9') {
1.100 + putChar(c);
1.101 + break;
1.102 + }
1.103 + literal_ += c;
1.104 + }
1.105 +
1.106 + return TOK_INTEGER_LITERAL;
1.107 +}
1.108 +
1.109 +TokenKind
1.110 +Scanner::identifier(char first)
1.111 +{
1.112 + literal_ = first;
1.113 +
1.114 + for (;;) {
1.115 + char c = getChar();
1.116 + if (!isascii(c) || (!isalnum(c) && c != '_')) {
1.117 + putChar(c);
1.118 + break;
1.119 + }
1.120 + literal_ += c;
1.121 + }
1.122 +
1.123 + if (literal_ == "new")
1.124 + return TOK_NEW;
1.125 + if (literal_ == "for")
1.126 + return TOK_FOR;
1.127 + if (literal_ == "return")
1.128 + return TOK_RETURN;
1.129 + if (literal_ == "if")
1.130 + return TOK_IF;
1.131 + if (literal_ == "else")
1.132 + return TOK_ELSE;
1.133 + return TOK_NAME;
1.134 +}
1.135 +
1.136 +TokenKind
1.137 +Scanner::next()
1.138 +{
1.139 + assert(!pushed_);
1.140 +
1.141 + for (;;) {
1.142 + char c = getChar();
1.143 + switch (c) {
1.144 + case '/':
1.145 + {
1.146 + c = getChar();
1.147 + if (c == '/') {
1.148 + eatSingleLine();
1.149 + break;
1.150 + }
1.151 + if (c == '*') {
1.152 + if (!eatComment())
1.153 + return TOK_ERROR;
1.154 + break;
1.155 + }
1.156 + putChar(c);
1.157 + return TOK_SLASH;
1.158 + }
1.159 +
1.160 + case ';': return TOK_SEMICOLON;
1.161 + case '{': return TOK_LBRACE;
1.162 + case '}': return TOK_RBRACE;
1.163 + case '(': return TOK_LPAREN;
1.164 + case ')': return TOK_RPAREN;
1.165 + case '[': return TOK_LBRACKET;
1.166 + case ']': return TOK_RBRACKET;
1.167 + case '*': return TOK_STAR;
1.168 + case '+': return TOK_PLUS;
1.169 + case '-': return TOK_MINUS;
1.170 + case ',': return TOK_COMMA;
1.171 + case '\0': return TOK_EOF;
1.172 +
1.173 + case '=':
1.174 + c = getChar();
1.175 + if (c == '=')
1.176 + return TOK_EQUALS;
1.177 + putChar(c);
1.178 + return TOK_ASSIGN;
1.179 +
1.180 + case '<':
1.181 + c = getChar();
1.182 + if (c == '=')
1.183 + return TOK_LE;
1.184 + putChar(c);
1.185 + return TOK_LT;
1.186 +
1.187 + case '>':
1.188 + c = getChar();
1.189 + if (c == '=')
1.190 + return TOK_GE;
1.191 + putChar(c);
1.192 + return TOK_GT;
1.193 +
1.194 + case ' ':
1.195 + case '\t':
1.196 + case '\r':
1.197 + case '\f':
1.198 + case '\n':
1.199 + break;
1.200 +
1.201 + default:
1.202 + if (c >= '0' && c <= '9')
1.203 + return integerLiteral(c);
1.204 + if (c == '_' || (isascii(c) && isalpha(c)))
1.205 + return identifier(c);
1.206 + return TOK_ERROR;
1.207 + }
1.208 + }
1.209 +}
1.210 +
1.211 +TokenKind
1.212 +Scanner::scan()
1.213 +{
1.214 + if (pushed_) {
1.215 + pushed_ = false;
1.216 + return lastToken_;
1.217 + }
1.218 + lastToken_ = next();
1.219 + return lastToken_;
1.220 +}
1.221 +
1.222 +TokenKind
1.223 +Scanner::peek()
1.224 +{
1.225 + if (pushed_)
1.226 + return lastToken_;
1.227 + scan();
1.228 + pushed_ = true;
1.229 + return lastToken_;
1.230 +}
