some refactoring in preparation for a preprocessor
[cparser] / lexer.c
diff --git a/lexer.c b/lexer.c
index 8cf2d71..3ba2c37 100644 (file)
--- a/lexer.c
+++ b/lexer.c
@@ -1,7 +1,27 @@
+/*
+ * This file is part of cparser.
+ * Copyright (C) 2007-2008 Matthias Braun <matze@braunis.de>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ * 02111-1307, USA.
+ */
 #include <config.h>
 
 #include "diagnostic.h"
 #include "lexer.h"
+#include "symbol_t.h"
 #include "token_t.h"
 #include "symbol_table_t.h"
 #include "adt/error.h"
 #define strtold(s, e) strtod(s, e)
 #endif
 
-#if defined HAS_SIGNED_CHAR
-typedef signed char char_type;
-#elif defined HAS_UNSIGNED_CHAR
-typedef unsigned char char_type;
-#else
-#      error signedness of char not determined
-#endif
-
 static int         c;
 token_t            lexer_token;
 symbol_t          *symbol_L;
@@ -686,7 +698,12 @@ static int parse_octal_sequence(const int first_digit)
        if (!is_octal_digit(c)) return value;
        value = 8 * value + digit_value(c);
        next_char();
-       return (char_type)value;
+
+       if(char_is_signed) {
+               return (signed char) value;
+       } else {
+               return (unsigned char) value;
+       }
 }
 
 /**
@@ -700,7 +717,11 @@ static int parse_hex_sequence(void)
                next_char();
        }
 
-       return (char_type)value;
+       if(char_is_signed) {
+               return (signed char) value;
+       } else {
+               return (unsigned char) value;
+       }
 }
 
 /**
@@ -886,14 +907,17 @@ end_of_string:
  */
 static void parse_wide_character_constant(void)
 {
+       const unsigned start_linenr = lexer_token.source_position.linenr;
+
        eat('\'');
 
-       int found_char = 0;
        while(1) {
                switch(c) {
-               case '\\':
-                       found_char = parse_escape_sequence();
+               case '\\': {
+                       wchar_rep_t tc = parse_escape_sequence();
+                       obstack_grow(&symbol_obstack, &tc, sizeof(tc));
                        break;
+               }
 
                MATCH_NEWLINE(
                        parse_error("newline while parsing character constant");
@@ -904,28 +928,31 @@ static void parse_wide_character_constant(void)
                        next_char();
                        goto end_of_wide_char_constant;
 
-               case EOF:
-                       parse_error("EOF while parsing character constant");
+               case EOF: {
+                       source_position_t source_position = lexer_token.source_position;
+                       source_position.linenr = start_linenr;
+                       errorf(source_position, "EOF while parsing character constant");
                        lexer_token.type = T_ERROR;
                        return;
+               }
 
-               default:
-                       if(found_char != 0) {
-                               parse_error("more than 1 characters in character "
-                                           "constant");
-                               goto end_of_wide_char_constant;
-                       } else {
-                               found_char = c;
-                               next_char();
-                       }
+               default: {
+                       wchar_rep_t tc = (wchar_rep_t) c;
+                       obstack_grow(&symbol_obstack, &tc, sizeof(tc));
+                       next_char();
                        break;
                }
+               }
        }
 
-end_of_wide_char_constant:
-       lexer_token.type       = T_INTEGER;
-       lexer_token.v.intvalue = found_char;
-       lexer_token.datatype   = type_wchar_t;
+end_of_wide_char_constant:;
+       size_t             size   = (size_t) obstack_object_size(&symbol_obstack);
+       const wchar_rep_t *string = obstack_finish(&symbol_obstack);
+
+       lexer_token.type                = T_WIDE_CHARACTER_CONSTANT;
+       lexer_token.v.wide_string.begin = string;
+       lexer_token.v.wide_string.size  = size;
+       lexer_token.datatype            = type_wchar_t;
 }
 
 /**
@@ -1002,19 +1029,23 @@ static void parse_character_constant(void)
 
        eat('\'');
 
-       int tc;
        while(1) {
                switch(c) {
-               case '\\':
-                       tc = parse_escape_sequence();
+               case '\\': {
+                       int tc = parse_escape_sequence();
                        obstack_1grow(&symbol_obstack, (char) tc);
                        break;
+               }
 
                MATCH_NEWLINE(
                        parse_error("newline while parsing character constant");
                        break;
                )
 
+               case '\'':
+                       next_char();
+                       goto end_of_char_constant;
+
                case EOF: {
                        source_position_t source_position;
                        source_position.input_name = lexer_token.source_position.input_name;
@@ -1024,10 +1055,6 @@ static void parse_character_constant(void)
                        return;
                }
 
-               case '\'':
-                       next_char();
-                       goto end_of_char_constant;
-
                default:
                        obstack_1grow(&symbol_obstack, (char) c);
                        next_char();
@@ -1040,7 +1067,7 @@ end_of_char_constant:;
        const size_t      size   = (size_t)obstack_object_size(&symbol_obstack);
        const char *const string = obstack_finish(&symbol_obstack);
 
-       lexer_token.type           = T_CHARS;
+       lexer_token.type           = T_CHARACTER_CONSTANT;
        lexer_token.v.string.begin = string;
        lexer_token.v.string.size  = size;
        lexer_token.datatype       = type_int;
@@ -1453,25 +1480,25 @@ void lexer_next_preprocessing_token(void)
                        ELSE('/')
                case '%':
                        MAYBE_PROLOG
-                       MAYBE('>', T_PERCENTGREATER)
+                       MAYBE('>', '}')
                        MAYBE('=', T_PERCENTEQUAL)
                                case ':':
                                        MAYBE_PROLOG
                                                case '%':
                                                        MAYBE_PROLOG
-                                                       MAYBE(':', T_PERCENTCOLONPERCENTCOLON)
+                                                       MAYBE(':', T_HASHHASH)
                                                        ELSE_CODE(
                                                                put_back(c);
                                                                c = '%';
-                                                               lexer_token.type = T_PERCENTCOLON;
+                                                               lexer_token.type = '#';
                                                                return;
                                                        )
-                                       ELSE(T_PERCENTCOLON)
+                                       ELSE('#')
                        ELSE('%')
                case '<':
                        MAYBE_PROLOG
-                       MAYBE(':', T_LESSCOLON)
-                       MAYBE('%', T_LESSPERCENT)
+                       MAYBE(':', '[')
+                       MAYBE('%', '{')
                        MAYBE('=', T_LESSEQUAL)
                                case '<':
                                        MAYBE_PROLOG
@@ -1497,7 +1524,7 @@ void lexer_next_preprocessing_token(void)
                        ELSE('|')
                case ':':
                        MAYBE_PROLOG
-                       MAYBE('>', T_COLONGREATER)
+                       MAYBE('>', ']')
                        ELSE(':')
                case '=':
                        MAYBE_PROLOG