6 files changed, 178 insertions(+), 72 deletions(-)

A c_stage1.c
M cc.h
M cc_globals.c
M cc_globals.h
M cc_reader.c
M cpp.c

@@ 0,0 1,118 @@
+/* Copyright (C) 2016, 2021 Jeremiah Orians
+ * Copyright (C) 2020 deesix <deesix@tuta.io>
+ * Copyright (C) 2020 Gabriel Wicki
+ * This file is part of M2-Planet.
+ *
+ * M2-Planet 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 3 of the License, or
+ * (at your option) any later version.
+ *
+ * M2-Planet 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 M2-Planet.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "cc.h"
+#include "cc_globals.h"
+
+int strtoint(char *a);
+struct token_list* new_token(char* s, int size, char* filename, int linenum, struct token_list* prev, struct token_list* next);
+struct token_list* reverse_list(struct token_list* head);
+void dump_stage_raw();
+void dump_stage_straight();
+
+struct pattern* new_pattern(char* s, struct pattern* next)
+{
+	struct pattern* i = calloc(1, sizeof(struct pattern));
+	i->s = s;
+	i->next = next;
+	return i;
+}
+
+int pattern_compress(struct token_list* i, struct pattern* p)
+{
+	if(NULL == p) return TRUE;
+	if(NULL == i) return FALSE;
+	if(!match(p->s, i->s)) return FALSE;
+	if(!pattern_compress(i->next, p->next)) return FALSE;
+	if(NULL == p->next) return TRUE;
+	struct token_list* hold = i->next;
+	i->next = i->next->next;
+	free(hold);
+	return TRUE;
+}
+
+void combine_common()
+{
+	struct token_list* i = global_token;
+
+	struct pattern* decrement = new_pattern("-", new_pattern("-", NULL));
+	struct pattern* increment = new_pattern("+", new_pattern("+", NULL));
+	struct pattern* addassign = new_pattern("+", new_pattern("=", NULL));
+	struct pattern* subassign = new_pattern("-", new_pattern("=", NULL));
+	struct pattern* mulassign = new_pattern("*", new_pattern("=", NULL));
+	struct pattern* divassign = new_pattern("/", new_pattern("=", NULL));
+	struct pattern* modassign = new_pattern("%", new_pattern("=", NULL));
+	struct pattern* equality = new_pattern("=", new_pattern("=", NULL));
+	struct pattern* inequality = new_pattern("!", new_pattern("=", NULL));
+	struct pattern* GTE = new_pattern(">", new_pattern("=", NULL));
+	struct pattern* LTE = new_pattern("<", new_pattern("=", NULL));
+	struct pattern* logicaland = new_pattern("&", new_pattern("&", NULL));
+	struct pattern* logicalor = new_pattern("|", new_pattern("|", NULL));
+	struct pattern* shiftleft = new_pattern("<", new_pattern("<", NULL));
+	struct pattern* shiftright = new_pattern(">", new_pattern(">", NULL));
+	struct pattern* arrow = new_pattern("-", new_pattern(">", NULL));
+	struct pattern* andassign = new_pattern("&", new_pattern("=", NULL));
+	struct pattern* orassign = new_pattern("|", new_pattern("=", NULL));
+	struct pattern* xorassign = new_pattern("^", new_pattern("=", NULL));
+	struct pattern* leftassign = new_pattern("<", new_pattern("<", new_pattern("=", NULL)));
+	struct pattern* rightassign = new_pattern(">", new_pattern(">", new_pattern("=", NULL)));
+
+	while(NULL != i)
+	{
+		if(NULL == i->next) break;
+
+		if(pattern_compress(i, leftassign)) i->s = "<<=";
+		else if(pattern_compress(i, rightassign)) i->s = ">>=";
+		else if(pattern_compress(i, decrement)) i->s = "--";
+		else if(pattern_compress(i, increment)) i->s = "++";
+		else if(pattern_compress(i, addassign)) i->s = "+=";
+		else if(pattern_compress(i, subassign)) i->s = "-=";
+		else if(pattern_compress(i, mulassign)) i->s = "*=";
+		else if(pattern_compress(i, divassign)) i->s = "/=";
+		else if(pattern_compress(i, modassign)) i->s = "%=";
+		else if(pattern_compress(i, equality)) i->s = "==";
+		else if(pattern_compress(i, inequality)) i->s = "!=";
+		else if(pattern_compress(i, GTE)) i->s = ">=";
+		else if(pattern_compress(i, LTE)) i->s = "<=";
+		else if(pattern_compress(i, andassign)) i->s = "&=";
+		else if(pattern_compress(i, orassign)) i->s = "|=";
+		else if(pattern_compress(i, xorassign)) i->s = "^=";
+		else if(pattern_compress(i, logicaland)) i->s = "&&";
+		else if(pattern_compress(i, logicalor)) i->s = "||";
+		else if(pattern_compress(i, shiftleft)) i->s = "<<";
+		else if(pattern_compress(i, shiftright)) i->s = ">>";
+		else if(pattern_compress(i, arrow)) i->s = "->";
+		else i = i->next;
+	}
+
+}
+
+void stage1_preprocess()
+{
+	if(1 == DUMP_STAGE) dump_stage_raw();
+
+	global_token = reverse_list(global_token);
+
+	if(2 == DUMP_STAGE) dump_stage_raw();
+	if(3 == DUMP_STAGE) dump_stage_straight();
+	combine_common();
+	if(4 == DUMP_STAGE) dump_stage_raw();
+	if(5 == DUMP_STAGE) dump_stage_straight();
+
+}

@@ 20,6 20,8 @@
 #include <stdio.h>
 #include <string.h>
 
+#ifndef _CC_H
+#define _CC_H
 
 #define FALSE 0
 #define TRUE 1


@@ 67,4 69,11 @@ struct token_list
 	};
 };
 
+struct pattern
+{
+	char* s;
+	struct pattern* next;
+};
+
 #include "cc_globals.h"
+#endif

@@ 16,6 16,8 @@
  * along with M2-Planet.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <stdio.h>
+
 /* What types we have */
 struct type* global_types;
 struct type* prim_types;


@@ 30,9 32,9 @@ struct token_list* globals_list;
 
 /* Make our string collection more efficient */
 char* hold_string;
-int string_index;
+size_t string_index;
 
-int MAX_STRING;
+size_t MAX_STRING;
 
 /* enable preprocessor-only mode */
 int PREPROCESSOR_MODE;


@@ 44,7 46,10 @@ int WORDSIZE;
 int ENDIAN;
 char* BASEADDRESS;
 int STDIO_USED;
-char* TEMPDIR;
+
+/* How we deal with output */
+char* DESTINATION_NAME;
+FILE* DESTINATION_FILE;
 
 /* So we don't shoot ourself in the face */
 int FUZZING;

@@ 16,6 16,10 @@
  * along with M2-Planet.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <stdio.h>
+
+#ifndef _CC_GLOBALS_H
+#define _CC_GLOBALS_H
 /* What types we have */
 extern struct type* global_types;
 extern struct type* prim_types;


@@ 30,10 34,10 @@ extern struct token_list* globals_list;
 
 /* Make our string collection more efficient */
 extern char* hold_string;
-extern int string_index;
+extern size_t string_index;
 
 /* Allow us to have a single settable max string */
-extern long MAX_STRING;
+extern size_t MAX_STRING;
 
 /* enable preprocessor-only mode */
 extern int PREPROCESSOR_MODE;


@@ 45,10 49,15 @@ extern int WORDSIZE;
 extern int ENDIAN;
 extern char* BASEADDRESS;
 extern int STDIO_USED;
-extern char* TEMPDIR;
+
+/* How we deal with output */
+extern char* DESTINATION_NAME;
+extern FILE* DESTINATION_FILE;
 
 /* So we don't shoot ourself in the face */
 extern int FUZZING;
 extern int DIRTY_MODE;
 extern int DUMP_STAGE;
 extern int DEBUG_LEVEL;
+
+#endif

@@ 85,7 85,7 @@ int preserve_keyword(int c, char* S)
 
 void clear_string(char* s)
 {
-	int i = 0;
+	size_t i = 0;
 	while(0 != s[i])
 	{
 		s[i] = 0;


@@ 118,8 118,7 @@ struct token_list* eat_token(struct token_list* token)
 	return token->next;
 }
 
-
-void new_token(char* s, int size)
+struct token_list* new_token(char* s, int size, char* filename, int linenum, struct token_list* prev, struct token_list* next)
 {
 	struct token_list* current = calloc(1, sizeof(struct token_list));
 	require(NULL != current, "Exhausted memory while getting token\n");


@@ 129,11 128,11 @@ void new_token(char* s, int size)
 	require(NULL != current->s, "Exhausted memory while trying to copy a token\n");
 	copy_string(current->s, s, MAX_STRING);
 
-	current->prev = token;
-	current->next = token;
-	current->linenumber = line;
-	current->filename = file;
-	token = current;
+	current->prev = prev;
+	current->next = next;
+	current->linenumber = linenum;
+	current->filename = filename;
+	return current;
 }
 
 int get_token(int c)


@@ 176,7 175,7 @@ int get_token(int c)
 			c = consume_byte(c);
 		}
 	}
-	else if(in_set(c, " \a\b\t\n\v\f\r<=>|&!^%*+-"))
+	else if(in_set(c, " \a\b\t\n\v\f\r<=>|&!^%*+-:;"))
 	{
 		c = consume_byte(c);
 	}


@@ 187,20 186,12 @@ int get_token(int c)
 	}
 	else if(in_set(c, "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_"))
 	{
-		c = preserve_keyword(c, "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_:");
+		c = preserve_keyword(c, "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_");
 	}
 	else if(in_set(c, "'\""))
 	{
 		c = preserve_string(c);
 	}
-	else if(c == '*')
-	{
-		c = consume_byte(c);
-		if(c == '=')
-		{
-			c = consume_byte(c);
-		}
-	}
 	else
 	{
 		c = consume_byte(c);


@@ 231,23 222,22 @@ void insert_file_header(char* name, int line)
 	strcat(hold_string, name);
 	strcat(hold_string, " ");
 	strcat(hold_string, hold_line);
-	new_token(hold_string, strlen(hold_string)+2);
-	new_token("\n", 3);
+	token = new_token(hold_string, strlen(hold_string)+2, name, line, token, token);
+	token = new_token("\n", 3, name, line, token, token);
 }
 
 struct token_list* read_all_tokens(FILE* a, struct token_list* current, char* filename)
 {
 	token = current;
 	input  = a;
-	line = 0;
+	line = 1;
 	file = filename;
 	insert_file_header(filename, 1);
-	line = 1;
 	int ch = grab_byte();
 	while(EOF != ch)
 	{
 		ch = get_token(ch);
-		new_token(hold_string, string_index + 2);
+		token = new_token(hold_string, string_index + 2, file, line, token, token);
 	}
 
 	return token;

@@ 17,15 17,13 @@
  * along with M2-Planet.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include"cc.h"
-#include <unistd.h>
+#include "cc.h"
+#include "cc_globals.h"
 
 int strtoint(char *a);
 struct token_list* read_all_tokens(FILE* a, struct token_list* current, char* filename);
-
-char* destination_name;
-FILE* destination_file;
-
+void dump_stage_straight();
+void stage1_preprocess();
 
 void prechecks(int argc, char** argv)
 {


@@ 109,10 107,10 @@ void prechecks(int argc, char** argv)
 		}
 		else if(match(argv[i], "-o") || match(argv[i], "--output"))
 		{
-			destination_name = argv[i + 1];
-			require(NULL != destination_name, "--output option requires a filename to follow\n");
-			destination_file = fopen(destination_name, "w");
-			if(NULL == destination_file)
+			DESTINATION_NAME = argv[i + 1];
+			require(NULL != DESTINATION_NAME, "--output option requires a filename to follow\n");
+			DESTINATION_FILE = fopen(DESTINATION_NAME, "w");
+			if(NULL == DESTINATION_FILE)
 			{
 				fputs("Unable to open for writing file: ", stderr);
 				fputs(argv[i + 1], stderr);


@@ 142,8 140,8 @@ int main(int argc, char** argv)
 	STDIO_USED = FALSE;
 	DIRTY_MODE = FALSE;
 	Architecture = NULL;
-	destination_name = "/dev/stdout";
-	destination_file = stdout;
+	DESTINATION_NAME = "/dev/stdout";
+	DESTINATION_FILE = stdout;
 	global_token = NULL;
 
 	/* Our fun locals */


@@ 157,12 155,12 @@ int main(int argc, char** argv)
 
 	if(0 == DUMP_STAGE)
 	{
-		fputs("{VARIABLE, PURPOSE, VALUE}\n", destination_file);
-		fputs("{DESTINATION_FILE, The target location of where to put the output, ", destination_file);
-		fputs(destination_name, destination_file);
-		fputs("}\nARCHITECTURE, The architectures for which to expand in #IFDEF blobcks, ", destination_file);
-		fputs(Architecture, destination_file);
-		fputs("}\n", destination_file);
+		fputs("{VARIABLE, PURPOSE, VALUE}\n", DESTINATION_FILE);
+		fputs("{DESTINATION_FILE, The target location of where to put the output, ", DESTINATION_FILE);
+		fputs(DESTINATION_NAME, DESTINATION_FILE);
+		fputs("}\nARCHITECTURE, The architectures for which to expand in #IFDEF blobcks, ", DESTINATION_FILE);
+		fputs(Architecture, DESTINATION_FILE);
+		fputs("}\n", DESTINATION_FILE);
 		exit(EXIT_SUCCESS);
 	}
 


@@ 274,30 272,7 @@ int main(int argc, char** argv)
 		}
 	}
 
-	if(1 == DUMP_STAGE)
-	{
-		char* hold;
-		fputs("{token,\n file name,\n line number,\n}\n", destination_file);
-		while(NULL != global_token)
-		{
-			fputs("{ \"", destination_file);
-			if(NULL == global_token) fputs("NULL", destination_file);
-			else if(NULL == global_token->s) fputs(":NULL:", destination_file);
-			else if(match("\n", global_token->s)) fputs("\\n", destination_file);
-			else fputs(global_token->s, destination_file);
-
-			fputs("\",\n ", destination_file);
-			if(NULL == global_token) fputs("::NULL::", destination_file);
-			else if(NULL == global_token->filename) fputs(":::NULL:::", destination_file);
-			else fputs(global_token->filename, destination_file);
-
-			fputs(",\n ", destination_file);
-			hold = int2str(global_token->linenumber, 10, FALSE);
-			if(NULL == hold) fputs("::::NULL::::", destination_file);
-			else fputs(hold, destination_file);
+	if(C_PREPROCESS) stage1_preprocess();
 
-			fputs("\n }\n", destination_file);
-			global_token = global_token->next;
-		}
-	}
+	dump_stage_straight();
 }