Cfff

1.txt

@@ -0,0 +1,2 @@
+const x=8,y=7;var a,b,c;begin a=2*x;;b=a+x+y;;c=a+3*b;end
+const x=8,y=7;var a,b,c;begin a=b+x;;if a>0 then begin c=y -1;;a=a+2;end else begin c=a+y;end;while a>0 do a=a-1;end

CPlab2_词法分析/词法分析.cpp → CPlab2_LexicalAnalysis/LexicalAnalysis.cpp

@@ -7,13 +7,13 @@ using namespace std;
 #define ID 15
 #define NUM 16
 
-//��Ŵ�������ַ��� 
+
 char tempstr[255] = {};
-//�ո��־ 
+
 bool temp = false;
-//��ʱ����
+
 char word[255] = {};
-//keyword�ؼ��� 
+
 string keyword[9] = {
 	"Const", "Var", "if", "else", "then",
 	"while", "do", "begin", "end"
@@ -23,28 +23,29 @@ int keyword_num[9] = {
 	1,2,3,4,5,
 	6,7,8,9 
 };
-//������������������ 
-char symbol[9] = { '+','-','*','/',';','(',')','{','}'};
-//��Ӧ������ֵ
+
+char symbol[9] = { '+', '-', '*', '/', ';', '(' ,')' ,'{' ,'}'};
+
 int symbol_num[9] = { 21,22,23,24,25,26,27,28,29 };
 
-//�ж��Ƿ�Ϊ��ĸ 
+bool err_flag = false;
+
 bool IsLetter(char ch)
 {
-	if ((ch >= 'a'&&ch <= 'z') || (ch >= 'A'&&ch <= 'Z'))
+	if ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z'))
 		return true;
 	return false;
 }
 
-//�ж��Ƿ�Ϊ���� 
+
 bool IsDigit(char ch)
 {
 	if (ch >= '0'&&ch <= '9')
 		return true;
 	return false;
 }
 
-//�ж��Ƿ�Ϊ�������
+
 int IsSymbol(char ch)
 {
 	for (int i = 0; i<9; i++)
@@ -55,7 +56,7 @@ int IsSymbol(char ch)
 	return -1;
 }
 
-//�ж��Ƿ�Ϊ�ؼ��� 
+
 int IsKeyword(string str)
 {
 	for (int i = 0; i<26; i++)
@@ -65,15 +66,15 @@ int IsKeyword(string str)
 			return i;
 		}
 	}
-	//���ǹؼ��ּ�ΪID
+
 	return ID;
 }
 
-//�ո���
+
 void HandleSpace(char a[])
 {
 	int j = 0;
-	memset(word, 0, 255);//��Ҫ��գ���Ȼ���ܲ����ϴε��ַ���
+	memset(word, 0, 255);
 	temp = false;
 	for (int i = 0; i < strlen(a); i++)
 	{
@@ -94,7 +95,7 @@ void HandleSpace(char a[])
 	}
 }
 
-//����"//"ע��
+
 void prePro()
 {
 	int j = 0;
@@ -119,45 +120,48 @@ void Scanner(char *str);
 
 int main()
 {
-	char instr[255] = {}; //���������ַ��� 
-	bool flag = false; //����ע�ͱ�־,falseΪδ����ע������ 
-	string Token;//����ַ��� 
-	char *str = NULL;//���ÿ�е��ַ��� 
-	char delims[] = " ";//�ָ��־ 
+	char instr[255] = {};
+	bool flag = false;  
+	string Token; 
+	char *str = NULL;
+	char delims[] = " ";
 	freopen("test.cpp", "r", stdin);
 	freopen("result.txt", "w", stdout); 
 
 	printf("�ʷ�����������£�\n");
 	printf("******************************\n");
 
-	while ((gets_s(instr)) != NULL)
+	while ((gets(instr)) != NULL)
 	{
 		HandleSpace(instr);
 		prePro();
 
-		str = strtok(tempstr, delims);//�ָ��ַ��� 
+		str = strtok(tempstr, delims);
 
 		while (str != NULL)
 		{
-			//ͷ�ļ����궨��
+
 			if (*(str) == '#')
 			{
 				printf("#\n");
 				break;
 			}
 
 			Scanner(str);
-
+			if (err_flag) {
+				break;
+			}
 			str = strtok(NULL, delims);
 		}
+		if(err_flag) break;
 	}
 
 	return 0;
 }
 
 void Scanner(char *str) {
-	bool flag = false; //����ע�ͱ�־,falseΪδ����ע������ 
-	string Token;//����ַ��� 
+	bool flag = false; 
+	string Token; 
 	int type = 0;
 
 	for (int i = 0; i<strlen(str); i++)
@@ -170,8 +174,8 @@ void Scanner(char *str) {
 				break;
 			}
 		}
-		//ע�ʹ���: */,ע��������� 
-		if (*(str + i) == '*'&&flag)
+
+		if (*(str + i) == '*' && flag)
 		{
 			if (*(str + i + 1) == '/')
 			{
@@ -181,7 +185,6 @@ void Scanner(char *str) {
 			}
 		}
 
-		//��ʶ�����ؼ��� 
 		if (IsLetter(*(str + i)) && (!flag))
 		{
 			while (IsLetter(*(str + i)) || IsDigit(*(str + i)) || *(str + i) == '_')
@@ -204,14 +207,28 @@ void Scanner(char *str) {
 
 		if (IsDigit(*(str + i)) && (!flag))
 		{
+
 			while (IsDigit(*(str + i)))
 			{
 				Token += *(str + i);
+				if(IsLetter(*(str + i + 1))) {
+					int tempIndex = 1;
+					while(IsLetter(*(str + i + tempIndex))) {
+						Token += *(str + i + tempIndex);
+						tempIndex += 1;
+					}
+					printf("<error: %s wrong identifier>", Token.c_str());
+					Token = "";
+					err_flag = true;
+					break;
+				}
 				i++;
 			}
+			if (err_flag) break;
 			type = NUM;
 			printf("< %s , %d >\n", Token.c_str(), type);
 			Token = "";
+
 		}
 
 		//<,<=,<>

CPlab2_LexicalAnalysis/LexicalAnalysis.py

@@ -0,0 +1,94 @@
+def scan(string):
+    KEY_WORDS = ["Const", "Var", "if", "else", "then","while", "do", "begin", "end"]
+    SYMBOLS = [',', '+', '-', '*', '/', ';', '(' ,')' ,'{' ,'}', '==', '<=', '>=', '<>', '<', '>', '=']
+    word_map = {key: i+1 for i, key in enumerate(KEY_WORDS)}
+    word_map.update({key: i+21 for i, key in enumerate(SYMBOLS)})
+    def process(string):
+        word = ''
+        if ' ' in string:
+            for s in string.split(' '):
+                word += s
+        else:
+            word = string
+        if '//' in word:
+            return word.split('//')[0]
+        return word
+
+    def have_keywords(word):
+        for keyword in KEY_WORDS:
+            if keyword in word:
+                return keyword
+        return None
+
+    def have_symbol(word):
+        for sy in SYMBOLS:
+            if sy in word:
+                return sy
+        return None
+
+    def is_alpha(ch):
+        return (ch >= 'a' and ch <= 'z') or (ch >= 'A' and ch <= 'Z')
+
+    def is_digit(ch):
+        return ch >= '0' and ch <= '9'
+
+    # def is_in_symbols(c):
+    #     return c in SYMBOLS
+
+    # def is_keywords(c):
+    #     return c in KEY_WORDS
+
+
+    queue = [process(string)]
+    if '/*' in queue[0] or '*/' in queue[0]:
+        return 
+    stop = False
+    while not stop:
+        changed = False
+        tempQueue = []
+        for item in queue:
+            key = have_keywords(item)
+            sy = have_symbol(item)
+            if key and key != item:
+                changed = True
+                item1, item2 = item.split(key)
+                if item1:
+                    tempQueue.append(item1)
+                tempQueue.append(key)
+                if item2:
+                    tempQueue.append(item2)
+                continue
+            elif sy and sy != item:
+                changed = True
+                sp = item.split(sy)
+                for i, it in enumerate(sp):
+                    if it:
+                        tempQueue.append(it)
+                    if i != len(sp) - 1:
+                        tempQueue.append(sy)
+
+            else:
+                tempQueue.append(item)
+
+            if item == queue[-1] and not changed:
+                stop = True
+        queue = tempQueue.copy()
+    for item in queue:
+        if item in word_map:
+            print("<", item, ",", word_map[item],">")
+        else:
+            if item and is_alpha(item[0]):
+                print("<", item, ",", 15,">")
+            if item and is_digit(item[0]):
+                print("<", item, ",", 16,">")
+
+def main():
+    with open('./test.cpp', 'r') as r_stream:
+        for line in r_stream:
+            if line:
+                if '\n' in line:
+                    scan(line[0:-1])
+                else:
+                    scan(line)
+
+main()

CPlab2_LexicalAnalysis/result.txt

@@ -0,0 +1,26 @@
+�ʷ�����������£�
+******************************
+#
+< int , 15 >
+< main , 15 >
+< ( , 26 >
+< ) , 27 >
+< { , 28 >
+< Const , 1 >
+< const_a , 15 >
+< = , 35 >
+< 10 , 16 >
+< ; , 25 >
+< Var , 2 >
+< var_b , 15 >
+< = , 35 >
+< 15 , 16 >
+< ; , 25 >
+< if , 3 >
+< ( , 26 >
+< const_a , 15 >
+< > , 33 >
+< 5 , 16 >
+< ) , 27 >
+< { , 28 >
+

CPlab2_LexicalAnalysis/test.cpp

@@ -0,0 +1,19 @@
+Const x=8,y=7;
+Var a,b,c;
+begin
+   a=b+x;
+
+    if a>0
+       then 
+       begin
+         c=y -;
+         a=a+2;
+       end
+    else
+       begi
+         c=a+y;
+       end
+
+    while a>0
+      do  a=a-1;
+end

CPlab2_LexicalAnalysis/test.py

@@ -0,0 +1,3 @@
+a = 'adsd;'
+if 'adsd;' in a:
+    print('vfs')