编译原理-词法分析器

#include<cstdio>
#include<cstring>
#include<map>
using namespace std;
typedef long long ll;
typedef unsigned long long ull;
const int N=35;
const char WORD[51][101]={
	"","PROGRAM","CONST","VAR","INTEGER","LONG","PROCEDURE","IF","THEN","WHILE","DO","READ","WRITE","BEGIN","END","ODD",
	"+","-","*","/","=","<>","<","<=",">",">=",",",".",";",":",":=","(",")","","","#"
};
char s[10010];
map <ull,int> mp;
int cnt_var,len;

const ull seed=13131;
ull gethash(int l,int r){
	ull re=0;
	for (int i=l;i<=r;i++){
		re=re*seed+s[i]-'0';
	}
	return re;
}

void print(int pos,int l=0,int r=0){
	if (pos==34){
		printf("<34,");
		for (int i=l;i<=r;i++) printf("%c",s[i]);
		printf(">\n");
	}
	else printf("<%d,%s>\n",pos,WORD[pos]);
}


void Divide_Expression(){
	int l=1;
	for (int i=1;i<=len;i++){
		if (s[i]=='+'){
			if (i-1>0) print(34,l,i-1);
			print(16);
			l=i+1;
		}else if (s[i]=='-'){
			if (i-1>0) print(34,l,i-1);
			print(17);
			l=i+1;
		}else if (s[i]=='*'){
			if (i-1>0) print(34,l,i-1);
			print(18);
			l=i+1;
		}else if (s[i]=='/'){
			if (i-1>0) print(34,l,i-1);
			print(19);
			l=i+1;
		}else if (s[i]=='='){
			if (i-1>0) print(34,l,i-1);
			print(20);
			l=i+1;
		}else if (s[i]=='<'){
			if (i-1>0) print(34,l,i-1);
			if (s[i+1]=='>') print(21),i++;
			else if (s[i+1]=='=') print(23),i++;
			else print(22);
			l=i+1;
		}else if (s[i]=='>'){
			if (i-1>0) print(34,l,i-1);
			if (s[i+1]=='=') print(25),i++;
			else print(24);
			l=i+1;
		}else if (s[i]=='.'){
			if (i-1>0) print(34,l,i-1);
			print(26);
			l=i+1;
		}else if (s[i]==','){
			if (i-1>0) print(34,l,i-1);
			print(27);
			l=i+1;
		}else if (s[i]==';'){
			if (i-1>0) print(34,l,i-1);
			print(28);
			l=i+1;
		}else if (s[i]==':'){
			if (i-1>0) print(34,l,i-1);
			if (s[i+1]=='=') print(30),i++;
			else print(29);
			l=i+1;
		}else if (s[i]=='('){
			if (i-1>0) print(34,l,i-1);
			print(31);
			l=i+1;
		}else if (s[i]==')'){
			if (i-1>0) print(34,l,i-1);
			print(32);
			l=i+1;
		}else if (s[i]=='#'){
			if (i-1>0) print(34,l,i-1);
			print(35);
			l=i+1;
		}else if (i==len){
			print(34,l,len);
		}
	}
}
void Check_Keyword(){
	len=strlen(s+1);
	int tag=0;
 	if(s[1]=='P')   // 1-PROGRAM or 6-PROCEDURE
	{
		tag=1;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==7) {
			print(1); return;
		}
		tag=6;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==9){
			print(6); return;
		}else {
			Divide_Expression(); return;//print(34); continue;
		}
	}
	if (s[1]=='C') // 2-CONST
	{
		tag=2;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==5){
			print(2); return;
		}else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='V'){ // 3-VAR
		tag=3;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==3){
			print(3); return;
		}
		else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='I'){ // 4-INTEGER or 7-IF
		tag=4;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==7) {
			print(4); return;
		}
		tag=7;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==2){
			print(7); return;
		}else {
			Divide_Expression(); return;//print(34); continue;
		}
	}
	if (s[1]=='L'){  // 5-LONG
		tag=5;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==4){
			print(5); return;
		}
		else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='T') {// 8-THEN
		tag=8;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==4){
			print(8); return;
		}
		else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='W'){ // 9-WHILE or 12-WRITE
		tag=9;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==5) {
			print(9); return;
		}
		tag=12;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==5){
			print(12); return;
		}else {
			Divide_Expression(); return;//print(34); continue;
		}
	}
	if (s[1]=='D'){ // 10-DO
		if (len==2&&s[2]=='O'){
			print(10); return;
		}else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='R'){ // 11-READ
		if (len==4&&s[2]=='E'&&s[3]=='A'&&s[4]=='D'){
			print(11); return;
		}else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='B') {  // 13-BEGIN
		tag=13;
		for (int i=1;i<=len;i++){
			if (s[i]!=WORD[tag][i-1]){
				tag=-1; break;
			}
		}
		if (tag!=-1&&len==5){
			print(13); return;
		}
		else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='E'){  // 14-END
		if (len==3&&s[2]=='N'&&s[3]=='D'){
			print(14); return;
		}else {
			Divide_Expression(); return;
		}
	}
	if (s[1]=='O'){  // 15-ODD
		if (len==3&&s[2]=='D'&&s[3]=='D'){
			print(15); return;
		}else {
			Divide_Expression(); return;
		}
	}
	Divide_Expression();
}


int main(){
	//freopen("read.txt","r",stdin);
	//freopen("out.txt","w",stdout);
	while (~scanf("%s",s+1)){
		Check_Keyword();
	}
}