Need help completing the C++ scanner program // A simple lexical analyzer for C/

Question

Need help completing the C++ scanner program

// A simple lexical analyzer for C/C++-style variable declarations.
// The grammar for the declarations is as follows:
//
// <declaration>    ::=    <type> <var> ’;’ | <type> <var> ’=’ <number> ’;’
// <type>        ::=    int | float
// <var>        ::=    A | B | C | D | E
// <number>    ::=    <integer> | <float>
// <integer>    ::=    <integer> <digit> | <digit>
// <digit>        ::=    0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9
// <float>        ::=    <integer> ‘.’ <integer>
//
// The task of this exercise is to write a lexical analyzer (or
// scanner) for the tokens used in the above grammar. The following
// is the regular expression that defines the tokens:
//
//       = | ; | int | float | A | B | C | D | E | [0-9]+ | [0-9]+.[0-9]+
//
// The program will read a declaration from the keyboard, and the
// scanner you design should recognize and print out all tokens
// included in the input. For example, given the following declaration:
//
//       int A = 123;
//
// your program should print:
//
// int
// A
// =
// 123
// ;
//
// Make sure your program print out the token one per line in the order
// they appear in the input. Once an erroneous token is encountered, your
// scanner should print out an error message and stopped scanning. For
// example, given the following input:
//
//       int A = 0#;
//
// your program should print:
//
// int
// A
// =
// 0
// #: Error: Unrecognizable token
//
// Note that tokens may NOT be separated by spaces. For example, the above
// input:
//
//       int A = 123;
//
// does not have a space to separate 123 and ;. Also, the following inputs
// are also legal and generate the same output:
//
//       int A=123;
//       intA=123;
//
// However, a whole token cannot be separated by spaces. For example, the
// following input will cause 12 and 3 to be regarded as two distinct tokens.
//
//       int A = 12 3;
//
// The ouput will look like:
// int
// A
// =
// 12
// 3
// ;
//
// Also note that the scanner doesn't check for syntactic errors. Therefore the
// above input is legal to this program.
//
// Important!!! Save your GetToken program. We are going to use it in future
// labs.

#include <iostream>
#include <string>

using namespace std;

string GetToken();
void error(int);

int main() {
   string token;

   cout << "Please enter a declaration in format "
       << "<type> <variable> [= number];" << endl;

   cout << "The following are the tokens in the input:" << endl;

   token = GetToken();

   while (token != "") {
       cout << token << endl;
       token = GetToken();
   }

   cout << "Done!" << endl;

   return 0;
}

string GetToken() {
   string token;
   char ch;

   // Write the code here. Read the next token and store it in variable "token".
   // The token must be read character by character. Use the regular expression
   // defined above to extract tokens from the input.
   //
   // To read a character from keyboard, use:
   //
   //       cin.get(ch);
   //
   // where "ch" is a character variable.

   return token;
}

Explanation / Answer

# include <iostream.h>
# include <fstream.h>
# include <string.h>
# include <stdlib.h>
# include <stdio.h>
# include <conio.h>
# include <ctype.h>
constint iColumns = 12;

int iTT[iRows][iColumns] = {0};

{
fstream File("tt.txt", ios::in|ios::nocreate);

if (!File)
{
cout << " Unable to open the input file." << endl;
cout << " Press any key to exit.";

getch( );
exit(0);
}

char sInput[100]={NULL};

for (int i = 0; i < iRows; i ++)
{
strset(sInput, NULL);

File.getline(sInput, 80);

char *sPtr=NULL;

sPtr = strtok(sInput, " ");

iTT[i][0] = atoi(sPtr);

for(int j = 1; j < iColumns; j ++)
{
sPtr=strtok(NULL, " ");

iTT[i][j] = atoi(sPtr);
}
}

File.close( );
}
{
if (isalpha(cChar))
return iTT[iState][1];

elseif (isdigit(cChar))
return iTT[iState][2];

elseif (cChar == '.')
return iTT[iState][3];

elseif (cChar == '"')
return iTT[iState][4];

elseif (cChar == '_')
return iTT[iState][6];

elseif (cChar == '+')
return iTT[iState][7];

elseif (cChar == '=')
return iTT[iState][8];

elseif (cChar == '-')
return iTT[iState][9];

elseif (cChar == '%')
return iTT[iState][10];

elseif (cChar == '!')
return iTT[iState][11];

elseif (cChar == '>')
return iTT[iState][12];

elseif (cChar == '<')
return iTT[iState][13];

elseif (cChar == '/')
return iTT[iState][14];

return iTT[iState][0];
}

{
if (strlen(sToken) > 16 || strlen(sToken) == 0)
return 0;

char sKeywords[44][15] = {
"asm","auto","bool","break","case","catch",
"char","class","const","const_cast",
"continue","default","delete","do","double",
"dynamic_cast","else","enum","explicit",
  
"static_cast","struct","switch","template",
"this","throw","true","try","typedef",
"typeid","typename","union","unsigned","using",
"virtual","void","volatile","wchar_t","while"
};

for(int iCount = 0; iCount < 64; iCount ++)
{
if (strcmpi(sKeywords[iCount], sToken) == 0)
return 1;
}

return 0;
}

{
clrscr( );

loadTransitionTable( );

fstream File("input.txt", ios::in|ios::nocreate);

if (!File)
{
cout<<" Unable to open the input file."<<endl;
cout<<" Press any key to exit.";

getch( );
exit(0);
}

char sToken[255] = {NULL};
int iTokenIndex = 0;

char cChar = NULL;
int iState = 0;
int iFlag = 0;

char cTemp = File.get( );

do
{
Start:

if (iFlag == 0)
{
cChar = cTemp;
cTemp = File.get( );
}

else
iFlag = 0;

if (cChar == '/' && cTemp == '/')
{
while(File.get( ) != ' ')
{
if (File.eof( ))
goto End;
}

cout<<' ';

cTemp = File.get( );

goto Start;
}

if (cChar == '/' && cTemp == '*')
{
cTemp = File.get( );

do
{
cChar = cTemp;
cTemp = File.get( );

if (File.eof( ))
goto End;
}
while(cChar != '*' && cTemp != '/');

cout<<' ';

cTemp = File.get( );

goto Start;
}

iState = getNextState(iState, cChar);

switch (iState)
{
case 0 : cout << cChar;

iState = 0;
iTokenIndex = 0;

strset(sToken, NULL);

break;

case 1 :
case 3 :
case 5 :
case 7 :
case 10 :
case 14 :
case 18 :
case 25 :
case 26 : sToken[iTokenIndex] = cChar;
iTokenIndex ++;

break;

case 2 : if (isKeyword(sToken))
cout << sToken;

else
cout << "<ID>";

iState = 0;
iTokenIndex = 0;
iFlag = 1;

strset(sToken, NULL);

break;

case 4 : cout << "<INT>";

iState = 0;
iTokenIndex = 0;
iFlag = 1;

strset(sToken, NULL);

break;

case 6 : cout << "<FLOAT>";

iState = 0;
iTokenIndex = 0;
iFlag = 1;

strset(sToken, NULL);

break;

case 8 : cout << "<STR>";

iState = 0;
iTokenIndex = 0;

strset(sToken, NULL);

break;

case 9 :
  
case 24 :
case 27 :
case 28 : cout << "<OPR>";

if (cChar != '+' && cChar != '-' && cChar != '/'
&& cChar != '>' && cChar != '<' && cChar != '=')
iFlag = 1;

iState = 0;
iTokenIndex = 0;

strset(sToken, NULL);

break;

case 30 :
case 33 : iState = 0;
iTokenIndex = 0;

strset(sToken, NULL);

break;
}
}
while(!File.eof( ));

End:

getch( );
return 0;
}

Related Questions

Navigate

• Browse (All)
• Browse N
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.