How to convert a text file from Windows to Unix - c++

Question

When converting from Unix to Windows, I get the correct output; however, when going from Windows to Unix I get some strange output. I thought all I had to allow for was the removal of the carriage return, '\r'. This isn't working though. When I open the text file after running the code, I get some strange results, the first line is correct, and then all hell breaks lose.
int main( )
{
bool windows = false;
char source[256];
char destination[256]; // Allocate the max amount of space for the filenames.
cout << "Please enter the name of the source file: ";
cin >> source;
ifstream fin( source, ios::binary );
if ( !fin ) // Check to make sure the source file exists.
{
cerr << "File " << source << " not found!";
getch();
return 1;
}//endif
cout << "Please enter the name of the destination file: ";
cin >> destination;
ifstream fest( destination );
if ( fest ) // Check to see if the destination file already exists.
{
cout << "The file " << destination << " already exists!" << endl;
cout << "If you would like to truncate the data, please enter 'Y', "
<< "otherwise enter 'N' to quit: ";
char answer = char( getch() );
if ( answer == 'n' || answer == 'N' )
{
return 1;
}//endif
}//endif
clrscr(); // Clear screen for neatness.
ofstream fout( destination, ios::binary );
if ( !fout.good() ) // Check to see if the destination file can be edited.
{
cout << destination << "could not be opened!" << endl;
getch();
return 1;
}//endif
// Open the destination file in binary mode.
fout.open( destination, ios::binary );
char ch = fin.get(); // Set ch to the first char in the source file.
while ( !fin.eof() )
{
if ( ch == '\x0D' ) // If ch is a carriage return, then the source file
{ // must be in a windows format.
windows = true;
}//endif
if ( windows == true )
{
ch = fin.get(); // Advance ch, so that at the bottom of the loop, the
}//endif // carriage return is not coppied into the new file.
if ( windows == false )
{
if ( ch == '\x0A' ) // If the file is in the Unix format..
{
fout.put( '\x0D' ); // When a new line is found, output a carriage
}//endif // return.
}//endif
fout.put( ch );
ch = fin.get();
}//endwh
if ( windows == true )
{
fout.put( '\x0A' );
}//endif
fout.close();
fin.close(); // Close yer files.
if ( windows == true ) // A little output for user-friendly-ness.
{
cout << "The contents of " << source << " have been coppied to "
<< destination << " and converted to Unix format." << endl;
}else{
cout << "The contents of " << source << " have been coppied to "
<< destination << " and converted to Windows format." << endl;
}//endif
cout << "Enter any key to quit.." << endl;
getch();
return 0;
}//endmn

Answer 1

*If* you only need to convert simple ascii (and perhaps utf-8) text files, you could read the source file line-by-line in a loop in translated mode (handles newlines for you enough for this case) with non-member getline() and then output the lines to the output file while inserting \n or \r\n after each line except the last.
Then, you can remove the original file and rename the temp file to have the original file's name. Or, if you want, you can instead push_back the lines into a vector<string>. Then, you could close the input handle to the file, do ofstream out("filename", ios_base::trunc) and write the elements of the vector to the file while separating them by the newlines you want.
It all depends on your requirements.
The following is an example with minimal error handling. But, it's really only the FOR loop and reading line-by-line that I want to show here as a different way of doing things.
convert_file.exe "test.txt" "linux"
convert_file.exe "test.txt" "win"
#include <iostream>
#include <string>
#include <fstream>
#include <ostream>
#include <cstdlib>
#include <cstdio>
using namespace std;
int main(int argc, char* argv[]) {
if (argc != 3) {
cerr << "Usage: this.exe file_to_convert newline_format(\"linux\" or \"win\")" << endl;
return EXIT_FAILURE;
}
string fmt(argv[2]);
if (fmt != "linux" && fmt != "win") {
cerr << "Invalid newline format specified" << endl;
return EXIT_FAILURE;
}
ifstream in(argv[1]);
if (!in) {
cerr << "Error reading test.txt" << endl;
return EXIT_FAILURE;
}
string tmp(argv[1]);
tmp += "converted";
ofstream out(tmp.c_str(), ios_base::binary);
if (!out) {
cerr << "Error writing " << tmp << endl;
return EXIT_FAILURE;
}
bool first = true;
for (string line; getline(in, line); ) {
if (!first) {
if (fmt == "linux") {
out << "\n";
} else {
out << "\r\n";
}
}
out << line;
first = false;
}
in.close();
out.close();
if (remove(argv[1]) != 0) {
cerr << "Error deleting " << argv[1] << endl;
return EXIT_FAILURE;
}
if (rename(tmp.c_str(), argv[1]) != 0) {
cerr << "Error renaming " << tmp << " to " << argv[1] << endl;
return EXIT_FAILURE;
}
}
As others have said though, there are already utilities (including text editors like Notepadd++) that do newline conversion for you. So, you don't need to implement anything yourself unless you're doing this for other reasons (you didn't specify).

Answer 2

Don't worry about checking for windows in the loop. Simply check for a carriage return. Set a variable 'carriage_return.' Next iteration, if 'carriage-return' and ch != linefeed, simply insert a linefeed. Then reset the carriage_return variable to false. It's a very simple and basic rule which won't send you wrong.
bool carriage_return = false;
const char linefeed = '\n'; // Is it? I forget.
const char cr = '\r'; // I forget again. Too late to check.
char ch = fin.get();
if (ch == cr) carriage_return = true;
while (!fin.eof()){
if (carriage_return) { // Check if we already have a newline
if (ch != linefeed) { // If we have a newline previously, we need a linefeed. If it's already there just leave it, if it isn't there put it in
fout.put(linefeed);
}
if (ch != cr) carriage_return = false; // Reset the carriage-return flag *if* we don't have another carriage return. This handles multiple empty lines in an easy way for us.
}
fout.put(ch);
ch = fin.get();
}

Answer 3

I have re-edit your code and it works fine for me..
Hope this helps !
#include <iostream>
#include <fstream>
#include <iostream>
#include<stdio.h>
using namespace std;
int main( )
{
bool windows = false;
char source[256];
char destination[256]; // Allocate the max amount of space for the filenames.
cout << "Please enter the name of the source file: ";
cin >> source;
ifstream fin( source, ios::binary );
if ( !fin ) // Check to make sure the source file exists.
{
cerr << "File " << source << " not found!";
return 1;
}//endif
cout << "Please enter the name of the destination file: ";
cin >> destination;
ifstream fest( destination );
if ( fest ) // Check to see if the destination file already exists.
{
cout << "The file " << destination << " already exists!" << endl;
cout << "If you would like to truncate the data, please enter 'Y', "
<< "otherwise enter 'N' to quit: ";
char answer;
cin >> answer;
if ( answer == 'n' || answer == 'N' )
{
return 1;
}
}
//clrscr();
ofstream fout( destination);
if ( !fout.good() )
{
cout << destination << "could not be opened!" << endl;
return 1;
}
char ch = fin.get();
while (!fin.eof())
{
if ( ch == '\r' )
{
windows = true;
}
if ( ch == '\n' && windows == false ) // If the file is in the Unix format..
{
// Don't do anything here
}
fout.put( ch );
cout << ch; // For Debugging purpose
ch = fin.get();
}
fout.close();
fin.close();
if ( windows == true ) // A little output for user-friendly-ness.
{
cout<<endl;
cout << "The contents of " << source << " have been coppied to "
<< destination << " and converted to Unix format." << endl;
}else{
cout << "The contents of " << source << " have been coppied to "
<< destination << " and converted to Windows format." << endl;
}//endif
cout << "Enter any key to quit.." << endl;
return 0;
}

Answer 4

Did you make sure you're reading in the data in the right format and saving it in the right format?
Trying to work with a different character encoding and just 'reading' it in leads to very bad things :|
And you then also have to account for different replacements that need to be done.
This may help link