checking of the stream result? - c++

Question

I have two questions on the following simple code:
/*
test.cpp
© BS.
Example of creation of a class of a flow for which it is possible to assign the additional
characters interpreted as separators.
11/06/2013, Раздел 11.7
*/
//--------------------------------------------------------------------------------------------------
#include <iostream>
#include <sstream>
#include <algorithm>
#include <exception>
#include <string>
#include <vector>
using namespace std;
//--------------------------------------------------------------------------------------------------
class Punct_stream
// it is similar to istream flow, but the user can independently set separators.
{
private:
istream& source; // source of chars
istringstream buffer; // buffer for formatting
string white; // whitespaces
bool sensitive; // case sensitive
public:
Punct_stream(istream& is): source(is), sensitive(true) {}
void whitespace(const string& s) { white = s; }
void add_white(char c) { white += c; }
void case_sensitive(bool b) { sensitive = b; }
bool is_case_sensitive() { return sensitive; }
bool is_whitespace(char c);
Punct_stream& operator >> (string& s);
operator bool();
};
//--------------------------------------------------------------------------------------------------
bool Punct_stream::is_whitespace(char c)
// is the c a whitespace?
{
for(int i = 0; i < white.size(); ++i) if(c == white[i]) return true;
return false;
}
//--------------------------------------------------------------------------------------------------
Punct_stream::operator bool()
// check the input result
{
return !(source.fail() || source.bad()) && source.good();
}
//--------------------------------------------------------------------------------------------------
Punct_stream& Punct_stream::operator >> (string& s){
while(!(buffer >> s)){ // try to read the data from the buffer
if(buffer.bad() || !source.good()) return *this;
buffer.clear();
string line;
getline(source,line); // read the line from the source
// if necessary we replace characters
for(int i = 0; i < line.size(); ++i)
if(is_whitespace(line[i])) line[i] = ' ';
else if(!sensitive) line[i] = tolower(line[i]);
buffer.str(line); // write a string line to the stream
}
return *this;
}
//==================================================================================================
int main()
// entry point
try{
Punct_stream ps(cin);
ps.whitespace(";:,.?!()\"{}<>/&$##%^*|~");
ps.case_sensitive(false);
cout << "Enter the words, please: ";
vector<string> vs;
string word;
while(ps >> word) vs.push_back(word); // enter words
sort(vs.begin(), vs.end());
// we delete counterparts
for(int i = 0; i < vs.size(); ++i) if(i == 0 || vs[i] != vs[i-1]) cout << vs[i] << endl;
}
catch(exception& e){
cerr << e.what() << endl;
return 1;
}
catch(...){
cerr << "Unknown exception." << endl;
return 2;
}
The checking in the Punct_stream::operator bool() function is not clear for me:
return !(source.fail() || source.bad()) && source.good();
My questions:
Why author checked the 'fail' and the 'bad'? Why it wasn't restricted the 'good' check only? Unless the positive 'good' automatically doesn't imply, what 'fail' and 'bad' are set in 'false' value?
Besides, often in a code write such construction:cin >> x; while(cin){//...}
Why the author didn't write by analogy so:
Punct_stream::operator bool()
// check the input result
{
// return !(source.fail() || source.bad()) && source.good();
return source;
}
The alternative option shown by me doesn't work for me (Windows crashed), it would be desirable to understand why, what I missed?
Thank you.

Answer 1

The istream object contains a bunch of bit flags that represent its internal status. Of them, the ones that interests you are:
eofbit End-Of-File reached while performing an extracting operation on an input stream.
failbit The last input operation failed because of an error related to the internal logic of the operation itself.
badbit Error due to the failure of an input/output operation on the stream buffer.
goodbit No error. Represents the absence of all the above (the value zero).
Their status is represented in the following table
good() eof() fail() bad()
goodbit No errors (zero value iostate) true false false false
eofbit End-of-File reached on input operation false true false false
failbit Logical error on i/o operation false false true false
badbit Read/writing error on i/o operation false false true true
Respectively, good() returns the goodbit, eof() checks the eofbit, fail() checks the failbit, and, surprisingly, bad() returns the badbit.
So depending on what you are doing, each function could be used in a different settings.
However, in your case, simply checking the trueness of the good bit would be enough, as it is true when the others are all false. Testing at the same time the failbit or the badbit is redundant.
source: http://www.cplusplus.com/reference/ios/ios/
EDIT:
Actually I'm not really sure why your alternative wouldn't work, as it works for me. What data did you exactly pass to the program?