Notes on file I/O:

          Predefined objects used with I/O operators:

                cin is instance of istream tied to standard input
                device (keyboard)

                cout is instance of ostream tied to standard output
                device (monitor)

                also cerr (unbuffered ostream object) for error output messages
                g++ -o mypgm.exe mypgm.cpp >& myout.txt get both outputs

                endl stream manipulator both issues '\n' and causes
                buffer output to be written immediately

                can also force write of output buffer with flush manipulator
                  cout << flush;

          get member functions (of istream) has three versions

                get with no arguments, inputs one character from
                  designated stream (even if whitespace), and returns
                  character (including EOF)

                get with char reference argument inputs next character and stores it
                  in the char argument

                get with three arguments (char array, size limit and delimeter with
                  default of '\n') reads characters from the input
                  stream up to one less than specified size limit or
                  until delimiter is reached, a null character is
                  added, and the resulting null-terminated string
                  stored in the char array;
                  delimiter remains in input stream

           getline
                like last version of get above, but delimiter is read
                from (and removed) from stream, but not added to the array


    Unformatted I/O

          Unformatted I/O performed with read and write member
          functions

          Bytes are read as bytes and stored in memory; a byte is not
          considered part of a larger unit (like an int)

          write requires address and length arguments:
            cout.write(myString, 15);

          read inputs designated number of bytes into address
          specified (usually an array)

                if fewer than designated number of bytes read, failbit
                is set

                member function gcount returns number of bytes read by
                the last operation


    Stream Manipulators

        * Stream manipulators perform formatting tasks
        * header file <iomanip> must be included for parameterized
          stream manipulators
        * Setting base of integers
              o hex, oct, and dec are manipulators
              o setbase is parameterized stream manipulator, requires
               argument of 8, 10, or 16 (for octal, decimal, and hex)
              o stream base is persistent
        * Floating point
              o determines number of digits to the right of the decimal point
              o precision is member function (cout.precision(n))
              o setprecision is parameterized manipulator
                  cout << setprecision(n))
              o stream precision is persistent
              o precision with no arguments returns current precision
                  n = cout.precision()

        * Field width
              o sets width of field for next input (if character array) or output
              o width is member function of ios
                  cin.width(n) or cout.width(n))
              o setw is manipulator
                  cout << setw(n))
              o field width is not persistent

    Stream Format States

        * There are format flags associated with each stream that
          determine the type of formatting to be performed during I/O
          operations
        * member functions setf, unsetf, and flags control flag settings
        * can also use stream manipulator setiosflags and resetiosflags
        * can set many flags at once by combining with the bitwise-or
          operator ( | ) into a long
              o flags() sets all flags on a stream and returns long
                value containing previous options, so that they can be saved
                and reset

                    ios_base::fmtflags originalFormat = cout.flags();
                    cout << ... ;               // change format
                    cout.flags(originalFormat); // restore format

              o setf can set one or more flags (ORs with existing flag settings)

        * Trailing zeros and decimal points (ios::showpoint):
              o forces display of decimal point (and trailing zeros
                according to current precision) for "integer" values (i.e.,
                the value 23.00 displays as just 23 without this flag set)
        * Justification within a field (ios::left, ios::right, ios::internal)
              o left-justify, right-justify, or fill number between
                sign and value with pad character
              o when using setf, must use ios::adjustfield as second
                argument: cout.setf(ios::left, ios::adjustfiled)
              o fill member function (or setfill manipulator)
                specifies pad character
        * Padding (fill, setfill)
              o  note also use of ios::showbase to force display of
                 integer base (none for decimal, 0 for octal, 0x for hex)
        * Floating point number format
              o use with precision
              o to reset to default: cout.setf(0, ios::floatfield);
        * Upper/lowercase
              o only for uppercase X and E in hex, scientific
                notation, and all hex numerals (A-F)

    Stream Error States
        * State of a stream can be tested through bits in class ios

                eofbit is set (=1) after EOF is encountered
                can use eof function to test this bit

                failbit is set for a stream when a format error occurs
                (e.g. a non-digit character is encountered while
                inputting an integer)
                      test with fail()

                badbit set when error occurs that results in loss of data
                      nonrecoverable
                      test with bad()

                goodbit is set if eofbit, failbit, and badbit are not set
                      test with good()

                rdstate member function returns error state of the
                stream, which can then be tested (using a switch that
                examines each bit)

                clear() member function resets specific bit, or sets
                goodbit if no argument given


    Tying an Output Stream to an Input Stream

          Normally each stream is treated independently

          If streams are buffered, input/output happens only when
          buffers are full

          In some cases, input and output should be "paired" (or
          synchronized) so than output happens before input,
          regardless of buffer state (as in a prompt for subsequent input)

          tie function used to do this:
            cin.tie( &cout );    // ties cin to cout

          Above done automatically in C++, but other istream/ostream
          pairs would need to be explicitly tied by user.