Tools

• You probably know this, but I want to review.
• A file stream is a connection between your code and another device.
  • cin, cout are connections to the terminal normally.
  • But we can change that if we wish.
• Quickly look at counter.cpp
• File redirection is a shell feature in most shells

  •  counter > myfile

  • This redirects cout to the file myfile
    • This is a little hard to use because you don't see the prompt.

  •  counter < data

  • This redirects the file data to cin
  • You can do both if you wish.

  •  counter < data > myfile

  • This is useful when you program produces a lot of data.
  • It is also useful to make sure you run your program with the same input every time.
• Pipes are a second useful feature of shells .
  • Pipes allow you to redirect output to another program.
  • more is a *nix program that shows output one screen at a time.
    • more file will show the file one line at a time.
    • more will show cout one line at a time.
    • Seems strange until the example.
  • Try

     more numberWord.cpp

  • Try

     counter | more

  • This takes the output of counter and runs it through the program more.
  • This lets you look at the output one page at a time.
  • This is useful when you program produces a lot of data.
• I would like to have a program that prints the following table

  0 is Zero.
  1 is One.
  2 is Two.
  3 is Three.
  4 is Four.
  5 is Five.
  6 is Six.
  7 is Seven.
  8 is Eight.
  9 is Nine.

  • Note, there is no space after the .
  • Each number starts with an upper case letter.
  • Each line ends in a newline
  • There are no blank lines.
  • I have stored this table in Standard
  • I have some proposed code in numberWord.cpp
• There are quite a few errors, I could solve this problem by checking by hand.
  • But I would probably miss something.
  • Another way I could do this is with the *nix program diff
  • diff compares two text files and
    • Indicates where the lines are different.
    • It event attempts to sync up with missing lines.
    • It is really useful

  • % numberWord > outfile
    % diff Standard outfile
    1,10c1,12
    < 0 is Zero.
    < 1 is One.
    < 2 is Two.
    < 3 is Three.
    < 4 is Four.
    < 5 is Five.
    < 6 is Six.
    < 7 is Seven.
    < 8 is Eight.
    < 9 is Nine.
    ---
    >
    > 0 isZero.
    > 1 isOne .
    > 2 istoo.
    > 3 isThree.
    > 4 isFore.
    > 5 isFive.
    > 6 isSeven.
    > 7 isSeven.
    > 8 isEight.
    > 9 isNine.
    > 10 isUnknown.
    \ No newline at end of file
    % 

  • Notice the lines with differences are printed.
  • Then a < indicates the line in the first file
  • And a > indicates a line in the second file

  • < 0 is Zero.
    ...
    > 0 isZero.  

  • These lines differ in spacing at least

  • < 4 is Four.
    ...
    > 4 isFore.  

    There is also a difference in spelling.
  • This is nice, but pipes provide a short cut

  •  numberWord | diff - Standard

  • Or even

  •  numberWord | diff - Standard | more

• Let's fix numberWord.cpp so it matches the standard.
• One more tool.
  • od stands octal dump.
  • It allows us to look at the individual characters that make up a text file.
  • I created filea.dat on linux with vi
  • I created fileb.dat on windows with notepad++
  • When I cat or more these two files they appear the same.
    • Well sort of, cat show a difference.
  • If I diff them, it shows a difference.
  • What is it?
  • od will allow us to look at all of the characters in the file.
  • The command line argument is -c
    • printable characters show up as themselves
    • non-printable characters show up with a \
  • Notice the *nix file
    • Each line ends in a newline (\n)
    • The last line ends with a newline as well
  • On the windows file
    • Each line ends with both a carriage return (\r) and a newline (\n)
    • The last line has no marker
  • This has an impact.
• Reading the files without an ending \n.
  • look at adder.cpp
  • Look at adder1.cpp
  • The missing \n causes a problem for adder1
  • How does cin work with an integer?
• Reading files with \r
  • Look at looker.cpp
  • Note how the \r messes with the strings.
• I type my data files on a linux system.
  • They do not contain \r
  • They do end with a \n on the last line
    • This is a semi-standard.
  • I expect your programs to run with these files.
• So what can I do
  • Download my data files, do not recreate them.
  • Put a newline at the end of your files, even if it looks strange.
  • Check things out with od -c
  • Run dos2unix on your files.
• dos2unix and unix2dos will convert a file back and forth by adding/removing carriage return (\r)