emilford has asked for the wisdom of the Perl Monks concerning the following question:

I've written a program that both reads from and writes to standard text files. The program creates the text file so that each line contains a specific chunk of information. The problem I'm experience is that some of the lines I want to write out to the file are longer than the, I guess, the Windows allowed lengths (if there is such a thing). When I open up the file in Notepad, the lines are not maintained and everything is overlapping. In Wordpad, however, the file looks how I expected it to, one long chunk of information per line.

This is what I want: 
information line 0000000000000000000000000
information line 1111111111111111111111111
information line 2222222222222222222222222

[download]
This is what I get: 
information line 000000
0000000000
information line 111111
1111111111
information line 222222
2222222222

[download]
The problem comes when I am attempting to read from the files. I do a @array = <FILE> to get all of the lines into the array, but the lines aren't seperated like they are suppose to be. Why aren't the lines coming in like the way they are when opened in Wordpad?

After the program writes the files, they are saved as standard txt extensions. I opened them up in Wordpad and saved them in the Unicode format. This made the file appear correctly in both Notepad and Wordpad, but the program still didn't interact with them correctly.

I tried one last thing and save them from Unicode back to a txt document using Wordpad. This suprisingly made everything work. They looked the way they should and the program worked correctly.

What is causing this problem? Does Windows tack on endlines where it sees fit or something? How do I get around this. I don't think I had this problem on Linux. How do I store the file so that when I want to store chunks of information seperated by lines, they are actually stored that way?!? Is there another file extension I should try? Thanks in advance. -Eric

Replies are listed 'Best First'.
Re: automatic endlines when saving txt file under windows?
by JayBonci (Curate) on Jun 06, 2002 at 20:00 UTC
    Traditional UNIX endlines are \n. In Windows, they are \r\n. You may need to tack on the \r\n for notepad to see them correctly (Wordpad always behaves as it should, seeing the \n as meaning a new line). Even if you open them up in Wordpad and save them again, they will not modify the newlines. There are some flukes with getting notepad to read UNIX newlined text files. Sometimes it will appear overlapped, othertimes you might see boxes. If you save something with a .wri extension, wordpad will be (by default) associated with it for opening because it's the old Windows Write extention. Hope that helps

        --jb
[reply]
      Thanks for the information. I tried adding the \r\n in place of just \n, but that didn't work either. I'm not really concerned with how the information looks when I open it in Notepad, I just want the program to interact with the files correctly. Shouldn't Perl recongnize the \n as endlines even though Notepad doesn't?
[reply]
Re: automatic endlines when saving txt file under windows?
by jsprat (Curate) on Jun 06, 2002 at 20:29 UTC
    You do have "word wrap" disabled in notepad, don't you? ;)
[reply]
      Yup, just double checked.
[reply]
Re: automatic endlines when saving txt file under windows?
by {NULE} (Hermit) on Jun 07, 2002 at 00:45 UTC
    Hi emilford,

    I work with a data interchange language called HL7, which normally isn't that bad except when some brain-dead integration code drops odd characters into it. Not to deride wordpad and notepad for the fine products they are (hey, they work exactly as advertised!) but other tools may help you make sure what you think you have in the file is really what you have.

    For example in one troubled message we received that was causing our engine to puke the bad record it looked normal in emacs but in gvim what looked like a pipe was actually something else. It looked like a control character in gvim and in command mode typing 'ga' with the cursor on the character give it's ascii and octal values. I'm not knocking emacs here since other tools can produce odd results when faced with unexpected data (like wc -l in dealing with mixed \r and \n).

    Anyway my long winded point is that you need something that can tell you what you really have in the files. Obviously changing their extension doesn't change the content (unless windows is dumber than I thought) so there is some kind of file type interpolation going on. I recommend using gvim for windows (and every other OS {G} - and while I'm p1mp1ng it, vim has great Perl syntax highlighting - http://vim.sf.net/ ) where if you type ':%!xxd' from the command mode will convert the file into hex and you can see exactly what the file contains. Any other hex editor or tools like xxd or od -c will do the trick as well. (Oh, and in gvim ':%!xxd -r' will convert back to plain ascii. Obviously the single quotes should not be used in typing any of the vim commands.)

    Sadly a coder is only as good as their tools, but with Perl on our side we are already fairly excellent. ;)

    Good luck,
    {NULE}
    --
    http://www.nule.org

    P.S. One more thing, (I honestly didn't intend to turn this post into 'The VIM Show™') but if your files are supposed to be identical you can use VIM to diff them by calling 'gvimdiff file1 file2' or 'gvim -d file1 file2'.

[reply]
Re: automatic endlines when saving txt file under windows?
by dda (Friar) on Jun 07, 2002 at 09:07 UTC
    Do you have binmode STDOUT in your script? If your script is being run under Windows, and you want to produce a windows-like file endings, you should not use binmode.

    Regards, Dmitry.

[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom