mr.dunstan has asked for the wisdom of the Perl Monks concerning the following question:

Hi - I'm 3 hours into this problem and I'm getting bummed out. Here's what I'm trying to do:

I have a cgi script I'm working on that allows me to upload a file into the perl script using a file input element and CGI.pm and a multipart form. No problem there, been there done that.

The problem is - once I have read the file contents into memory, I want to parse it line-by-line, hopefully by reading each line into an array (this is fixed-width data mixed in with some unstructured data, thanks silly vendor, and the document itself is only vaguely structured, but if I can read it line by line I'm good to go.) The thing is - once I've read it into a scalar, the line feeds disappear BUT if I read it out to a file on the local filesystem (not an option in production BTW), the line feeds are there?!

I've tried split on the line feeds but they're not there - see example below. Way confused. Anyway here's what I'm trying to do, in short. Yes I am using strict ... 


    use strict, CGI, blah blah ... 

    my $cgi = new CGI;
    my $filename = $cgi->param('uploaded_file');

    my $file = $filename;        # can't remember why I do this, alway
+s have
    $file =~ s!^.*(\\|\/)!!;    # something about cleaning stuff up
    
    my $buffer;
    my $content_to_parse;        
    
    while (my $bytesread = read($filename,$buffer,1024)) {
        $content_to_parse = $content_to_parse . $buffer;     
    }

    # ok now file contents are in $content_to_parse
    
    # try splitting on ^M (windoze newlines) 
    my @contentarray = split /\015$/, $content_to_parse;
    # doesn't work - @contentarray is empty 
    
    # scratch head, get another diet coke

    # now do line by line parsing ... the end

[download]
I'm sure there's a better way, like reading the file into an array somehow. Oh perlmonks deliver me from this pain!!!!

-mr.dunstan

Replies are listed 'Best First'.
Re: splitting on win32 linefeeds aka pass the crackpipe!
by little (Curate) on Feb 14, 2002 at 01:53 UTC
    Better split this way, so you can split also files saved in Unix format
    # try splitting on ^M (windoze newlines) 
    my @contentarray = split /\r?\n/, $content_to_parse;

    [download]

    Have a nice day
    All decision is left to your taste
[reply]
[d/l]
      Wow. I feel dumb. Thanks!

      -mr.dunstan
[reply]
Re: splitting on win32 linefeeds aka pass the crackpipe!
by buckaduck (Chaplain) on Feb 14, 2002 at 02:25 UTC
    Update: Mea culpa. I wasn't paying attention to the .* in the regex. Sigh. But I'll leave my post intact, to preserve my mistake for posterity...

    Off topic, but: 

      my $file = $filename;        # can't remember why I do this, always 
+have
  $file =~ s!^.*(\\|\/)!!;     # something about cleaning stuff up

    [download]
    ...is pretty ugly. If you're trying to trim a trailing slash, it's easier to read like this: 
      my $file = $filename;
  $file =~ s![\\/]$!!;

    [download]
    If you can't tell what your own regex is doing, it's time to clean it up.

    buckaduck

[reply]
[d/l]
[select]
      The original substitution doesn't actually remove a trailing slash; rather, it removes the directory path leaving just the name of the file. If $filename is '/usr/local/bin/perl', $file will be 'perl'.

      This is better done with File::Basename.

[reply]
      chipmunk++ for pointing out File::Basename.

      mr.dunstan, you'd probably be better off with this: 
      my ($actualFileName, $filePath) = fileparse ($filename);

      [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom