Visually identical strings not matching against 'ne' ...?

[PerlIsGood]

Here's an excerpt of the script:

use strict;
use CGI;

my $q = new CGI;
my %query = $q->Vars();
my $userdir = "./ns_forum/profiles";

if ( -e "$userdir/$query{'user'}.dat" )
{
  open USR, "$userdir/$query{'user'}.dat" ||
    &err_sub('Error', "Failed to open user file: $!");
  chomp ( my @user_file = <USR> );
  close USR;

  foreach ( @user_file )
  {
    my ( $key, $value ) = split /\|/;
    $user{$key} = $value;
  }

  if ( $user{'pass'} ne $query{'pass'} )
  {
    &err_sub('Error', 'Password invalid.',
      &quot;\$query{'pass'}=$query{'pass'}&quot;,
      &quot;\$user{'pass'}=$user{'pass'}&quot;);
  #the added strings passed to sub for debugging
  }

} else {
  &err_sub('Error', &quot;Complain 'cause it's fun: $!&quot;);
}

So if a user file has the passwd abc123, and the user enters abc123 into the form field, the script dumps into the $err_sub routine for 'Password Invalid'. I printed the values for debugging and they're the same. What am I missing?

(I tried using != just to see what happens, but that has a bad habit of letting incorrect passwds through...I also tried tossing an extra chomp around $query{'pass'} and $user{'pass'} without luck.) Notorious P.I.G.

 

[icrf]

Yeah, != is for numeric only.

abc123 != 0
123abc != 123

are both false. It converts the scalar string to a number left-to-right. The extra chomp is a good call, would have been my first suggestion. Only things I can think of now is to first print out the length of each string, make sure there aren't any non-printable hidden charcters. Then next would be to dump them out with specific formatting using perl's pack function. Maybe each character as an integer? Just some thoughts. ----------------------------------------------------------------------------------
...but I'm just a C man trying to see the light

 

[PerlIsGood]

Ok, I spit out the length of the 2 strings and found that there was indeed a newline in there (i overlooked it when reviewing the source spit out by the $err_sub routine).

After inserting a s/\n|\r// line before the user check everything ran fine. My question is: why did chomp miss this newline?

Anyway, thx for the suggestions. That pointed me in the right direction. Notorious P.I.G.

 

[tanderso]

It might have been out of scope...

my @user_file = <USR>; chomp @user_file;

Or perhaps chomp only takes a string as input...

my $user_file; while (<USR>) {chomp; $user_file .= $_;}

Sincerely,

Tom Anderson
CEO, Order amid Chaos, Inc.

http://www.oac-design.com