meta parsing problems

[Anonymous Monk]

This snippet below isn't parsing meta tags the way it should. This is using LWP::Simple and the source code (which works) is in $content.

This needs to match ANY meta tag in the format shown in the script, and yes I know I shouldn't do it this way but that's the question isn't on using a different module, it's how to fix this problem. Problem is, meta tags aren't always on the same line, they can all be bunched together like a paragraph and that's where this must be messing up.

Nothing prints at all, when I use an array that's split on /n it doesn't work because not all meta tags are separated by new lines.

Where is the error in this, anyone know?

my @meta_results;
my $count = 0;

my @lines = split /\n/, $content;

while(<$content>)
{
  if (/<meta name=\"(.+?)\" content=\"(.+?)\">/gi)
  {
    $count++;
    $meta_results[$count] = "$1::$2";
  }
}

foreach (@meta_results) {print "$_\n";}

[Joost]

Quick & dirty solution:

  while($content =~  /<meta\s+name=\"(.+?)\"\s+content=\"(.+?)\">/ig) 
+{
      $count++;
      $meta_results[$count] = "$1::$2";
  }

It won't work with reversed attributes, single quoted attributes, and lots more... Use a module. I hear HTML::Parser is nice.

Good luck

Updated: thought $content really was a filehandle, but it wasn't.

[exussum0]

You treat $content as a string. 5th line, you treat it like a filehandle, reading it line by line. You can't have it both ways.

What if name and content were switched in positions. XML allows that. regexp's dont' work easily on context free languages.

[Anonymous Monk]

So what's the solution for the string problem? The array didn't work so I was forced to use a scalar, which obviously doesn't work either.

I'm prepared for the reversed meta tags, I thought about that already but that's to be worked on after the meta tags work the first time around with this.

thanks

[Ctrl-z]

this isnt fool proof, but it'll probably do what you want...

foreach( $content =~ m#<meta (.*?)>#sgoi )
{
  my $name = $1 if( $_ =~ m#name\s*?=\s*?["'](.*?)["']#sgoi);
  my $cont = $1   if( $_ =~ m#content\s*?=\s*?["'](.*?)["']#sgoi);
}




time was, I could move my arms like a bird and...

[Anonymous Monk]

Eventhough the thread's a bit old... There are problems with this approach. You should rellay consider using HTML::TreeBuilder, it's as easy as

  use HTML::TreeBuilder;
  my $tree = HTML::TreeBuilder->new()->parse($data);

  for my $tag ($tree->look_down( _tag => "meta")) {
    $kWords{$tag->attr("name")} = $tag->attr("content");
  }

The above code takes care of spaces/linebreaks &s.o. And its fast and widely used. Just my 5cents. FJ

[mrpeabody]

This needs to match ANY meta tag in the format shown in the script, and yes I know I shouldn't do it this way but that's the question isn't on using a different module, it's how to fix this problem. Problem is, meta tags aren't always on the same line, they can all be bunched together like a paragraph and that's where this must be messing up.

You don't want to hear about using a module, but then you complain about having to solve precisely the nontrivial problem that the module is designed to solve. "I'm having trouble mowing my lawn with these scissors. I don't want to use my lawnmower, so don't suggest that. How can I make these scissors work better?"

You can either use the module and be done with a robust solution in ten minutes, or spend hours re-implementing the module and have a weak, brittle solution. Your choice.