Getting rid of ereg

Teach

Hi guys

I have been reading up and see that ereg isn't wise to use now; and that it's being dumped soon

I used my favorite WinGREP program to find all instances of it, and I only use it in once place. I have no idea what I should use instead though as I always seem to make the script stop working

Basically, I'm using this on a form for a IRC bot lending channel to submit a request for a bot to be added in their channel. We use a 'badword' table with "type=0" for bad words, or type=1 for exceptions which are allowed through

I want to get all the type=0 words from the table and compare it against the $_POST['CHANNEL'], and if found (and no type=1 to except it), then it should block...

Currently we use:

$words=mysql_query2("Select * from badwords where type=0");
$counter_badlist=0;
while ($word=mysql_fetch_array($words))
{
  $badlist[$counter_badlist]=$word['badword'];
  $counter_badlist++;
}

$words=mysql_query2("Select * from badwords where type=1");
$counter_exception=0;
while ($word=mysql_fetch_array($words))
{
  $exception[$counter_exception]=$word['badword'];
  $counter_exception++;
}

$count=count($badlist);
for ($x=0;$x<$count;$x++) {
  if (ereg("".strtolower($badlist[$x])."",strtolower($_POST['CHANNEL'])) ) {
    $found=0;
    for ($y=0;$y<count($exception);$y++) // we found it in badlist array
    {
      if (ereg("".strtolower($exception[$y])."",strtolower($_POST['CHANNEL'])) ) {
        $found=1; // we found it in exception
      }
    }
    if ($found==0)
      $msg .= "                <li><span>Your channel name is bad, no thanks</span></li>\n";
  }
}

I'm sure there is a much tidier easier way to do it with in_array. We've used this for ages, and a long time ago I posted regarding it when we didn't use a database, instead we had a hardcoded list.

Any help in sorting this out would be appreciated

Weedpacket

There's no point using a regular expression function here, since you're not using any regular expressions. You could have, but you wrote the code to avoid doing so, preferring instead to have a bunch of loops and test each string one at a time (not even using a case-insensitive search to search case-insensitively, but instead lowercasing everything).

$really_bad_list = array_diff($badlist, $exception);
$really_bad_expression = join('|', $really_bad_list);
if(preg_match("/$really_bad_expression/i", $_POST['CHANNEL'])
{
   $msg .= "Bad channel name.";
}

Teach

Thanks for that Weedpacket!

How would I get my type=0 words from the table to insert into those first two lines? This is the bit that confuses me 😃

Weedpacket

Er.... what's wrong with the way you do it now?

Teach

Hi there 🙂

Yeah, the current way works, but would I still need all that "counter" stuff with the code you provided?

Thanks!

Weedpacket

Teach wrote:
would I still need all that "counter" stuff with the code you provided?

If you look at what my code is doing (see the manual's sections on the functions used if they're unfamiliar) you'll see what it does and from that you should be able to infer what it needs (arrays called $badlist and $exception, obviously - they have to come from somewhere and I even kept their names the same).

Check the manual's section on arrays to see why it wasn't needed in your code either.

Teach

Thanks Weedpacket

I've worked a little more with this, how does this look?

I don't have a testing station (yet) so I thought I'd double check before trying it. Hope you don't mind

Basically I went to test "CHANNEL" against type 0 first (which are the bad words...)
If one is found, then it should test against type 1 (which are exceptions)

If an exception exists, no error. If not, show the error


  // Look for bad words first...
  $words=mysql_query("Select * from badwords where type=0");
  while ($word=mysql_fetch_array($words))
  {
    if (preg_match ("/(.*)".$word['badword']."(.*)/", strtolower($_POST['CHANNEL'])))
    {
      // Found it
      // Set $found to 0 first
      $found=0;

      // Check if a type=1 exists
      $exceptions=mysql_query("Select * from badwords where type=1");
      while ($except=mysql_fetch_array($exceptions))
      {
        if (preg_match ("/(.*)".$except['badword']."(.*)/", strtolower($_POST['CHANNEL'])))
        {
          // Set $found to 1 to prevent the error
          $found=1;
        }
      }

      if ($found==0) {
        $msg .= "<li><span>I found the bad word...</span></li>\n";
      }
    }
  }

Thanks

Weedpacket

The (.*) you've added won't improve anything, just require more time and memory to be spent matching everything that isn't what you're looking for.

You've lost the ability to match strings regardless of their case (you've forgotten to use strtolower on the words you fetch).

Once you correct that, you're back in exactly the same situation I observed in my original reply.

Teach

Excellent, thanks!

I have put a strtolower on the actual word when it's inserted into the database, so now all the words in the database are lowercase - so I only need to lowercase the $_POST['CHANNEL'] to compare now

I'll give this a try, thanks!