Problem returning file from fgets()

tubedogg

I'm using PHP4.2.1 on Win32, Apache 1.3. I've got some code that I'm using to open a connection using fsockopen and then grab a file (this is borrowed and modified from someone's code that I found on these forums):

function openURL($host,$port=80,$page,$user_agent='',$user='',$pass='',$referer='',$withHeaders=FALSE) {
// Open a connection to the webserver
$fp = fsockopen($host,$port,$errno,$errstr,10);
if (!$fp) {
return false;
}

// Request the page
fputs($fp,"GET $page HTTP/1.0\r\n");
if (!$user_agent) {
	$user_agent = 'Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Q312461; rant.idlewords.net)';
}
fputs($fp,"User-agent: $user_agent\r\n");
if ($user) {
	fputs($fp,"Authorization: BASIC ".base64_encode($user.':'.$pass)."\r\n");
}
if ($referer) {
    fputs($fp,"Referer: $referer\r\n");
}
fputs($fp,"\r\n");

// Read the response
while(!feof($fp)) {
    $document .= fgets($fp,2048);
}

// Close the Connection
fclose($fp);

if (!$withheaders) {
	$pos = strpos($document, "<!doctype");
	$document = substr($document,$pos);
}

return $document;

}

Generally the code works. However sometimes instead of returning the file intact, it replaces random newlines with (supposedly) ASCII char 42 (PHP is telling me this is a backslash - \ - though I can't prove it as it appears as a little block in my text editor). Any idea why that would happen?

Anon

That's not right for three reasons:

1 - backslash is ASCII 92
2 - ASCII 42 is the asterisk *
3 - Any text editor can show ASCII 42 correctly - therefore if it's a little block, then it's not ASCII 42.

Diego

tubedogg

You're right - it's ASCII 92. I misremembered the error I received.

However, I don't know why a backslash would have problems showing in my text editor. I also don't know why my efforts to str_replace() it into a newline character are failing.

Any more ideas? 🙂

Anon

The files that you read using fgets, are they ASCII or binary? If they are binary, you can't use fgets, you have to use fread.

Diego

tubedogg

It's an ASCII file - an HTML page actually. And there's no pattern to the lines it's doing this to, and it doesn't always happen at the end of lines. It does however always happen in the same places when I'm retrieving the same file. It also happens too frequently in some places for it to be a byproduct of reading the file 2048 bytes at a time.

I appreciate your help. 🙂 Anything else you might suggest?

tubedogg

I should add, it doesn't appear to be a problem with the file itself as I can retrieve the file through other methods and it does not have this problem.

tubedogg

I spent some more time (OK, the last 6 hours :eek🙂 playing with it and I've downloaded two other text editors. After a ton of messing around I came the conclusion that it was a CR/LF problem which I sort of suspected but I couldn't figure out how to fix. I've managed to fix up the problem by running these two lines on my returned file before doing anything else with it:

$hfile = preg_replace("/" . chr(13) . chr(10) . "/", "chr10", $hfile);
$hfile = str_replace("chr10", chr(10), $hfile);

This seems to sort out any rogue \r\n into just \n which seems to fix it, at least to PHP's satisfaction as my string matching functions are now running flawlessly on it.

Thanks for your help. 🙂