php+XML - escaped ampersands not escaping correctly

veritas · 06-19-2007, 05:21 PM

hello

I'm writing a script that parses an XML file. Right now I have the php code echoing whatever is inside a segment tag. Sometimes the tag content will contain ampersands (which are escaped in the xml file like &, but when they are encountered by the parser, something strange happens..

what the XML in the file looks like:

Code:

<segment>part1of36.54oUmWqvRiP6jU&amp;0JoVS@powerpost2000AA.local</segment>

What the output looks like (I just echo whatever is in <segment>):

Quote:

part1of36.54oUmWqvRiP6jU
&
0JoVS@powerpost2000AA.local

The parser seems to split the segment data into 3 parts and treats each part like it just found a new segment. I know this because its adding a <br> to the end of each line, which only happens when it finds a new segment tag. Why would it be doing this? BTW - every other piece of data between <segment> displays correctly if it doesn't have an ampersand.

PHP 4.4.4 slack11

graemef · 06-20-2007, 06:06 PM

It might help if you showed the appropriate bits of code that you are using.

veritas · 06-20-2007, 07:27 PM

Code:

if (! ($xmlparser = xml_parser_create()) )
{ 
   die ("Error: cannot create XML parser");
}

$current = "";

// called when an open tag is found
function open_tag($parser, $tagname, $attribs)
{
  global $current;
  $current = $tagname;
}

// the function called when a closing tag is found
function close_tag($parser, $tagname)
{
// I dont need it to do anything right now
}

// setting up the tag handler
xml_set_element_handler($xmlparser, "open_tag", "close_tag");

// function called when data between tags is found
function tag_contents($parser, $tagdata)
{ 
  global $current;
  if($current == "SEGMENT") //if it detects a segment tag, echo whats inside
  {
    echo $tagdata."<br>"."\n";
  }
}

// setting up the handler for tag contents
xml_set_character_data_handler($xmlparser, "tag_contents");

// set the filename var
$filename = $_FILES['nzbfile']['tmp_name'];

//open the file for reading
if(!($fp=fopen($filename, "r")))
{
  die("Cannot open the nzb entitled ".$filename);
}

// read the file, then send the data to the parser
while($data = fread($fp, $_FILES['nzbfile']['size']))
{
  if (!xml_parse($xmlparser, $data, feof($fp)))
  {
    $reason = xml_error_string(xml_get_error_code($xmlparser));
    $reason .= xml_get_current_line_number($xmlparser);
    die($reason);
  }
}
// destroy parser
xml_parser_free($xmlparser);
fclose($fp);

graemef · 06-21-2007, 12:05 AM

The XML parser that you are using is calling the tag_contents() function multiple times for those lines. You might want to consider using the simpleXML functions. Insert the following code, it should get you going.

PHP Code:



$xmlobj = simplexml_load_file($filename);
var_dump($xmlobj);

Guttorm · 06-21-2007, 01:53 AM

Hi
"tag_contents" can be called many times. The way the parser is designed, you don't need to read the whole XML in memory at once, even if you do in your example. The parser works on a stream and buffers can run empty any time.

Change it to:

PHP Code:



function close_tag($parser, $tagname)
{
  global $current;
  if($current == "SEGMENT")
    echo "<br>\n";
}

function tag_contents($parser, $tagdata)
{ 
  global $current;
  if($current == "SEGMENT") //if it detects a segment tag, echo whats inside
  {
    echo $tagdata;
  }
}

veritas · 06-21-2007, 08:50 PM

Guttorm -- thank you, that fixed it the glitch. Could you explain why placing the breaks in close_tag() changed it? To me it seems either way would do the same thing, but obviously that isn't so

Guttorm · 06-22-2007, 04:31 AM

Hi again.

Well I tried to say, the SAX XML parser is made to operate on streams - not just memory.

PHP Code:



// read the file, then send the data to the parser 
while($data = fread($fp, $_FILES['nzbfile']['size'])) 
{ 
  if (!xml_parse($xmlparser, $data, feof($fp))) 
  { 
    $reason = xml_error_string(xml_get_error_code($xmlparser)); 
    $reason .= xml_get_current_line_number($xmlparser); 
    die($reason); 
  } 
}

In your while loop, you read the entire file at once - passing $_FILES['nzbfile']['size'] as how much you want to read. You could have used a smaller buffer, like 100 bytes, and it still works - even if the segment tag contained more than 100 bytes.

The disadvantage is that tag_contents can be called many times, but it's not a problem since open_tag and close_tag is called so you know when the tag starts and finishes.

The advantage is that the XML file can be gigabytes long, and you can parse it without gigabytes of memory.

veritas · 06-22-2007, 05:51 PM

Thanks, I think i'm all cleared up now.