[RESOLVED] Getting htmlentities to do the parsing

[Tim_Silva]

Hi,

after spending the last three days trying all conceivable ways to get an xml fopen page right, I am really at lost. I am able to get a beautiful xml page except for some disgusting looking characters that won't allow ie7 to load it. So that:

Now, a UK team of scientists believe they may be able to use these cells -- known as Müller glial cells -- to regenerate

Gives me:

Now, a UK team of scientists believe they may be able to use these cells � known as M�ller glial cells � to regenerate

I am using something like:

$_xml .="<title>".htmlentities($info['title'])."</title>\r\n";
$_xml .="<description>".htmlentities($info['content'])."</description>\r\n";

I've also tried many nice preg_replace based functions (such as xmlentities) from php.net with no positive results.

Any hints would be appreciated.

Tim

[NogDog]

See if using a CDATA tag helps:

$_xml .="<description><![CDATA[".htmlentities($info['content'])."]]></description>\r\n";

[Tim_Silva]

No luck.

It gives: Parse error: parse error, unexpected '[' in /homepages

[NogDog]

That's a PHP syntax error. Can you show us the actual PHP code that is generating it?

[Tim_Silva]

$data = mysql_query("SELECT id, title, content, date FROM $usertable  WHERE audio = 'Y' ORDER BY date DESC LIMIT 7") 
or die(mysql_error()); 
while($info = mysql_fetch_array( $data ))

{
$_xml .="<item>\r\n";
$_xml .="<title>".htmlentities($info['title'])."</title>\r\n";
$_xml .="<description>"<![CDATA[".htmlentities($info['content'])."]]>"</description>\r\n";
$_xml .="</item>\r\n";
}

[Tim_Silva]

The weirdest thing. Just for testing purpose, I did this preg_replace and I've got an 'x' between every word (very good), but I still had the same problem. The "it's", for example came as "it?s".

It?sxNotxFunnyxAndxit?sxNotxMine

$patterns[0] = '/-/';
$patterns[1] = '/\"/';
$patterns[2] = '/―/';
$patterns[3] = '/ /';
$patterns[4] = '/\'/';
$replacements[4] = 'x';
$replacements[3] = 'x';
$replacements[2] = 'x';
$replacements[1] = 'x';
$replacements[0] = 'x';

$_xml .="<description>".preg_replace($patterns, $replacements, $info['content'])."</description>\r\n";

[NogDog]

Just a bit of a quoting problem. The CDATA tag is just more text to output, as far as PHP is concerned:

$_xml .="<description><![CDATA[".htmlentities($info['content'])."]]></description>\r\n";

[Tim_Silva]

The Parse error is gone but unfortunately the entities problem remain. I am beginning to think I will have to do this parsing some other way.

[NogDog]

Try it without the PHP htmlentities(). (Sorry I'm being a bit random here -- I ran into something like this before quite awhile ago, but don't have access to the script I edited.)

$_xml .="<description><![CDATA[" . $info['content'] . "]]></description>\r\n";

[Tim_Silva]

Still no luck.

Here is another piece to this puzzle.

If I do:

$_xml .="<description>".htmlentities($info['content'])."...</description>\r\n";

I can't get the punctuations (such as single quote) to work. So that I have: Crohn�s disease.

But if I use another db field:

$_xml .="<description>".htmlentities($info['title')."...</description>\r\n";

Then everything works fine and I am able to get: Crohn's disease.

Because of that, I got both, 'content' and 'title' as the same field type (text). But I still have the same bad result.

[NogDog]

Only way I could get it to work was to use ISO-8859-1 encoding, CDATA tags, and no htmlentities():

<?php
header("Content-Type: text/xml");
?>
<?xml version='1.0' encoding='ISO-8859-1'?>
<test>
<?php
$text = '"Parlez vous Français?" René asked.';
echo "<text><![CDATA[$text]]></text>\n";
?>
</test>

[Tim_Silva]

YES!

Changing the encoding to 'ISO-8859-1' made all the difference. What I will have to find out now is if there is any implication as far as how the outside world can interpret the feed, such as iTunes podcasting, for example.

Tim