Question

I wrote a small PHP script to edit the site news XML file.
I used DOM to manipulate the XML (Loading, writing, editing).

It works fine when writing English characters, but when non-English characters are written, PHP throws an error when trying to load the file.

If I manually type non-English characters into the file – it’s loaded perfectly fine, but if PHP writes the non-English characters the encoding goes wrong, although I specified the utf-8 encoding.

Any help is appreciated.

Update: with the helpful answers, it is solved (read below).

Errors:

Warning: DOMDocument::load()
[domdocument.load]: Entity ‘times’ not
defined in filepath

Warning: DOMDocument::load()
[domdocument.load]: Input is not
proper UTF-8, indicate encoding !
Bytes: 0x91 0x26 0x74 0x69 in filepath

Here are the functions responsible for loading and saving the file (self-explanatory):

function get_tags_from_xml(){
// Load news entries from XML file for display
    $errors = Array();

    if(!$xml_file = load_news_file()){
    // Load file
        // String indicates error presence
        $errors = "file not found";
        return $errors;
    }
    $taglist = $xml_file->getElementsByTagName("text");
    return $taglist;
}
function set_news_lang(){
// Sets the news language
    global $news_lang;

    if($_POST["news-lang"]){
        $news_lang = htmlentities($_POST["news-lang"]);
    }
    elseif($_GET["news-lang"]){
        $news_lang = htmlentities($_GET["news-lang"]);
    }
    else{
        $news_lang = "he";
    }
}
function load_news_file(){
// Load XML news file for proccessing, depending on language 
    global $news_lang;

    $doc = new DOMDocument('1.0','utf-8');
    // Create new XML document
    $doc->load("news_{$news_lang}.xml");
    // Load news file by language
    $doc->formatOutput = true;
    // Nicely format the file

    return $doc;
}
function save_news_file($doc){
// Save XML news file, depending on language 
    global $news_lang;

    $doc->saveXML($doc->documentElement);
    $doc->save("news_{$news_lang}.xml");
}

Here is the code for writing to XML (add news):

<?php ob_start()?>
<?php include("include/xml_functions.php")?>
<?php include("../include/functions.php")?>
<?php get_lang();?>
<?php
//TODO: ADD USER AUTHENTICATION!
if(isset($_POST["news"]) && isset($_POST["news-lang"])){
    set_news_lang();

    $news = htmlentities($_POST["news"]);

    $xml_doc = load_news_file();
    $news_list = $xml_doc->getElementsByTagName("text");
    // Get all existing news from file

    $doc_root_element = $xml_doc->getElementsByTagName("news")->item(0);
    // Get the root element of the new XML document
    $new_news_entry = $xml_doc->createElement("text",$news);
    // Create the submited news entry

    $doc_root_element->appendChild($new_news_entry);
    // Append submited news entry
    $xml_doc->appendChild($doc_root_element);

    save_news_file($xml_doc);

    header("Location: /cpanel/index.php?lang={$lang}&news-lang={$news_lang}");
}
else{
    header("Location: /cpanel/index.php?lang={$lang}&news-lang={$news_lang}");
}
?>
<?php ob_end_flush()?>

Update: with the helpful answers you provided, its solved:
The value submitted by form is non-English, and it contains some HTML entities,
I used htmlentities() on the POST, that made the non-English string unreadable.
Replaced htmlentities() with htmlspecialchars(), and it works like magic.

Conclusion: htmlentities() can ruin non-English strings.

Answer 1

Character encoding is always a hassle. Make sure the page containing your form, the xml you load into $dom, and the php file itself are also utf-8 encoded, or translate accordingly. Otherwise all your strings won’t be, and handling them as utf-8 won’t work.

Try this: echo your original news XML onto an empty page. Then switch page encoding in the browser to see which one displays the characters correctly. Repeat this for $news after retrieving the input from the form. This usually provides a clue on where the encoding goes wrong.