Whether use CDATA to handle XML

[lampnick]

When we use setValue() function,if we pass the $replace parameter with special character such as &,<, etc.The generated word document can't open with the error message like "illegal character" in /word/document.xml,line：2，column：925564.So,I suggest to use default CDATA to surround the $replace param.

[FBnil]

@lampnick
How does MSWord/LibreOffice handle cdata? Is it still editable (wordbreaks, paragraphs) and does the spellchecker still work?
How about a subclass that overwrites protected static function ensureUtf8Encoded($subject) to do cdata instead of UTF8 encoding? What is wrong with the current Settings::isOutputEscapingEnabled() implementation?

[lampnick]

@FBnil It's passed by php5.3,5.4,5.5, but in php5.6 throwed the following exception:
"1) PhpOffice\PhpWord\Element\ImageTest::testUnsupportedImage
Failed asserting that exception of type "PhpOffice\PhpWord\Exception\InvalidImageException" matches expected exception "\PhpOffice\PhpWord\Exception\UnsupportedImageTypeException". Message was: "Invalid image: http://samples.libav.org/image-samples/RACECAR.BMP"."

[FBnil]

@lampnick in ./tests/PhpWord/Element/ImageTest.php you have this function:

 /** * Test unsupported image * * @expectedException \PhpOffice\PhpWord\Exception\UnsupportedImageTypeException */ public function testUnsupportedImage() { //disable ssl verification, never do this in real application, you should pass the certiciate instead!!! $arrContextOptions = array( "ssl" => array( "verify_peer" => false, "verify_peer_name" => false, ), ); stream_context_set_default($arrContextOptions); $object = new Image('https://samples.libav.org/image-samples/RACECAR.BMP'); $object->getSource(); } 

See that @expectedException in the comments? It is interpreted as "this function will throw an exception of this type" during the testing.

Now, in ./src/PhpWord/Element/Image.php we have two nested "trow" commands in __construct() (which is what you call when you do a new Image() ) and checkImage() has the

 if (!is_array($imageData)) { throw new InvalidImageException(sprintf('Invalid image: %s', $this->source)); } 

Which could happen, when, for example, the website samples.libav.org is too busy and does NOT give you an image. So, During NOT USA daytime hours, edit "certiciate " into "certificate" and push it again (in the hope, this time, it goes all ok).

[lampnick]

@FBnil Knowbie, Thank you very much!

[FBnil]

I DO see one 'but', and that is the following:

First we do a

 $replace = '<![CDATA[' . $replace . ']]>'; 

And then we do a

$replace = $xmlEscaper->escape($replace); 

This basically means:

(The escape() function does this: return htmlspecialchars($input, ENT_QUOTES, 'UTF-8');)

'<![CDATA[' . $replace . ']]>'; becomes '>![CDATA[' . $replace . ']]<'; BEFORE the insertion into the document.

You solve this by: if (!$useCDATA && Settings::isOutputEscapingEnabled()) {

What you also need is a TestCase,
I added reflections in my testcases (see the peek and poke functions), this allows me to rewrite the private tempDocumentMainPart variable that holds the document xml:
$this->poke($templateProcessor, 'tempDocumentMainPart', $newXmlStr);
And read it out with peek:
$this->peek($templateProcessor, 'tempDocumentMainPart')
so you can make many testcases and combinations without creating yet another word document.

[lampnick]

@FBnil Thank you! I'll try.

[troosan]

All you need to do is call

Settings::setOutputEscapingEnabled(true);