Languages

Ciao Marco, avevo già visto i tuoi contributi a proposito.

As I said, the html-parser in 1.9.x does not touch my complex forum posts with mutlilanguage <span> tags. But in 2.x it changes my source to the example shown above, I think when parsing and storing the text I type in. So the multilanguage filter is not the problem I think but the html-parser before it stores everything. WHY is my source modified??

Of course your code-snippet is a good way to handle things on a larger scale. But it must be a bug if something does not work as it used to work in 1.9.x

If this is worth a tracker-issue, and you experience the same misbehaviour, I will post it there.

Rosario

Hi Rosario. The implementation of this feature has not changed between 1.9 and 2.0. I believe the lack of the legacy behaviour is caused by a different setting of your 2.0 site.

The reason of your problem is that the code you use is not valid XHTML. Inline element <span> can't contain any nested block elements like <p>. Therefore the HTMLpurifier (a library that Moodle uses for cleaning the output) modifies the code to make it valid - which in turn breaks the multilang filter syntax.

What I believe is that in your 1.9 site, the HTMLpurifier was not enabled as it used to be an experimental feature for PHP5 servers only that had to be enabled explicitly. In Moodle 2.0, it is enabled by default unless you turn it off.

The only solution I can see for you at the moment is to disable HTMLpurifier. That is definitely not an optimal choice from other points of view but until someone writes a new multilang filter that supports nested block elements, there is probably no other way ho to achieve what you describe.

Hope this helps.

Those are known limitations of the current multilang filter. It was actually designed to support just short texts without HTML, not full blocks of HTML text. You can't include the block element P inside the inline element SPAN as that is not valid HTML. And even the example 3) in your post is tricky because the filter actually stops at the first closing tag while it should do it at the second.

There is a proposal for an improved version of the filter somewhere in the tracker that goes around these limitations.

Ciao Rosario,
I've not still played enough with 2.3 to give you a specific hint but... did you check the PHP configuration for any magic_quotes* setting change in the deploy after the initial 2.3 installation?

HTH,
Matteo

Hi guys,

I have been using moodle for a very short time, and had come up against the problem of providing multi-language content. I'd found a couple of references to enclosing content within formatted span tags for each language, but this kept getting filtered out. I had not found a lot of references to this problem and until now, no solutions.

I am using moodle version 2.2.2

The reason for the filtering of the span tag, has to do with the xhtml standard, which does not allow block tags (such as paragraphs <p>) to be inside a span section. Block tags are however allowed inside divs.

What I've done, is edited the file filter/multilang/filter.php and changed the following:

$search = '/(<span(\s+lang="[a-zA-Z0-9_-]+"|\s+class="multilang"){2}\s*>.*?<\/span>)(\s*<span(\s+lang="[a-zA-Z0-9_-]+"|\s+class="multilang"){2}\s*>.*?<\/span>)+/is';
} else {
// old syntax
$search = '/(<(?:lang|span) lang="[a-zA-Z0-9_-]*".*?>.*?<\/(?:lang|span)>)(\s*<(?:lang|span) lang="[a-zA-Z0-9_-]*".*?>.*?<\/(?:lang|span)>)+/is';
}

to

$search = '/(<(?:div|span)(\s+lang="[a-zA-Z0-9_-]+"|\s+class="multilang"){2}\s*>.*?<\/(?:div|span)>)(\s*<(?:div|span)(\s+lang="[a-zA-Z0-9_-]+"|\s+class="multilang"){2}\s*>.*?<\/(?:div|span)>)+/is';
} else {
// old syntax
$search = '/(<(?:lang|div|span) lang="[a-zA-Z0-9_-]*".*?>.*?<\/(?:lang|div|span)>)(\s*<(?:lang|div|span) lang="[a-zA-Z0-9_-]*".*?>.*?<\/(?:lang|div|span)>)+/is';
}

AND

$searchtosplit = '/<(?:lang|span)[^>]+lang="([a-zA-Z0-9_-]+)"[^>]*>(.*?)<\/(?:lang|span)>/is';

to

$searchtosplit = '/<(?:lang|div|span)[^>]+lang="([a-zA-Z0-9_-]+)"[^>]*>(.*?)<\/(?:lang|div|span)>/is';

These changes now allow me to use the div as well as the span tag. I generally use span in single line input fields such as Course Names, and div for the mulitline textareas such as descriptions. 

I've not tested it extensively, but it has worked for any content I have translated into more than one laguage so far. Hope this can be of help to anyone else.

Hi Sebastian,

I've got to say I am amazed at the number of people who resort to bypassing the issue rather than fixing it. To simply disable the html purifier is asking for trouble. There's a good reason the developers put it in, including filtering out any malicious code.

I admit to being a novice when it comes to moodle, but as I see it, the problem is that at the moment the way you distinguish the different language blocks is by enclosing your text in specially formatted spans, one for each languge, <span lang="en" class="multilang">....</span><span lang="de" class="multilang">....</span>

The problem with span is that it cannot enclose block tags such as <p>...</p>, it would constitute invalid html. The solution to this is to include the <div> tag as an alternative to using <span>, which will happily enclose any other type of block tags and would remain valid html. The code that does the parsing is in the file filter/multilang/filter.php, where regular expressions currently filter out anything but the span tag for use in the language blocks. All that needs to be done is to add the div element to this (see my post directly above yours).

Adding the div tag is the correct solution, removing the html purifier is a dangerous solution.

Regards