Mircodata empty content handling.

[Tim Potter <te...@yahoo-inc.com>]

Hi,
   We've noticed that while extracting Microdata propItems, elements with missing, empty or whitespace string values results in an IllegalArgumentException being thrown.

HTML that trigger this could be one of:

<span itemprop="nameA"></span>


<meta itemprop="nameB">

<meta itemprop="nameC" content="  ">

The code explicitly check for this condition in ItemPropValue.itemPropValue():


if(content instanceof String && ((String) content).trim().length() == 0) {

    throw new IllegalArgumentException("Invalid content '" + content + "'");

}

Is this correct behavior?  Is it not possible that some properties could legally have an empty string as a value?  I looked at the RFC and it seems that having no value should be treated as an empty string, but it doesn't state that empty values are illegal.

http://www.w3.org/TR/html5/microdata.html#values

Regards,
  Tim P.

Re: Mircodata empty content handling.

[Michele Mostarda <mi...@gmail.com>]

Hi Tim,


On 30 April 2012 13:15, Tim Potter <te...@yahoo-inc.com> wrote:

> Hi,
>    We've noticed that while extracting Microdata propItems, elements with
> missing, empty or whitespace string values results in an
> IllegalArgumentException being thrown.
>

Thanks for reporting.


> HTML that trigger this could be one of:
>
> <span itemprop="nameA"></span>
>
> <meta itemprop="nameB">
>
> <meta itemprop="nameC" content="  ">
>
> The code explicitly check for this condition in
> ItemPropValue.itemPropValue():
>
> if(content instanceof String && ((String) content).trim().length() == 0) {
>
>     throw new IllegalArgumentException("Invalid content '" + content + "'"
> );
>
> }
>
> Is this correct behavior?  Is it not possible that some properties could
> legally have an empty string as a value?  I looked at the RFC and it seems
> that having no value should be treated as an empty string, but it doesn't
> state that empty values are illegal.
>

As stressed by [1] the Microdata extractor must be updated to the latest
RDF mapping specification (which evolved after the first MicrodataExtractor
implementation).
 Adding a note to [1].


> http://www.w3.org/TR/html5/microdata.html#values
>
> Regards,
>   Tim P.
>


The best.

Mic

[1] https://issues.apache.org/jira/browse/ANY23-67

-- 
Michele Mostarda
Senior Software Engineer
skype: michele.mostarda
twitter: micmos
mail: me@michelemostarda.com
site : http://www.michelemostarda.com