GSP
Quick Navigator

Search Site

Unix VPS
A - Starter
B - Basic
C - Preferred
D - Commercial
MPS - Dedicated
Previous VPSs
* Sign Up! *

Support
Contact Us
Online Help
Handbooks
Domain Status
Man Pages

FAQ
Virtual Servers
Pricing
Billing
Technical

Network
Facilities
Connectivity
Topology Map

Miscellaneous
Server Agreement
Year 2038
Credits
 

USA Flag

 

 

Man Pages


Manual Reference Pages  -  XML::FILTER::CHARACTERCHUNK (3)

.ds Aq ’

NAME

XML::Filter::CharacterChunk - SAX Filter for Wellballanced Chunks

CONTENTS

SYNOPSIS



  use XML::Filter::CharacterChunk;

  my $filter = XML::Filter::CharacterChunk->new();



DESCRIPTION

From time to time it happens that there are XML informations in a SAX stream, that are not available as events, but as a string. This filter offers an interface to transfrom such strings into ordinary SAX events. It will collect all charaters() calls within a element and build a data string from it. As soon the element ends, this data string will be processed as it would be well balanced XML data: the following handler will recive the appropriate SAX events instead of simple characters() calls.

The following example could be produced as a result of database -> XML transformation:



   <foo>
      <expand>
         expand the <bar value="foobar"/>
      </expand>
      <other>
         data
      </other>
   </foo>



While data in databases usually is stored as strings, this string will appear as a XML encoded string instead of structured XML data. When processed with SAX, this filter can do the expansion for the tag that should contain XML data rather that Text data. Because not all data may be expanded, the filter can be configured only to attend to certain tags.

For our example the filter would be set up as following:



   $filter = XML::Filter::CharacterChunk->new( TagName=>["expand"] );



This will cause the filter to wait for <B>expandB>-elements to be started before it collects the data. Therefore the filter will leave the data inside the <B>otherB>-element untouched. The result a SAX chain would be the same as if the following document was processed:



   <foo>
      <expand>
         expand the <bar value="foobar"/>
      </expand>
      <other>
         data
      </other>
   </foo>



In our example one a single tag is expanded by the filter. It is possible to defined as many tag names as required for expansion. If in our example the other-element should be processed as well, it has to be added to the observed tag names. Either



  $filter = XML::Filter::CharacterChunk->new( TagName=>["expand", "other"] );



or



  $filter->set_tagnames("other");



set_tagnames() will add a list of additional tag names to the TagNames specified with the constructor.

Another thing is to observe a special namespace. This lets the filter wait for tags within this namespace. Currently the filter allows only to test a single namespace.



  $filter = XML::Filter::CharacterChunk->new(
                   TagName=>["expand", "other"],
                   NamespaceURI=>"foo" );



will cause that the example will not be processed at all. If a namespace uri is set, the filter will only attend to that namespace and will ignore any global tags (or prefixes, if set).

    Constructor Extensions

XML::LibXML’s parse_xml_chunk() will throw exceptions if the parsing fails. By default XML::Filter::CharacterChunk will note this, but generate any warnings. This can be changed if XML::Filter::CharacterChunk->new() will recieve the extra parameter Warnings => 1. This will cause the class to warn() some information about the failure.



   my $filter = XML::Filter::CharacterChunk->new(Handler => $some_handler, Warnings => 1);



As shown this flag can get passed to the filter as ordinary XML::SAX construction parameters are passed.

    General Pitfalls

XML::Filter::CharacterChunk will <B>onlyB> collect the data from characters() calls. It is not designed for mixed data. The collected information will be processed if the containing element ends. The filter assumes silently that the containing element has only text data. Therefore the document



   <foo>
      <expand>
         this <bar> is <tag/> wrong </bar>
      </expand>
   </foo>



will <B>notB> result



  <foo>
     <expand>
        this <bar> is <tag/> wrong </bar>
     </expand>
  </foo>



but



  <foo>
     <expand>
        <tag/>this <bar> is wrong </bar>
     <expand>
  </foo>



because the tag-element is not processed by the filter but imediatly passed to the filters handler. XML::Filter::CharacterChunk will process the data at the last possible point, which is right before the element is closed. With mixed content this will lead to data restructuring and may cause some confusion. There is only one solution: <B>Do not use XML::Filter::CharacterChunk for mixed data!B>

Another problem appears with nested tags of the same name, such as



  <foo>
    <expand>
        test <expand>this <foobar/></expand> extra <foobar/>
    </expand>
  </foo>



It will cause the filter to find all data until the <B>firstB> closing tag. Doing so, the result will be unexpected:



 <foo>
    <expand>
        <expand>test this <foobar/></expand>  extra <foobar/>
    </expand>
 </foo>



I am not shure if this is bad behaviour or just bad data design. Actually, this will cause loss of data, if the chunk is not well balanced when the first end_element() is caugth.



  <foo>
      <expand>
          <foobar/><expand> test this </expand> extra <foobar/>
      </expand>
  </foo>



will result as a side effect of this problem:



  <foo>
     <expand>
          <expand/> extra <foobar/>
     </expand>
  </foo>



BUGS

Namespaces used in the chunk are not expanded properly.

AUTHOR

Christian Glahn, christian.glahn@uibk.ac.at, Innsbruck University

SEE ALSO

XML::Filter::GenericChunk
Search for    or go to Top of page |  Section 3 |  Main Index


perl v5.20.3 CHARACTERCHUNK (3) 2016-04-03

Powered by GSP Visit the GSP FreeBSD Man Page Interface.
Output converted with manServer 1.07.