PM66374: XCIDYNAMICERROREXCEPTION PARSING ERROR FOR LARGE XML INPUT

APAR status

• Closed as program error.

Error description

• A parsing error may occur for particular XML inputs larger
  than 64K if the end element tag for the parent of a skipped
  subtree is split across input buffers. This is a rare condition
  that will occur only for certain input data when using lazy
  parsing mode in products such as WebSphere Process Server.
  
  When this occurs, an exception such as the following
  will be thrown:
  
  com.ibm.xml.xci.errors.XCIDynamicErrorException:
  The content of elements must consist of well-formed
  character data or markup.
   at com.ibm.xml.xci.adapters.xlxp.scanner.XCIScanner
       .produceFatalErrorEvent(XCIScanner.java:636)
   at com.ibm.xml.xlxp.internal.s1.scan.DocumentScanner
       .reportFatalError(DocumentScanner.java:4980)
   at com.ibm.xml.xlxp.internal.s1.scan.DocumentScanner
       .reportFatalError(DocumentScanner.java:1333)
   at com.ibm.xml.xlxp.internal.s1.scan.DocumentScanner
       .scanMarkup(DocumentScanner.java:2294)
   at com.ibm.xml.xlxp.internal.s1.scan.DocumentScanner
       .nextEvent(DocumentScanner.java:1464)
   at com.ibm.xml.xci.adapters.xlxp.scanner.XCIScanner
       .toNextElementNode(XCIScanner.java:868)
   at com.ibm.xml.xci.adapters.xlxp.scanner
       .DelegatingSubtreeScanner.toNextElementNode
       (DelegatingSubtreeScanner.java:258)
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All IBM WebSphere Application Server V7     *
  *                  Feature Pack for XML users                  *
  ****************************************************************
  * PROBLEM DESCRIPTION: Error occurs for rare scenarios         *
  *                      involving                               *
  *                      skipped subtree content that is split   *
  *                      across buffer boundaries.               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  The issue occurs in skipped subtrees where the end tag of the
  parent element (which owns the skipped content) is split across
  buffer boundaries.   More specifically, it occurs when the '<'
  character of the end tag is the last character in one buffer,
  and the '/' character of the end tag is the first character of
  the next buffer.
  

Problem conclusion

• The algorithm for processing split skipped subtree content has
  been corrected to properly handle these edge cases.
  
  The fix for this APAR is currently targeted for inclusion in the
  XML Feature Pack 1.0.0.13. Please refer to the Recommended
  Updates page for delivery information:
  http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  PM66572

Fix information

• Fixed component name

  XML FEATUREPACK

• Fixed component ID

  5724J0856

Applicable component levels

• R700 PSY

     UP