PM48795: MATHMATICAL CALCULATIONS USING BINARY NUMBER FIELDS AS A SOURCE MAY PROVIDE INCORRECT VALUES ON THE LINUX PLATFORM

APAR status

• Closed as program error.

Error description

• A typetree object defined as a binary number item will display
  correctly, but any mathmatical operations using that typetree
  object when executed on the Linux platform may be missing or
  incorrect.
  
  For example, the following item:
  
  <ITEM SimpleTypeName="BinaryNumberItem"
  CategoryOrItemParent="NumberItems root" partition="NO"
  OrderSubtypes="ADDLAST"><BinaryNumber><BinFloat Length="8"/>
  <ValueRestrictions IgnoreCase="NO"
  Rule="INCLUDE"></ValueRestrictions>
  </BinaryNumber>
  
  and the following rule:
  
  = 7.12345
  
  would produce the output of:
  
  ÊÃB­i~@
  
  The following rule, (using the previous result)
  
  =10*TONUMBER(TEXT(BinaryNumberItem NumberItems:Out))
  
  ...on any other platform, would report:
  
  71.2345
  
  ...but on Linux, you would get:
  
  0.
  
  Additionally, the following rule:
  
  =10*BinaryNumberItem NumberItems:Out
  
  ...should report:
  
  71.2345
  
  However, on Linux you would get a zero length response.
  

Local fix

• WTXCQ - WTX00038929
  PB / PB
  No known workaround.
  

Problem summary

• Users Affected:
  WebSphere Transformation Extender customers using float
  precision numbers on Linux
  
  Problem Description:
  Float numbers are not converted correctly on Linux platform.
  "Output invalid" error may happen.
  
  Platforms Affected:
  Linux, zLinux
  

Problem conclusion

• Correct code to properly convert float numbers on Linux
  
  Delivered in:
  8.1.0.6, 8.2.0.6, 8.3.0.5
  

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:

Fix information

• Fixed component name

  WSTX

• Fixed component ID

  5724M4000

Applicable component levels

• R810 PSY

     UP

• R820 PSY

     UP

• R830 PSY

     UP