PK82916: GARBAGE INVALID CHARACTERS WHEN USING DBCS EBCDIC

APAR status

• Closed as program error.

Error description

• Receiving file contain garbage character.
  Customer set 'LANG=Ja_JP.IBM-939' at USS setting for DBCS code
  conversion.
  Sending DBCS file from Windows side to Unix file on zOS,
  with a 'LF(line feed)' at the end of file.
  It is the problem of dealing with 'LF' on DBCS(using
  0E/0F(shift in/out) environment.
  .
  This looks to be an FTE problem with DBCS EBCDIC. Essentially
  when the end of line character is added we're not correctly
  taking account of the SI/SO state and hence the corruption
  of the data.
  

Local fix

Problem summary

• Text transfers to DBCS EBCDIC codepage files may generate
  corrupted output for new line characters. Essentially FTE did
  not account for the shift in shift out state of the DBCS EBCDIC
  data when adding new line characters.  The result is that when
  in the shift out (double byte mode) state a single byte new line
  character (0x15) is inserted, where the sequence should be to
  write a shift in (single byte mode) then the new line character.
  For example: sequence generated by FTE: 0x0E 0x44 0x8A 0x15 0x0F
  (for data from a file containing text of a single double byte
  character followed by a new line) should be: 0x0E 0x44 0x8A 0x0F
  0x15
  
  USERS AFFTECTED:
  Users that perform text transfer to a DBCS EBCDIC codepage file.
  e.g text transfers to a z/OS system with locale Ja_JP (Japan)
  
  PLATFORMS AFFECTED:
  Z/OS
  

Problem conclusion

• This fixes the problem such that new line characters are
  correctly added for text transfers to DBCS EBCDIC data files.
  

Temporary fix

• Ifix Available
  

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

  WMQ FTE Z/OS

• Fixed component ID

  5655U8000

Applicable component levels

• R700 PSY

     UP