Program Status Data Structure

Program Status Data Structure

The PSDS (PSDS) keyword on a DCL-DS declares the Program Status Data Structure. Its subfields are placed at the positions IBM i defines, with POS, and hold the same information there: code written for IBM i reads the same fields.

DCL-DS PgmInfo PSDS QUALIFIED;
  ProcName CHAR(10)   POS(1);     // procedure name
  Status   ZONED(5:0) POS(11);    // status code
  ExcType  CHAR(3)    POS(40);    // exception type, e.g. RNX
  ExcNbr   CHAR(4)    POS(43);    // exception number, e.g. 0100
  UserName CHAR(10)   POS(254);   // user
  JobNum   ZONED(6:0) POS(264);   // job number
  PgmName  CHAR(10)   POS(334);   // program
END-DS;

POS	Type	Content	OpenRPG supplies
1	CHAR(10)	Procedure name	the program's name
11	ZONED(5:0)	Status code	✅
16	ZONED(5:0)	Previous status code	✅
37	ZONED(3:0)	Number of parameters	✅
40	CHAR(3)	Exception type (RNX, MCH, CPF...)	✅ for errors with an ID
43	CHAR(4)	Exception number	✅
91	CHAR(80)	Exception data: the message text	✅
244	CHAR(10)	Job name	the program's name
254	CHAR(10)	User	the login name, upper case
264	ZONED(6:0)	Job number	the process ID (last six digits)
276	ZONED(6:0)	Date the program ran	MMDDYY
282	ZONED(6:0)	Time the program ran	HHMMSS
334	CHAR(10)	Program name	✅
344	CHAR(10)	Module name	the program's name
358	CHAR(10)	Current user profile	the login name

Positions IBM i defines that have no equivalent here -- the program's library (81), the statement number (21), file information -- read as blanks or zero. Numeric subfields are zoned, as on IBM i: a status declared PACKED(5:0) at position 11 reads bytes that are not packed decimal data there.

The PSDS is refreshed when an error is handled -- at the start of every ON-ERROR and of the *PSSR -- so inside a handler it holds that error's status, the previous status and the exception:

MONITOR;
  x = 10 / zero;
ON-ERROR;
  DSPLY ('status ' + %CHAR(PgmInfo.Status));   // status 102
ENDMON;