Add parameter to CHGPF to ignore CPA32B2

Our CI/CD software uses SSH to communicate with IBM i. In addition to installing ILE program objects, IFS files, ... it can also execute CHGPF.

When removing a column from a PF IBM i will send the message CPA32B2 which will be automatically answerded with the default answer C (cancel).

To avoid this IBM describes to use "CHGJOB INQMSGRPY(*SYSRPYL)" and "ADDRPYLE SEQNBR(17) MSGID(CPA32B2) RPY('I')" (https://www.ibm.com/support/pages/handling-inquiry-messages-database-server-jobs). But we don't want to change the system reply list on a customer system.

As a workaround we wanted to execute the following CL commands using the bash buildin cl, which will execute the commands in the same job:

CRTMSGF MSGF(QTEMP/X)

MRGMSGF FROMMSGF(QSYS/QCPFMSG) TOMSGF(QTEMP/X) SELECT(CPA32B2)

CHGMSGD MSGID(CPA32B2) MSGF(QTEMP/X) DFT(I)

OVRMSGF MSGF(QCPFMSG) TOMSGF(QTEMP/X)

CHGPF ...

OVRMSGF will not work in bash because the bash buildin command cl will internally execute QCMDEXC in a default activation group. Because of the default activation group the scope of the override will be changed to call level. Therefore the override is automatically removed after QCMDEXC (OVRMSGF) ends.

To fix this we must call a CL-program from bash which includes the above commands.

This workaround is not very nice. So we need some extension in IBM i. For example:

• A new parameter in CL command CHGPF to ignore the message CPA32B2 (like option *IGNINQMSG in CL command DLTJRNRCV)

• An environment variable to ignore that message (should be easier to deliver in a current IBM i release via PTF)

• A new parameter OVRSCOPE (*JOB) in CL command OVRMSGF (like in all other OVR-commands), so that we can call the above commands directly from bash cl buildin