For compile commands what causes the extension to see compile error status?

[m-tyler]

When I create with CTRSQLRPGI I get a notification about the compile failure. But if I compile with our custom compile command, which runs our CMS product's compile command, I don't get compile failures.

What is it that triggers off the notification about the compile being in error? My custom compile command halts with a job completion error level of 20.

When my custom compile command ends with level 20 it also causes the CMS product compile to produce a PL/1 source dump (because of the default reply of PL/1 message IDs). So there is an extra spooled report.

I can see the extension getting the EVFEVENT data for both methods.

[worksofliam]

The extension checks the exit code that is returned from the command. The code can be determined with the exit C function.

[m-tyler]

Okay, that helps. Researching further I can see the issue is with my CMS product's compile command that I wrap in my custom compile command.

I have to use the CMS compile command because it sets the *LIBL correctly which might be a little different than the connection I am currently using.

[Justin-E-Taylor]

Have you found anything on this? We use Alan Campin's COMPILE utility. When I compile, the IBM i Output view shows the joblog. Compile errors aren't shown in the Problems view or inline in the source.

[m-tyler]

Yes, but at the moment I have to use the standard CRT commands and then I end up using our custom build command when all code compiles clean.

[worksofliam]

@Justin-E-Taylor

We use Alan Campin's COMPILE utility. When I compile, the IBM i Output view shows the joblog. Compile errors aren't shown in the Problems view or inline in the source.

This is because the command needs to contain *EVENTF in the string, which will then fetch the errors into the problems view for you.

[Justin-E-Taylor]

Here's the Action I defined (I copied it from a working RDi and changed the variable names):
COMPILE/COMPILE OBJ(&OPENLIB/&OPENMBR) SRCFILE(&OPENLIB/&OPENSPF) SRCMBR(&OPENMBR) ENV(*GUI) BATCH(*NO) /* *EVENTF SRCMBR(&N) */

Here's a simple example I threw together:

**free
//    *_> ADDLIBLE LIB(QGPLT) POSITION(*BEFORE QGPL)
//    *_> CRTSQLRPGI OBJ(QGPLT/@4) SRCFILE(@2/@1) SRCMBR(@3) +
//    *_>   OBJTYPE(*MODULE) OPTION(*EVENTF) DBGVIEW(*SOURCE) +
//    *_>   RPGPPOPT(*LVL2) COMPILEOPT('TGTCCSID(*JOB)') CLOSQLCSR(*ENDMOD)
//    *_> CRTPGM PGM(QGPLT/@4)
//    *_> DLTMOD MODULE(QGPLT/@4)
//    *_> RMVLIBLE LIB(QGPLT)
       Ctl-opt
         Debug(*Yes) Option(*SrcStmt: *NoDebugIO :*NoUnref) Optimize(*Basic)
         ExtBinInt(*Yes)
       ;

       x += 1;

       *inLR = *on;
       Return;

[worksofliam]

@m-tyler Does this actually work in the Action with /* *EVENTF SRCMBR(&N) */ in the command?

[m-tyler]

This is the command I run WBUILD OBJ(pgmr/pgmnam) SRCFILE(pgmr/srcpf) SRCMBR(pgmsrc) DBGVIEW(*SOURCE) OPTION(*EVENTF) USECMS(*YES)

This command is a copy of the tool built by Scott Klement, modified to add logic for our CMS product (Rocket LM(i)). When the LM(i) create command fails, it sets the job status to 0 even though it ends in error.

The issue is with our CMS create tools. I just needed to figure out what this extension looks for to know why my create action failed. I have to modify our WBUILD command to compensate for the CMS create commands.

Until I get time to fix our WBUILD command I have to use the default packed create commands then I can use our command.

[Justin-E-Taylor]

The &N is an RDi variable that I forgot to change. Now that you ask, I don't think that will do anything. That looks like a CL comment. I removed it, but the behavior didn't change. It still runs the compile but fails to bring in the compiler errors.