Re: RDi Compile and CMD Analyzer

Michael,

Re: the PTFs -- Yeah, several of my customers reported this weird thing
where stuff would randomly stop working (in several of my products that sit
in that exit point for various but very different reasons). You don't get
any error message or any indication at all that anything is wrong, just
things that don't quite work like they should. Once we determined it was
an IBM issue, getting a PTF only took 6 months. Usually we were able to
make it stop working when we executed a PROXY command with a different name
than the real command in a library other than QSYS (like, QGPL/SQL as a
proxy to QSQL/STRSQL). The Command Analyzer, after that, would NEVER
trigger that exit point again until you destroyed it and re-registered (I
think because the command, as they figured it, was "QGPL/STRSQL" and it
broke something since there's no such command -- they never replaced the
real LIBRARY, assuming it was QSYS, they just replaced the name of the
command with the real one!)

The late evening (CST) phone calls to China were a hoot!

Stu









On Mon, Jan 25, 2016 at 1:11 PM, <MichaelQuigley@xxxxxxxxxx> wrote:

Stuart,

Thanks for clarifying the limitation stated in the documentation. The exit
program will be called, but the command cannot be changed. If his CMD
Analyzer is being called and tries to change the command, that part won't
work. However, you're right it should still call it to do other
pre-processor work such as file overrides, etc. Buck's suggestion to check
what command is actually being executed would be wise. Also, it could be
related to your note that the QSYS/CRTBNDRPG is an alias for
QDEVTOOLS/CRTBNDRPG. Perhaps de-registering and re-registering the exit
points would do the trick?

I would also highly suspect the PTF you point out could have an impact.
It's not on any Cum so it very well could be missing. The description on
the PTF would make me want to always register the exit program on the
QDEVTOOLS command. I don't see how (with the PTF applied), someone could
work around the exit program.

Michael Quigley
Computer Services
The Way International
www.TheWay.org

"WDSCI-L" <wdsci-l-bounces@xxxxxxxxxxxx> wrote on 01/25/2016 01:00:26 PM:

----- Message from Stuart Rowe <rowestu@xxxxxxxxx> on Mon, 25 Jan
2016 08:58:58 -0600 -----

To:

"Rational Developer for IBM i / Websphere Development Studio Client
for System i & iSeries" <wdsci-l@xxxxxxxxxxxx>

Subject:

Re: [WDSCI-L] RDi Compile and CMD Analyzer

Gad, MichaelQ, et al,

A program running in system state has naught to do with CALLING the exit
program, only what you can do to the command string DURING the exit

program.

The exit program is called, but the command *cannot be changed* if any

of

the following occur:

- The command was qualified with a specific library name.
- The command has a parameter defined as RTNVAL(*YES), which allows

the

command processing program (CPP) to return a value to the calling

program.

- The command has a parameter defined as either DSPINPUT(*NO) or
DSPINPUT(*PROMPT), which does not allow the value to be shown in the

prompt

display or joblog.
- The command was used in a program running in system state.


Gad:

What you are doing should work just fine. Do not give up.

I just registered a little "tattle-tale" program to that exit point and

ran

some compiles through RDi -- it triggered the exit program EVERY TIME.
There was no difference in the parameters passed to the called program
whether I compiled in batch or inline (in RDi, the opposite of BATCH is
INLINE, not INTERACTIVE), the exit program was triggered with the same

data

every time. I registered QSYS/CRTRPGMOD, QSYS/CRTBNDRPG, etc. I also
tried it with QDEVTOOLS/CRTRPGMOD and it worked just fine every time. I
cannot imagine what your problem is.

Have you set a service break in the pre-processor program to be SURE it

is

not being called (and not just doing nothing)? You should, first thing,
and let us know if it IS being called. Don't just judge by end results.

Other stuff to check:

1. If your exit program is checking the command name and/or library
before doing its job, make sure it is checking for the correct

command name

and/or library. When I get the exit point triggered (either by RDi

or

PDM), the "command being executed" (in the command information

structure

passed as paramter 1) is QDEVTOOLS/CRTRPGMOD while the proxy chain
indicates it really started with QSYS/CRTRPGMOD.
2. There are settings on each Connection in RDi that specify which

*JOBD

is used for batch compiles -- make sure that if the jobd specifies a

user

profile that it has authority to call the pre-processor program or

make

sure your pre-processor has PUBLIC *USE authority on it.
3. Check the joblog of your compile job for errors in calling or

while

running your pre-compiler.
1. if in batch, it will be named after your job description,
2. if inline it will be in the remote command server job that is
processing your requests (QZRCSRVS with your user ID on it)

Recently (the past year or so) I have had several issues with the

Command

Analyzer not working correctly when using the QIBM_QCA_CHG_COMMAND exit
point wikth proxy commands (several bugs were introduced in V5R4 with

the

proxy command implementation). When a proxy command is used (CRTBNDRPG,
CRTCPPMOD, etc are PROXY commands in QSYS), I find it wise to register

the

root command (QDEVTOOLS/CRTBNDRPG) instead of the more obvious
QSYS/CRTBNDRPG (I believe Bruce Vining pointed this out to you last
August). Not saying this is your issue, just pointing out that if you
register QSYS/CRTRPGMOD, someone can dodge your exit program by

executing

QDEVTOOLS/CRTRPGMOD.
One of the issues I had fixed lately was the exit program would just

STOP

BEING CALLED until you de-registered and re-registered the exit program.
We never figured out exactly why it sould stop being called, even after

a

couple hours on the phone with China (the command analyzer is maintained

by

Chinese contractors now). At any rate, here is a PTF that may (or may

not)

help:
V6R1 = SI58104
V7R1 = SI58105
V7R2 = SI58106

Good luck.

Stu







On Mon, Jan 25, 2016 at 4:04 AM, Gad Miron <gadmiron@xxxxxxxxx> wrote:

Hello Mike

I've cheked,

The Library list is correct and the (prompted) invoked compile request

is

correct
CRTBNDRPG PGM(GAD/BB) SRCFILE(GAD/QRPGLESRC) SRCMBR(BB)

OPTION(*EVENTF)

DBGVIEW(*SOURCE) REPLACE(*YES)
*But*
the CMD Analyzer PGM is not invoked

Gad


On Sun, Jan 24, 2016 at 8:07 PM, Mike Hockings <hockings@xxxxxxxxxx>
wrote:

RDI does not do anything special, it just invokes the compile

command so

I

would expect that any exit programs should run. Make sure that the

library

list is as expected and you could try prompting the compile command

to

make

sure that any specific settings are defined the way you expect as

well.

Mike Hockings, M.Eng., P.Eng.
IBM Rational Developer for System z and Power Systems Software

Technical

Support
IBM Canada Ltd. Laboratory
hockings@xxxxxxxxxx

--
This is the Rational Developer for IBM i / Websphere Development

Studio

Client for System i & iSeries (WDSCI-L) mailing list
To post a message email: WDSCI-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/wdsci-l.

----- Message from Buck Calabro <kc2hiz@xxxxxxxxx> on Mon, 25 Jan
2016 11:44:38 -0500 -----

To:

wdsci-l@xxxxxxxxxxxx

Subject:

Re: [WDSCI-L] RDi Compile and CMD Analyzer

On 1/24/2016 10:16 AM, Gad Miron wrote:

I'm afraid I wouldn't know where to begin creating a CMD in RDi as I'm

a

newbei novice indeed
when it comes to RDi.

Can you send me a lead please

Compile > Work With Compile Commands...

In light of Stu's experience, perhaps you should check to see that your
RDi is in fact using the command you think it should. Is it possible
that it's been set up to use a custom command?

--
--buck

--
This is the Rational Developer for IBM i / Websphere Development Studio
Client for System i & iSeries (WDSCI-L) mailing list
To post a message email: WDSCI-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/wdsci-l.