Code Inspector - CALL TRANSACTION

[Peter_Inotai]

Hi,

In my program I use 'CALL TRANSACTION' statement, and when I run a Code Inspection on it, I got the following message:

<i>CA CL_CI_TEST_CRITICAL_STATEMENTS0002

____________________________________________________

Code Inspector

Critical Statements

Call of Transaction &1

&1 = Name of Transaction

For CALL TRANSACTION there must already be a suitable transaction authorization with the calling transaction.

Message can be hidden using pseudo-comment "#EC CI_CALLTA

</i>

Even if I check the authorization before I still get the same message.

I do know that with pseudo-comment I can hide this message, but it's not the way I'm looking for.

Can it be the problem that it's a 46C version of Code Inspector?

What is the correct way of calling transaction?

Thanks in advance,

Peter Inotai

My relevant code fragmant:

...
*-----> Vendor - XK03
        ELSEIF gwa_alvouttab-objectclas = c_objectclas_vend.

          CALL FUNCTION 'AUTHORITY_CHECK_TCODE'
               EXPORTING
                    tcode  = 'XK03'
               EXCEPTIONS
                    ok     = 1
                    not_ok = 2
                    OTHERS = 3.

          IF sy-subrc > 1.
*           No authorization for transaction &1
            MESSAGE i063(vo) WITH 'XK03'.
          ELSE.
            SET PARAMETER ID 'LIF' FIELD gwa_alvouttab-kunnr .
            SET PARAMETER ID 'KDY' FIELD '/130/120/111'. " General data
            CALL TRANSACTION 'XK03' AND SKIP FIRST SCREEN."#EC CI_CALLTA
          ENDIF.

        ENDIF.
...

[Former Member · Accepted solution]

Hi Peter,

When Using a Call Transaction Method, I think You need to handle certain exceptions like Error MEssages,Mode of Update and so on,Any way am sending you the syntax for calling a Transaction.

CALL TRANSACTION 'LQ02' USING git_bdcdata

OPTIONS FROM git_ctu

MESSAGES INTO git_messtab.

Where the structure git_ctu is of type ctu_params and

git_messtab of type bdcmsgcoll.

Regards,

Rajeev.N

[Peter_Inotai · Accepted solution]

Hi Rajeev.N ,

Thanks for your answer.

I think the problem is not with the message handling, Code Inspector was complaining about Authorization check.

From the other hand I can use SET/GET parameters in the mentioned case, so I don't need bdcdata table for the transaction call.

Peter

[former_member183804 · Accepted solution]

Hello Peter,

as you already know 'CALL TRANSACTION xxx' does not check permissions and requires a prior call to FuBa 'AUTHORITY_CHECK_TCODE' (as described in the online docu).

For a static check it would be hard to check if this is done. And if one looks at the implementation of the CI test CL_CI_TEST_CRITICAL_STATEMENTS=>Run( ) its easy to discover that the test only scans for the tokens 'CALL TRANSACTION'. So I assume the pseudo comment '"#EC CI_CALLTA' should be used to tag the coding as okay.

Kind Regards

Klaus

[Peter_Inotai · Accepted solution]

Hi Klaus,

Thanks a lot for your help.

I didn't know method CL_CI_TEST_CRITICAL_STATEMENTS=>RUN and all the relevant classes, but it's good to know that the checks are based on these classes.

I'm just wondering pseudo-commets why don't work for the 46C version, however within CL_CI_TEST_CRITICAL_STATEMENTS=>RUN seems to be considered. Probable the answer is in another method.

Regards,

Peter

[former_member183804 · Accepted solution]

Hello Peter,

as far as I understand is the pseudo code checking done within the Inform() method. I set a break-point within the case of the Run() method, maybe you can try too.

First the Error description was read from the error code:

  read table SCIMESSAGES into SMSG
       with table key TEST = P_TEST
                      CODE = P_CODE.

The SMsg structure contained a component with the pseudo comment. This string was checked against the tokenized source.

Kind Regards

Klaus

[Peter_Inotai · Accepted solution]

Hi Klaus,

Thanks a lot for your help again.

I checked the method INFORM and it always went to 'raise event MESSAGE ...' statement in line 54.

I check in a WAS 6.20 system this method, and I found the following difference:

In line 9

4.6C:

  if P_POSITION > 0.

6.20:

  if P_POSITION > 0 and P_SUPPRESS is not initial.

I might make difference, but I'm not sure.

I don't think it makes sense to modify this line, it's better for wait to the Upgrade.

When I have time I'll debug these methods, as it's quite interesting to understang on this level that how Code Inspector works.

Regards,

Peter