Conference azur::mcc

1410.0. "dyn child with unsigned8 ident wildcard ?" by UTES09::GOES (Paul Goes - T&N Digital Utrecht) Wed Aug 28 1991 09:40

Hi I'm facing an problem:

I want to have a dynamic child entity class that can be used with the
iconic map PM. I know that I have to support a 
SHOW class <inst> child * ALL IDENTIFIERS directive in my AM.

The problem that I have, is that the IDENTIFIER of the child entity
class is of datatype UNSIGNED8, and according to the SRM (chap 9) it is
not possible to wildcard an UNSIGNED8 datatype.

Is the following possible:

In the validation table of ...SHOW_IDENTIFIER.H:

static dt_valid_in_entity_args valid_in_entity_args[] =
	{ MCC_K_CLASS_PARENT, CAV_V_INST_NO_WILDCARD, MCC_K_DT_FULL_NAME,
          MCC_K_CLASS_CHILD, CAV_INST_FULL_WILDCARD, MCC_K_DT_UNSIGNED8,
	  CAV_K_END_OF_LIST };

In the do_directive of ...SHOW_IDENTIFIER.C:

/* note that this is no real code */

/* create descriptor */
mcc_desframe_create_descriptor( ...input_dsc... );

/* get instance descriptor from in_entity */
mcc_aes_get(&p_callargs->p_in_entity, level_2 , input_dsc, ...);

/* test for wildcard */
mcc_aes_wild( input_dsc, level_2, type);

Is type now MCC_K_AES_INSTANCE_FULL when I issue:
	SHOW PARENT <inst> CHILD * ALL IDENTIFIER        ???


Regards Paul

    Paul,
    
    The PMs always allow full entity instance wildcarding, so your
    SHOW class <inst> child * ALL IDENT will be passed down to your AM.
    
    I believe the wildcarding rules specified in Chapter 9 of the SRM only
    applies to attribute/argument values, and maybe partial entity
    wildcarding.  For example, you can't have SHOW class <inst> child 9*
    ALL IDENT.
    
    Christine
    

The validation tables look fine.

In your code, the call to 'mcc_aes_get' will return the wildcard_type field.
This would indicate what type of wildcard was entered - in your case a
FULL_WILDCARD.

The 'mcc_aes_wild' searches through the AES spec looking for the first
wildcard, and returning the level.

>> Is type now MCC_K_AES_INSTANCE_FULL when I issue:
>>   SHOW PARENT <inst> CHILD * ALL IDENTIFIER        ???

Yes - I believe this is true.

I think the Bridge team allows wildcarding on the PORT child entity instance,
which is some type of integer value (probably unsigned8 too)

/keith