| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2529.1 | | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Wed Mar 11 1992 16:27 | 32 |
| RE: .0
> I am running into a small problem when using the Circuit AM in conjunction
> with a home-brewed AM written with the desing Framework.
>
> When setting the admin state of the circuit, my AM gets called at the SET CHAR
> entry point, but the In_Parameter validation fails, because the Datatype field
> of the MCC Descriptor pointing to In_P is not set.
>
> The workaround we found is to set it to MCC_K_DT_ILV at the AM entry point,
> (just like it is when our AM is called by the FCL PM).
>
> Something tells me this may be a bug, I'm surprised it hasn't cropped up with
> other AM's... Or maybe it's fixed in T1.2.4 ?
Philippe,
I just double checked the Design Framework source and it indeed checks
that the In-P Descriptor datatype is set to MCC_K_DT_ILV. Now, the question
is why isn't the datatype field set up correctly (?)
Although the Sample AM doesn't have a set directive, the Example FM does
and uses the Design Framework too...without this problem.
(Q) Are you using the FCL or IMPM?
(Q) Could you post the IN-P sections of your Set Directive .H file here:
MCC_<mm-name>__SET_CHAR.H (or, what ever you have named it.)
Thanks,
Keith
|
| 2529.2 | | CCIIS1::ROGGEBAND | | Fri Mar 13 1992 06:49 | 15 |
| Keith,
When I use the FCL PM, the datatype is set correctly. I only have
problems when I use the Circuit AM, so I would tend to think that the
problem lies in the way the Circuit_AM encode In_parameters...
When using DECnet IV endpoints, there is no problem, so perhaps the
DECnet AM does not validate in_p the same way as the design Framework ?
Anyone from the Circuit AM team listening in ?
Regards,
Philippe.
|
| 2529.3 | Its all getting much clearer now | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Fri Mar 13 1992 14:41 | 19 |
| RE: .2
Philippe.
> When I use the FCL PM, the datatype is set correctly. I only have
> problems when I use the Circuit AM, so I would tend to think that the
> problem lies in the way the Circuit_AM encode In_parameters...
>
> When using DECnet IV endpoints, there is no problem, so perhaps the
> DECnet AM does not validate in_p the same way as the design Framework ?
I suspect then that the Circuit AM is not setting the In-P Datatype field
correctly before they call you AM.
The DNA4 AM is not based on the Toolkit Design Framework -- I know they
have a lot of validation (as the Design Framework code was mostly taken
from their algorithms).
/keith
|
| 2529.4 | QAR 2487 | TOOK::MINTZ | Erik Mintz, DECmcc Development, dtn 226-5033 | Fri Mar 13 1992 16:11 | 3 |
| T1.2.3 is getting rather old now. However, I have filed
QAR 2487 describing this problem
|