T.R | Title | User | Personal Name | Date | Lines |
---|
1040.1 | notify and getevent are different | TOOK::CALLANDER | | Wed May 22 1991 14:21 | 15 |
| From the iconic map, when you enable a rule there is also a NOTIFY command
running. The NOTIFY command makes subsequent GETEVENT calls to get the
event (ex. rule fired) information from the alarms module. The GETEVENT
generated from the NOTIFY will have a scope of interest that says I
want all events from now, duration=forever. From the FCL my guess is
you are simply entering the GETEVENT command without adding on any time
qualifiers (FOR clause). If you try it from the FCL with the FOR clause
I bet you get the same problem. Remember that GETEVENT is NOT the same as
a SHOW command and that a time spec is supported, where as if you got
the same time spec on a SHOW you don't support it.
Keith if you are out there, what would be the appropriate desframe call to
make to validate a scope on a GETEVENT command?
jill
|
1040.2 | | PHONE::ALLAIN | | Wed May 22 1991 14:47 | 12 |
| You're right jill,
I get the same problem from the FCL when I try to enable the alarm.
Actually I don't get this problem only when I run my getevent like this:
MCC> getevent cmeagt papa testobj fils bip
....
Francois
|
1040.3 | I removed the mcc_desframe_validate_callargs | PHONE::ALLAIN | | Thu May 23 1991 11:49 | 12 |
| Thanks Jill,
I removed the mcc_desframe_validate_callargs call and it's better, i can get in
my getevent directive and I receive the event...
That's good, but I don't understand why the validate_callargs is wrong... Do you
think, i could have to add a time_spec to my local context, in order to provide
the validate_callargs routine something to compare with?...
Francois.
|
1040.4 | validation for a show | TOOK::CALLANDER | | Thu May 23 1991 12:10 | 4 |
| my understanding of it is that the validation routine you are using is
for validation of call arguments on a SHOW. They take a different time
spec then the GETEVENT. You can modify the validation to not validate the
time and have a separate one for that....any other ideas out there?
|
1040.5 | valid_time_spec | PHONE::ALLAIN | | Thu May 23 1991 13:22 | 12 |
| static
dt_valid_time_spec
valid_time_spec = {
CAV_V_TIME_NOW
};
This is my validation table for the time_spec, does that help?
For the time beieng I'm gonna put a null pointer for the time spec field
in the validation table...
Francois
|
1040.6 | time now | TOOK::CALLANDER | | Thu May 23 1991 14:59 | 2 |
| time now means ST START=NOW (in terms of the FCL qualifiers), that is NOT
what we pass from notification (NOTIFY xxx FOR START=NOW DURATION=FOREVER).
|
1040.7 | The fix is to disable DesFrame Time Validation | NANOVX::ROBERTS | Keith Roberts - DECmcc Toolkit Team | Thu May 23 1991 17:10 | 26 |
| I just check the Alarms GET EVENT code & the Call Argument Validation Tables.
What I did was to disable (by putting in a 0) validation of time:
static dt__ValidCallArgs Valid_CallArgs = {
&valid_verb,
valid_in_entity,
valid_attribute,
0, /* Disable Time Validation - Uses FOR clause */
&valid_in_q,
&valid_in_p, /* EVENTIDLIST Validation done by GET_EVENT code
&valid_handle,
&valid_out_entity,
&valid_time_stamp,
&valid_out_p,
&valid_out_q
};
Currently the DesFrame Validation routines only handle validation of
'time = now' -- but for your get event directive you can process (by
passing the timespec to the event manager) scope-of-interest times.
Putting the 0 in the table is the proper fix (at this time)
/keith
|
1040.8 | Ok. | PHONE::ALLAIN | | Thu May 23 1991 17:31 | 5 |
| Thanks Keith,
This is what I did this morning and it worked fine...
Francois
|