[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference azur::mcc

Title:	DECmcc user notes file. Does not replace IPMT.
Notice:	Use IPMT for problems. Newsletter location in note 6187
Moderator:	TAEC::BEROUD

Created:	Mon Aug 21 1989
Last Modified:	Wed Jun 04 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	6497
Total number of notes:	27359

513.0. "How to ILV-encode a GETEVENT reply ?" by TENERE::LAVILLAT () Tue Dec 04 1990 14:45

I am currently trying to encode a GETEVENT verb for an AM with
the new T1.1.0 kit.

I have some problems to encode the EVENT_REPORT argument in out_p.

I thought I could ILV-encode it as an AttribList, so I tried the
following :

	mcc_put_param_begin   /* with mcc_l_id field = 1 */
	mcc_put_cons_begin    /* with IdCode = 1, Mode = list_value */
	mcc_put_cons_begin    /* with IdCode = EventCode, 
				      DataType = EVENT_REPORT,
				      Mode = native */	
	mcc_put	              /* first arg */
	mcc_put		      /* second arg */
	mcc_put_cons_end
	mcc_put_cons_end
	mcc_put_param_end

In fact, I got this (well known to happy AM developers) :

ILV dump of OUT_P:

[  1 ] (
    [  1 ] (
        [  1 ] (
            [  1 ]             50  -- P
            [  3 ]             00
            [  4 ] (
                [  2 ]                 01
                [  1 ]                 01
                )
            )
        )
    )

GMA_TEST 0 UAL toto
AT  4-DEC-1990 16:33:04 Any Event

Getevent Success

%MCC-E-NOENTITY,  no corresponding entity instance exists

Any Help ?

Thanks

Pierre.


PS: Here is some pieces of my MSL :

{...}

   EVENT PARTITION Ual = 1 :

      EVENT State Change = 1 :
	 DISPLAY = TRUE,
         COUNTED AS = (State Changes),
	 SYMBOL = EVT_CHANGE_STATE,
	 TEXT = "UAL State has change from  !<Previous State> to !<New State>",
         (* The State Change event report is generated each time the
            state of the Ual entity changes. *)
         ARGUMENT Previous State = 1 : UalState
            DISPLAY = TRUE,
            SYMBOL = ARG_EVENT_PREVIOUSSTATE,
         END ARGUMENT Previous State;

         ARGUMENT New State = 2 :  UalState
            DISPLAY = TRUE,
            SYMBOL = ARG_EVENT_NEWSTATE,
         END ARGUMENT New State;  

      END EVENT State Change;

{...}

      DIRECTIVE GetEvent = 65 :      
	DIRECTIVE-TYPE = EVENT, 
        DISPLAY = TRUE,
        SYMBOL = GETEVENT_DIRECTIVE,
        CATEGORIES = (ALL),


      REQUEST
      ARGUMENT Events Wanted = 1 : EventIdList
	   ECHO = TRUE,
	   DISPLAY = TRUE,
	   DEFAULT = "All Events",
	   SYMBOL = ARG_EVENTS_IDS
      END ARGUMENT Events Wanted ;
      END REQUEST ;

      RESPONSE Success =  1 :
	    SYMBOL = GETEVENT_SUCCESS,
	    TEXT = "Getevent Success ",
	    ARGUMENT Arg Getevent = 1 : Event_Report
		SYMBOL= ARG_GETEVENT
	    END ARGUMENT Arg Getevent;
      END RESPONSE Success;

    END DIRECTIVE GetEvent;

{...}

T.R	Title	User	Personal Name	Date	Lines
513.1	how to encode an event reply	GOSTE::CALLANDER		`Tue Dec 04 1990 17:22`	121
	an event report is not an attrib list. The following is a quick run, using alarms FM to show you what the getevent encoding looks like: MCC> create mcc 0 alarm rule foo1 exp=(node4 goste state=on, at every ::15) MCC 0 ALARMS RULE foo1 AT 4-DEC-1990 14:13:26 Entity created successfully. MCC> enable mcc 0 alarm rule foo1 MCC 0 ALARMS RULE foo1 AT 4-DEC-1990 14:14:14 Normal operation has begun. MCC> getevent mcc 0 alarm rule foo1 any notification event ***************************************************** * FCL PM Arguments before call_function: * *************************************************** VERB code: 65 PARTITION code: 16 AES dump of ENTITY IN: depth=3 class code= 7 instance = 0 class code= 1 instance = no curlen class cod e= 5 instance = foo1 ILV dump of IN_P: in_p is NULL ILV dump of IN_Q: in_q is NULL TIME SPEC is: 0, NOW ******************************************** FCL PM Arguments on return from call_function: CVR returned: %MCC-S-RESPONSE, success with response reply ILV dump of OUT_P: [ 1 ] ( ! response code (EVENTRECEIVED) [ 1 ] ( ! argument code (EVENTRESPONSE) [ 25 ] ( ! event code (RULEFIRED) [ 27 ] ( ! event arg (EVIDENCE) [ 1 ] ( [ 1 ] ( di [ 0 ] ( [ 1 ] 01 [ 2 ] 0c [ 3 ] 02 [ 4 ] 19 [ 5 ] 11 fb ) ) [ 2 ] 53 74 61 74 65 20 3d 20 4f 6e 20 -- State = On [ 3 ] 00 43 f2 86 8f 93 c9 01 ff ff ff ff ff ff d4 1e ) ) [ 21 ] 28 6e 6f 64 65 34 20 67 6f 73 74 65 20 73 74 61 74 65 3d 6f 6e 2c 20 61 74 20 65 76 65 72 79 20 3a 3a 31 35 29 00 [ 29 ] 00 ) ) ) AES dump of ENTITY OUT: depth=3 class code= 7 instance = 0 class code= 1 instance = no curlen class cod e= 5 instance = foo1 MCC 0 ALARMS RULE foo1 AT 4-DEC-1990 14:15:01 NOTIFICATION EVENTS Alarm Event Received Alarm Rule Fired EVIDENCE = { ( entity = Node4 4.507 , data = "State = On ", timestamp = 4-DEC-1990 14:15:01.04 ) } Expression = (node4 goste state=on,at every ::15) Perceived Severity = indeterminate MCC> From the dictionary this is the RULEFIRED event definition: DAP> SHO EVENT RULEFIRED Class (1) : MCC Subclass (2) : ALARMS Subclass (2) : RULE -------> Event (10) : RULEFIRED Definition (3) : PRESENTATION_NAME Definition (3) : DISPLAY Definition (3) : REPLY_TEXT Definition (3) : CATEGORIES Argument (4) : EVIDENCE 27 Argument (4) : EXPRESSION 21 Argument (4) : PERCEIVEDSEVERITY 29 DAP> I hope that this helps. The basic statement is that you should encode it the same as you do a standard reply. The major change is where you put in the event code. /jill
513.2	Pb was with my MSL definition of events	TENERE::LAVILLAT		`Wed Dec 05 1990 08:31`	45
	Thanks for your answer. I had already tried this ILV-encoding before, but got this : ILV dump of OUT_P: [ 1 ] ( [ 1 ] ( [ 1 ] ( [ 1 ] 01 [ 2 ] 01 ) ) ) AES dump of ENTITY OUT: depth=2 class code= 9501 instance = 0 class code= 1 instance = toto GMA_TEST 0 UAL toto AT 5-DEC-1990 10:42:43 Any Event Getevent Success %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=0100000C, PC =0023B3B0, PSL=03C00000 %SYSTEM-E-ACCVIO, access violation, reason mask=00, virtual address=0000000A, PC =0000000A, PSL=0000000F $ So because encoded as an AttribList the error message was slightly less discouraging I thought it was a better way to do it... The problem comes from my MSL definition of the event. It seems that the FCL does not handle properly text substitution for events : This works fine : TEXT = "UAL State has change", This will crash : TEXT = "UAL State has change from !<Previous State> to !<New State>", Regards Pierre.
513.3	This ILV output looks correct	WAKEME::ROBERTS	Keith Roberts - DECmcc Alarms Team	`Wed Dec 05 1990 10:45`	23
	RE: .2 >> [ 1 ] ( ! Response Id >> [ 1 ] ( ! Argument Id >> [ 1 ] ( ! Event Code >> [ 1 ] 01 ! Event Argument 1 >> [ 2 ] 01 ! Event Argument 2 >> ) >> ) >> ) You had your Event Report encoded right, the first time. When I looked at your MSL in 513.0 the first thing I noticed were the arguments in the reply text ... I don't know if that is allowed (you obviously got an ACCVIO). None of the events produced by Alarms have Arguments 'in' the text, nor any I've seen generated by the DECnet AM. ?? /Keith
513.4	FCL has limited support for argument substitution in reply text (for ALL reply texts)	TOOK::DITMARS	Pete	`Wed Dec 05 1990 19:28`	24
	Hi, Thanks for the problem report. FCL does support argument substitution in the reply text of an event. It works fine with arguments of the simplest datatypes, but I'm sure it doesn't work with complex beasts like attribute lists, and it probably doesn't work with records, sets, enumerations, etc.. These same restrictions apply to argument substitution for the reply text of RESPONSEs and EXCEPTIONs as well. Can you let us know what the datatypes of the arguments you wanted to substitute were? Were they enumerations? Also, FCL (and the iconic map) is picky about incorrectly spelled argument names. You'll get the "entity not found" message (indicating that there was no information found in the dictionary for that misspelled argument). Again, this goes for EVENTs, RESPONSEs and EXCEPTIONs. In any "event", FCL probably shouldn't have accvio'd. I'll QAR this once I understand better what we're doing wrong. regards, Pete
513.5	They were enumerations datatype	TENERE::LAVILLAT		`Thu Dec 06 1990 06:27`	16
	Both args were enumeration datatypes. I was a little confused since text substitution worked correctly for RESPONSEs and EXCEPTIONs (the point was that arguments were Latin1String or Simplenames in this cases). Maybe you should document this restriction in release notes (sorry if they are, and I did not see it) and try to avoid ACCVIOs ... Anyway I removed my !<xxx> in the TEXT definition and then it worked fine, it is not a great inconvenient. Thanks and regards Pierre.
513.6	enum should work	TOOK::HAO		`Thu Dec 06 1990 12:40`	7
	Text substitions for enumeration datatype is supported. If it doesn't work, it should be qared. The datatypes that aren't supported for text substitution are the constructor/constructed ones like Set Of, Sequence Of, Record, etc. Christine
513.7	this has been QAR'd	TOOK::DITMARS	Pete	`Thu Dec 06 1990 19:08`	1
	It is QAR 84 in the MCC_INTERNAL QAR database.
513.8	SPACES gone	GOSTE::CALLANDER		`Wed Jan 09 1991 12:32`	13
	The problem with extra blank lines should be rectified in the V1.1 final release. Ex: (using 30 day minimum ship kit--prebuild) MCC> register bridge ditmars line 1 BRIDGE KISHKA_NS:.ditmars LINE 1 AT 9-JAN-1991 09:30:14 No such entity: BRIDGE KISHKA_NS:.ditmars LINE 1 Unknown Entity = BRIDGE KISHKA_NS:.ditmars LINE 1 MCC>
513.9	Event with no arg	DGOSW0::GUESDON		`Mon Apr 22 1991 15:01`	3
	When there is no argument associated with the event (For example, the initialisation of the entity is only associated with an event code), how can I encode the event ? Do I build an empty ILV construction ?
513.10	you got it	TOOK::CALLANDER		`Mon Apr 22 1991 16:14`	3
	Simple do a put param begin for the getevent_success response, then do a put cons begin for the event argument then a cons end followed by a param end.
513.11	closed qar 84 fixed	TOOK::CALLANDER		`Thu May 16 1991 16:27`	4
	RE .6, .7 QAR 84 has been fixed and closed, as of the V1.1 SSB release.