[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

4567.0. "MTU converts enum data from "-" to "_"" by VCSESU::WADE (Bill Wade, VAXc Systems & Support Eng) Thu Feb 18 1993 09:05

    
    MTU converts all instances of "-" in  enumerated datatypes to "_".  
    Is this the indended behavior? 
    
       The object in the mib looks like this -
    
      	  transferAction OBJECT-TYPE
		SYNTAX	INTEGER {
				none(1),
				do-mop(2),
				do-tftp(3)
			}
    
    
       But the dictionary shows -
     
	DAP> show class snmp subclass DEC subclass ema subclass gigaswitch 
    	subclass giga upgradesoftware subclass dotransfer attribute 
    	transferaction definition CONSTRUCTOR_DATA_TYPE

   	Definition Name = CONSTRUCTOR_DATA_TYPE
   	Type = T   Length = 27   Count = 1   Defined = TRUE   Class = S   
    								Usage = ILV

         (
          1 = (
              1 = %X    6E6F6E65 -- "none"
              2 = %X    646F5F6D6F70 -- "do_mop"
              3 = %X    646F5F74667470 -- "do_tftp"
            )         )


    Why does this conversion take place?
T.RTitleUserPersonal
Name		DateLines
4567.1VCSESU::WADEBill Wade, VAXc Systems & Support EngFri Feb 19 1993 16:4315
    I guess I can assume that either
    
    	A) no one has given this a second thought and it isn't seen
    	   as a problem or 
           
    	B) its a know problem, its documented, there is a very good reason 
    	   to do this  and there are no plans to make a change
    
    My concern is that someone with zero or very little experience with mcc
    looks in the MIB sees do-mop as a possible value, tries to enter it and 
    then spends half a day trying to figure out why mcc won't accept it as
    input.    
    
    Bill
4567.2Please be patient...MOLAR::DFLAT::PLOUFFEJerryFri Feb 19 1993 17:437
Bill:

  The MTU developer is on vacation and will not be back for 2 weeks.  I have
  forwarded your note to one of his co-workers and hopefully you will get an
  answer to your question.  

                                                                      - Jerry
4567.3The Vacations over!VCSESU::WADEBill Wade, VAXc Systems & Support EngMon Mar 08 1993 16:202
    Any word on this?
4567.4He's baaaaaaaaacccccccccckkkkkkkkkkkk...MOLAR::DFLAT::PLOUFFEJerryMon Mar 08 1993 18:256
Bill:

  Rahul just came back from vacation today.  I just sent mail to him and a 
  co-worker as a reminder.

                                                                     - Jerry
4567.5MOLAR::YAHEY::BOSEMon Mar 15 1993 10:4215
	Bill,

		Sorry for taking so long to answer your question. The reason
	MTU converts "-" to "_" is that "-" is a continuation character in
	FCL and in DAP commands. I thought that leaving the hyphens in 
	attribute and enumeration names might cause complications.

		I understand your problem with this conversion. However,
	with the iconic map interface this problem is alleviated to some
	extent since the user can get a list of possible choices by clicking
	on MB3.

	Rahul.
.
4567.6hyphens should be okayTOOK::GUERTINMCC: Legend or Nightmare?Mon Mar 15 1993 13:008
    Hyphens should not be a problem in that context.  In general, parsing
    algorithms remove comments first, then check if there is a hyphen at the
    end.  In any case it is often valid to have hyphens in data.  Take
    ethernet addresses for example.  If hyphens don't work for enumerations,
    I would QAR dap or fcl.

    -Matt.
4567.7MOLAR::YAHEY::BOSEMon Mar 15 1993 15:167
	I guess the right thing to do would be to keep the hyphens and then
	make sure that everything works correctly with DAP,PTB,FCL,IMPM,
	Alarms,Historian,Exporter, PA ... I'll do it when I can find some
	spare cycles. Till then one will have to live with the underscores.

	Rahul.
4567.8hyphen as last character onlyGOSTE::CALLANDERMon Mar 15 1993 17:216
    Matt is correct. The FCL strips comments first and then 
    checks to see if the last (non-blank) character on the line
    is a hyphen. Only hyphens in that position are accepted as
    line continuation characters.
4567.9VCSESU::WADEBill Wade, VAXc Systems & Support EngTue Mar 16 1993 16:1237
    
    MTU SHOULD NOT HAVE MODIFIED OBJECTS IN THE MIB.  GIVEN THAT MTU DID 
    CHANGE "-" TO "_", THE USER SHOULD HAVE BEEN ALERTED THROUGH OUTPUT FROM 
    MTU.

    Okay, now that I got that off my chest.
	                      
    
    RE .5  
    	
    	MCC> cre domain test-test
    
    	Domain LOCAL_NS:.test-test
    	AT 16-MAR-1993 16:06:19
    
   	Create Successful
    
    
    re .7 Don't change it yet!  Think about the potential impact on
	  the users!   This is the type  of problem where the fix is probably 
	  worse than leaving it as it is.  What happens to the 1.2/1.3 
	  customers who have figured out that there is a man behind the 
	  curtain and they have written startup/load scripts that use the 
	  "_" object names?  
           
    I think we all agree that this isn't like a critical problem.  Although it
    does hurt the ease of use of the product and I can't think of anything 
    that's more important especially when that's the biggest complaint against
    DECmcc.  (Not my complaint, I'm accustomed to the DECmcc way and would 
    probably find SunNet Manager and OpenView Node Manager hard to use.)
              
    At this point I would suggest leaving it as it is and documenting the 
    quirk.  But, that's not necessarily the correct answer either.  These kinds
    of decisions need more input than the opinion of one person.