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 |
V1.1 SSB I know this sounds impossible, but FCL isn't always encoding the default flag in ILV for a SET operation, depending on the order of the attributes specified. If the command is SET BL CART AA0000 SIDE 0 desc = "Test text", readonly the ILV is [ 0 ] ( [ 1 ] ( [ 62 ] ( [ 1 ] 03 [ 2 ] 54 65 73 74 20 74 65 78 74 -- Test text [ 3 ] 00 ) [ 67 ] ( [ 1 ] 01 [ 2 ] ( [ 5 ] ) [ 3 ] 00 ) ) ) and all is as it should be, but if the order of the attributes on the command line is reversed SET BL CART AA0000 SIDE 0 readonly, desc = "Test text" [ 0 ] ( [ 1 ] ( [ 67 ] ( [ 1 ] 01 [ 2 ] 00 [ 3 ] 00 ) [ 62 ] ( [ 1 ] 03 [ 2 ] 54 65 73 74 20 74 65 78 74 -- Test text [ 3 ] 00 ) ) ) FCL has now encoded the actual default value (False) into the ILV stream. As long as the PM and the entity implementation agree on the default value, the two ILV stream give the same results. But it is a curiosity nonetheless. MSL for the two attributes is: ATTRIBUTE Description = 62 : Latin1String (* Free text string that can be used to describe the media, its use, or its contents. The value of this attribute does not influence the operation of the Media Library. *) SYMBOL = ATTR_DESCR, DEFAULT = "" END ATTRIBUTE Description; ATTRIBUTE ReadOnly = 67 : Boolean (* Used to control whether the Media Library is to allow writes to the Media. *) SYMBOL = ATTR_READONLY, DEFAULT = False END ATTRIBUTE ReadOnly;
T.R | Title | User | Personal Name | Date | Lines |
---|---|---|---|---|---|
1185.1 | Yeah, its a bug. | TOOK::CAREY | Tue Jul 02 1991 18:34 | 27 | |
Right now, there is a bug in FCL (Jill has mentioned this one before) where sometimes it includes the displayed default value instead of encoding the "Implementation Default" value that is the architected approach for DECmcc V1.1. That was a change from DECmcc V1.0, where the Director encoded the actual value included in the dictionary (such as Default = False) and the AM told the entity that that is what it wanted. The fix apparently slips under some conditions (which I don't understand entirely). Which for V1.1 is a bug (it wasn't supposed to pass down the values). However, the new Entity Model is out, and the section about Director specified defaults has been clarified or changed (I don't have an old copy handy). And now, the behavior that is a bug for 1.1 (encoding the actual value) is definitely the preferred behavior according to the new entity model. So its sort of six of one, half a dozen of the other right now, and we'll figure out which mode of operation is really a bug for V1.2. -Jim Carey | |||||
1185.2 | TOOK::STRUTT | Management - the one word oxymoron | Wed Jul 03 1991 16:28 | 12 | |
I talked to Jim about this. The REAL entity model (v1.0) ha the corrected text - this was fixed at the last minute (thanks Mark S.) As documented in the SRM, when the user does not specify the value, the flag default is sent down, rather than the default value itself. Thus, the only use (I can think of) for the Director to know the default is to show it to the user to let them know what value will likely be chosen by the agent if the user does not supply the value. Colin | |||||
1185.3 | slippage.. | TOOK::CALLANDER | Jill Callander DTN 226-5316 | Fri Jul 12 1991 12:19 | 9 |
There are a number of cases as stated by Jim, and they get a bit confusing, but suffice it to say that in your case my guess is that the one's you are getting the flag for are the ones that have defaults without specifying a default value (only the flag) in the dictionary. Hopefully this will be fixed in V1.2 (it actually didn't make it onto the fix list, so I will see what i can do to get it in). jill |