T.R | Title | User | Personal Name | Date | Lines |
---|
1548.1 | empty record value. | TOOK::CALLANDER | MCC = My Constant Companion | Wed Sep 25 1991 19:46 | 9 |
| Off the top of my head I can't think of a way. What we currently do in
similar situations is to have a value called NULL or None or something
like that that can be entered. Yes I know that isn't great. Let me think
about it some more. I am simply not ccertain how from a parse stand point
I can solve it. !! unless the v1.2 null set support will allow a null
bitset!! let me get back to you (send mail if you don't hear from me
in the near future).
jill
|
1548.2 | Surely its a bug | MARVIN::COBB | Graham R. Cobb (Wide Area Comms.), REO2-G/H9, 830-3917 | Mon Sep 30 1991 09:59 | 10 |
| This is just a straight bug, right? BITSETs are no different from any other
sets as far as the user is concerned (the *only* difference is in the way
they are encoded on the wire).
A null set is a null set is a null set. The problem here isn't a missing
record field but a field with a perfectly legal set value (that happens to
be the empty set). It *will* be fixed, won't it? BITSETs are used in Phase
V management and need to work!
Graham
|
1548.3 | | MAVIC::D_MOORE | | Tue Oct 08 1991 16:01 | 18 |
|
Yes, this is simply a bug. Jill, it is not a matter of sending a "null"
value, as in an mcc_a_pointer of null. Taking an example like Nestor's,
if you have:
Days_of_week : (Sunday, Monday, ..., Saturday)
Days : BITSET OF Days_of_week
A null value for Days is simply a bitset with all bits set to zero and
the user input is "()". "A null set is a null set is a null set", as
Graham said. Wherever this is broken in MCC, it should not be a big deal
to fix it. The various pieces were put together with this in mind. It is
no harder to parse a null BITSET than it is to parse a null SET OF -- they
can both be represented by "()". However, the internal representation of
those values are different because the data types are different.
- Dave
|
1548.4 | null bitsets in v1.2 | TOOK::CALLANDER | MCC = My Constant Companion | Thu Oct 17 1991 18:42 | 5 |
| I made the changes, Dave is taking care of the SRM (BNF doesn't allow for
null sets to be entered for the bitset). But, I had better double check
that I checked the file back in!!
Thanks for pointing it out.
|
1548.5 | Care to test it? | MARVIN::COBB | Graham R. Cobb (Wide Area Comms.), REO2-G/H9, 830-3917 | Tue Oct 22 1991 17:01 | 7 |
| Thanks for do ing that.
It would be useful to test out the BITSET support. Care to try some
experiments with (DNA5) MOP CIRCUIT FUNCTIONs? We can probably help (e.g.
suggest some things to try and provide a node to try them to).
Graham
|
1548.6 | request for exampls | TOOK::CALLANDER | MCC = My Constant Companion | Wed Nov 06 1991 16:37 | 4 |
| sure, give me some examples to try.
jill
|