[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| 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 |
175.0. "PROBLEM: Changes to the MSL translator" by FORTY2::MELLING (Chris Melling @REO. 830-4528) Mon Jul 09 1990 12:44
The new version of the MSL translator which comes with EFTU has introduced some
things which we're tripping over, which I'd appreciate some comment on.
1. It is now mandatory to have an identifier attribute partition for an entity.
Our entity is a module of NODE, and as such does not have an identifier
attribute.
So, I create an empty identifier partition. This compiles, but I havent
yet seen if it works yet. The problem for us though is that the DNU
(DECnet-Ultrix) compiler will not accept empty attribute partitions.
2. The keyword SUBORDINATE has been replaced with CHILD. Again, this is a
simple enough mod, but it stops the file compiling under DNU.
I find it odd that SUBORDINATE has been changed to CHILD, but SUPERIOR has
not been changed to PARENT!
What's the right thing to do with all this MSL. Which version do we submit to
the DSSR, the one which compilers under DNU's CMSL, or MCC's MCCMSL? Last time
I looked, the copies of the Phase V MSL files in the DSSR CMS library were the
ones which compile under DNU. Is there some marvy piece of code somewhere which
converts them to compile under MCC? (If so, can I have a copy please?) Or is
MCC maintaining a separate source stream for them?
Will the empty partition mentioned above work? Is this what's been done with
the rest of the Phase V definitions? If not, what is the correct solution?
It seems to me that the whole problem of having two compilers for the MSL is
getting worse. It was bad enough that there were small differences which had
to be worked around by always choosing the lowest common denominator of the two,
but surely we're not heading for completely different syntaxes. Whats the
future for these tools in the long term and the short?
Maybe its just that the language is evolving, and the two compilers are moving
towards the goal at different speeds?
Chris
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 175.1 | Consider yourself lucky... | MARVIN::COBB | Graham R. Cobb (Wide Area Comms.), REO2-G/H9, 830-3917 | Mon Jul 09 1990 14:30 | 17 |
| ...that you aren't trying to get your MSLs built into DECnet-VAX NCL parse
tables (yet!). That introduces a third variation on everything!
Wally Matthews has tried, every so often, to get these problems resolved but
to no avail, so far.
>What's the right thing to do with all this MSL. Which version do we submit to
>the DSSR, the one which compilers under DNU's CMSL, or MCC's MCCMSL?
Come to the meeting next week in REO with the DSSR Registrar (I will mail
you details). However, the DSSR MSL library is an NCL MSL library. You
should submit MSLs which compile under one or (preferably) both of the DNU
and DECnet-VAX MSL compilers.
Welcome to the wacky world of inconsistent MSLs.
Graham
|