Re: Tag values in 9X

[news.intercon.com!psinntp!calspan!westley@louie.udel.edu (Terry J. Westl]

In article <1s8p3h$sud@huon.itd.adelaide.edu.au> 
  andrewd@cs.adelaide.edu.au writes:

>The draft LRM for 9X says (14.6.4) "The value used to represent each
>tag is determined at the time the corresponding type declaration is
>compiled." 
>
>I have a couple of questions about this.
>
>Scenario 1. Jane creates some tagged types, writes them to a file
>using stream io, and sends the file to Fred, with the message "Here
>are the sources I used to create the file." Can Fred, using the same
>architecture/compiler be guaranteed to read correctly the file he
>receives from Jane? That is, will the same tag values be generated at
>both ends?

I have a question along these lines.  We have built a TCP/IP
communication system which sends messages between heterogeneous
architectures (SPARC and MC68030) using Verdix' compiler.

We have adopted a convention whereby we MUST fully specified the
message with representation clauses to assure it has the same
structure on both architectures.

I am evaluating this system for conversion to 9X, primarily because
we can use a class of messages rooted at the default message which
is a header (source, destination, message id, etc.) only.  This avoids
the current problem of converting to/from arrays of bytes where
multiple message types must be handled, thereby losing type checking.

The question is this:  How can I specify the tag in the record
representation clause?  And, as Mr. Dunstan has pointed out, how can I
be sure that the tag will identify the same type in both systems?

>
>andrew
>
>#  Andrew Dunstan                   #   There's nothing good or bad   #
>#  net:                             #                                 #
>#    adunstan@steptoe.adl.csa.oz.au #   but thinking makes it so.     #
>#  or: andrewd@cs.adelaide.edu.au   #                                 #
-- 
Terry J. Westley, Principal Computer Scientist
Calspan Corporation, P.O. Box 400, Buffalo, NY 14225
westley@calspan.com
Let's hear it for smart mailers that cut off long signa