Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!apple!sun-barr!cs.utexas.edu!uunet!cimshop!davidm
From: davidm@cimshop.UUCP (David Masterson)
Newsgroups: comp.databases
Subject: Re: SQL-schema to C generation
Message-ID: <454@cimshop.UUCP>
Date: 4 Aug 89 17:40:17 GMT
References: <1910007@hpmcaa.mcm.hp.com>
Organization: Consilium Mt. View CA
Lines: 28
In-reply-to: nacer@hpmcaa.mcm.hp.com's message of 3 Aug 89 19:03:16 GMT

>Before the schema becomes stable every time it is modified the
>corresponding C declarations need to be updated manually I would like
>to know if there is a stand-alone SQL-schema to C declaration
>translator.
...
>Has any body used or attempted a similar automation?
>Can one use awk to produce such a tool?
>
The closest thing I've seen that is supported is VAX/VMS CDD+.  It is a data
dictionary product that will read the database and understand the declarations
of all tables there in.  VAX C (amongst the other VAX languages) has hooks in
it that allow it to ask the data dictionary for the current definition of
tables within the database.

However, this should be a fairly trivial thing to develop through AWK and the
like.  All the information that is needed should be in the relational
database's data dictionary.  All that is needed is to develop an appropriate
query to get the information out and then a shell script to massage the
information into C syntax.  The hard part would be how to make it automatic so
that MAKE could recognize definitions as having changed and rebuild the code
appropriately.  A possibility there would be to have MAKE always execute this
shell script, then DIFF the output with the previous output and only continue
if there are differences.  Just make sure that the script produces one file
for each table in the database (so you don't MAKE everything if only one table
changed).

David Masterson
uunet!cimshop!davidm		or		DMasterson@cup.portal.com