home *** CD-ROM | disk | FTP | other *** search
- --::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
- --NOTES.DOC
- --::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
- The following notes relate to the installation and use of the Rapport
- Ada interface on a VAX computer using Telesoft's Ada (subset) compiler versions
- 1.3 or 1.5. The user should read the entire set of notes before beginning the
- installation.
-
- Examples and additional information may be obtained by contacting:
- Tim Porter
- SAIC Comsystems, San Diego
- UNITREP@NOSC-TECR
- 619-293-7500
- or
- Logica Database Products Inc.
- 666 third Avenue
- New York, N.Y. 10017
- 212-682-9344
-
- ========================================
- Implementing the Rapport - Ada Interface
- ========================================
-
- 1. The Rapport DBMS must be installed and available to the user.
-
- 2. Compile USERCALL.FOR and link with the appropriate Rapport
- interface modules referenced in the FORTRAN code.
-
- 3. Edit the V23USR.MAR to change MESSAGEPIPE to reflect the location
- of USERCALL.EXE.
-
- 4. Assemble V23USR.MAR utilizing AV23.COM and link into Telesoft
- utilizing LV23ROS.COM. (Refer to Telesoft documentation).
-
- 5. Compile R8LIB.TXT.
-
- 6. Follow the procedures detailed below to create applications.
-
-
- Production of a Database and Program
- ====================================
-
- The early stages using RAPPORT are documented in the RAPPORT 'Facilities'
- manual and the manual 'Designing a Database'. The writing of an Ada program
- is described in the Ada/RAPPORT Interface User Manual. The stages in the
- implementation of the hand generic method of database production is given
- below.
-
- - The logical names for Ada and RAPPORT and their help and command libraries
- are set up in the "LOGIN.COM" file for RAPPVMS3. RAPPORT might be
- protected from access by other users.
-
- - Define the structure of the database in a RAPPORT Database Definition File.
- The name of this file would normally be used for the package name of the
- file independent RAPPORT procedures. The filename extension should be
- ".DDF". If RAPIER or RAPIDE are to be used the standard ddf include
- statements must be appended and a sort records statement added to the
- beginning of the ddf.
-
- - 'Compile' the ddf to give a Database Common File by using the Database
- Definition Processor (DDF). This produces a file with the same name and
- an extension of ".DCF".
-
- - The database files must then be created. The RAPPORT utility DBS creates
- files with the correct structure, prompting for the filenames to be
- created.
-
- - If RAPIER or RAPIDE are to be used the program PREPARE must be used to
- initialize their data areas.
-
- - The contents of the other files can be initialized by RAPIDE (RAPPORT's
- prototyping language) or RAPIER (Screen Data Entry package).
-
- - The DCF checksums must be obtained. This can be done by setting up a
- dummy FORTRAN/RAPPORT program called "TEST.CPI". It can have the form
-
- PROGRAM TEST
- FILES ( filename )
- STARTRAPPORT( JTASK, JERROR )
- ENDRAPPORT( JTASK )
- END
-
- The program must be pre-processed using the RAPPORT Command Processor
- (RCP). The resulting program, "TEST.FOR", contains the basic calls to
- RAPPORT. The values for LAST_FILE and CHECKSUM are given in the subroutine
- call to R0STAR. Eg
-
- CALL R0STAR(' ',1,.FALSE.,
- 14,9295,jtask,jerror)
-
- The LAST_FILE is the fourth argument (14), the CHECKSUM the fifth (9295).
-
- - create a new subdirectory to [ADARAPP] for the 'instantiated' Ada packages.
-
- - Instantiate the top level RAPPORT package using the editor. This 'generic
- package' is called PERDDF.TXT and resides in the directory
- [adarapp.generics]. If it is edited and substitutions made the new copy
- must be placed in the new application subdirectory. (NB EDT takes its
- default directory to be that of the file it is editing). The instructions
- for instantiating the package are in the package header.
-
- - Compile the top level package.
-
- - Instantiate and compile the file level 'generic package' for each file to
- be used in the application. The file number required in this package can
- be found from the order of definition of files in the ddf. (The first
- file being 1, the next 2, and so on). The record structure is defined in
- this package and its translation from the ddf is defined in the User
- Manual. (See below)
-
- - Instantiate and compile the field level 'generic package' for each field to
- be used in a retrieval condition or ordering condition by the application.
- The field number required in this package can be found from the order of
- definition of fields in the ddf. (This is an absolute count of fields
- starting from the first field in the ddf.)
-
- The order of compilation of these packages is important since they use the
- context of packages compiled at previous levels. You are now ready to write
- an application program.
-
- Once it has been written and compiled a nucleus must be started. It can be
- either run on a terminal or submitted to the batch queue. Running a nucleus
- on the terminal can be done by the RAPPORT DCL command NUCLEUS. The command
- file to run a nucleus in the batch queue is
-
- NUCLEUS
- <a mailbox name used to identify the nucleus, eg ADANUC>
- <the dcf file name>
- ENTER
- CONT
- EXIT
-
- A nucleus run in the batch queue by this method may be stopped by the service
- command. Eg
-
- SERVICE/NAME=<mailbox name>
- Differences between the User Manual and the Prototype Implementation
- ====================================================================
-
- are listed below.
-
- - RAPPORT on a VAX uses 32 bit integers so, where the type INTEGER is
- used in the User Manual, the Telesoft-Ada program needs a declaration
- of LONG_INTEGER.
-
- - Normally record type definitions would be placed in the same package
- as the instantiations. In the prototype they are defined in the package
- for each file and is named DB_RECORD. A declaration of a database
- record would then be
-
- REC : file.DB_RECORD;
-