Introduction to TAS V2.0 for EOS

The new aspects of V2.0 TAS for EOS, including CVS code management, are:

General description

Moving from TAS V1 to TAS V2

This describes some of the transition issues in moving from TAS V1 to TAS V2 for EOS.

At present (21-Oct-94) the TAS V2 library set up is functioning on sseos.lbl.gov and the V2 items listed below pertain to the present setup.

The list below gives an approximate translation of development procedures in TAS V1 and TAS V2. Procedures which include VMS are yet to be implemented or described.

              TAS V1                        TAS V2
1.        ~eostpc/login_eos             ~eostpc/login_eos.csh
                                            (in your .cshrc file)
2.         tas_edit                      tas2_menu
                                            1. TAS table editer
3.         gen_tas_table_reps            tas2_menu
                                            5. Generate .F format TAS table report files
4.         tas_update                    like 3. above, followed by cvs checkin
5.         eos_all_mms.bld               make install
6.         @eosx.bld                     make tas

Besides these differences in development procedures, there are differences in using CVS for code management compared with using CMS. An approximate comparision is given below but there are significant differences between CVS and CMS show you should become familiar with the CVS procedures and how different they are than CMS and not just assume that you can type cvs rather than cms. One primary difference between CMS and CVS is that CVS operates on many files or directories at once while CMS oeprates on single files or named groups of files.

    CMS command                CVS with some equivalence to CMS
1.  cms create element         cvs add
2.  cms reserve element        cvs checkout
3.  cms replace element        cvs commit
4.  cms delete element         cvs remove
5.  cms insert element         cvs tag, cvs rtag
6.  cms create class           cvs tag, cvs rtag
7.  cms fetch                  cvs export
8.  cms fetch/gen=             cvs export -r

Login setup

The setup for TAS V2.0 for EOS exists on the LBL UNIX machines, sseos, nsdssf, nsdssi, nsdssk, ...
To define your environment you must place

source ~eostpc/login_eos.csh

in your .cshrc file. This should replace "source ~eostpc/login_eos" that you already had. For this to be effective, you should not set your path in your .cshrc file, but set it instead in your .login file. It will be necessary for you to put a "source ~/.cshrc" in your .login file in this case.

As before (with ~eostpc/login_eos) when you type "newgrp nsdtpc" to change your group to the EOSTPC group id, the environment for TAS V2.0 development is set up from the ~eostpc/login_eos.csh" file. This basic environment setup is required for all other aspects of the TAS V2.0 for EOS discussion.

CVS code management

As mentioned above under "General description" there are some differences between CMS and CVS, besides the VMS-UNIX difference. One thing you should do is read the man page on cvs (man cvs).

CVS manages a whole directory tree structure in addition to individual files and most cvs actions automatically descend through all subdirectories. In this sense, CVS is considerably more powerful than CMS; a lot can happen even if you do not specify a particular file.

The cvs commands that you would normally use, and should read up on are:

You can also browse through this FAQ on cvs.

Module interface definition (.mid) files

The module interface definition (.mid) file is the central difference between TAS V1 and TAS V2. This is where the various sets of tables and modules that are partitioned in TAS V2 are brought together to make a single executable program. The TAS V2.0 manual has a section describing this which you should read. at present (11/14/94) the TAS manual describes .mid files and the interface to analysis modules in the context of .f (dot little f) format files (normal Fortran). The procedures that exist now for TAS V2.0 use the .F (dot big F) source format for Fortran (see section on "Source Format" below).

Source Format (.F)

The .F source format means that the Fortran compiler first runs the C preprocessor before running the Fortran compiler. This is used to handle include file specifications as

#include "some_include_file.inc"

rather than the Fortran INCLUDE statement. The reason for using the C preprocessor has to do with features allow specifying directories to search for include files more useful and permit a more reliable development process with general purpose make files.

Converting a file between the .f and .F format is done with the ftoF and Ftof commands. At the moment, the specifications for the source format of TAS V2.0 code to run under VMS is not finalized and so all the code used on UNIX machines and managed under CVS should be in the .F format.

Tutorial examples


Last modification: 8-March-1995

Doug Olson
dlolson@lbl.gov
Up to EOS home page