PROGRAM NAME:  splitf

AUTHOR:  Jim Besemer

DATE WRITTEN:  Jan [2-3], 1977

SOURCE LANGUAGE:  C

LOCATION OF SOURCE:  /usr/source/new/splitf.c

LOCATION OF BINARY:  /usr/bin/splitf

SYNTAX:
splitf <files>

DESCRIPTION:
This program is to help get around one of the problems with
our idiot-fortran compiler, which refuses to accecpt more than
a single fortran program or subroutine per input file.
Splitf reads data off of each of the input files, which are
presumed to contain FORTRAN subroutines.  It splits the file
up into seperate files, one for each FORTRAN module.  It tries
to create the output file name so that it is the same as the
subroutine's name with a ".f" appended to it.  If it cannot
figure out what the name is, it uses "sub000", "sub001", etc.
The only reason that it might not recognize a name is if there
are comments or garbage between an end card and the following
header statement.  
For "block data" subroutines, it uses the name "bdata.f" for
the output file.

MESSAGES & DIAGNOSTICS:
Splitf outputs a single line for each module encountered.
It consists of:
the selected output file name followed by the card it used
to deduce this name.
It issues the usual, self explanitory messages for files which it
can't open, etc.
If it picks a name for an output file, tries to "creat" it, and
can't, it will try again with the "sub000" business until it finds
one which works.

SEE ALSO:
fc(I)

FILES USED:
All sorts of possible output file names are possible.
It is recommended that fortran files be split up in an
empty directory all their own, to avoid possible conflicts.
