|
|
1.1 root 1: .\" Copyright (c) 1983 Regents of the University of California.
2: .\" All rights reserved. The Berkeley software License Agreement
3: .\" specifies the terms and conditions for redistribution.
4: .\"
5: .\" @(#)fsplit.1 6.1 (Berkeley) 4/29/85
6: .\"
7: .TH FSPLIT 1 "April 29, 1985"
8: .UC 5
9: .SH NAME
10: fsplit \- split a multi-routine Fortran file into individual files
11: .SH SYNOPSIS
12: .B fsplit
13: [
14: .B -e
15: efile] ... [ file ]
16: .SH DESCRIPTION
17: .B Fsplit
18: takes as input either a file or standard input containing Fortran source code.
19: It attempts to split the input into separate routine files of the
20: form
21: .I name.f,
22: where
23: .I name
24: is the name of the program unit (e.g. function, subroutine, block data or
25: program). The name for unnamed block data subprograms has the form
26: .I blkdtaNNN.f
27: where NNN is three digits and a file of this name does not already exist.
28: For unnamed main programs the name has the form
29: .I mainNNN.f.
30: If there is an error in classifying a program unit, or if
31: .I name.f
32: already exists,
33: the program unit will be put in a file of the form
34: .I zzzNNN.f
35: where
36: .I zzzNNN.f
37: does not already exist.
38: .PP
39: Normally each subprogram unit is split into a separate file. When the
40: .I -e
41: option is used, only the specified subprogram units are split into separate
42: files. E.g.:
43: .nf
44: fsplit -e readit -e doit prog.f
45: .fi
46: will split readit and doit into separate files.
47: .SH DIAGNOSTICS
48: If names specified via the
49: .I -e
50: option are not found, a diagnostic is written to
51: .I standard
52: .I error.
53: .SH AUTHOR
54: Asa Romberger and Jerry Berkman
55: .SH BUGS
56: .I Fsplit
57: assumes the subprogram name is on the first noncomment line of the subprogram
58: unit. Nonstandard source formats may confuse
59: .I fsplit.
60: .PP
61: It is hard to use
62: .I -e
63: for unnamed main programs and block data subprograms since you must
64: predict the created file name.
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.