. \" Definitions of F, L and LR for the benefit of systems
. \" whose -man lacks them...
.de F
.nh
.if n \%\&\\$1
.if t \%\&\f(CW\\$1\fR
.hy 14
..
.de L
.nh
.if n \%`\\$1'
.if t \%\&\f(CW\\$1\fR
.hy 14
..
.de LR
.nh
.if n \%`\\$1'\\$2
.if t \%\&\f(CW\\$1\fR\\$2
.hy 14
..
.TH F2C 1
.CT 1 prog_other
.SH NAME
f2c \- Convert Fortran 77 to C or C++
. \" f\^2c changed to f2c in the previous line for the benefit of
. \" people on systems (e.g. Sun systems) whose makewhatis cannot
. \" cope with troff formatting commands.
.SH SYNOPSIS
.B f\^2c
[
.I option ...
]
.I file ...
.SH DESCRIPTION
.I F2c
converts Fortran 77 source code in
.I files
with names ending in
.L .f
or
.L .F
to C (or C++) source files in the
current directory, with
.L .c
substituted
for the final
.L .f
or
.LR .F .
If no Fortran files are named,
.I f\^2c
reads Fortran from standard input and
writes C on standard output.
.I File
names that end with
.L .p
or
.L .P
are taken to be prototype
files, as produced by option
.LR -P ,
and are read first.
.PP
The following options have the same meaning as in
.IR f\^77 (1).
.TP
.B -C
Compile code to check that subscripts are within declared array bounds.
.TP
.B -I2
Render INTEGER and LOGICAL as short,
INTEGER\(**4 as long int.  Assume the default \fIlibF77\fR
and \fIlibI77\fR:  allow only INTEGER\(**4 (and no LOGICAL)
variables in INQUIREs.  Option
.L -I4
confirms the default rendering of INTEGER as long int.
.TP
.BI -I dir
Look for a non-absolute include file first in the directory of the
current input file, then in directories specified by \f(CW-I\fP
options (one directory per option).  Options
\f(CW-I2\fP and \f(CW-I4\fP
have precedence, so, e.g., a directory named \f(CW2\fP
should be specified by \f(CW-I./2\fP .
.TP
.B -onetrip
Compile DO loops that are performed at least once if reached.
(Fortran 77 DO loops are not performed at all if the upper limit is smaller than the lower limit.)
.TP
.B -U
Honor the case of variable and external names.  Fortran keywords must be in
.I
lower
case.
.TP
.B -u
Make the default type of a variable `undefined' rather than using the default Fortran rules.
.TP
.B -w
Suppress all warning messages, or, if the option is
.LR -w66 ,
just Fortran 66 compatibility warnings.
.PP
The following options are peculiar to
.IR f\^2c .
.TP
.B -A
Produce
.SM ANSI
C (default, starting 20020621).
For old-style C, use option \f(CW-K\fP.
.TP
.B -a
Make local variables automatic rather than static
unless they appear in a
.SM "DATA, EQUIVALENCE, NAMELIST,"
or
.SM SAVE
statement.
.TP
.B -C++
Output C++ code.
.TP
.B -c
Include original Fortran source as comments.
.TP
.B -cd
Do not recognize cdabs, cdcos, cdexp, cdlog, cdsin, and cdsqrt
as synonyms for the double complex intrinsics
zabs, zcos, zexp, zlog, zsin, and zsqrt, respectively,
nor dreal as a synonym for dble.
.TP
.BI -d dir
Write
.L .c
files in directory
.I dir
instead of the current directory.
.TP
.B -E
Declare uninitialized
.SM COMMON
to be
.B Extern
(overridably defined in
.F f2c.h
as
.B extern).
.TP
.B -ec
Place uninitialized
.SM COMMON
blocks in separate files:
.B COMMON /ABC/
appears in file
.BR abc_com.c .
Option
.LR -e1c
bundles the separate files
into the output file, with comments that give an unbundling
.IR sed (1)
script.
.TP
.B -ext
Complain about
.IR f\^77 (1)
extensions.
.TP
.B -f
Assume free-format input: accept text after column 72 and do not
pad fixed-format lines shorter than 72 characters with blanks.
.TP
.B -72
Treat text appearing after column 72 as an error.
.TP
.B -g
Include original Fortran line numbers in \f(CW#line\fR lines.
.TP
.B -h
Emulate Fortran 66's treatment of Hollerith: try to align character strings on
word (or, if the option is
.LR -hd ,
on double-word) boundaries.
.TP
.B -i2
Similar to
.BR -I2 ,
but assume a modified
.I libF77
and
.I libI77
(compiled with
.BR -Df\^2c_i2 ),
so
.SM INTEGER
and
.SM LOGICAL
variables may be assigned by
.SM INQUIRE
and array lengths are stored in short ints.
.TP
.B -i90
Do not recognize the Fortran 90 bit-manipulation intrinsics
btest, iand, ibclr, ibits, ibset, ieor, ior, ishft, and ishftc.
.TP
.B -kr
Use temporary values to enforce Fortran expression evaluation
where K&R (first edition) parenthesization rules allow rearrangement.
If the option is
.LR -krd ,
use double precision temporaries even for single-precision operands.
.TP
.B -P
Write a
.IB file .P
of ANSI (or C++) prototypes
for definitions in each input
.IB file .f
or
.IB file .F .
When reading Fortran from standard input, write prototypes
at the beginning of standard output.  Option
.B -Ps
implies
.B -P
and gives exit status 4 if rerunning
.I f\^2c
may change prototypes or declarations.
.TP
.B -p
Supply preprocessor definitions to make common-block members
look like local variables.
.TP
.B -R
Do not promote
.SM REAL
functions and operations to
.SM DOUBLE PRECISION.
Option
.L -!R
confirms the default, which imitates
.IR f\^77 .
.TP
.B -r
Cast REAL arguments of intrinsic functions and values of REAL
functions (including intrinsics) to REAL.
.TP
.B -r8
Promote
.SM REAL
to
.SM DOUBLE PRECISION, COMPLEX
to
.SM DOUBLE COMPLEX.
.TP
.B -s
Preserve multidimensional subscripts.  Suppressed by option
.L -C
\&.
.TP
.BI -T dir
Put temporary files in directory
.I dir.
.TP
.B -trapuv
Dynamically initialize local variables, except those appearing in
.SM SAVE
or
.SM DATA
statements, with values that may help find references to
uninitialized variables.  For example, with IEEE arithmetic,
initialize local floating-point variables to signaling NaNs.
.TP
.B -w8
Suppress warnings when
.SM COMMON
or
.SM EQUIVALENCE
forces odd-word alignment of doubles.
.TP
.BI -W n
Assume
.I n
characters/word (default 4)
when initializing numeric variables with character data.
.TP
.B -z
Do not implicitly recognize
.SM DOUBLE COMPLEX.
.TP
.B -!bs
Do not recognize \fIb\fRack\fIs\fRlash escapes
(\e", \e', \e0, \e\e, \eb, \ef, \en, \er, \et, \ev) in character strings.
.TP
.B -!c
Inhibit C output, but produce
.B -P
output.
.TP
.B -!I
Reject
.B include
statements.
.TP
.B -!i8
Disallow
.SM INTEGER*8 ,
or, if the option is
.LR -!i8const ,
permit
.SM INTEGER*8
but do not promote integer
constants to
.SM INTEGER*8
when they involve more than 32 bits.
.TP
.B -!it
Don't infer types of untyped
.SM EXTERNAL
procedures from use as parameters to previously defined or prototyped
procedures.
.TP
.B -!P
Do not attempt to infer
.SM ANSI
or C++
prototypes from usage.
.PP
The resulting C invokes the support routines of
.IR f\^77 ;
object code should be loaded by
.I f\^77
or with
.IR ld (1)
or
.IR cc (1)
options
.BR "-lF77 -lI77 -lm" .
Calling conventions
are those of
.IR f\&77 :
see the reference below.
.br
.SH FILES
.TP
.nr )I 1.75i
.IB file .[fF]
input file
.TP
.B *.c
output file
.TP
.F /usr/include/f2c.h
header file
.TP
.F /usr/lib/libF77.a
intrinsic function library
.TP
.F /usr/lib/libI77.a
Fortran I/O library
.TP
.F /lib/libc.a
C library, see section 3
.SH "SEE ALSO"
S. I. Feldman and
P. J. Weinberger,
`A Portable Fortran 77 Compiler',
\fIUNIX Time Sharing System Programmer's Manual\fR,
Tenth Edition, Volume 2, AT&T Bell Laboratories, 1990.
.SH DIAGNOSTICS
The diagnostics produced by
.I f\^2c
are intended to be
self-explanatory.
.SH BUGS
Floating-point constant expressions are simplified in
the floating-point arithmetic of the machine running
.IR f\^2c ,
so they are typically accurate to at most 16 or 17 decimal places.
.br
Untypable
.SM EXTERNAL
functions are declared
.BR int .
.br
There is no notation for
.SM INTEGER*8
constants.
.br
Some intrinsic functions do not yet work with
.SM INTEGER*8 .